PR

Loughborough University
Institutional Repository
Construction labour
productivity analysis and
benchmarking: the case of
Tanzania
This item was submitted to Loughborough University's Institutional Repository

by the/an author.
Additional Information:
A Doctoral Thesis. Submitted in partial fulllment of the requirements

for the award of Doctor of Philosophy of Loughborough University.
Metadata Record: https://dspace.lboro.ac.uk/2134/7444
Publisher: c Ninatubu Mbora Lema

Please cite the published version.

This item is held in Loughborough Universitys Institutional Repository
(https://dspace.lboro.ac.uk/) and was harvested from the British Librarys
EThOS service (http://www.ethos.bl.uk/). It is made available under the
following Creative Commons Licence conditions.
For the full text of this licence, please go to:

http://creativecommons.org/licenses/by-nc-nd/2.5/
CONSTRUCTION LABOUR PRODUCTIVITY
ANALYSIS AND BENCHMARKING - THE CASE OF
TANZANIA
by
Ninatubu Mbora Lema, BSc. (Hons), MSc.
A Doctoral Thesis submitted in partial fulfilment of the

requirements for the award of Doctor of Philosophy of the
Loughborough University
September, 1996
Ninatubu Mbora Lema, 1996

TABLE OF CONTENTS
Page
Abstract ii
Certificate of originality
Acknowledgements iv
List of chapters V
List of appendices XII
List of tables XIII
List of figures xv
Chapters 1
References 317
Appendices 343
ABSTRACT
This research aimed at investigating strategies for construction performance

improvement in Tanzania. The research established that the Total Quality
Management (TQM) philosophy provides a feasible long term performance
improvement strategy. Benchmarking was identified as a tool for initiating
and sustaining the TQM programme.
Labour productivity was selected as the key construction performance

indicator. A framework for labour productivity benchmarking was
developed, on the basis of current mean productivity (CMP) and target
mean productivity (TMP). Construction labour productivity at macro-
economic level and site level were also investigated.
Analysis at macro economic level over a twenty five year period between
1969 and 1993 indicated a continuous decline in productivity expressed in
value added per person engaged. Site labour productivity was investigated
for eight construction activities on 46 sites belonging to 23 different
contractors. Two significant findings emerged in the analysis: first, the
variability quantified by coefficient of variation was considerably higher than
in similar studies elsewhere; and secondly the distribution was skewed to
the left suggesting that productivity was low for most of the operatives.
These characteristics were indicative of the productivity improvement
potential in the Tanzanian building construction industry. A distribution
modelling exercise established that Johnson SB distribution (with shape
parameters, 11=1 and y= 1) model well represented productivity distribution
for most activities. From this distribution, it was established that about 85
per cent of operatives productivity was below the median, which provided a
basis for quantifying the potential for improvement. The benchmarking
model established that there was an improvement potential of about 133 per
cent. This potential was verified through an opinion survey of operatives.
Factors that influence operatives productivity were identified through an

opinion survey. Factor related to motivation were ranked highest in the
survey . Possible influence of various factors on productivity was quantified
through regression modelling based on actual construction operation
observations. This analysis indicated that productivity depended on
productive time which is largely influenced by operative motivation,
supporting the results of the opinion survey.
ACKNOWLEDGEMENTS
I wish to express my most sincere gratitude to Dr. Andrew Price, my

supervisor, who was very keen to see this work through in time. His special
academic qualities inspired me throughout the research period. I am very
grateful for all I have acquired in the course of this research. Many thanks
also to my Director of Research, Prof. Ronald McCaffer for his support.
I wish to acknowledge the support from Association of Commonwealth

Universities, National Construction Council, Faculty of Engineering at the
University of Dar es Salaam and the National Board of Architects, Quantity
Surveyors and Building Contractors who facilitated this research in different
ways.
This research was made possible by a number of people and organisations.

Many construction workers from numerous companies were involved. I wish
to express my sincere gratitude to all of them as well as the many company
executives who participated. Special thanks go to construction executives
who provided invaluable information and opinion. Many thanks to the civil
engineering students at the University of Dar es Salaam who assisted me in
conducting the site studies. The contributions made by Ramadhani Mlinga,
Declare Mushi, Hillary Towo, E. Lima and F. Muimbula are acknowledged
with thanks.
Many friends and colleagues at Loughborough contributed to the research

through informal discussions, advice and criticisms. I wish to express my
thanks to Francis Edum-fotwe, Abdul Kadir, Brian Wakley, Bashir Assalah,
Vian Ahmed, Mathew Chemelil, Jawahar Nesan, Tarek Hassan and Zaimi
Abdul Majid.
I have received invaluable support from my family during this research. My

very special thanks to my wife, Loyce Lema who sacrificed her career to
accompany me to UK. My children Lolande and Eonike who endured
without realising. My brothers and sisters who took over my responsibilities
in Tanzania. No words can express enough my thanks to my parents Mr.
and Mrs Mbora Lema, for their dedication and sacrifice.
Thanks to the Lord for this work, for without His guidance, we labour in vain.
iv
LIST OF CHAPTERS
Section page
1. Introduction
1.1 Background 1
1.2 Justification and significance 5
1.3 Significance of building sector 7
1.4 Objectives of the research 8
1.5 Research methodology 10
1.6 Guide to the thesis 11
1.7 Summary of conclusionsand recommendations 17
2.0 Construction industry characteristics and performance

improvement
2.1 Introduction 20
2.2 Construction industry - definition and characteristics 20
2.2.1 The dilemma - defining what construction industry is 20
2.2.2 The search for a working definition 21
2.2.3 Characteristics of the construction industry 26
2.2.4 Special characteristics of construction industries in
developing countries 28
2.3 Tanzanian construction industry structure and
characteristics - the formal sector 31
2.3.1 Formal sector contracting capacity 31
2.3.2 Professional design services 33
2.3.3 Costs and cost trends 34
2.3.4 Other characteristics 35
2.3.5 Recent developments -a diagnosis 36
2.4 Economic significance of the construction industry 38
2.5 Economic performance indicators and relationships 40
2.5.1 Performance indicators 40
2.5.2 Productivity and economic growth 43
2.5.3 Project level performance evaluation 46
2.6 Challenges in the construction industry performance
improvement 48
2.7 Summary 49
V
3.0 Modern approach to business performance improvement
3.1 Introduction 51
3.2 Business performance measurement 51
3.2.1 Purpose of measurement 51
3.2.2 Role of financial indicators 53
3.2.3 Limitations of financial performance measures 53
3.3 Total Quality Management (TQM) and associated
philosophies 54
3.3.1 Emergence of Total Quality Management (TQM) 54
3.3.2 Linkages between productivity and quality 59
3.3.3 The role of TQM in construction performance
improvement 61
3.3.4 Benchmarking as a tool for TQM implementation 62
3.4 Emergence of the benchmarking concept 64
3.4.1 Background 64
3.4.2 Objectives and definition of benchmarking 66
3.4.2.1 Objectives of benchmarking 66
3.4.2.2 Definitions 66
3.4.2.3 Terms used in benchmarking 68
3.4.3 Brief history of benchmarking 69
3.4.4 TQM and benchmarking 70
3.4.5 The strength of benchmarking concept. 71
3.5 Scope, procedures and success of benchmarking
applications 72
3.5.1 Scope 72
3.5.2 Types of benchmarking 74
3.5.3 Benchmarking models 77
3.5.4 Successful applications of benchmarking 80
3.6 Potential for TQM and benchmarking in construction 81
3.6.1 Philosophy 81
3.6.2 Applications of benchmarking in the UK construction
industry 82
3.6.3 Other construction benchmarking initiatives 86
3.6.4 A framework for TQM and benchmarking applications
in construction 88
3.7 Relevance of the conceptual framework in this research 89
3.8 Summary 91
vi
4.0 Construction productivity
4.1 Introduction 93
4.2 Productivity - origin, concept and basic definition 93
4.2.1 Origin of the word 'productivity' 93
4.2.2 The search for a comprehensive definition of
productivity 95
4.2.3 Dictionary definition 95
4.2.4 Other definitions 95
4.2.5 Productivity concepts 96
4.2.6 Basic definitions of productivity
101
4.3 Construction industry definitions of productivity and
associated measures 104
4.3.1 The need for construction industry specific
definitions 104
4.3.2 Labour productivity -a partial definition of
productivity 104
4.3.3 Multi-factor productivity (total factor productivity) 105
4.3.4 Industry level productivity 106
4.3.5 Project level productivity 107
4.3.6 Crew level productivity 107
4.4 Productivity measurement 109
4.4.1 Significance of productivity measurement109
4.4.2 General obstacles to effective measurement110
4.5 Work study 113
4.5.1 Brief history 113
4.5.2 Objectives of work study 115
4.5.3 Work measurement (time study) 116
4.6 Activity sampling 117
4.6.1 Background 117
4.6.2 Theory 118
4.6.3 Activity sampling procedure 121
4.6.4 Activity sampling in construction 121
4.6.5 Activity sampling as a work measurement tool - the
dilemma 123
4.7 Factors that affect construction labour productivity 126
4.7.1 Significance of the study of productivity factors in
construction 126
vii
4.7.2 Classifications of productivity factors 128
4.7.3 Factors affecting construction labour productivity 133
4.7.4 Summary of factors affecting productivity 138
4.8 Summary 139
5.0 Research methodology and benchmarking framework
5.1 Introduction 142

5.2 Research methodology 143
5.2.1 Literature on research theory and methodology 143
5.2.2 General framework 143
5.3 Research processes 146
5.3.1 Problem genesis 146
5.3.2 Problem definition 146
5.3.3 Research strategy 146
5.3.4 Discussion of strategies used in the research 149
5.3.5 Reliability and validity of survey results 149
5.4 Special aspects of research in developing countries 150
5.5 The formulation of benchmarking framework 151
5.5.1 The research problem 151
5.5.2. Approach to formulating a benchmarking
framework 152
5.5.3 Description of labour productivity benchmarking
framework 153
5.5.3.1 Planning 155
5.5.3.2 Analysis 156
5.5.3.3 Implementation 158
5.6 Review of labour productivity models 159
5.6.1 Significance of productivity models 159
5.6.2 Review and evaluation of labour productivity
models 159
5.7 Choice of a suitable model 167
5.8 Productivity performance gap concept 170
5.9 Summary 172
viii
6.0 Data collection

6.2 Labour productivity data collection process - Phase I 175
6.2.1 Operative level productivity 175
6.2.2 Training and deployment of site observers 176
6.2.3 Activity descriptions 177
6.2.4 Data collection procedure 178
6.2.5 Data quality control 182
6.3 Data c ollection - Phase II 183
6.3.1 Scope for data collection 183
6.3.2 Tanzanian construction industry performance data 183
6.3.3 Opinion survey - labour productivity factors 184
6.3.4 Site labour productivity studies and productivity
factors assessment 189
6.3.5 Opinion survey on labour productivity 191
Operatives structured interviews
6.3.6 192
6.3.7 Interviews with prominent persons in the Tanzanian
construction industry 193
6.4 Summ ary 194
7.0 Tanzanian construction industry performance

7.2 Performance measures 196
7.3 Tanzanian construction industry performance197
7.3.1 History 197
7.3.2 Emergence of formal construction industry 198
7.3.3 Post independence construction industry 199
7.3.4 Construction output quantification 199
7.3.5 Construction output between 1969 and 1993. 201
7.3.6 Performance trends analysis 1969 - 1993 202
7.3.6.1 Analysis framework, 202
7.3.6.2 Output trends 202
7.3.6.3 Contribution to the GDP and GFCF 205
7.3.6.4 Creation of employment 206
7.3.6.5 Productivity trends 207
7.4. Construction industry performance comparison 208
7.4.1 Scope of comparison 208
ix
7.4.2 Output comparison - value added comparison 208
7.4.3 Creation of employment 212
7.4.4 Construction industry productivity 214
7.5 Wider international productivity analysis 218
7.5.1 Significance of relationshipbetween productivity and
GDP per capita 218
7.5.2 Productivity and GDP relationship for 1979 data 218
7.5.3 Productivity and GDP relationship for 1992 data 221
7.5.4 Construction contribution to GDP 227
7.5.5 Productivity comparison with selected developed
economies 229
7.6 Summary 230
8.0 Analysis of labour productivity at site level

8.2 Productivity data analysis 234
8.2.1 Objectives of the analysis 234
8.2.2 Review of previous studies 235
8.3. Variability analysis 237
8.3.1 Significance of variability and statistical distribution 237
8.3.2 Summary and analysis of productivity study results 237
8.3.3 Analysis of the variability of the results 239
8.3.4 Comparison of productivity variability with
other studies 240
8.4 Frequency distributions of the productivity data 242
8.4.1 Graphical method - histograms 242
8.4.2 Normality test 245
8.4.3 Fitting data on empirical distributions 247
8.4.4 Justification for Johnson SB distribution 248
8.5 Comparison with other research findings 255
8.6 Quantification of productivity benchmarking gap 255
8.7 Validation of benchmarking gap 258
8.7.1 Operatives' percieved productivity potential 258
8.7.2 Contractors' quantification of productivity potential 260
8.8 Summary 261
X
9.0 Construction labour productivity factors
9.2 Summary and evaluation of productivity factors rating 267
9.2.1 Summary of factors ratings 267
9.2.2 Evaluation of degree of agreement amongst
operatives 268
9.2.3 Validation of productivity factor ranking 271
9.2.4 Comparative factor ranking by contractors' senior
management 274
9.2.5 Ranking of factors 276
9.3 Productivity and factors evaluation on construction sites 277
9.3.1 Concreting process and productivity observation 277
9.3.2 Basic concreting productivity statistics 282
9.3.3 Exploration of some productivity relationships 283
9.4 Site evaluation of factors that influence productivity 293
9.4.1 Summary of evaluation of factors 293
9.4.2 Productivity and productive time comparisons 295
9.5 Regression modelling 297
9.5.1 Basis of modelling 297
9.5.2 Concrete mixing regression models 298
9.6 Discussion of regression results and productivity factors 301
9.7 Summary 302
10. Conclusions and recommendations for further research and

actions
10.2 Conclusions 305
10.2.1 Continuous construction productivity improvement
philosophy 305
10.2.2 Tanzanian construction industry performance 306
10.2.3 Labour productivity at operative level 307
10.2.4 Factors that influence operative productivity 311
10.2.5 Evaluation of productivity factors on sites 311
10.2 Recommendations for further research and actions 313
10.2.1 Performance measurement in the construction
industry 313
10.2.2 Framework for implementation of research findings 314
10.2.3 Widened scope of performance comparison 315
10.2.4 Detailed studies on labour productivity factors 315
xi
LIST OF APPENDICES
Appendix 1: Construction operatives' and contractors' questionnaires

Phase I
Appendix 2: Cost breakdown of selected building projects in Dar es

Salaam (based on the winning tenders)
Appendix 3: Contractors' questionnaires - Phase II
Appendix 4: Operatives' site questionnaires
Appendix 5: Construction executives' interview guide
Appendix 6: Construction industries performance statistics - Tanzania,

Kenya and United Kingdom
Appendix 7: Construction output statistics for 87 countries
Appendix 8: Site productivity data for eight construction activities in

Tanzania
XII
LIST OF FIGURES
Figure Page
Figure 1.1 Thesis logical framework 12

Figure 1.2 Detailed thesis framework 13
Figure 2.1 Conceptual construction industry scope 25
Figure 2.2 Registered building contractors in Tanzania
1980 " 1993 32
Figure 3.1 A TQM framework 56
Figure 3.2 The evolution of technique, practice and philosophy
leading to total quality management in construction
in the 1990's 57
Figure 3.3 Juran's triple role concept applied to construction 58
Figure 3.4 Efficiency matrix 61
Figure 3.5 Deming's quality: productivity chain reaction 61
Figure 3.6 Mapping progress against the survival, prevention and
continuous improvement assessment criteria 63
Figure 3.7 Benchmarking types 77
Figure 3.8 The benchmarking process compared to Deming Cycle 78
Figure 3.9 Process of benchmarking 42
Figure 3.10 Xerox benchmarking model 79
Figure 3.11 Benchmarking for TQM implementation 89
Figure 3.12 Benchmarking metrics and construction productivity
measures 90
Figure 4.1 Conceptual production process illustration 98
Figure 4.2 A conceptual process production for a car 99
Figure 4.3 A conceptual concreting process illustration 100
Figure 4.4 Work study ' 115
Figure 4.5 Normal distribution for activity sampling application 119
Figure 4.6 Classification of construction working day 120
Figure 4.7 Conceptual representation of UN Report 129
Figure 4.8 Factor model 130
Figure 4.9 Construction as an open conversion process 132
Figure 5.1 Research methodology adopted on this study 145
Figure 5.2 Labour productivity benchmarking research
methodology 148
Figure 5.3 Generic benchmarking model 152
Figure 5.4 Labour productivity benchmarking framework 154
xiii
Figure 5.5 Labour productivity models 160
Figure 5.6 Closed conversion construction process 160
Figure 5.7 Delay model time breakdown 162
Figure 5.8 Graphical representation of the activity model 162
Figure 5.9 Application of site factors to basic times 163
Figure 5.10 Performance determinants 165
Figure 5.11 Hybrid labour productivity study model 170
Figure 5.12 Labour productivity benchmarking concept 171
Figure 7.1 Construction industry output in Tanzania - 1960 - 68 200
Figure 7.2 Construction output in Tanzania 204
Figure 7.3 Construction output in 1987 constant prices 204
Figure 7.4 Construction contribution to the GDP and GCFC 205
Figure 7.5 Total and construction employment trends 206
Figure 7.6 Output and value added per employee trends 207
Figure 7.7 Comparison of construction industry contribution
to GDP 209
Figure 7.8 Construction value added in 1987 constant prices 210
Figure 7.9 Value added and total output in 1987 constant prices 210
Figure 7.10 Construction value added per capita comparison 212
Figure 7.11 Construction employment and population comparison 213
Figure 7.12 Construction industry contribution to employment 214
Figure 7.13 productivity in output per person engaged (OPE) 215
Figure 7.14 Productivity in value added per person
engaged (VAPE) 215
Figure 7.15 Productivity trend in 1987 constant prices 216
Figure 7.16 Productivity trends in output per person
engaged (OPE) 217
Figure 7.17 Relationship between VAPE and GDP (in current US $
based on 1979 statistics) 220
Figure 7.18 Value added per employee VAPE and GDP per capita 225
Figure 7.19 GDP and construction value added per capita 228
Figure 7.20 International construction productivity trends
comparison 230
Figure 8.1 Flow chart for the analysis methodology 238
Figure 8.2 Box-and-whiskers plot for productivity data 240
Figure 8.3 Frequency histograms for selected construction
activities 244
Figure 8.4 Plots of (b1, b2) for productivity distributions on (l, 02)
plane for various distributions 249
xiv
Figure 8.5 Plots of b1, and b2 for various productivity
distributions 250
Figure 8.6 Typical Johnson SB distribution with y=il = 1;
E=0; a,=1 253
Figure 8.7 Actual and expected frequencies using Johnson SB
distribution 254
Figure 8.8 Current labour distribution 257
Figure 8.9 Target labour productivity distribution 257
Figure 8.10 Labour productivity benchmarking gap 257
Figure 9.1 Analysis framework and regression modelling 266
Figure 9.2 Ranking of productivity factors by operatives 268
Figure 9.3 Ranking of productivity factors by operatives 271
Figure 9.4 Ranking of productivity factors by contractors' senior
management 275
Figure 9.5 Schematic representation of concreting activity 279
Figure 9.6 Frequency histograms concreting productivity 283
Figure 9.7 Gang size ratio and productivity for ground floor slab
(GFS) 285
Figure 9.8 Gang size ratio vs productivity for FFS and SFS 285
Figure 9.9 Pour size and mixing gang size relationship 287
Figure 9.10 Pour size and placing gang size relationship 288
Figure 9.11 Mixing productivity vs gang size 289
.
Figure 9.12 Placing productivity vs gang size 290
Figure 9.13 Mixing productivity vs productive time 291
Figure 9.14 Placing productivity vs productive time 292
Figure 9.15 Mixing productivity rate vs productive time 292
Figure 9.16 Placing productivity rate vs productive time 292
xv
LIST OF TABLES
Table Page
Table 2.1 Number of registered building contractors - December 1993 33

Table 2.2 Registered consultancy firms in Tanzania as of 1990 34
Table 2.3 Supply and demand of construction technical personnel 35
Table 2.4 Average contribution to GDP, GFCF and employment in
countries grouped according to income 42
Table 2.5 Project success factors 47
Table 2.6 Construction project soft indicators 47
Table 3.1 The quality evolution - stages of development for the world
class company 63
Table 3.1 Comparison of benchmarking definition by different
authors 75
Table 4.1 Chronology of some important definitions of productivity 94
Table 5.1 Subjective evaluation of productivity models for
benchmarking suitability 169
Table 6.1 Sample activity sampling study sheet 180
Table 6.2 Sample activity sampling study summary sheet 181
Table 6.3 Operatives' productivity factors ranking matrix 189
Table 6.4 Site productivity factors assessment form 190
Table 6.5 Contractors' opinion survey list and class of registration 192
Table 7.1 Construction value added per person engaged and GDP
for selected countries (based on 1979 statistics) 219
Table 7.2 Construction employment, GDP and VAPE for selected
countries (based on 1992 statistics) 222
Table 7.3 International construction industry productivity
comparison 229
Table 8.1 Productivity data sources - contractors and sites 234
Table 8.2 Productivity results and basic statistics 239
Table 8.3 Productivity variability from various studies 242
Table 8.4 Normality tests using standard error of skewness 246
Table 8.5 Johnson SB distribution parameters and chi-squared test 253
Table 8.6 Current and target mean labour productivity for selected
activities 258
Table 8.7 Comparisons of operative productivity potential 259
Table 8.8 Productivity comparisons - Nigeria and Tanzania 260
Table 8.9 Contractors' quantification of productivity potential 260
Table 9.1 Rating of productivity factors by construction operatives 267
xvi
Table 9.2 Validation factors rating by contractors' operatives 271
Table 9.3 Factor ranking comparison in two independent surveys 273
Table 9.4 Rating of productivity factors by contractors' senior
management 274
Table 9.5 Comparisons of factors ranking between contractors and
operatives 276
Table 9.6 Concreting productivity study results 280
Table 9.7 Concrete mixing and placing productivity 281
Table 9.8 Concreting operation productivity basic statistics 282
Table 9.9 Rated site productivity factor scores 294
Table 9.10 ANOVA tests on concrete mixing and placing productivity
for different slab types 296
Table 9.11 General regression models - analysis of variance 299
Table 9.12 Critical concrete mixing productivity model 300
Table 10.1 List of proposed actions and responsibilities 316
xvii
CHAPTER ONE
INTRODUCTION
CHAPTER ONE
INTRODUCTION
1.1 Background
Construction productivity has been an issue of concern in most countries for

some time. In the developed world, proof exists of a scientific approach to
productivity studies and improvement for over two and a half centuries (Price
1986). On the African continent there is, however, very little documentation in
spite of some of the most outstanding construction achievements, such as the
Egyptian pyramids. Prior to commercialisation of the building process in the
African continent, which came about during the colonial period, construction,
especially housing, remained a community responsibility. This process is still
predominant in rural areas where there is little influence of the western
civilisation. Men and women would gather all materials for a preconceived
standard thatched house. The construction of which normally takes only one
day. This would be difficult to achieve in an urban setting under current social
and economic conditions. Clearly, this would require thorough pre-planning to
ensure that all materials and manpower would be sufficient to undertake the
task in one day. However, little literature is available on the form of planning
nor of productivity achievable in these conditions. There is also little
documentation of the construction industry during the colonial period, in spite
of some major achievements such as the railway lines, roads, airports and
colonial government buildings. Construction industries in the economies of
post-colonial African states have had no significant efficiency improvements in
spite of efforts at national level (Wells 1986; Ofori 1990). The application of
scientific approaches to the study of crew productivity is one approach aimed
at analysing root causes of poor performance. This research focuses on the
Tanzanian construction industry performance analysis and improvement.
The construction industry in Tanzania is still not well developed nor are there
clearly set development objectives (Ministry of Works 1992). Building
construction is mostly labour based with basic hand tools and equipment.
Labour is characterised by low wages, low productivity and a poor level of
skills. Basic construction estimating and planning data are generally not
2
available. The gap covers the whole range of basic labour cost data, output
data to historical tender rates. Contractors' construction cost and time
estimating at the time of tender is based on personal experience whether
good or poor. In a number of cases, it would seem as though time and cost
estimates at the time of tender are unrealistically optimistic for the purpose of
winning the contract, irrespective of the likely site performance and actual site
management is still very poor resulting in gross time and cost overruns (Lema
and Mutabazi 1988). With respect to the building construction industry in
Tanzania, very little has been done in terms of productivity studies.
Historically, the Tanzanian construction industry suffered major setbacks soon

after independence in 1961. It was recorded that there was no Tanzanian in
the construction industry above the level of a general foreman at the time of
independence (Wells 1986). ' This was followed by an exodus of foreign
companies along with its experts during the next three years following a slump
in the construction activity. Studies performed in 1968 showed that with the
exception of simple warehouses and simple repetitive construction, all other
areas of the construction market were characterised by a serious lack of
competition, high contract prices and excessive profits (Bienefeld 1968).
Contractors of Asian origin dominated the construction industry during this
time. This situation deteriorated even further in 1970 when, following the
passing of the Acquisition of Building Act, 1971, whereby all rented property
above a certain value was nationalised, a substantial number of contractors of
Asian origin left the country. Initiatives were made to improve the local
construction industry capacity. Tanzanian public owned construction
organisations were established as stimulants for performance improvement.
Wells (1986) noted that these failed to perform for a variety of reasons. Such
construction industry development schemes have become part of national
programmes in a number of developing countries. However, most of these
have achieved little because many have concentrated on the symptoms
without considering the root causes resulting in the implementation of broad
based improvement policies (Ofori 1991). The case in Tanzania should
therefore not be seen as an isolated one.
In 1977, a major study was performed aimed at identifying factors contributing

to the problems inherent in Tanzanian construction and recommended steps
to improve the situation (Ministry of Works 1977). The study established that
the industry was characterised by:
3
" low productivity;
" unavailability of required resources, skilled manpower, materials,
equipment;
" unsatisfactory capacity utilisation;
" imbalances of capacities of the different sub-sectors; and
" low competition, high contract prices and high profit rates.
As a result of this study, the National Construction Council (NCC), was

established by an Act of Parliament (No. 20; 1979), to bring order to the
construction industry. One of the major tasks of the NCC was to address a
priority problem of low productivity identified in the study. The study pointed
out that the cost of construction in Tanzania were level with those prevailing in
Europe, while wages were significantly lower.
In a recent study focusing on cost trends of both construction inputs and

outputs, the ratio between the cost of basic inputs and prices of the outputs
was found to vary between three to five times (Lema et al. 1992).
Furthermore, the study established that between 1978 - 1990, construction
output costs were increasing at a higher rate than the basic input costs such
as labour, materials and equipment. This suggested that the delivery process
was in effect becoming more and more inefficient. Actual site management is
still very poor, resulting in gross cost and time over-runs. The ratio of actual
construction period and contract period at the time of tender was of the order
of three to four, and cost over-runs are of the order of 1.5 to 2.5 times (Lema
and Mutabazi 1988). These indicated poor performance at project level.
More recently, the government unveiled a policy document aimed at guiding

the development of the construction industry (Ministry of Works 1992). The
document highlighted the construction development strategy in which the
following objectives amongst others were set:
" to promote the economic and efficient utilisation of local capacities;

" to ensure increased efficiency, productivity and economic design,
construction, manufacture and distribution of construction materials;
" to establish realistic construction costs for the industry and monitor trends;
and
9 promotion of research and application of research findings.
4
This work aims to evaluate construction labour productivity in Tanzania with
the objective of establishing labour productivity benchmarks. The work is
based on earlier initiatives in Tanzania to establish a'basic schedule of rates'
which reflect a high degree of efficiency in the industry and thereby stimulate
performance improvement (Parker, et al. 1986). In this respect, a large
volume of information was gathered between 1985 and 1993 mainly on labour
productivity, labour costs and material costs in the building construction
industry. These data were collected by the researcher, as the principal
consultant to the NCC sponsored project, to establish a schedule of rates in
Tanzania. In 1992, initial discussions were held with the supervisor of this
research on the feasibility of using some of the data gathered. This research
has utilised the data gathered to analyse the characteristics of labour
productivity in Tanzania with the objective of developing a framework for
quantifying productivity gap and identifying labour productivity benchmarks.
Several data gaps were identified in the initial discussions and subsequently
three visits were made to Tanzania in which several surveys and site studies
were made. It was necessary to first justify the notion that labour productivity
is low in developing countries by using the case of Tanzania as a basis for the
perceived need for improvement. The use of labour productivity as a
construction industry performance indicator has been subjected to some
scrutiny with the objective of justifying its significance in the overall
performance.
1.2 Justification and significance
Two issues had to be addressed in justifying this research:
" the significance of the research with respect to other research in this area;
and
" the likely benefits of the outcome of this work.
This research focused mainly on the benchmarking of construction labour

productivity. Numerous construction labour productivity studies have been
performed, although there have been only a few which have addressed the
issue in developing African countries. Olomolaiye and Ogunlana (1989)
briefly studied labour productivity on construction sites in Nigeria where they
observed that construction operative output data was largely absent. In their
study, they concluded that there was a need for establishing output figures on
various construction sites through time study techniques. It was also
5
concluded that method studies and research results should be disseminated
not only to large firms but also to small firms so that the most productive
working methods (or best practices) could be adopted by operatives resulting
in an increase in output without necessarily exerting more physical effort.
This study was very much based on a similar philosophy and relied on work
study methods to gather output data of construction operatives.
Parker et al. (1987) observed that labour productivity data were also not
available from the Tanzanian construction industry and that reliable
productivity data could only be established on the basis of actual site
observations. On the basis of limited data, he concluded that labour utilisation
on construction sites was less than 30 per cent in Tanzania. This conclusion
was based on observations of workers engaged in specific assignments
during a working day. Utilisation levels would be considerably lower if start
delays and early stops were to be included. Labour utilisation levels of
between 20 - 30 per cent have been reported even in Britain (Williams 1991)
although the definitions of utilisation in both cases are not necessarily
concurrent. Available literature suggests that construction productivity in
developing African countries is low (Wells 1986; World Bank 1984; Miles and
Neale, 1992). Some of these suggestions were based on information
available nearly ten years ago and even then, these suggestions were not
well supported with actual observations. It is one of the objectives of this
research to seek evidence of low productivity in the construction industry in
developing countries by using the case of Tanzania.
It is clear from the background given in the previous section that productivity
improvement in the Tanzanian construction industry is important as reflected
in the development strategy. This research directly addresses some of the
objectives of the development of the construction industry in Tanzania. The
research addressed several weaknesses pointed out in previous studies with
the following potential benefits.
" Identify, test and establish a practical methodology for construction

productivity data collection suitable for construction environments in
developing countries.
" Provide reliable labour output data for construction planning, estimating
and control which is a prerequisite for an efficient and competitive industry.
6
" Investigation on scientific basis of the root causes of low productivity which
may lead to specific actions rather than broad based policy actions which
have been proved ineffective in performance improvement.
Establish productivity benchmarks, best practices and productivity gaps

that can provide a motivation to improve competitive performance among
contractors.
Other benefits may include initiation of a research culture in the construction

industry which is normally seen as too rigid to adopt changes.
Although serious construction productivity research in developed countries

has been going on for more than forty years now, literature is still full of
discussions on the meaning of productivity, let alone methodologies for
quantifying productivity. It follows, therefore, that even methodologies for
quantifying construction productivity and justification for using labour
productivity as a construction industry performance indicator are very much
still a subject of further research. This research has, therefore, sought to
reinforce the notion that labour productivity is a good indicator of construction
productivity and that efforts concentrated on enhancing site labour productivity
are likely to result in a significant improvement of the construction industry
performance. In this respect, the study has established a methodology for
benchmarking labour productivity, that is, establishing both labour productivity
metrics and proposing methodologies for both identifying and adopting best
practices.
The concept of benchmarking, which has its roots in the Japanese

manufacturing industry, is fairly new and research into its applicability in the
construction industry is very topical. Other modern performance improvement
philosophies such as Total Quality Management (TQM) are seen as the main
framework upon which to initiate and sustain performance improvement
efforts.
1.3 Significance of the building sector
The focus of this research was on the building construction industry due to its
significance both it terms of its share of the construction volume and its
contribution as a seed corn for the growth of the local construction industry
capacity. Official statistics indicate that building construction consumes about
7
70 per cent of the construction investment in developing countries (World
Bank 1984). This is in spite of the unrecorded share of the building related
activities in the informal sector. Also, large civil engineering projects are
normally executed by foreign contractors who also undertake large building
projects. Small and medium size projects are mainly executed by local
contractors who gradually acquire the expertise to compete with foreign
contractors. The building construction industry provides excellent
performance improvement opportunities for local contractors and therefore
forms the initial basis for research aimed at enhancing local contractors
performance. Further, most civil engineering projects executed by foreign
contractors in developing countries are of a one-off nature. During the
execution of these projects, substantial technological and management
expertise is passed onto local personnel. However, this is of little value
unless there is continuity of work providing opportunity of utilising this newly
acquired expertise. In contrast, there is better continuity in building projects,
thereby providing a better chance of utilisation and consolidation of expertise
gained in a building project.
1.4 Objectives of the research
The main objective of the research was to: establish benchmarks based on
labour productivity studies; identify factors that are percieved as important by
construction operatives in improving their productivity; and explore the effects
of these factors on actual construction process. Labour productivity formed
the focal area of the study because that is where production essentially takes
place. Identification of factors that influence construction productivity was
based on an opinion survey of construction operatives as they are the main
construction process owners at site level. Their opinion was considered of
importance in improving productivity. A labour productivity model based on
actual site observations was then developed in an attempt to quantify the
effects of practices or factors identified. In order to explore a diversity of
construction environments, building construction projects formed the main
focus of the study in view of the significance of the building construction
sector in local capacity and efficiency improvement in developing countries.
Productivity characteristics of eight building activities were investigated, with
the objective of quantifying the potential for labour productivity improvement.
Concreting was then selected for more in depth analysis because of its
significance in building construction and its specific characteristics which
8
makes it more interesting to study. The sub-objective of the research were
therefore to:
" develop a framework in line with modern performance improvement

philosophies which can be applied in the construction industry in
developing countries. In this respect linkages between TQM,
benchmarking and productivity have been examined;
" justify the use of labour productivity as a measure of construction site

performance and therefore industry performance;
" confirm that the apparent need for improved construction industry
performance in Tanzania by identifying practical indicators and evaluating
performance trends over a reasonable period; I
" develop a simple methodology for labour productivity data collection on

construction sites suitable for conditions in developing countries;
" examine the productivity data variability with the objective of identifying
productivity improvement potential;
" identify factors that may have some influence site operatives' productivity;
" explore the effects of selected factors on productivity., Concreting will be

used as a mode towards exploring these factors as they affect productivity
such as:
" pour size and type;

" gang size;
" level and quality of supervision;
" level of mechanisation
" materials layout and quality; and
" payment and incentives.
" create a basis for labour output database development in Tanzania.
The achievement of these objectives would result in identifying some actions

that may contribute to the productivity improvement in the construction
industry.
9
1.5 Research methodology
To meet the requirements of the objectives set above, the following approach
was devised.
The initial stage of this research involved a literature review to confirm

the research objectives. The need to develop a performance
improvement and evaluation framework in the construction industry
was identified. Modern management philosophies such as TQM were
explored, leading to the formulation of a viable research framework
within these concepts. In particular, benchmarking was identified as a
viable tool for embarking on a continuous improvement programme in
the construction industry. Labour productivity was identified as a key
performance indicator. A labour productivity benchmarking framework
was developed.
Macro-construction industry performance indicators were identified and

used to evaluate the performance trends of the Tanzanian construction
industry. Performance comparisons were made with selected
construction industries.
Extensive labour productivity data had been gathered under a separate

project. In this research, these data were analysed to obtain output
levels, statistical characteristics of labour productivity such as variation
and distribution patterns. The results were compared with
observations in studies elsewhere. The data were used to test the
benchmarking framework developed which led to the quantification of
the labour productivity potential. Opinion data were used to validate
the potential identified through data analysis.
A high degree of variability was observed in the above data. This

required some explanation. General factors that influence labour
output on construction sites were identified on the basis of the literature
survey. Key construction process owners were consulted on the
validity of these factors, and then interviewed on the basis of a
structured questionnaires to establish their perceived importance of the
predetermined productivity factors. Emphasis was placed on the
operatives' views because they are the main process owners. The
10
results of the interviews have been tested for consistence, validated
and some comparisons made with contractors' views.
" Further site productivity studies were performed to analyse the extent
of the influence of the factors identified. An exploration of relationships
between productivity and various factors was performed. A multiple
regression model that related productivity and various productivity
factors was proposed.
Major outputs for this work include: 4
"a list of current mean productivity, CMP, and target mean productivity,
TMP, for selected activities; and
9
" some proposed actions that can be taken by various participants in the
industry in order to realise the productivity potential identified.
1.6 Guide to the thesis
Figure 1.1 presents the general thesis framework. A detailed description of

the framework is presented in Figure 1.2.
11
OBJECTIVES
CONSTRUCTIONPERFORMANCE
IMPROVEMENT
PRODUCTIVITY II BENCHMARKING
PRODUCTIVITY
BENCHMARKING
FRAMEWORK
DATA AND INFORMATIONREQUIREMENTSAND COLLECTION
Publisheddata
Interviews and questionnaires
Site productivity studies
Executive interviews
ANALYSIS OF DATA
VALIDATIONS
DISCUSSIONSOF THE RESULTS
CONCLUSIONSAND RECOMMENDATIONS
Figure 1.1: Thesis logical framework
12
d
N C
cN .
CO) O
CO) >
0U
(6 a)
0+ N .
"U (1) V3 O C U fn U
Cc C -o dL 2
U W O
n d
'rnd c' c - 3C0
a) E Ca) n ca
c E d
3 E N cu C 0 N N V1 N > d fn
O c U L c6
Co V)
d is 2 dU " T a v d o
3 cN
O
0 E CL C
.
ca - yE my cn YV o
+" 0 d C p "(Z Q t E (n N
-y U) - 6
M Co () ` -O co 0xa u) pC
. CO E
VN N C 'C O 2
CU "0 NLWNC 1O
O E
Co h I c W ci flhll 0.
Co
toia
t)
1 a
c
Y
c0
E U) ccmon
o

o
E n
co
rn o2
v

p
È
O
M
r_
0
0
m
N U v
c c
E
L
E
O
"0 D
(D
0
O
..

.r-
t
o
O
F
-5
LL)
C
O
a cm
(1)
u)

0C
v)
OO
nCc2.10 (1) E0
0 u)
"-
'rn s3
.C
C a)
OC.C
NO"y
L
c
U) -N
(0
t ' 0r
a>
N _.
U
CV
c
mo
-
c
c
E
3 .-
OC
"V
c w

y
fl. cu ai ca '
cc c
nrny
caaoZ . U, v .- 3-
a oZ pia c c a) c
ri y :3 :e
' cSfQ UE
v d
C E QU >2
O >g. a
to a> 0 O : o C (D ooLoc
L oC 2 I- O cn CI) Z a m nCi. o; ._ L
V s nE a) no n O
1 8 8 1
'v N O
U)
W
E
Co O
(/) CN7
(lS
Z
>
4) cc
. OEU
T n d O
C ") EE 2e c N
C) tu
C
U "NOOCO N
d C Uf j ro C C C 'O
tLd
4) is G)
Co c c (0"c (D E y E - c
C c C 2 E CF- rr
O (0 (U caCU)WW
CC ro U -' O Ctl
2 > C
U d WCO 'C
.2
r- CO) .A a) Co c)
0 0
N Q
.a o to c: = oan rr

0 M- c c G)
w j-- 0 zi Co c o
(0 0 r- C
v v -2
Co `' Z' -i Z'Z' (CSa)
*'k c c v 'r' mE
s c o
c ai 3
ci
'n jc >
Ot v `c in im N
9 c( yC x'vaoa .b d
Uo c ; fn ; NyL>C
O
"
L
Ny 0

Z' E E 1
(V V
N
T5
U
N
c N
n o "Z(
l V t`

0)
M a) L
+% . 0
o 0 2 v, 2 y v U (D
c C ,
C n a)
0 L a o m v, C
U) w y>
LLNOO ci c: U.
U I ' F- U 'O U "O U 4) C
LN
L 2O
(a d) OU- C
N NNO O
` U L T
v
O o "_
ai
U) C d rs iA p
. cd
L N
` 0 ) ro L to O-
0 +
C r_ (D ,_ :3
(Zi c w
0 N
'0 .0 > U)
= T vv(Dvi
O v rn O
(a 0 ? (DCo E >
0 - 0
r_ 9) (D o
C _ o t
'- ` (1)
N fC N v m E
1.. . 0.0 O O
O O O -iz 7
0) E (D
V . " N
O E
a(0 c L12) N
j
CD
(Up N .
O
O 7 UE fd
CO Ui ir CD CI) N ' N "p t
LLCU
cu (U
1
Q0C Off'^dV+'">dN LL E030Y

*- dC0 Co
13

(D Co
(' coo d
O Co 15 L i 1
(U 0 0 e)
Co .-
r_
y+
CJ (Z
Qi (D tu -0
E >
j 0 2 " N NC W N
O 7 c
+ -"
V v c Ed l) o
x w "a :g
; cV C
c)
in oc
O -
-
cu v) c i .
c Cl) v c
U) U) C
'
N N Cl)CO :3
U
0 E D _
Co O
7 - 7
C) c2" N n CL N N 06
'O o Z "Z-
N N O N
N N 7 N
a0 3
n 2 clt-a
u, ca ai
a5 92
Q ) O O ) O c
n m
CO NN
co
(D
( .C
f0 N cn
c
p
NT
0 fA .N 7
~c :3y a
d
' n.
00 o
Y O
C rn a t :EZ7 v C
0 h- C Y o L> a--
` 3 3 a 3 00 N n 5 a C
O1 v,
oa> (D m O O
. roc
O N- cY
E aim O
E 3 O > O N m "d O
E 0
No
am 13 (U m O NJ : (D Q E
O. L
. cri
C d E E -
0 o CL 17 c N
(0 -
41 C T
C > to m
o
rn> ou E 0)
i c m 0
m cm m d Y voi'3n.. 0u - L L
o 43) cmro c
v
o cai Cl) 0 3r> = vE n
t`Cr v O ic i e . '
L o E .a
00
CO) . t do =3o- (0 0 n LL
CE ro
G-)G)
a a y F- a a>
13)0 (D
E c v 9) CL
N N
c ? :3 Q` N`
- 3w O N
.. m c 3 E '15 CLapt020 m >
O E O 2
O C U "O LeNNN d
y.r
N m
C.
Co v
(D 5 o ca
' o
oC a
oC U) LL GC
uioc E on
V 4 F- -
.
(D (0 C C
C X ? 'a
C C- C
O ; '
Co ,,, O N >, 7 N
C p) ->,
nCC Y "C f/1 5
.NON
r- o (1)
(0 im 0
0. to 7 O
. u
' N NNV '
"'
N a o13 "0ta 3 c
Mv ai i 3
ui
0-0 0 A'LE .
o o Co i
ac 0S"
- N t
o o c ova>N mo Q
U a Xy 073 y M. G o
. c '0 c
as ..c
)=c c 00
et co d'OONO coc- > ai To
Co "O
.C f0 x W E -> >V
L
:3 0
c 2 ` = 2 gr 2c cm " N <C
CL N .` cU fflUU
' 2 1
-
V0Cr.. 0C *+ 0 O="- Wu -> d0 V- a. Co Ed30h. Y
14
c
OY C
C A
0 0 N (0 > NC
, C
N C.
pU 7
CO) c2
O
Co oZ
CNOy o 0 rn
C
C" y
L c Q! y
(U m ca
y
0 a' j
2-1r- c
Q (0 ycN.2 0
c
o 'p 'O
=6 vp
c c cC
C > - cN.. 0 0 c
c
0 7 y
vN' y o8o LC?
oc
aoia OC
C 3c
y N c (0 :3 :3 - maco 5 E rn o
0 > -5 0 _cu
c3) m a? Z
22 C c v y
f; M cn 0 - 0 :3
to
p 7p y (U cj .- jp
0
yO
am
p . Cd (6 y to
O ( " 'p T
UV QyO . O O C
UUL (13 O C Q O
- ' OO 'O (lf '
O Co dOy EO
o c c ' c a) o n '>
(1) in ai c) co c V
G IC
Nc aiy Q ayiv
. L 7> V7 N
m O- V 2 c
M tu E
V >? >8y > E
m y a
c .
uwo "y Uw CC Q a.. EpC
o
pN C3
>IM
YN
0
0' o-A
t ,O
U!
a 10
072- c 0) C)
Ò 'CM CM
C
1-- U
N :C
V
C
;r E Cu
0)
-0 0)
OY
v NN
v 5 E> C d C
3 0 -
CC Q
VV0
N
:3 L
0
p 5
> M
10
O tA F-
C "O
Cu
OU
LC
U,
12
.
0)
o , E Nd
JL N
0
y
V
03 c
O t0 0 T
T O (U Z, C
N
C
U O 7 m C)
dN :=
.
CO
C E c -
U) M CC 0
i..
(U Co E Y (1) (U :3 ' C C tu C)
N
Co Nmm O
d U cm
7 Mn "
O
QW > 0 C C -
--
C
co LL
a
'c C 0 `n 7
c N 'v' -
ci)
co m c CL 0
Cu
O COC
C
a j : it
"- - to (U U !AC>
C rYCN c'Ji
L M Cl) (0 = U
V J (0 m> NNN
fn U O
I-(13 .vO
0E

U
oc .. C
UC
e, " N

ECN to.o c
O y
,n
dO 0
O
U
O ~Ca
j5 O
E 0 in 0 U)
Co
9 Cl) G) CCUE
r, w to C- p 0
j VC
E 0 C
d0CU
0 f0 0) n
"t%1 U
C 0 C tu
y
E 2L. * T7 E E
> O O
U
:3 3m' C C
v CL c v r V
O
uni c U
- tu
y>
N
N Z. om (0 N u)
.3 C Om O
.o CC
2
p
m c v N
i-
Vv ~ c 0 -0 0
13 c (U y d)
O Om ti! Q)
U 0
to U 2 NOO Ca O_ E

N -in . ON
: " (0 n v CO CLIP
(U u5t) m A
3 a
0
O d)
E cu
ZCL0 c: O U O. 'p
la)
1-- C
c< C U0 Oj
U0 0 CO
CZE77CN c Kl-
d > O'D'O o w
0 m
UOC.. 4) !Ld3Oi
.y .C
15
C)
N
T
C)
L
LL
O Ca0
. (T~ r-
cc D
V Ul. CO
_ fA
(D a) N
E O to +. Q. ro
E co U)
CdN C
N
OL. U) Oo = ca N
Cl) L
Nd yOO O
L. N O =Cyp
V N
E O
O
CL
(L)
' 3 - al s
i, 3 c
y
po
c. -F= LL.
U o E'3 (1)
v, O
O aci
E
vs 3
C
N CO Co fn <.f
O 2 E C
1-
a) C V
OwNtO
U 'y NC MD)
-,
2
0 T E5 as ,P - u)
m OC
. D cc m
o
U0C*. O'^ Ed3OiY
F7 - 41 C r. + 0 d V+'"-> dN LL i 19
16
1.7 Summary of conclusions and recommendations for further
research
The performance trends of the Tanzanian construction industry, analysed

over a period of 25 years between 1969 and 1993 led to the conclusion that
performance was poor on the basis of the following:
" construction industry output from late 1980s was about 40 per cent of the
output levels in 1970s;
" its contribution to the Gross Fixed Capital Formation (GFCF) declined from
about 60 per cent in 1969 to less than 30 per cent in 1993;
" productivity had dropped to about 72 per cent of 1969 level in 1993.
" its contribution to employment declined from about 15 per cent to less than
5 per cent of the total employment; and
9 performance comparisons with selected developed countries indicated
contrasting trends.
These observations formed a firm basis for research on the strategies for
reversing these trends. Productivity was identified as the key performance
indicator. This research argued that productivity improvement efforts should
focus at operative level because they are the key process owners. An
extensive study of productivity at operative level had significant findings.
" Analysis of variability of labour productivity of eight building activities on 46

sites belonging to 23 contractors indicated a coefficient of variation of
between 45 - 100 per cent in comparison to similar studies in USA and UK
where the coefficient of variation was between 5- 35 per cent (Noor 1992).
This wide variation indicated a significant improvement potential.
" Statistical modelling of the data established that the Johnson SB

distribution model was found to represent the productivity distribution fairly
well. The model represents a distribution which is skewed to the left.
" The statistical distribution model was used to quantify the productivity
improvement potential, computed from the difference between current
mean productivity, CMP, and the target mean productivity, TMP. Using
the model, a potential for improvement of about 133 per cent was
established.
17
" This potential was confirmed through an opinion survey of construction
operatives. A similar survey for contractors' senior management indicated
that contractors were generally unaware of the productivity improvement
potential.
The low productivity levels were linked to factors related to motivation on

construction sites. These factors were identified on the basis of opinion
survey amongst operatives. These were then evaluated on actual
construction operations along with operation specific factors. These formed
the basis of concrete mixing productivity regression modelling. Critical factors
in the regression model were: pour size; gang size; and productive time.
Limited variability of labour productivity factors across the sites studied had an
influence on the critical regression model results.
The study has proposed some actions that could lead to labour productivity
improvement in the the construction industry. A number of areas of further
research.have also been identified. These include:
" widening the scope and applications of performance indicators in the

construction industry in line with the Total Quality Management (TQM)
philosophy;
" adoption of a wider definition of productivity in order to capture the

influence of key inputs, specifically the influence of mechanisation;
" relate between performance at various levels in the construction industry,

that is operative, project, corporate and industry level;
" identification of modes of implementing benchmarking study results at

national level; and
" detailed studies of productivity factors where influence of variability of such

factors on productivity can be quantified with focus on motivation factors in
relation to existing theory.
It is hoped that further research into these issues will provide more specific
knowledge for productivity improvement in the construction industry.
18
CHAPTER TWO
CONSTRUCTION INDUSTRY PERFORMANCE

IMPROVEMENT
19
CHAPTER TWO
CONSTRUCTION INDUSTRY PERFORMANCE IMPROVEMENT
2.1 Introduction
The need for performance improvement in the construction industry world-

wide is well accepted. This research was motivated by the need to improve
construction industry performance in Tanzania. The main objectives of this
chapter are to examine a framework for: construction industry performance
measurement in general; and challenges to its improvement.
Methodologies for evaluating performance in the construction industry are
examined. Also discussed are the associated difficulties, along with some
philosophical discussions regarding the definitions, characteristics and
scope of the industry in generic terms and how these can be applied to
Tanzania. Specific characteristics of construction industries in developing
countries that may have some bearing on this work are reviewed.
Tanzanian construction industry structure and characteristics are examined.
The significance of the construction industry in the national economy is
discussed in some detail as a justification for applying national efforts to its
improvement. The need for performance improvement in African
developing countries, and specifically Tanzania is justified within the
traditional framework. Some indicators of performance at project level are
discussed.
2.2 Construction industry - definition and characteristics
2.2.1 The dilemma - defining what construction industry is
A literature search for a precise definition of construction industry is likely to

be frustrating. Literature lists what the industry is concerned with. There is
no clear boundary of the industry's involvement. Indeed, there are some
discussions not only on what the construction industry is, but also whether
there is such thing as the 'construction industry' (Groak 1994; Hillebrandt
1985; Drewer 1980 and Bon 1990)). Ofori (1990) stated that there is, as
yet, no acceptable definition of the construction industry. Some consider it
as involving only site activities, others include planning and design
20
functions, and yet others extend it to cover the manufacturing and supply of
materials and components, finance of projects or management of existing
construction items (Turin 1975; Hillebrandt 1985). While some argue for
viewing the 'industry' as 'sector' that comprises of a number of related but
disparate 'industries' (Drewer 1980; Bon 1990), Groak (1994) argued that
construction activities should not be treated as belonging to an industry with
definable boundaries, specific skills and using specific resources, but that
the focus should be more on the end products and services recognising
increasingly external linkages. In this sense, he argued that, terms such as
'construction capacity', as used by Hillebrandt (1975), have little meaning.
The industry boundary dilemma was further highlighted by Edmonds and
Miles (1984). There is also no universal agreement on whether
construction is a service or a production industry. Channon (1978) and
Hillebrandt (1985) have referred to the construction industry as a service
industry. However, most consider construction as a production activity.
While such critical philosophical discussions are part of the academic

process, workable construction industry definitions and boundaries have
still to be sought if solutions to existing problems are to be found. In this
respect, Ofori (1994) argued that the construction industry should be
defined by considering the objective of improving its efficiency, recognising
practical administrative difficulties and drawing boundaries. It should
embrace planning, design, construction, maintenance and demolition (Ofori
1990).
2.2.2 The search for a working definition
The Standard Industrial Classification 1968 of United Kingdom (Fleming

1980) for construction lists the following product areas.
" Erecting and repairing buildings of all types. Construction and repairing
roads bridges: erecting steel and reinforced concrete structures; other
civil engineering work such as laying sewers, gas or water mains, and
electric cables, erecting overhead lines and line supports and aerial
masts, extracting coal from open cast workings etc. The building and
civil engineering establishments of government departments, local
authorities and new town corporations and commissions are included.
On-site industrialised building is also included.
21
" Establishments specialising in demolition work or in sections of
construction work such as asphalting, electrical wiring, flooring, glazing,
installing heating and ventilation apparatus, painting, plastering,
plumbing, roofing. The hiring of contractors' plant and scaffolding is
included.
While the above list is extensive, it still excludes some important types of
construction products (such as construction and maintenance of railways,
airports, ports and harbours) that take up a substantial share of civil
engineering works. The construction industry includes professional
services and to some extent suppliers of inputs into the industry. However,
the boundary of the construction industry, with respect to inputs in the
construction process, is fairly vague. Some may argue that a stone quarry
is part of the construction industry as well as other input manufacturing
industries such as brick manufacturing factory, cement manufacturing plant,
steel rolling mill etc. It is doubtful whether the manufacture of building
inputs, such as electrical and plumbing items, would qualify as part of the
construction industry yet these are manufactured to the specifications and
requirements of the industry. These are sometimes referred to as
associated industries. It is also unclear with regard to the classifications of
other inputs such as energy, manpower and some equipment.
Furthermore, the classification based on practice in UK may not necessarily
apply elsewhere.
The national construction industry as defined by the United Nations for

statistical purposes is not based along the product lines but rather in terms
of who performs the work and who pays for it (United Nations, 1985, pp. 7).
A national construction industry can be described as follows.
Construction industry proper: contract construction by general

builders, civil engineers, and special trade contractors.
Contract construction performed for others by establishments or

organisations classified to industries other than construction.
Own-account construction performed by independent units of

enterprises or other organisations not classified to the construction
industry proper.
22
Own account construction performed by establishment or other
organisations not classified to the construction industry with no
independent construction unit.
Own account construction performed by individuals.
This classification does not state what should constitute construction

activities. Further, the terminology used is not very clear, leading to a
possibility of confusion.
The National Construction Industry Development Strategy document

published by the Ministry of Works (1992, pp. 1) in Tanzania defined the
construction industry as follows:
'The sum of all economic activities related to civil and building works;
their conception, planning, execution and maintenance. Such works
normally comprise capital investment in form of roads, railways, airports
and harbours, dams, irrigation schemes, health centres and hospitals,
educational institutions, offices, warehouses, factories and residential
premises. '
This is a very wide definition, especially with respect to economic activities

related to construction which vary extensively and some of which are
remote from actual construction activities. Indeed, there could be specific
reasons for adopting a wide definition, although this is not stated in the
document. It has to be pointed out that although the definition does not
explicitly include the informal sector, practice in Tanzania and indeed the
policy itself does not address this sector.
A recent report on the evaluation of the construction industry in Tanzania

(Bjorklof 1992) argued that there was a need to extend the definition of the
construction industry, in countries such as Tanzania, to include not only the
technical and economic resources and their organisation, but also
institutions connected to building; research institutes, professional
associations of different kinds, as well as parts of the educational system.
The main reason for this view is mainly as a result of the close
interdependency between the various parts of the industry for the health
performance of the whole. Indeed, this view has been echoed by other
researchers such as Edmonds and Miles (1984), Wells (1986) and
23
Groak (1994), although not expressed in such a direct form. The dilemma
is the complex relationships which are inevitable with an adoption of such a
wide view. As pointed out by Bon (1990), this form of dilemma is likely to
persist as research in construction industry is still fairly new.
In spite of the difficulty in having a clear definition of the construction

industry, it is reasonable to define the industry with respect to the scope as
synthesised from various literature sources. This may be referred to as a
working definition. First, literature is fairly clear on what constitutes the
products of the construction industry, in spite of the philosophical
arguments and discussions mentioned earlier. It is also clear as to what
constitutes a construction industry process and what constitutes direct
inputs in these processes. It is, however, unclear as to what constitutes
inputs that may be considered part of the industry. There are inputs which
cannot be considered part of the construction industry, such as energy,
finance and transport (distribution systems). These become part of the
industry once they are made use of in the construction process, and indeed
any other innovative adoption of existing technology as pointed out by
Groak (1994). Some inputs consumed by the industry are planned,
manufactured and distributed outside the industry, although these are
produced to the requirements and specifications of the industry. An attempt
has, therefore, been made to diagramatically represent the industry in form
of PREPARATION OF INPUTS, INPUTS, PROCESSES and OUTPUTS in
Figure 2.1.
24
onstruction industry conceptual boundary
Inputs Processes Outputs

Prep ation Buildings
Design
Roads
Manufacture
Manpower Assembly
Dams
Process
Materials Erection
40 - p op- Airports
Train
Equipment Construct
Harbours
Distribute
Expertise Maintain
Water supply
Rehabilitate sytems
Repair I Sewerage
systems etc.
CONTROL INSTITUTIONS AND MECHANISMS
Figure 2.1: Conceptual construction industry scope
The preparation of inputs falls partly within the construction industry. This
could be for practical or political reasons. For practical reasons, the above
stone quarrying example can be considered as part of the construction
industry, while the manufacture of electrical fittings is part of the
manufacturing industry. This is illustrated by the situation in Tanzania
where practically every large contractor owns a stone quarry and in most
cases, specific stone quarries are opened up just for one-off large projects.
It is difficult to separate between the quarrying operations and actual
construction within the contractors' organisation. The same is true for
inputs such as building blocks where again practically all contractors
manufacture on site almost all blocks required for the project from basic
materials. Other inputs, such as electrical fittings and plumbing
components, are bought in from suppliers. In such situations, policy
25
emphasis on the development of the construction industry, could lead to the
manufacture of the bulk of building materials being considered part of the
construction industry:"This is particularly so for-resources balancing within
the construction industry in relatively poor economies. This is exemplified
by Wells (1994), where a decision was made to manufacture and
incorporate some locally produced materials into a project in order to
reduce the import content of the project. Such action could possibly explain
the adoption of a very wide definition of the construction industry by the
Ministry of Works (1992). Further, the national construction industry
boundaries can be interpreted differently for different purposes. The
National Construction Council of Tanzania (NCC), for example, restricts its
definition of the Tanzanian construction industry as comprising local
indigenous consultants and contractors in order to facilitate demarcation of
the boundaries within the industry, within which development efforts should
be directed towards building local capacity (Mawenya 1995). The NCC
otherwise views the participation of others as transitional and not
constituting a permanent reliable long-term local capacity. It is also
appropriate to consider the construction control institution and mechanisms
as being partly within the construction industry. This appreciation is
appropriately represented in Figure 2.1.
2.2.3 Characteristics of the construction industry
The construction industry is one of the few industries whose products tend
to increase in value over time, unlike the majority of other industries whose
products begin to depreciate immediately from the time of purchase
(Harvey and Ashworth 1993). Other characteristics that separate the
construction industry from other industries have been extracted from
various authors and listed below (Harvey and Ashworth 1993; Edmonds
and Miles 1984; Wells 1986; Hillebrandt 1988 and World Bank 1984).
These characteristics are common to both developing and developed
countries, and include:
" the physical nature of the product;

" the wide variation in terms of materials used in the production and the
standards in terms of space, quality, durability and aesthetic
considerations; -
" the product is normally manufactured and left fixed on the client's
premises, i.e. the construction site;
26
9 many of its projects are one-off designs and lack any prototype model
being available;
" the separation of the responsibility for design from that of construction;
" the location of production area changes constantly and is subject to
interference from weather;
" the organisation of the construction process which comprises of a large
number of diverse specialised trades often casually employed;
" the method used for price determination;
" the final product is often large, expensive and slow to produce and in
some cases required over a large geographical area such as road
schemes, railways, etc.;
" the susceptibility of the industry with respect to the general level of
economic activity;
" most of the products of the construction industry cannot be produced in
advance of demand but rather have to be sold before they are
produced, or made to order;
" the construction industry is often used as an economic regulator
because it can be technologically flexible, it can be used to create
employment where necessary; and
the industry is considered a low prestige industry, because it is mainly
an intermediate service industry and secondly because of the relatively
crude and basic resources and processes.
The characteristics of the construction industry listed above are to some

extent common to most countries. Edmonds and Miles (1984) state that the
construction industry with these characteristics has been reproduced the
world over. This they attribute to historical influences of the British which
has formed the basis of regulatory and organisational frameworks
especially in developing countries.
Among the economic sectors, the construction industry is the least studied
and understood despite its importance in the economy even in developed
countries, and the situation is not likely to change soon (Bon 1990). In
some studies, this has been attributed to the economists' lack of interest in
construction (Construction Industry Development Board (CIDB) 1988), that
may be a result of construction industry's image as a traditional and
technologically backward industry (Ranbird and Syben 1991). Existing
knowledge is, therefore, not of a high quality, hence, the basis for further
research and conceptualisation is weak. This is more so with construction
27
industries in developing countries where there is a dearth of information
upon which to base research, business and development decisions (Wells
1986; Parker et al. 1987; Ogunlana and Olomolaiye 1989 and Lema 1995).
Besides this aspect, construction industries in developing countries have
other special characteristics that are examined below.
2.2.4 Special characteristics of construction industries in

developing countries
Construction industry characteristics listed above are common to both

developing and developed countries. However, construction industries in
developing countries have other specific characteristics. Firstly, unlike in
developed countries, the structure of the industry is different. There is a
bulk of the construction volume undertaken outside the formal sector (World
Bank 1984; Edmond and Miles 1984 and Wells 1987). This work is mostly
related to the provision and maintenance of basic shelter. It is estimated
that in Tanzania, for example, as much as 95 per cent of all rural homes
and 85 per cent of urban houses are built through the informal sector (Nuru
1990). Major projects are undertaken by foreign companies as local
companies lack the management expertise to handle such big projects. In
some countries, the proportion of construction work undertaken by foreign
firms is substantial. The World Bank advocated that this proportion can be
as much as 95 per cent of all contracted work (e. g. 95 per cent of all
contracted work in Benin in 1976 was done by foreign companies).
However, in spite of this, most of the work is still done by either local firms,
firms within the region or the neighbourhood for most countries (Wells
1995). In most developing countries, local companies are left to execute
small to medium size building projects. Most construction relates to new
work, unlike in developed countries in which a substantially higher share is
repair and maintenance.
Secondly, construction activity tends to be more labour intensive in

developing countries mostly because labour is cheap. Whereas unskilled
labour is readily available, skilled labour is generally scarce. Edmond and
Miles (1984) argued that construction needs to be even more labour
intensive in consonant with the performance of the economy of these
countries. The evidence supporting this argument is weak in that it is based
on the observation that the ratio between investment per work place in
developing countries to their GNP is higher than a similar ratio for
28
developed countries. In other words, there is overcapitalisation in the
construction industry in these countries. Further, this argument is based on
fairly crude data.
Thirdly, governments and other public agencies are the major clients of
construction projects and have direct or indirect control of the construction
demand, most of which is on new work. This can account for as much as 80
per cent of the demand in the formal construction sector (World Bank 1984).
In Tanzania, for example, a considerable portion of the construction
demand came from public bodies, accounting for over 70 per cent of the
work in the formal sector, excluding direct labour public works, or'force
account'. A more recent report estimated this to be about 75 per cent
between 1974 and 1982 (Bjorklof 1992). The World Bank made similar
estimates for Burma, Nepal,-Papua New Guinea, and several other African
countries. Wells (1986) estimated that the public sector share of the
investment in the monetary economy in Tanzania was around 96 per cent
in the early eighties. It could be argued that governments are in a better
position to influence the industry more than is possible in developed
countries where the public share of construction tends to decrease. In
United States, for example, the demand generated by public bodies in
1982 accounted for 50 per cent of the total demand excluding the
construction of single family homes.
The structure of the construction industry in developing countries is slightly

different from that of developed countries. The construction process in
developing countries is undertaken by about four distinct groups in
proportions dissimilar to those in developed countries (World Bank 1984).
These include:
Small builders - these are principally concerned with building and

improving single family shelter. Informal, self-help building work is
the predominant construction activity in developing countries,
catering for essential shelter and related requirements. The informal
construction activities rely heavily on self-employed and family
labour. They are small enough to escape most legal regulation and
elude statistical enumeration and therefore there is dearth of
information on these. The United Nations' definition of construction
industry for statistical purposes, for example, excludes the informal
sector (United Nations 1984). As a consequence, this part of
29
construction activity tends to receive less attention from governments
than that dedicated to formal construction projects.
" Communal and self help organisations - these comprise all members
of the community who can contribute either their skills or labour for
the construction of communal projects. The projects undertaken are
normally labour intensive. In remote areas of Sub-Saharan Africa, for
example, the most economic way of building a small school and
sanitary facilities is by employing local self-help builders or by
mobilising the local population (World Bank 1984). A good example
is in Malawi where small projects such as wells, schools, health
clinics, and houses for teachers and health workers are constructed
through self-help with central government assistance channelled
through local development committees. One self-help scheme to
supply piped water has operated since 1968 and by 1982 3,000
kilometres of pipeline and 4,160 village taps had been installed
serving about 640,000 people (World Bank 1984).
0 State owned construction organisations are established for various

purposes. In countries where there is a good supply of competitive
contractors, the public sector usually.takes charge of construction
repair, maintenance, and emergency work which are generally
unattractive-to contractors. In other cases, governments may start up
a domestic construction industry by creating a public organisation to
act as a seedbed for developing construction skills or supporting
incipient domestic companies. Wells (1986), Miles and Neale (1992)
and Ofori (1994) observed that these efforts were not necessarily
successful.
" The private sector of the industry covers a wide spectra of

enterprises and degrees of contractual responsibility. The simplest
form is the interface with those who work for the informal sector.
They carry limited contractual responsibilities and risks within the
goods or services supplied without time, cost or quality obligations
regarding the final product. Intermediate levels include small
subcontractors and specialist subcontractors (for example plumbers
and electricians, painters etc. ). The last group is that of contractors
who take full responsibility for winning the tender competitively by
assuming all the risks implied and enters into a contract and is
30
responsible for price, time limit, volume, and quality of work. The last
group can be sub-divided further into two groups:
" local contractors - those who are registered locally and are not
less than 51 per cent locally owned; and
" foreign contractors who are locally registered but less than 51
per cent locally owned and those not locally registered but
who come into the country for a one-off project.
Most large projects are undertaken by foreign companies who either come
in for one-off projects or maintain a skeleton presence during periods of
construction recession. Fluctuations in construction activity hinders a
development of a stable construction capacity which can only be ensured
through long-term competitiveness amongst local contractors. This aspect
will be discussed further, in Section 2.3, when examining the Tanzanian
construction industry. This research focused on the formal sector of the
construction industry, not only because of its recognised significance to the
economy, but also because it is can be easily identified. The structure and
characteristics of this sector are discussed in the next section.
2.3 Tanzanian construction industry structure and

characteristics - the formal sector
2.3.1 Formal sector contracting capacity
The formal construction sector plays a significant role in the economy. The
formal construction sector in Tanzania is not well developed due to
historical constraints leading to the low level of performance discussed in
Chapter 1. The formal sector has mainly been involved in public projects
because investment in the private sector was mostly in the subsistence, or
non-monetary sector of the economy due to the state-controlled economic
policy in Tanzania between early 1970s to mid 1980s. This was partly
responsible for the low level of development in the industry, because there
was very little inflow of private capital into the country that would stimulate
capacity expansion and competitiveness. Entrepreneurial initiatives were
also discouraged during this period. Both contracting and consulting
capacity has remained low. Between 1980 and 1990, the number of
registered contractors increased from 479 to 1078 as shown in Figure 2.2.
31
1100
1000
O
U
900
O
U
800
700
0
60C
50C
1980 1985 1990 1995
Figure 2.2: Registered building contractors in Tanzania 1980 -1993
The number decreased slightly between 1990 and 1993 because a number
of contractors, especially in the lower classes, were struck from the list
because they were not active. In spite of the apparent large increase in the
number of contractors registered between 1980 and 1990, this increase
was mainly concentrated in the lower classes of contractors. The
classifications of contractors as of end of 1993 are presented in Table 2.1 in
which a total of 1021 contractors are indicated. Seventeen out of forty eight
registered in Class I, were foreign contractors. Some of these were
undertaking large one-off projects in the country while some had only a
skeleton presence in Tanzania. Of the locally owned contractors in this
class, only two were owned by African Tanzanians, the rest were owned by
Tanzanians of Asian origin. For sustained development of competitive
contracting capacity, the government emphasised the improvement of
performance of African contractors in Tanzania through training and
preferential treatment in awarding contracts (Ministry of Works 1992). This
was based on the experience of the early sixties when most competent
contractors of Asian origin left the country as a result of a slump in the
construction.
At the other end of the classification scale, there was a large number of
Class VII contractors (632 or about 60 per cent of the total). A large number
of these played little role in the commercial sector of the construction
32
industry. Most were active in the non-financial sector of the construction
industry and a number were only involved in construction only as
secondary or even tertiary business activity. Even when they were active,
their maximum contract limit was only about 25 million Tanzanian Shillings
in 1993 (or about 30,000), a sum enough for a medium sized residential
house only. However, a number of these were managed by qualified
engineers whose major constraints were capital.
Table 2.1: Number of registered building contractors - December 1993
Class Number registered emitof project cost in

1000000 Tshs
48 unlimited
11 20 600
III 53 400
IV 85 200
V 105 100
VII 78 50
VIII 632 25
IL- Total 1021

Source: National Board of Architects Quantity Surveyors and Building Contractors (1994)
The development of contractors in Tanzania is faced with a number of

obstacles such as lack of education and training, fluctuating work load,
delays in payment, lack of equitable contract documents, problems of
bonding, weak financing, restrictions on import of equipment and the lack of
a fair competivive environment.
2.3.2 Professional design services
Local design capacity is still very limited and major building and civil works
are still designed by foreign consultants outside the country (Bjorklof et al.
1992). This does not help in the local capacity development, coupled with
the fact that projects are based on design standards from different
countries. It is not unusual for the same stretch of road to be based on
different design standards, or to have building drawings labelled entirely in
Chinese or Japanese. The capacity of local consultants remains small and
recent efforts to develop joint ventures with foreign consultants have not
been successful. Recent policy proposals have addressed this (Ministry of
33
Works 1992). The different types of construction consulting firms in
Tanzania as of 1990, are shown in Table 2.2.
Table 2.2: Registered consultancy firms in Tanzania as of 1990
Type Domestic Foreign

Architecture only or architectural and planning 20 5
Engineering only 27 12
Architecture and engineering 12 9
Engineering,architectureand quantitysurveying 4 2
Quantity surveying firms 6 4
Total number of firms 69 32
source: tjorKIot et at. (i vuz).
2.3.3 Costs and cost trends
Construction inputs comprise mainly of labour, materials and equipment.

Different types of projects have differing ratios of these inputs. In general,
building construction tends to be more labour intensive while civil
engineering projects are more equipment intensive. In Tanzania, medium
sized building projects are almost labour exclusive. The cost of labour is
relatively low, and materials comprise the major share of cost. There have
been few studies to estimate the actual cost of labour cost in a typical
building project in Tanzania. Some studies have indicated that the labour
costs are between 12 - 15 per cent of the project direct cost (Mbwambo
1988; Kishimbo et al. 1992). A survey performed as part of this research,
based on contractors estimates, put this figure at between 25 - 30 per cent.
Project overhead costs are estimated at between 20 - 30 per cent of project
costs. Again contractors estimated this to be between 25 - 30 per cent. It
could be argued that if labour costs are low, there would be no real
incentive to improve productivity. This argument is weak because by
improving productivity, project duration is reduced which has a direct effect
on overhead costs. The second most important reason for productivity
improvement is the reduction of inflationary effects on the project costs.
This factor becomes very significant where inflation rates are high. A recent
study established that building materials costs increased at an average of
between 25 - 66 per cent per annum for the period 1978 - 1990 ( Maro et al.
1992). Construction equipment hire rates increased at an average of 47
per cent and labour costs increased at an average of 12 per cent per
34
annum. The motivation to improve labour productivity is therefore not
necessarily to reduce labour costs but to reduce project duration. Operative
remuneration could therefore be significantly increased if this can result in
reduced project duration, without necessarily increasing total project costs.
2.3.4 Other characteristics
There have been few detailed studies of manpower requirements in the

construction industry. This was partly caused by the fluctuations in the
construction workload in the country. The demand for the informal sector,
which may also be significant, is difficult to estimate. Previous manpower
demand estimates were mainly based on the public sector requirements. In
general, labour is abundant, but there is scarcity of skilled labour,
technicians and engineers. According to the Ministry of Education in
Tanzania, the demand and supply of qualified personnel at the disposal of
the construction industry were as shown in Table 2.3. However, these
figures include all types of engineers, technicians and artisans. The figures
fail to allow for those who find employment outside the construction
industry. In reality therefore, trained manpower available to the
construction industry is far less than claimed by the Ministry of Education.
Table 2.3: Supply and demand of construction technical personnel
Category Supply Demand Output per year
1988 1988 2000

Engineers 4507 7253 11344 369
Technicians 16380 23131 36176 845
Artisans/ Craftsmen 103060 122327 191315 11165
Source: bJorKior,ei ai. 1UUZ.Pp.b5
Other problems that affect the performance of the construction industry

include: lack of equipment and finance; corruption; low level of
competitiveness and ineffective regulatory set-up. This research does not
attempt to address all these problems. By focusing on productivity and the
potential for its improvement, it is hoped that this will stimulate competitive
performance improvement which may then lead to addressing the other
institutional problems. Some recent changes in the country in the last ten
35
year can only help to stimulate competitive performance improvement.
These are briefly examined in the following section.
2.3.5 Recent developments -a diagnosis
The last ten years have seen some changes in Tanzania both as a result of
external as well as factors within the industry. These can be categorised
under various headings.
Political-economic changes:
" Political orientation has shifted from mainly socialist to market economy
leading to direct changes in the economic policy from a generally state
controlled economy to a market economy. This has had direct
consequences on the local construction industry which was mainly
dependent on public funded projects. Private sector funded building
projects are gradualy becoming more predominant although major civil
engineering projects are still public funded.
" The lifting of restrictions on importations in general has enabled wide

availability of basic construction materials and tools in the private market
at competitive prices in contrast to the situation prevailing during the
previous ten years when importation of construction inputs was severely
affected by lack of foreign currency. This is likely to have some impact
on the industry's performance.
Local construction capacity
" Severe materials constraints from mid 1970s to mid 1980s resulted in
the deterioration of both quality and productivity. Some local contractors
diversified into other economic ventures such as production of
construction inputs, agriculture, transportation etc. Others left the
country altogether. While the late 1980s and early 1990s have
generally seen an increase in the construction activity especially in the
private sector, this has come about when the formal construction
industry is very weak. This led to two forms of reactions.
36
a) Local private investors have tended to mistrust formal contracting
procedures and bypassed the system by hiring individuals instead of
companies for the design and supervision of projects.
b) Donor funded projects have generally insisted on foreign designers

and contractors. A number of the foreign contractors who came into
Tanzania tended to stay and execute projects traditionally done by
the local contractors. This has gradually led to overpenetration.
Manpower
" Skilled manpower has always been a major setback in the Tanzanian
construction industry as discussed in the previous section. Again, the
slump in the construction activity between mid 1970s and 1980s did little
to alleviate this problem. Variations in crossborder construction activity
has had both negative and positive effects on the problem. A notable
change in the construction manpower is the availability of qualified
engineers from the University of Dar es Salaam over the last eighteen
years. The last ten years have seen a gradual increase of small
construction firms managed by qualified engineers. This trend is likely
to remain and with deliberate assistance, these firms have the scope for
growth. This has slowly initiated an awareness for competitiveness and
the scope of improvement both in quality and productivity.
Institutional support
" Several institutions exist from which construction industry support can
be sought. The main of which is the National Construction Council
(NCC) established in 1979 to organise the construction industry. The
framework for the development of the industry has been set. The
industry is yet to take full advantage of what it has to offer.
"A number of seminars and short courses have been held by various
professional bodies and educational institutions and these have had
little impact on performance improvement. Generally, the feeling is that
the need for such support has not been well appreciated where the
competitiveness through quality and productivity are not yet
performance drivers. A top executive in the industry pointed out that the
main bottleneck for competition driven performance, especially in public
37
projects was corruption. The situation is changing as private sector
construction volume gradualy increase.
These factors make this research very timely, as contractors, some of whom
are qualified engineers face the challenges of performance improvement
through competition.
2.4 Economic significance of the construction industry
The construction industry is an essential contributor to the social-economic

development process. The industry has as its own customers, if not for new
work then for repair and maintenance either through the formal or the
informal sector. This includes virtually every industry and public or private
sector organisation, and many households. The contribution of the industry
to the national economy is measured in several ways. Three common
indicators include: its contribution to the Gross Domestic Product (GDP); the
Gross Fixed Capital Formation (GFCF); and the amount of manpower
employed by the industry (Wells 1986, Bjorklof 1992, Hillebrandt 1984).
The United Nations (1988) defined GDP and GFCF as follows:
" The contribution of the construction industry to the GDP is referred to as

value added by the industry which is determined by the gross product of
the industry at producers' prices less the value of their intermediate
consumption at purchasers' prices.
The GFCF is the gross expenditure on fixed assets construction

products such as buildings, roads, railways, etc. It also includes outlays
on reclamation and improvement of land and development of extension
of timber tracts, mines, plantations, orchards, vineyards etc. Machinery
and equipment is also included.
Other indicators of its importance include the consumption of public

investment and value added by the construction industry. The extent of
public investment indicates the extent of control the government has over
the industry's work load. It has already been mentioned that there is a
tendency for higher public investment for less developed countries. The
value added is an important indicator of the increased value of materials
and services bought into the industry in the production process. The
industry typically consumes 50 70 per cent of public investment,
-
38
contributes about 5- 10 per cent of the Gross National Product (GNP), and
provides employment to a comparable proportion of the labour force
(Edmond and Miles 1984). In both developed and developing countries,
construction usually accounts for over 50 per cent of the fixed capital
formation (Edmonds and Miles 1984; Wells 1995). In spite of its
importance, the assessment of total value of the construction output in any
economy is not easy.
The construction industry's contribution to the GDP is the difference

between the value of the construction industry products at market prices
and the market value of all inputs, therefore it excludes value of purchased
building materials and components, fuel, transport, professional services,
insurance and legal fees (Wells 1986). This measure is therefore affected
by the profit levels in the construction industry. The assessment of value
added in developing countries is particularly difficult as obtaining reliable
data is a major bottleneck. In a single accounting period, inflationary effects
(which are normally high in a number of developing countries) of these
records have significant effects on the accuracy of the information. The
problem is compounded by lack of both the necessary records and
motivation for the contractors (especially medium to small sized contractors)
to provide the required information in developing countries.
While capital formation in construction is a measure of the gross output of

the construction sector, and therefore does include the value of goods and
services supplied to the construction from other sectors, it unfortunately
excludes those not considered to accrue to the formation of new capital,
mainly repair and maintenance works that can amount to as much as one
third of the total output of the sector. Further, in most developing countries,
the problem is compounded by the fact that a large percentage of the
construction activity is not executed within the formal sector of the
construction industry. It is, therefore, frequently omitted altogether from
national statistics and even when it is included, the methods of assessment
are, of necessity, crude. For example, the United Nations Yearbook of
National Accounts Statistics neglects the contribution of the informal sector
(World Bank 1984). Other major problems that affect comparisons at
international level are:
39
" the fact that a large proportion of construction activities take place
outside the formal sector and the likely variations from one country to
another;
" the fluctuation of the value of the national currencies and especially the
over-valuation of the local currencies in some developing countries; and
" the fluctuations of the construction industry's work load and the casual
nature of employment which necessitates a large labour turnover within
any one organisation.
Considerable deviations from the standardised systems remain in the data

of many countries, despite the efforts of the international organisations to
bring consistency in the definition and presentation of the national statistics.
Thus, all types of data relating to the construction industry, especially in less
developed countries should be interpreted with caution (Edmonds and
Miles 1984; Wells 1986). This problem has been examined in Chapter 7,
with specific reference to Tanzanian construction industry data.
The above mentioned setbacks not withstanding, construction industry data

for a large number of countries are available from the United Nations
Yearbook of National Accounts Statistics, Construction Industry Statistics
Yearbook also published by the United Nations, and data on employment
in construction can be found in the International Labour Offices Yearbook of
Labour Statistics. There is, however, discontinuity of data for some
developing countries. For example, Tanzanian construction industry
statistical data are very sketchy in Construction Industry Statistics Yearbook
for 1985 (the last and most recent of the publication in the series) and
construction employment data for Tanzania for the last ten years are
missing altogether from the relevant ILO publication for the year 1994. In
any case, the data available should be analysed with caution.
2.5 Economic performance indicators and relationships
2.5.1 Performance indicators
Various attempts have been made to correlate construction industry and

economic growth. Such relationships were investigated in the late 1960s
by Turin (1969) who compiled data on construction output, population and
GDP from 87 countries at different levels of development. Since then,
similar analysis using the same methodology for different sets of data have
40
essentially produced the same general results (Edmonds and Miles 1984;
Wells 1986). The methodology involves correlation of economic growth
and the various indicators of construction industry economic performance
(i.e. contribution to the GDP, GFCF and employment). The results of
investigations by Edmonds and Miles (1984) are discussed here as a
representative study. They investigated the relationship between GNP and
value added in construction, value added per capita and employment
based on world wide statistics for 1974 (97 countries) and 1979 (116
countries). Although the data were somewhat outdated, there is no strong
reason to doubt the validity of the results because of their consistency over
three separate studies using different sets of data. Edmonds and Miles
(1984) established that the most striking thing was the data variability and
their range. The following relationships were established:
The percentage value added in construction is generally higher in

countries with a high GNP per capita than in countries with low GNP
per capita.
The relationship between value added per capita (VAC) and GNP
per capita (G) is very strong, 90 per cent of the variation being
explained by the changes in GNP per capita, thus
Log VAC = 1.13 log G-1.66 (R2 = 0.901) for 116 countries in 1979.
The relationship between GNP per capita (G) and investment per
capita, I, is also very strong
Log I= 1.12 log G-1.17 (R2 = 0.896) for 59 countries.
The relationship between GNP per capita and employment (E) is

also fairly strong and GNP per capita explains nearly 60 per cent of
the variation in the employment per 1000 population.
E= 14.84 log G- 29.55 (R2 = 0.574).
Similar relationships were investigated by Wells (1986) using a different set

of data and her observations were indicative of trends more or less similar
to those observed by Edmonds and Miles (1984). Wells (1986) further
established that time series data for particular countries which have been
41
through a period of rapid economic growth show that construction output
grows faster than the economy as a whole during the early stages of
growth. These trends were observed in the industrialising economies of
Malaysia and Thailand. Typical indicators of the economic significance of
the construction industry that show general conformance with the above
relationships are given in Table 2.4 below (Wells 1986). Average figures
for Tanzania for the period between 1969 - 1993 have been added for
comparison.
Table 2.4: Average contributionto GDP, GFCF and employment

in countries grouped according to income
Group I Group II Group III Group IV Tanzania

Variable under US $ 35 US $ 350-700 US $ 700-2000 over US $ 2000 US $ 249
(per capita) mean 1969 - 93
Value added in
construction as a 3.6 (23) 5.2 (21) 5.4 (31) 7.3 (27) 4.0
percentage of
GDP
Capital formation
in construction as 8.9 (13) 10.6 (10) 13.6 (27) 13.5 (23) 9.0
percentage of
GDP
Construction
employment as a 3.1 (9) 3.4 (14) 6.6 (22) 8.1 (26) 8.9
percentage of the
total
Construction as
percentage of 56 (13) 53 (10) 55.4 (26) 57.5 (23) 44
GFCF
Source: Wells (1986) - These refer to statistics of the period between 1977-1980.
Tanzanian data obtained from Bureau of Statistics, Dar es Salaam, (1995).
In general, as indicated in the relationships by Edmonds and Miles (1984),

the percentage construction value-added in GDP was found to increase
with increasing per capita GDP. The figures in Table 2.4 show that
construction's contribution to GDP was only about 3.6 per cent for 23 low
income countries compared to an average of 7.3 per cent for high income
countries for the period 1977-1980. There was also an increasing trend of
percentage contribution as GDP per capita increases. Average value
added by construction industry in Tanzania, over a period between 1969
and 1993 of 4.0 per cent is very close to the low income group average of
3.6 per cent. Similarly, capital formation in construction, as percentage of
42
GDP, follows similar trends with the high income countries with a higher
percentage. Once again, the average percentage contribution of the
construction to capital formation in Tanzania for the period 1969 - 1993 was
very close to the low income group average. While percentage
employment in construction is generally expected to increase as the
economy grows, the average figure for Tanzania are above the general
norms for countries in her income group. This indicator may already
suggest lower than average productivity in Tanzania. Similarly, average
contribution of construction to the GFCF was lower than the average for its
group. The indicators discussed above present a very general picture from
which specific conclusions cannot be drawn. A detailed examination of the
Tanzanian construction industry performance trends over a 25-year period
(1969 -1993) is presented in Chapter 7.
2.5.2 Productivity and economic growth
It is important that the different contributing sectors of the economy be

productive to enable the national economy to grow. The measurement of
productivity in any sector is both a quantitative and a technical problem.
The factor productivity concept has been used to give the contribution made
by each or all of the factors of the production. The factor is reflected in the
ratio between output and input in the production process. The concept can
be used to evaluate the variation of productivity over time (e.g. productivity
trends). The methodology, although widely and frequently used, does
represent only partial productivity in the sense that not all factors are
considered. This may give a misleading trend. For example, construction
productivity is usually measured in terms of labour productivity. Capital
investment over time, which may have productivity improvement
consequences in the industry, is not reflected in the labour productivity
measure. This increase comes about through substitution of capital for
labour and the reduction in numbers of workers employed. If account was
taken of the increasing capital input in the construction process, such that
the measure of productivity embraced a combined labour and capital input
in the denominator, the observed increase might not be so marked. The
concept of total factor productivity (TFP) is used to relate the various factors
of production. The concept of total factor productivity is defined as the ratio
of output to a weighted combination of inputs (Kumari 1993). The method is
used to evaluated the relative contributions of labour and capital towards
productivity improvement. One example of the methodologies of measuring
43
TFP for construction will be considered here. A detailed discussion of
productivity concept and construction productivity has been presented in
Chapter 4. The methodology is based on the following general
relationship:
Output index
Total Factor Productivity (TFP) .
Factor input index
One such methodology is the Kendrick Index denoted TFPK which defines
total factor productivity as (Kumari, 1993):
V'
TFPKr =
WoLr+ roKr
Where Vt = indices of real value added;

Lt = indices of labour;
Kt = indices of real gross fixed capital;
Wo = share of labour in the value added in base year; and
ro = share of capital in value added in base year.
The methodology requires both labour and capital input in the construction
industry to be quantified over a duration of time and their relative shares of
contribution to the value-added be identified. Initially, there is the problem
of defining the construction industry -a difficulty which has been discussed
earlier in this chapter. The input side of the construction industry processes
differs when developed and developing countries are compared. There is
a significant portion of semi-processed construction inputs in developed
countries compared with developing countries where most materials are in
raw form. In principle, the value added concept is supposed to cater for this
difference appropriately. Furthermore, construction processes in
developed economies are much more capital intensive. It could be argued
that the approach would be more suitable where capital plays a more
significant role in enhancing productivity. In any case, the measurement of
capital investment in construction is difficult and some methods are very
subjective, for example, Edmonds and Miles (1984). A simplified
methodology which considers labour input has often been used. It has to
be mentioned that a strict analysis of trends in labour productivity or
international comparisons should not ignore capital investment trends or
differences between different countries in the industry. However,
comparisons of productivity amongst countries of similar income can ignore
44
capital input based on the assumption that capital investment levels are the
same and therefore similar productivity effects apply.
Productivity can be measured in absolute terms as value added in the

economy by the construction industry, as well as labour input can be
estimated. It is also possible to explore the relationship between GDP per
capita and productivity using construction statistics for different countries.
Two values of productivity can be computed: value added per person
engaged (VAPE) in construction; and output per person engaged (OPE).
VAPE is computed by dividing the annual contribution of the construction to
the GDP by the number of persons engaged in construction in the year.
Likewise, OPE is determined by dividing total value of the output by the
number of persons engaged. However, the concept of industry output is not
as simple as it first appears. Briscoe (1988) summarised the difficulties of
measuring construction industry output as follows.
9 Physical work done must first be valued at an appropriate set of market

prices to determine recorded output.
" While private sector firms include their profit margin in their valuation,
public sector organisations commonly do not, and their valuation is often
at cost.
" It is also quite likely that some part of the work done by construction
firms is not fully recorded, as some small firms and self employed
workers do not make accurate statistical returns if any.
" Output is usually measured for the calendar year, but this does not
necessarily correspond with the accounting year for many construction
firms, so there is a further element of error in estimating the timing of the
output.
" When output valued in current prices has to be converted into constant
prices, difficulties may arise over the appropriate price indices to apply.
Construction labour input is gauged by the number of persons engaged in

the construction industry in a particular year. This figure is usually an
annual average estimate for the whole construction sector and takes no
account of whether the person was employed over the whole year or just
part of it. Indeed, when the number of workers is used to measure labour
input, no recognition is given to the number of hours actually worked. Over
time, the average number of hours worked each week varies, and when
man-hours are used as the productivity denominator, different results of
45
productivity are achieved than when number of men alone is used (Briscoe,
1988). Other setbacks include:
" the measure of the number of persons engaged fails to distinguish

between the various qualities of labour input; and
" the casual nature of engagement in the industry leads to under reporting
the actual number of persons engaged. This problem will be discussed
in some detail with respect to output data in Tanzania.
OPE does not differentiate between the differences in the state of inputs in
the industry (i. e. semi-processed or raw), whereas VAPE only considers
the value added by the construction industry. These productivity measures
are utilised in Chapter 7 for construction industry productivity analysis in
Tanzania.
2.5.3 Project level performance evaluation
Construction industry performance, measured in terms of labour productivity,

is only normally used for comparative purposes. A more relevant
measurement of construction performance has traditionally been project
based focusing on efficiency of execution with respect to cost and duration.
In most cases, the quality of the final product comes into the equation,
although the quantification of this aspect presents some difficulties. Besides
the more obvious traditional performance indicators - quality, cost and time
performance, - there is no agreement as to what the universal indicators are.
Sink (1986) has stated that performance consists of seven dimensions:
effectiveness, efficiency, quality, productivity, quality of work life, innovation
and profitability. Each of these dimensions must be of interest to the
construction organisation. The extent to which these measures were used
for project performance was not stated. Freeman and Beale (1992)
identified seven criteria of evaluating project success as shown in Table 2.5,
in which cost, duration and quality play a significant role in measuring
project performance.
46
Table 2.5: Project success factors (Freeman and Beale 1989, pp. 10)
Success Criterion Description Frequency

of Mention*
1. Technical performance To what extent the technical requirementsspecified at 93%
the commencement of the execution phase were
achieved.
2. Efficiency of the The degree to which target of time and cost were met. 93%
project execution
3. Managerialand A measure of client, parent, and user satisfaction 43%

organisational incorperating the degree to which the project was
implications performed without disturbing corporate culture or values.
4. Personal growth The satisfactionof the project team, particularly in terms 29%
of interest, challenge, and professionaldevelopment.
5. Project termination The completenessof the termination,the absence of 14%

post-projectproblems, and quality of post-audit
analysis.
6. Technical The success in identifyingtechnical problems during the 14%
innovativeness project and solving them.
7. Manufacturability The ease with which the product resultingfrom the project 43%
and technical can be manufactured,and its commercialsuccess.
performance
* Percentage of mention in the review of 14 papers.
Another study evaluating project performance identified efficiency indicators

(i.e. cost and time, as the main focal factors). Softer indicators related to
clients' satisfaction were identified by Bresnen (1990) as shown in Table 2.6.
Table 2.6: Construction project soft indicators
Main indicator Sub-indicators

Project objectives " Overall cost
" Time taken
" Value for money
" Type of contract
Contract performance " Contract performance

" Commitment and involvement
" Communications and co-ordination
Quality of work " Suitability for user

" Functional specification
" Overall quality
Client organisation " Communication and co-ordination

" Involvement and teamwork
Others Disruption caused

" Professional team performance
47
The above serve to illustrate the diversity of traditional measures of
construction performance, most of which focus on historical data and
information. Modern performance measures focus on continuous process
improvement which relies on on-going measurement and feedback. The
construction industry is faced with the challenge of adopting modern
performance improvement philosophies. Some of the limitations and
challenges facing the construction industry are discussed in Section 2.6.
2.6 Challenges in the construction industry performance

improvement efforts
Construction involves the interaction of equipment, personnel, environment,

materials and methods. Traditional construction project performance
measures have been examined in Section 2.5. The performance of a
construction organisation is a function of the individual's performance. This
generalisation pre-supposes that the performance vectors of individuals are
concurrent with organisational performance vectors. In other words, that
what constitutes an improvement in individual performance has a
maximised effect in the organisation's performance. Indeed, since the
emergence of the division of labour, the underlying assumption has been
the alignment of objectives in an organisation to maximise the performance
of the whole. While this is generally easily understood, its practical
implementation has been somewhat difficult particularly in the construction
industry. Disparities between project objectives and the permanent
objectives of the participating organisations play a major role in this.
Accordingly, each interacting task organisation has an interest on any given
issue, that is to say, an interest in seeing that the issue is resolved in such a
way that the outcome is most advantageous to its organisational objectives
(Mohsini and Davidson 1992). The situation is compounded when
individuals in the interacting organisation have objectives different from that
of the organisation. This has been attributed to the fragmentation in the
industry (Burati et al. 1992; Sanders and Eskridge 1993 and Latham 1994).
Sanders and Eskridge (1993) established that successful companies have

achieved an alignment of interests between primary decision makers,
workers at all levels, and the established mission statement and objectives
and that when people have their individual goals, corporate successes will
be short-term, even though people may be hardworking and successful in
48
their own particular area. This approach is captured in the Total Quality
Management (TQM) philosophy whose objective is to ensure maximum
effectiveness and efficiency (performance) within an organisation to
achieve customer satisfaction and continuous improvement. This view
widens the scope for construction performance that has , for a long time,
been limited to productivity and profitability. Maloney (1989) argued that
the understanding of the multidimensional nature of performance in the
construction industry is necessary for effective improvement. This is now
supported by research findings that show that greater adherence to the
TOM principles, was amongst the prominent trends on quality and
productivity improvement in the construction industry and has led to TQM
philosophy being described as the prescription for survival in the 1990's
(Construction Specifier 1992). In Chapter 3, a detailed examination of TQM
and its implication in construction industry performance improvement efforts
has been made.
2.7 Summary
This chapter has examined the definition, characteristics and boundaries of

the construction industry. This has led to the conclusion that there are still
differences as to the definition, scope, and roles of the construction industry.
The need to adopt a wide definition of the construction industry in a
developing economy such as Tanzania has been emphasised. The
structure and characteristics of the Tanzania construction industry have been
examined to appreciate the scope for improvement and some of the
constraints. Of significance are recent changes in Tanzania which may help
in fostering competitive performance improvement. The economic
significance of the construction industry and macro-economic performance
indicators have been discussed with emphasis on macro-productivity.
General relationships between these indicators and economic performance
have been explored. Diversity of traditional measures of construction
performance, most of which focus on historical data and information, have
been examined with a view to identifying their limitations. Challenges facing
the construction industry in adopting modern performance improvement
philosophies, and in particular TQM have been examined. A detailed
examination of modern business performance improvement philosophies
and their applications within in the construction industry have been
examined in Chapter 3. The relevance of these philosophies in this
research has also been discussed in Chapter 3.
49
CHAPTER THREE
MODERN APPROACH TO BUSINESS PERFORMANCE

IMPROVEMENT
50
CHAPTER THREE
MODERN APPROACH TO BUSINESS PERFORMANCE IMPROVEMENT
3.1 Introduction
This chapter examines the concept of business performance measurement

and improvement and some of the limitations of the traditional approach. This
concept is discussed within the context of the emerging modern management
philosophies such as Total Quality Management (TQM) and benchmarking.
The role of modern management philosophies and the general framework of
its application in the construction industry is also discussed. ' A detailed
account of benchmarking is presented which focus on: its meaning; historical
background; its scope of application and procedures; the potential for
application in construction; and recent applications. A framework for its
application in this research has also been presented.
3.2 Business performance measurement
3.2.1 Purpose of measurement
People have been counting and recording results since time immemorial. It is
a natural inclination to want to know'how many? ' and 'how much? '. William
the Conqueror elevated such questions to become matters of national
importance (Norman and Stoker 1991). Once the answers are recorded, it is
a straight forward step to compare one set of numbers with another. These
aspects are encountered in everyday life when we make comparisons such
as 'which brand is cheaper? ', 'which route is longer? ', 'who scored more
goals? ', and so on. These matters are not given much thought at this level.
At national, industrial and company levels, a great deal of thought and effort
go into: counting, measuring and comparing in order to detect changes from
one period to another. When such changes are central to the survival of an
economy or an organisations in a competitive environment, they play a key
role in influencing strategic, policy and operational decisions. The decision of
what to measure, what form of comparison to adopt, how the results have to
51
be interpreted at this level is not a straight forward issue either. A wrong
choice of indicators, incorrect measurement, or comparison may lead to
wrong decisions being made. This is very much the case for the construction
industry and organisations within the industry.
Performance measurement in a business organisation presumes that there is

a standard against which comparison can be made, whether internal or
external. The process involves the decision on what to measure, the actual
data collection process, data processing, data evaluation and data
interpretation. In spite of its apparent clarity and importance, performance
measurement is not without dilemma. In Zairi (1994 pp. 3) Smith stated that :
'The reality of measuring unconstrained human behaviour is

that the act of measuring a particular indicator will induce
behaviours which have as their objective to maximise the
performance of the indicator, virtually regardless of its effect on
the organisation as a whole. '
Zairi (1994) clarified the views presented above by arguing that at the heart of
the problem of performance measurement is the human component. People
involved in this complex task have to:
" set the objectives in the first place;

" design the system;
" help implement the system,
" conduct measurement exercise;
" bear the responsibility to interpret the measures obtained; and
" act on improving the measures.
The above serve to illustrate some of the problems and bias that can be
inherent in any business performance measurement system. Solving the
mystery of performance measurement does not become a question of how to
do it, but how to make it succeed, because at end of the exercise it must
achieve desired objectives in the organisation for it to succeed. The design
and implementation of any performance measurement system has to
acknowledge these observations and views so as to minimise subjectivity in
the results.
52
3.2.2 Role of financial indicators
Financial measures and indicators have traditionally been the cornerstone of

the performance measurement systems (Lothian 1987). These are widely
accepted for measuring corporate performance. They are well tried, easily
understood and produce apparently clear signals. Costs and revenues,
income and expenditure, profit and loss are universally familiar concepts to
even a non-financially untrained manager. These measures have played an
important role in shaping corporate management decisions and are likely to
continue doing so for some time to come. The overriding financial measure of
performance is profit (Norman and Stoker 1991). However, the calculation of
profit is hardly ever straightforward since it depends on sets of accounting
conventions. In addition, the measurement of profit perse gives no indication
of the potential for improvement within an organisation even in profit terms.
Further, the attainment of profit in turn depends on a number of other financial
and operational outcomes, and management has to encourage and monitor
performance across a wide front. Organisational management is a whole
process. Limiting performance measures to financial indicators (and
specifically to profit alone) is outdated because the indicators are themselves
derived from cost accounting information which is often based on outdated
and arbitrary principles. Emphasis on financial indicators and associated
derivatives had a considerable influence on management thought for the last
100 years. The evidence for this is clear when the history of management
thought is examined starting with the role of Frederick W. Taylor (1856 -
1915) and Frank B. Gilbreth (1868 - 1924) who applied their mind to method,
machines, and materials involved in time study in order to increase outputs in
comparison to inputs (productivity). Early work and time study initiatives led
to conflicts between unions, representing the direct interests of labour and
management, who were seen by the unions as champions of the
shareholders who received the greater benefits derived from more profitable
operations (Hellard 1993). Such conflicts are not uncommon even today
where profit centred performance improvement is seen to benefit a few in an
organisation.
3.2.3 Limitations of financial performance measures
Financial indicators are generally outcomes of what happened in a business

process. They cannot fully explain what happened in the process in order to
53
form a basis for future actions. They are not capable of explaining what is
happening because they are final outcomes of the business process, nor are
they in a position of forecasting what is likely to happen. These traditional
measures have not been linked to the process where the value-adding
activities take place. A continuous feedback to the process for improvement
has been lacking, thus providing little motivation to support attempts to
introduce continuous improvement programmes because of their inability to
map process performance. In an organisation that has to maintain
competitiveness, it has become evident that performance must begin to be
seen by the customer (Oakland 1993). This has necessitated a shift of
emphasis from financial figures to a broader perspective because business
competition is now on the basis of product quality, delivery, reliability, after-
sales services and customer satisfaction (Zairi 1994). None of the above
variables are measured by the traditional financial measures, despite the fact
that they represent the major goals of the world-wide companies. These
principles, which are well rooted in the manufacturing sector, have found little
application as yet in the construction industry. The fundamental difference
between manufacturing and construction has played a role in this. Many
construction projects are unique either in terms of design or site location.
Each project presents challenges and new learning processes. Even where
the project is repeated, its sheer duration generally means that the project
team, both management and workforce will have undergone considerable
turnover. This hinders increased efficiency that repetition and familiarity
ought to produce as in manufacturing. Fragmentation of the construction
industry has been cited as the main reason for the failure to implement
continuous improvement principles. However, there is wide agreement that
the industry needs to implement these principles if it is to remain competitive
(Burati et al. 1992).
3.3 Total Quality Management (TQM) and associated philosophies
3.3.1 Emergence of Total Quality Management (TQM)
Organisations, for many years, have considered quality to be an

independent support function and have often pursued quality and
productivity as two independent issues. This is probably attributed to the
practice of limiting corporate performance to financial indicators based on
the widely held view that in order to achieve quality, higher costs have to be
incurred which may eat up part of the profits. The quality revolution, brought
54
about by the Japanese in
management philosophy 1970's and 1980's, has
led to the integration of quality and productivity leading to new management
concepts. In order to meet the challenges of the new global competitive
environment, companies had to consider quality as an integral part of their
strategic business plans through a new management philosophy - Total
Quality Management (TQM). Over the past decade, TQM has been one of
the hottest issues in North America management circles and has gained a
firm foothold in Western Europe too (Macdonald 1993a).
The recently issued BS7850 Total Quality Management (1992, pp. 2) defines
TQM as:
A technique that assures maximum effectiveness and efficiency

within an organisation by putting into place processes and
systems which ensure that every aspect of its activities are aligned
to satisfy customer needs and all other objectives without waste of
effort and using the full potential of every person in the
organisation. '
The philosophy recognises that customer satisfaction, health, safety,

environmental considerations and business objectives are mutually
dependent. These objectives can only be achieved through management
involvement at all levels, continuous improvement of products, services and
processes, education and training of employees and participation of all
employees in problem solving (Chandra 1993; Macdonald 1993a). TQM is
often termed as a journey, rather than a destination. This is because the
continuous improvement process goes on for the life of the enterprise. The
improvement process never ends, therefore, no destination is ever reached
(Construction Industry Institute Annual Conference 1992). The critical
elements of TQM initiatives have been recognised by a number of authors
(Pender 1993; Chandra 1993 and Kearney 1992) as:
" customer focus;

" employee involvement;
" continuous improvement; and
" innovative leadership.
Figure 3.1 presents the TQM framework in which the above factors are
related to overall objectives of TQM.
55
Total Quality Management
- Customer focus - Employee involvement
- Continuous improvement - Innovative leadership
Enhanced quality of
Products Services Decisions Operations
Higher customer Improved productivity Better work

satisfaction Higher profitability environment
Figure 3.1: A TOM framework (Chandra1993, pp. 20)
TQM has been described as the third industrial revolution, that has emerged
from a rapid development in the third quarter of the twentieth century
(Hellard 1993). It has two main streams contributing to its development: that
of scientific development and that of quality. The progression of the two
streams over the last 100 years has been well summarised by Hellard
(1993), as shown in Figure 3.2. The merging of the two streams in the
1980's and the emergence of TQM in the 1990's, have created a whole new
philosophy and has brought back the quality concept from the product into
the process. The European Quality Award (EQA) self-evaluation model
represents the current management thinking brought about by the merger
(Davies 1993). The model illustrates: a shift away from the traditional
performance indicators such as business results to a more integrated
approach with priority on customer satisfaction; a shift in management style
from control to leadership; increased employee empowerment through
teamwork and decision making; and an emphasis on process and statistical
process control.
56
N
C
r- O CL
0 CL
7N2
Q(C
C
0
R Vi
rn
rn
F' E L
Q>
2

O1
C a)
O
CD-
C) C)
2 V
N
O
N
O
E
0) ON V
0 .2 O OD
I et aj O CY)
0) N
Ca C
0 l 0
0
-mai
C_
r+
C
d
"
E
o N
0)
LO C
m
m
l4
rn
Cy E
EN
0o
Y
OU O
U
O O'
C Co
O , !C
.y .r
0
N U
G) im
0
i N
a, r+
O O)
U
C
a
C
C) d
E C
C) .
O
0
TC
CC CO
cp .C
rn la CL
CU I 0
C H
eo m 0
E t
'4-
0 CL
N V
I. -3 C
CU 0
G) C) ia c
:3o 0) m
O QU T t)
0 w.
c)
0 CL
0 d
Co
to
Q
m
U
v
0
2 d
wO
c
0
L C
C
N 0
a_ iO
0 W
0
L 0
dvii F) v
d
O C
rn
2
cu
N
J M
d
luawa6euevi - UnD a,
U-
57
The quality being subjected to these techniques defies concise definition.
Oswald and Burati (1992) argued that it includes:
" satisfied customers;

" reduction of variation and re-work in the work process;
" reduction of production-service cycle times;
" the absence of disputes;
" greater alignment of individuals and corporate objectives; and
" repeat business, including long-term business relationships.
Indeed, all aspects of process dimensions are included in this quality

concept. Approaches such as Quality Function Deployment (QFD) were
being used to initiate a quality product development from the customers'
point of view, and tailor the design and manufacturing process towards
addressing the customers' needs (Karlof and Ostblom 1993).
The adopted concept of the customer in the TQM context is wider than the
end user of the product. There are internal customers who receive, process
and supply semi-finished products in the process chain in addition to the
external customers who receive the final products. Indeed, each party in a
process chain has three roles: supplier, processor and customer, whether
they be internal or external. Burati et al. (1992) refers to this as the triple
role concept (after Juran 1988). Each party plays a role of receiving
products processes these products, and supplies products of higher value to
the next customer in the chain, whether it be physical products or
information. The triple role concept is illustrated in Figure 3.3.
Processor
Customer of the Supplier Designer
' 1
'000" Design
Supplier
Requirements
Processor Plans and
Owner of the Specifications
Operation
Customer
Processor
Facility Supplier of the Customer
Construction Constructor
Figure 3.3: Juran's triple role concept applied to construction (Burati et

al. 1992, pp. 115)
58
This view has now led to productivity being considered as an aspect of
process quality. BS7850 recognises this view by offering the following
definitions for process, quality improvement and quality losses (BS7850: Part
2,1992: pp. 3).
Process: Any activity that accepts inputs and adds value to these
inputs for customers. The customers may be internal or
external to the organisation.
Quality Action taken throughout the organisation to increase the

Improvement: effectiveness and efficiency of activities and processes
to provide added benefits to both the organisation and
its customers.
Quality losses: Losses caused by failure to utilise most effectively and

efficiently the potential of human, financial and material
resources in a process.
The last definition leaves no doubt that BS7850 views productivity as part of
the process quality. Likewise product quality is another dimension of
process quality.
Productivity improvement efforts have to be viewed as an integral part of

process quality improvement. It has, therefore, been necessary to explore
current management trends in performance improvement in this research,
which is primarily aimed at labour productivity improvement and
development of an integrative framework in which the role of labour
productivity measurement and improvement is clearly identified. The
linkages between productivity and quality are discussed in the following
section.
3.3.2 Linkages between productivity and quality
Traditional definitions of productivity and quality do not betray any linkages

between the two concepts. Quality has been defined as the totality of
features and characteristics of a product or service that bear on its ability to
satisfy the stated or implied needs (Hellard 1993). On the other hand,
productivity has been defined, at a very fundamental level, as a ratio of
59
output to input in a production process (Drewin 1985). Until very recently,
the two have been pursued separately. The traditional organisation of the
construction industry places higher responsibility for productivity on the
contractors, while the consultant has to ensure that contractor's work product
is of the required quality. This separation of responsibilities have led to
some adversarial relations. The contractor is often viewed as the party who
is inclined to provide poor quality, if it is possible to get away with it. The
general view is that in order to achieve high quality, productivity has to be
sacrificed and vice versa. Mefford (1991) argued against the traditional view
that there is a trade-off between quality and productivity, and that efforts to
achieve ever higher quality levels of both are likely to be uneconomic. He
offered the following three reasons for a positive correlation between
productivity and quality.
Direct linkage: This is a direct result of how productivity is defined and

measured. Productivity is measured by dividing output by
input; it follows that the numerator should not include
defective output, thereby reducing productivity.
Indirect linkage: Quality improvement programmes are likely to result in

better and smoother production processes that are likely
to have fewer breakdowns and lesser defects.
Motivation: There is increasing evidence that improvements in quality

and productivity have a salutary effect on employee
motivation; through empowerment, implementation of
performance related bonuses, and the pride or security of
working for a growing organisation.
Karlof and Ostblom (1993) offered a viable model relating productivity and
quality through the concept of efficiency, which they proposed is made up of
four basic components: quality; price; production volume; and cost.
Value is the quotient of quality and price, and determines the number of units
sold in a free market economy. The quotient of production volume and cost
is productivity, and the motive force for efficiency in a market economy is
that the delivered value must be higher than the cost of producing a unit of
product or service. Thus, efficiency is a function of value and productivity as
shown in Figure 3.4. Thus, quality and productivity are determinants of
efficiency.
60
High
d
7 'C
>o
o
d

de
o
Na
Low High
Productivity
(Units of output per unit of input)
Figure 3.4: Efficiency matrix (Karlof & Ostblom, 1993, pp. 5)
This concept agrees well with Deming's claims that quality benefits both the
worker and the organisation, as argued by Deming (Hellard 1993). This
argument can be illustrated as shown in Figure 3.5.
costs decrease
Improve quality productivity improves
better quality/lower
more jobs prices capture market
created
business survives and
Figure 3.5: Quality, productivity, business growth chain
3.3.3 The role of TQM in construction performance improvement
Comparatively low level of productivity in the construction industry compared

to the manufacturing industry world-wide has resulted in calls for
improvement in performance in the construction industry. The
manufacturing industry has implemented TQM concepts for sometime now,
and obvious performance improvements have been observed, some of
which can be directly attributed to the application of the TQM philosophy.
61
Chase (1993) suggested that TQM implementation in the construction
industry in the United States is at least ten years behind the manufacturing
industry that has successfully formalised quality management since the early
1980s. While TQM concepts have been accepted as a tool for performance
improvement in the construction industry (Burati et al. 1992), the full
advantage of these concepts has not been accrued. The reasons for this
include lack of clear understanding of the practical applications of the
concepts within an industry which is fragmented, traditionally lacking
alignment of objectives of participants, and the one-off nature of the
construction projects. If TQM concepts are to be successfully and widely
applied in the construction industry, some of these barriers have to be
overcome. Lack of alignment of objectives of participants in the industry is
often cited as one such major barrier. - To overcome this barrier, a clearer
understanding of the construction process, process variables, and process
performance measures is necessary. This will lead to the development of
better customer-supplier relationships in the construction process chains as
customers' expectations are reflected in the suppliers' objectives.
The implementation of quality programmes in the construction industry,

based on TQM principles, would require a careful and systematic evaluation
of performance, to identify problems, establish priorities and monitor the
results achieved. This applies not only at corporate but also at project, site
and crew level. To achieve this, performance improvement tools, such as
benchmarking can be utilised.
3.3.4 Benchmarking as a tool for TOM implementation
Several models for TOM implementation in have been proposed. What is

lacking in these models is a guide to the initial steps required to launch a
TOM programme in an organisation. The Construction Industry Institute's
model (1992) mentioned the 'perceived need for change' as a starting point.
However, a tool for quantification of the 'need for change' is missing in this
model. Both the Malcolm Baldridge Award (Research Technology
Management 1990) and the European Quality Award (Davies 1993) models
serve as tools for progress evaluation in the TOM implementation. They
offer little in terms of kick-starting the programme. The same can be said of
'Total Quality in Construction' published by the European Construction
Institute (1992). Davies (1993) proposed a progress evaluation criteria
based on the European Quality Award (EQA) model from which he extracted
62
a progress checklist. His progress mapping proposal is based on the norm
that the quality evolution of an organisation goes through several stages of
development: survival, prevention, and continuous improvement, each of
which has its sub-stages, as presented in Table 3.1. The progress
evaluation mapping can then be made against criteria extracted from the
EQA model illustrated in Figure 3.6. This model offers an acceptable
framework for progress evaluation, but lacks practical tangible measures.
Table 3.1: The quality evolution

- stages of development for the
world class company (Davies 1993, pp. 37)
1. Survival 2. Prevention 3. Continuous

improvement
" Recognising " " Businessas an

Bringingthe business
competitive threat integrated process
under control
and the need for
improvement " Management " Total customer
ownership orientation
" Isolating key " Challenging existing " Empowering
problems
roles and methods managementstyle
"' Organising to solve
" Building quality " Controlled
them into the business
improvementsand
" Solving them " Developing capable change is the norm
and motivatedpeople
Continuous
Criteria Survival Prevention improvement
Leadership
People management
Policy/strategy
Resources
Processes
People satisfaction
Customer satisfaction
Impacton society
Business results
Figure 3.6: Mapping progress against the survival, prevention and

continuous improvement assessment criteria (Davies1993,pp.37)
There is still the lack of a motivational tool that could be used to encourage
an organisation to embark on a TQM programme. Benchmarking is a
63
performance comparison tool and could be used to quantify the performance
gap which would give the company management the motivation to embark
on a performance improvement programme. It is a powerful tool for
performance improvement through comparison with competitors both at the
initial stages of a TQM programme and for its sustenance through
continuous comparison and improvement. Benchmarking is examined in
detail in Sections 3.4 - 3.6.
3.4 Emergence of the benchmarking concept
3.4.1 Background
Benchmarking is the search for the best practices that will lead to superior
performance of an organisation (Camp 1989). It is a relatively new quality
concept that has captured the interest of many businesses, and has been
gaining popularity amongst executive and senior managers, mainly in the
manufacturing industry, of late. The subject has triggered considerable
interest although there is still some confusion relating to the true meaning of
benchmarking. Its relevance to business organisations, whether
manufacturing or service sectors, and how it could be successfully
implemented are not yet fully understood (Zairi 1992). Watson (1993),
reported that the first book on the subject by Robert C. Camp appeared only
in 1989. The book was based on the author's benchmarking experience
whilst working for Xerox Corporation in the United States. A search through
the Business Periodicals Index, published by H. W. Wilson Company
indicated that between July 1990 and June 1991, there were no articles on
benchmarking, however, for the same period in 1994/95, over 100 articles
were listed. An internet keyword search in July 1996, returned more than
2300 scores some of these linked to specialist consulting firms offering
benchmarking services. Zairi (1992) reported that, a study performed in
1990, concluded that, benchmarking was a little known management
technique in United Kingdom, however, it represents a novel and stimulating
topic for most managers. Further, the study concluded that:
" benchmarking applications were rapidly growing;

" its uses were becoming more widespread; but ,
" there was general lack of awareness of what benchmarking meant.
64
These observations were supported by similar studies performed by Oak
Business Developers in UK (Codling 1992). The term 'benchmarking' was
first used by Xerox in the United States of America which may well have
adopted the philosophy from the Japanese, who have used the term
'dantotsu' which means striving to be 'the best of the best' (Camp 1989).
This business philosophy has been applied in Japan since the end of World
War II (Taiichi 1990). Indeed, the first western world company to adopt
benchmarking practices for its products and processes, Xerox Corporation,
made its initial comparisons with its Japanese affiliate Fuji-Xerox and later
with other Japanese competitors in 1979 (Camp 1989). The results were
very revealing. Xerox established that its Japanese competitors were selling
photocopy machines at what it cost Xerox to produce them. This marked the
start of a series of benchmarking exercises in Xerox which, combined with
other performance improvement practices, enabled the company to
recapture its market leadership in the photocopies business. Since then,
there has been a surge, especially in the United States of competitive
performance benchmarking. This was first and foremost directed at
Japanese competitors, who were seen to be the major threat to the United
States companies. Since the Xerox benchmarking decade, 1976 to 1986,
the technique has been adopted by a number of firms in United States
leading to the development of benchmarking codes of conduct, making it
more acceptable as a management tool (Watson 1993). Based on its early
successes, and the concept of learning of best practices from business
leaders, benchmarking has led to the formation of benchmarking clubs in
which partner organisations (even industry competitors! ) offer to learn from
each other to improve performance (Main 1992).
In Britain, several such steps have been taken. A Benchmarking Centre was
formed in 1993 with the objective of co-ordinating benchmarking efforts. The
centre acts as a 'dating agency' to pair together companies which need to
benchmark similar processes (Costanzo 1993). As at October 1993, the
centre had 21 subscribers none of which were construction companies.
There was also little evidence of benchmarking application in the
construction industry at the time. In his recent book on Total Quality in
Construction Projects, Hellard (1993), admitted that the concept of
benchmarking is an excellent one. However, he argued that, by the nature
of the construction industry, with its essentially project-based activities in
different locations the concept and principles of benchmarking are difficult to
apply, and the lessons to be learnt more difficult to deduce. Since then,
65
there have been government driven initiatives to improve competitiveness in
British industries. These were first directed towards manufacturing
companies but have since been directed at the construction industry as well.
A number of benchmarking projects have since been initiated. These have
been examined in Section 3.6.2.
The construction industry can learn from the wealth of experiences available
throughout the manufacturing industry. A common understanding of
benchmarking has to be reached, and a conceptual framework that would
cater for the industry's primary objectives at various levels has to be
developed.
3.4.2 Objectives and definition of benchmarking
3.4.2.1 Objectives of benchmarking
Benchmarking is a positive, proactive process that can change business

operations in a structured fashion in order to achieve superior performance.
Benchmarking aims at ensuring that the best practices are followed in an
ever changing environment. The process provides a management tool for
measuring and comparing any part of an organisation's operation, product or
service against the best, that leads to superior performance on a continuous
basis. It necessarily involves investigating practices in and outside the
industry for incorporation into its own operations (Thamhain 1991).
3.4.2.2 Definitions
Dictionary definition
A'benchmark' is the term used in land surveying which, in the Concise

Oxford dictionary (1990), is defined as a 'mark or cut in rock etc. by surveyor
to mark point in line of levels making a criterion or point of reference'.
Benchmarking in land surveying is therefore the process of establishing a
benchmark which is a reference point against which others can be
compared. Allan et al. (1968) defines benchmark as a'point of known height
to which surveys are referred'. Benchmarking has also been used in the
computer industry to mean a standard process for measuring the
performance capabilities of software and hardware systems from various
vendors (Watson 1993). Benchmarking, as used in land surveying, aims at
66
establishing a rigid standard, unlike in business where it is to be under
continuous change to reflect responsiveness to competition.
Working definitions
A working definition is one that is simple enough to be easily understood and

put into practical action. Camp (1989, pp. 5) advanced a working definition
of benchmarking as follows:
'Benchmarking is the search for industry best practices that lead to

superior performance. '
The author, in this definition, portrays a very generalised view of

benchmarking. The focus is on adopting best practices or methods to
achieve superior performance. There is no emphasis on searching for best
practices from direct product competitors, nor is there an inherent implication
of where the search for best practice should concentrate unlike in the formal
definition above. The definition implies that the best practices are to be
pursued regardless of where they exist to identify the performance gap
which a business manager wants. Other definitions such as offered by
Liebfried and McNair (1992), have been criticised for implying that
benchmarking is internally focused only (Lema and Price 1995b).
A more refined definition was given in the document titled Planning,

Organising and Managing Benchmarking: Users Guide, (Houston TX:
APQC 1992 pp. 4) which stated that:
'Benchmarking is a systematic and continuous measurement

process; a process of continuously measuring and comparing an
organisation's business process against business leaders anywhere
in the world to gain information which will help the organisation to
take action to improve its performance. '
This definition, developed by International Benchmarking Clearing House

(IBC) Design Committee (United States), is said to represent a consensus
amongst 100 companies (Watson 1993). This definition answers the typical
questions of what benchmarking is, how it is to be performed, with whom
comparison is to be made and of what use the information is to the
organisation. These qualities make it a more suitable working definition than
67
one offered by Camp. The business process, in this definition, is to be
interpreted to incorporate products, processes and services.
3.4.2.3 Terms used in benchmarking
In the course of the development of benchmarking as a formal management

tool, associated terminology have also emerged. Some of the key
terminology and their definitions are as follows (Watson 1993):
Benchmark: A measured best-in-class achievement, a reference or

measurement standard for comparison; a
performance level recognised as the standard of
excellence for a specific business practice.
Best practice: Superior performance within an activity, regardless of

industry, leadership, management, or operational
approaches, or methods that leads to exceptional
performance; a relative term that usually indicates
innovative or interesting business practices that have
been identified during a particular benchmarking study
as contributing to the improved performance at the
leading organisation.
Enabler: The process, practice, or methods that facilitate the

implementation of a best practice and help to meet a
critical success factor; characteristics that help to
explain the reasons for the achievement of
benchmark performance.
Entitlement: The best that can be achieved in the process

performance using current resources to eliminate
waste and improve cycle time; obvious improvements
that are identified during benchmarking and may be
accomplished as short-term goals.
Best-in-class: Outstanding process performance within an industry;

term used as synonymous is best of breed.
68
Benchmarking gap: A difference in performance, identified through a
comparison, between the benchmark for a particular
activity and other companies; the measured
leadership advantage of the benchmark organisation
over other organisations.
World class: Leading performance on a process, independent of

industry or geographic location, as recognised using
process benchmarking for comparison to other world
contenders.
3.4.3 Brief history of benchmarking
In the year 500 BC., Sun Tzu, a Chinese general, wrote, 'If you know your
enemy and you know yourself, you need not fear the result of a hundred
battles'. This statement represents very well the competitive environment in
which businesses operate and illustrates the very early applications of the
benchmarking concept (Camp 1989, pp. 3). As early as the late 1800's,
Frederick Taylor's work on the application of the scientific method of
business had encouraged the comparison of work processes. This was also
within the concept of benchmarking. During World War II, it became a
common business practice for companies to check with other companies to
determine standards for pay, work loads, safety, and other business hygiene
factors (Watson 1993). Walter Chrysler, an early American entrepreneur in
the car industry, used to tear apart new models of Oldsmobile to determine
what went into a car, how much it cost and how it was made (Shetty 1993).
This was an early example of 'reverse engineering' which helped Chrysler to
understand his competitors.
In the early seventies', the Japanese seemed to be outstripping the world in

the ability to be productive while at the same time outstripping the
competitors' quality. Since that time, a growing readiness has developed on
the part of business to examine the management practices of other
organisations in order to regain market shares from the Japanese
competitors (Lake and Ulrich 1993; Karlof and Ostblom 1993). New
management tools and methods were developed to help improve
competitive performance in delivering products and services to customers.
Terms and techniques such as Management by Objectives (MBO),
Statistical Quality Control (SQC), Quality Assurance (QA), Total Quality
69
Management (TQM), Quality Function Deployment (QFD), Just-in-Time (JIT)
and many more, emerged to fill the quest for new competitive tools. The
business world of the late 1970's and early 1980's found themselves shifting
their focus more and more on quality (Macdonald 1993a). The wave of
quality movement led by Philip Crosby developed into the adoption of TQM
in the business processes of many companies including industrial giants
employing more than half a million employees to small businesses with well
under one hundred employees (Macdonald 1993b). Early recognition of the
significance of the quality movement in Britain led to the establishment of
British Standard BS5750 (1979) Quality Systems. The International
Standard ISO 9000 Quality Systems (1987) considered experience gained
from the use of BS5750. This was later adopted as British Standard BS5750
(1987) without modification. The quality standard is aimed at providing
guidelines that would enable organisations to adopt quality practices. It is
now becoming necessary for organisations to obtain accreditation to these
standards and large business and the public sector are increasingly
stipulating that their suppliers obtain accreditation to BS 5750 or ISO 9000
standard if they are to have their contracts renewed or retain their place in
the bidders list (Pengelly 1993). The recently issued British Standard
BS7850 (1992) Total Quality Management underlies further importance of
the quality movement.
3.4.4 TOM and benchmarking
In the TQM journey, goal setting is undergoing significant changes. Instead

of targeting incremental improvements, TQM companies are benchmarking
their performance, processes and products not only against the 'best-in-
class' among their competitors but against the best in class for a particular
function or activity measured world-wide. Benchmarking can be viewed as
an accelerator in the TQM journey. Instead of a companies fully depending
on evolvement from within, it adopts proven successful practices from
elsewhere (i. e. not re-inventing the wheel! ). Indeed, the emergence of
benchmarking can be viewed as a widened scope of competitive analysis.
The application of TQM supported with competitive analysis was for some
time seen to be the main success drivers in a competitive environment.
Karlof and Ostblom (1993) argue that the analysis framework should not
stop at end products and financial strengths as is normally the case in
competitive analysis, but should extend further. Benchmarking is now seen
as a powerful concept of extending competitive analysis to the underlying
70
operative content and the leadership skills that lay the foundation for
success amongst competitors. Benchmarking is now viewed as an
inevitable step within the TQM concept. Zairi (1992) advocated that the
quality movement would inevitably lead to benchmarking as it moved from
inspection to control to assurance to management. Zairi views
benchmarking as a powerful concept that now abandons focus on individual
steps of the quality movement but provides a holistic view of the system in
which business operates. In other words, the quality revolution has spawned
a new approach which seeks to address the whole way in which work is
organised. These sentiments are relevant irrespective of the type of
business.
3.4.5 The strength of benchmarking concept.
TQM has been referred to in the United States as 'the method that has
started the recovery with national co-operation as one of its greatest
strengths' (Karlof and Ostblom 1993, pp. 37), the co-operation mainly being
in the form of benchmarking. While benchmarking has been applauded as a
powerful performance management concept, so far there have been few
plausible explanations for its success. Karlof and Ostblom (1993) attempted
to explain the success of benchmarking by viewing it within the context of
differences between a planned and a free market economy. They viewed
the existing market economy as being only partially exposed to market
forces. The majority of organisations internal activities and operations in a
free market economy, operate under conditions of a planned economy
where they are not exposed to market pressures. The value of
benchmarking is that it provides an opportunity to open up these activities
and operations to the pressures of market forces. When this concept is
viewed within the context of the TQM philosophy, benchmarking is seen as a
perfect vehicle to ensure that the customer gets the best quality under
competitive conditions that ensure lowest prices. The three key roles that
benchmarking plays are to:
" widen the competition base by exposing internal organisational

processes to external market forces;
" accelerate the TQM process and therefore increase efficiency by
providing the opportunity to learn from others, adopt and improve; and
" act as a tool for co-operation to improve overall industry performance.
71
3.5 Scope, procedures and success of benchmarking applications
3.5.1 Scope
In order to fully appreciate the scope of benchmarking, it is necessary to

revisit its definition and objectives. Benchmarking aims at the improvement
of performance by incorporating best practices in the business process.
Camp (1989) states that benchmarking is, first and foremost, a goal setting
process. This objective is very broad, so is the definition. While, within the
context of business management, the understanding of performance may be
very different from that of a public service organisation, the underlying
principles are still the same. Karlof and Ostblom (1993) argued that the
purpose of any organised activity is to create value which is greater than the
cost of creating it. This statement embodies the profound truth that is the
driving force of all economic activity, and embraces all the components that
the management of an organisation can influence. It applies to all organised
activities and not just to economic activities. The value concerned can be of
any kind. This explanation greatly widens the scope of applicability of
benchmarking. While the scope is very wide, there are preconditions for its
successful applications, these are:
" the recognition for the need for continuous performance improvement;
" the recognition and acceptance that there are lessons to be drawn from
others that can lead to improved performance;
" the willingness and capability to change for better performance; and
" the accessibility to the best practices.
Most benchmarking applications have tended to focus attention on world

class performers. Karlof and Ostblom (1993) referred to this as being the
American way of confronting the industrial might of Japan. It is now
recognised that this powerful concept can also be applied where the
objective is not necessarily to attain world class position (Karlof and Ostblom
1993). Comparisons can be"made with a better organisation but not
necessarily a world class performer, provided it is done continuously in small
incremental steps. Indeed, this view is now supported by the pioneers of
benchmarking, Xerox (Main 1992) and there are a number of reasons for not
benchmarking against the world's best; some of which are listed below.
72
" It may take a long time to complete a full benchmarking exercise; Xerox
performed a detailed benchmarking study related to the their electronic
typewriters. Two years later, when the study was completed, the
personal computer had rendered the study worthless (Main 1992).
" The resources required to perform a full benchmarking study for smaller
organisations may be prohibitive. In the development of the Ford Taurus,
a market success when it was unveiled in 1986, the manufacturers had to
benchmark 400 features. To do this, the company had to acquire 50
vehicles, some of them specially imported for that purpose (Watson
1993).
" Even if substantial savings are identified, an organisation may find it

impractical to fully implement the findings owing to the size of the gap
involved. Ford, the American car manufacturer, believed that 500
employees processing accounts payable were far too many. When this
function was benchmarked against its Japanese partner, Mazda, it was
established that the job could actually be accomplished with fewer than
ten people. Ford introduced a partial implementation of the results and
reduced personnel to 200, a 60 per cent reduction against the 98 per
cent identified in the study (Main 1992).
" Finding a suitable partner who is a world leader is not easy. A survey by
the Benchmarking Centre in Britain established that 89 per cent of
companies who would like to benchmark cannot find suitable partners
(Costanzo 1993).
In spite of this, strong criticisms have been directed against those British
companies who do not benchmark against the world class competitors
(Costanzo 1993). In this study, benchmarking construction productivity
against world class performance may not be practical because of the vast
differences in construction environments. It is also likely that the
performance gap may be too large for practical implementation. An
appropriate benchmarking approach has to be sought by exploring the
different types of benchmarking.
73
3.5.2 Types of benchmarking
Literature does not suggest consensus on the types of benchmarking. A

number of authors seem to agree on four different types. Classifications
have mainly been based on approaches to benchmarking as follows (Camp
1989; Zairi 1992; Watson 1993):
Internal benchmarking: Performed within one organisation by

comparing performance of similar
business units or business processes.
Competitive benchmarking: A measure of an organisation's

performance compared to competing
organisations; studies that target specific
product designs, process capabilities or
administrative methods used by
acompany's direct competitors; practices
or services.
Functional benchmarking: An application of process benchmarking

that compares a particular business
function in two or more organisations
irrespective of the industry type.
Generic benchmarking : Benchmarking that is aimed at uncovering

best practices that can be applied in own
business process irrespective of the
source or type of industry.
While the authors generally agreed on the classification, it is surprising that

there was no agreement on the meaning of each type. Table 3.2 compares
the differences in meanings posited by Camp, Zairi and Watson.
74
Table 3.2: Comparison of benchmarking definition by different authors
Author Definitions Comments
Internal Benchmarking
There are no conflicting definitions

Camp (1989) Performance comparison of units or of internal benchmarking among
departments within one organisation authors
Zairi (1992) Same as above
Watson (1993) Same as above
Competitive benchmarking
Camp Direct product competitor Zairi's definition seems to deviate

benchmarking looking at processes from the other two definitions.
and products
Zairi Comparing specific models or
functions with competitors
Watson Product oriented comparisons with
processes involved
Functional benchmarking
Camp Specific function comparison with best Zairi's definition limits comparison
practice to best in class, a view not shared
by both Camp and Watson
Zairi Comparison of specific function with
best in industry and best in class
Watson Comparison of particular business
functions at two or more or anisations
Generic benchmarking
Camp Search for best practice irrespective Camp's definition is best here
of industry because the search should not be
limited to industry and the aim
should be to identify the best of the
best.
Zairi Comparison of all functions of
business operations with those of best
in class
Watson Comparison of particular business
functions at two or more organisations
without regard to indust .
Furthermore, there are a number of authors who suggest different

benchmarking categories Karlof and Ostblom (1993) suggest three
categories, internal, functional and external benchmarking. This
classification does not come as a surprise from these authors, given their
general view of benchmarking. They maintain that the objective of any form
of benchmarking is to be competitive. It follows that a separate classification
of competitive benchmarking as suggested by Zairi, Watson and Camp is
75
irrelevant. Karlof and Ostblom suggested a concept of internal
benchmarking which is very much in agreement with the others. They
define functional benchmarking as the comparison of products, services and
work processes with those of the top performing companies regardless of
the business they are in. This definition combines the functional and generic
benchmarking concepts described by Camp, Zairi and Watson in Table 3.2.
Karlof and Ostblom defined external benchmarking as a comparison of
organisations with similar or identical organisations elsewhere. This could
involve competitors or non competitors provided there are similarities in their
functions. This definition coincides with an overlap of definitions of
competitive and functional benchmarking as presented in Table 3.2.
However, Thamhain (1991) and Shetty (1993) suggested three principle
categories, which they described as follows:
" Strategic benchmarking involves the comparison of different business

strategies and their potentials for success.
" Operational benchmarking involves the evaluation of specific activities,

processes, methods, or functions.
" Business management benchmarking involves the analysis of support

functions or services and their comparison with the best in industry.
Thaimhain suggested further that benchmarking can employ comparisons of

internal practices, external competitive company practices and global
practices. This form of classification, although different, is not in direct
conflict with the classifications offered by Camp (1989), Zairi (1992) and
Watson (1993). Singh and Evans (1993) suggest five types of
benchmarking:
" internal;
" competitive;
" consultant study;
" functional; and
"' generic.
Most of the classifications and the definitions offered by these authors were
in agreement with Camp, Zairi and Watson. Consultant study, as a type of
benchmarking suggested by Singh and Evans, is not coherent with the other
76
types, but can be considered as a method of performing a benchmarking
study. Other different classifications have been suggested by Codling
(1992), Fisher et at. (1995), ECI (1994), Garnett and Pickrell (1995).
Benchmarking has only been recognised as a formal process in the last four
or five years. Current classifications of types of benchmarking are thus
inconsistent and confusing. It is, however, not entirely unexpected that there
would be deviations in the use and interpretations of relevant terminology. A
standardisation of both the process and the terminology is opportune.
Approaches for the classification of benchmarking types seem to be on the
basis of target comparison group(s). Two basic classifications can be
identified: internal benchmarking and external benchmarking. External
benchmarking can be further sub-classified into various forms. Figure 3.7
presents the various proposed forms and their relationships.
Benchmarking
External Internal
Reverse engineering
Competitive
Functional
Generic
Figure 3.7: Benchmarking types
3.5.3 Benchmarking models
A number of benchmarking models have been proposed. These indicate

benchmarking procedures in generic terms. Watson (1993) used the
Deming Cycle to model the benchmarking procedure. The cycle used the
Plan, Do, Check, Act functions. The benchmarking procedure
superimposed on the Deming Cycle is shown in Figure 3.8.
77
Adapting, improving Planning the
and implementing study
Plan
Check 20
Analysing the Conductthe
data research
Figure 3.8: The benchmarking process compared to Deming Cycle

(Watson 1993, pp. 4)
Shetty (1993) proposed a benchmarking model that comprises five basic

steps as shown in Figure 3.4. The diagram shows a process which is
deceptively simple. In practice benchmarking requires several iterations and
it is a judgmental process. For example, it may be difficult to identify
functions and firms to be benchmarked. Note also the feedback loops
require data to set new performance goals for continuous improvement.
Identify the function to be benchmarked
I Select the superior performers

(Competitiveor non-competitive)
Collect data and analyse for pinpointing

gaps in performance,processes and practices
Set performancegoals for improving and

surpassingthe best in class
I Implementing plans to bridge the gap

and monitor results
Figure 3.9: Process of benchmarking (Shetty 1993, pp. 42)
78
The process has been adopted and modified for application in different
companies. Xerox, the benchmarking pioneers, used a ten-step approach
(Camp 1989), but involving all the four steps of the Deming Cycle. The
Xerox model is reproduced in Figure 3.10. Note that there is a continuous
looping process to ensure continuing improvement. This model has been
adopted by Du Pont in benchmarking construction management (Du Pont
1992). Several other organisations have used the model for various types of
benchmarking projects. These include National Cash Register (NCR)
(Karsnia 1991), British Telecom (Pera International 1991) and the Chevron
Research and Technology Company (Merrow and Crocker 1994) just to
mention a few.
1. Identify benchmarking subject
Planning 112. Identify benchmarkingpartners
3a. Determine data collection methodology

3b. Collect data
4. Determinecurrent competitive gap

Analysis
5. Project future performance
6. Communicatefindings & gain acceptance

Integration
7. Establishfunctional goals
8. Develop action plans
Action 119. Implement plans & monitor progress
10. Recalibrate benchmark
Maturity
Leadership position attained
Practice fully integrated into process
Figure 3.10: Xerox benchmarking model (Camp 1989, pp. 17)
79
3.5.4 Successful applications of benchmarking
The Xerox Corporation is widely publicised as one of the first successful

pioneers of benchmarking as part of their TQM initiative. Indeed, Xerox was
one of the early winners of the prestigious Malcolm Baldridge National
Quality Award which was instituted in 1987 (Research Technology
Management 1990). The award is given to those companies which,
according to its statutes, have been most successful in pursuing their quality
programmes. The award is connected to benchmarking in two interesting
ways. Firstly, all prize winners undertake to share their knowledge with other
American companies, and secondly, a full quality points score is conditional
upon the use of benchmaking (Karlof and Ostblom 1993). Recent studies of
competitiveness have shown that a common characteristic of the best
performing companies in North America, regardless of size, is that they
apply benchmarking to their products, functions and practices, using the
leading companies in their respective fields as a reference point (Karlof and
Ostblom 1993).
The European Quality Award (EQA) instituted in 1992, provides a framework

against which progress and achievement can be mapped for any
organisation applying for the award. The model provides a balance between
inputs and results which form the basis for evaluation. In the results criteria,
the self-assessment include key parameters of a company's measure of
results benchmarked against both competitors and best-in-class
performance. It is not surprising that, the first award winner was Rank
Xerox, a British based organisation representing Xerox Corporation interests
in Europe. Miliken, a 1992 EQA prize winner, was also a Malcolm Baldridge
Quality Award winner in 1989.
There are numerous successes achieved by different organisations through

benchmaking. For example, Xerox had the following achievements (Bendell
et al. 1993):
" inventory reduced by two thirds;

" engineering drawings doubled per person;
" marketing productivity improved by a third;
" service labour cost reduced by thirty per cent; and
" distribution productivity improved by eight to ten per cent.
80
Organisations such as AT and T, Du Pont, Ford Motor, IBM, Eastman
Kodak, Miliken, Motorola, Xerox are reported to be using the benchmarking
as a standard tool. A report on the state of American industry, by
Massachusetts Institute of Technology, concluded that the most successful
firms shared an emphasis on competitive benchmarking (The Economist
1991). Management Today (1992) reported that a study on quality
performance, performed by O'Brien and Voss of the London Business
School, indicated that on the basis of the Malcolm Baldridge Quality Award
criteria, scores for British companies ranged between 10-40 per cent, a
score level not even enough to go in for the award let alone to win, thus
concluding that most British companies are a long way from TQM and
finding it very difficult to get there. The first half of the 1990s has seen a
significant growth of organisations using various forms of benchmarking in
the UK (CPN 1996). This is partly due to a benchmarking initiative launched
by the government in 1994 (DTI 1994). The initiative was initially directed
towards manufacturing companies.
3.6 Potential for TOM and benchmarking in construction
3.6.1 Philosophy
TQM assures maximum effectiveness and efficiency within an organisation

by putting in place processes and systems which ensure that every aspect of
its activities are aligned to satisfy customers' needs and all other objectives
without wasting effort and using the full potential of every person in the
organisation (British Standard BS7850: 1992). BS7850 further defines a
process as any activity that accepts inputs, adds values to these inputs for
customers both internal and external. This statement holds true for all
organisations including construction organisations (Karlof and Ostblom
1993). Indeed, TQM concepts are now accepted for performance
improvement efforts in the construction industry after successful applications
in service and manufacturing industries (Burati et at. 1991). Benchmarking
is also widely accepted as a tool for quality and productivity improvement
opportunity identification through comparison with recognised leaders. As
the practice of TQM becomes part of the organisations management
philosophy, benchmarking both internally and externally becomes part of the
process as a means of setting targets and identifying best practices. Both
the Malcolm Baldridge Quality Award and the European Quality Award have
recognised the importance of learning from each other (Chandra 1993;
81
Cottrell 1992). The Swedish Quality Institute emphasise building up quality
organisations through learning from one another as its key qualities (Karlof
and Ostblom 1993). This learning process can be achieved through
benchmarking.
Based on recent studies in the United States, Burati et al. (1992) argued that
the industry has no choice but to adopt TQM which fosters innovation and
team-work while focusing on continuous process improvement and long
range planning, a new approach for an industry characterised by low
productivity, fragmentation, divided responsibility and conflicting objectives.
Within the TQM framework, research has still to address the potential for
benchmarking in the construction industry as a tool to:
" identify and prioritise areas for performance improvement potentials;

" identify sources of best performance and best practices;
" set out a methodology for adopting and improving the best practices in an
organisation for quality and productivity improvement; and
" develop a framework for performance comparisons and target setting in
an organisation, within the industry and outside the industry; i.e. internal
competitive and functional benchmarking.
This research has addressed some of these issues with respect to the
Tanzania construction industry. Some of the demonstrated applications of
benchmarking in the UK construction industry are examined in the following
section.
3.6.2 Applications of benchmarking in the UK construction industry
A survey conducted in Tanzania, as part of this research, indicated that there

was little knowledge of benchmarking among top construction executives.
No organisation was cited to have implemented this concept. A study of the
applications of benchmarking in the UK was performed in the hope that
some lessons could be drawn.
There was little interest in both research and application of formal

benchmarking in the construction industry in the UK prior to 1994, except for
some work done by Stanhope Properties Plc. The company published
benchmarks for private sectors office development in some key areas of
82
construction performance and made some comparisons with USA,
continental Europe and Japanese data (Stanhope 1993). The following
findings emerged from the Stanhope study.
a) Quality was found to be poor with too many defects and that:
" twenty per cent of defects were caused by operatives while 80 per
cent were caused by management;
" cost of rework was 12 - 15 per cent of construction cost; and
" cost of waste was 10 - 20 per cent of construction cost.
b) Inadequate management led to poor productivity and high construction

costs with:
" less than 50 per cent of labourhours being classified as productive;

" 50 per cent more labour hours were required to complete a project
in England than in the USA;
" output per person in both Germany and France were 10 - 20 per
cent higher than in UK; and
" the number of staff needed on continental sites was only half of that
required in UK.
c) Construction output on site was 15 - 20 per cent below that of the USA.
d) Buildings were over specified in some areas of design, for example in

floor loading.
These initial efforts have been widely publicised within the industry although
they have little value as benchmarks mainly because, firstly, they only
constitute metrics and they do not contain relevant practices, and secondly,
they only offer some idea on the performance gaps. However, these studies
have played a significant role in catalysing the use of benchmarking in the
construction industry.
The need for benchmarking in the construction industry is now widely

accepted. Three major events initiated wider interest and applications of
benchmarking in the UK construction industry. The publication of the widely
publicised report titled 'Constructing the Team' by Michael Latham in July
1994, which identified a 30 per cent construction cost reduction potential by
83
the year 2000 (Latham 1994), and almost a simultaneous launching of a
Construction Benchmarking Challenge by the government (Construction
Monitor 1994) in which the UK government offered 100,000 to trade
associations to facilitate setting up four benchmarking clubs to make the
construction industry more competitive. Thirdly, the Innovative
Manufacturing Initiative (IMI) launched by the UK government under the
Engineering and Physical Sciences Research Council (EPSRC) in July 1994
added impetus to research into the applications of benchmarking in the
construction industry. The IMI for the construction industry was
appropriately termed 'Construction as a manufacturing process' with the
following global research and development targets in the next ten years:
" achieve a 50 per cent reduction in construction time;

" increase current contribution to GDP by 25 per cent;
" reduce energy consumption in buildings by 25 per cent;
" reduce imports of building materials by 50 per cent;
" double the current rate of improvement in productivity; and
" reduce the cost of reworking by 50 per cent.
These events presented clear challenges to the construction industry, the

success of which required the applications of benchmarking. Since then, a
number of benchmarking efforts have been made by academics, research
bodies and professionals. A recent meeting organised jointly by IMI and
Construction Productivity Network (CPN) focused on benchmarking in the
construction industry. It presented an excellent forum to assess the general
status of benchmarking in the construction industry in UK. A selection of
these have been examined in this work.
'LINK IDAC project - Benchmarking for construction' research project
This is a joint research project involving academic and research institutions

collaborating with industry partners as well as clients and management
consulting companies. The major objectives of the research were to:
" the develop a generic benchmarking methodology;

" establish time benchmarks through time study;
" study the construction process; and
" European comparisons.
84
The project, which was initially very ambitious, has not yet come up with
significant results, and has chosen to concentrate on construction time
benchmarks only, with the objective of stimulating the application of
benchmarking in the construction industry.
'Continuing Professional Development (CPD) in the Steel

Construction Industry'study
This project aimed at optimising the contribution of education and training

to the business competitiveness of the steel construction sector through
the:
identification of mismatches between business process objectives and

the CPD menu available;
identification of effective forms of CPD as benchmarks to support the
adoption of advanced manufacturing techniques; and
examination of the relevance of competence-based approaches to the
development of CDP courses.
To achieve these objectives, the research performed an opinion survey

amongst major CPD clientele to identify any weaknesses of the current
approach and the favoured form, content and delivery of future training
programme. Although the term 'benchmarking' was used in this project, its
relevance was not obvious. No clear metrics and practices were identified,
and the use of the term seemed unnecessary.
'The AGILE Construction Initiative' research project
The AGILE Construction Initiative is a research project which was aimed

at establishing world-wide civil engineering metrics and benchmarks based
on information collected from participating companies. The project draws
from experiences of a similar study based on The International Motor
Vehicles Programme (IMVP) which has been running for over a decade.
The project has just been initiated and was in the process of identifying key
focal areas, and relevant performance indicators in collaboration with some
industry partners.
85
Energy Consumption on Construction Sites'study
This is a benchmarking project by the Building Employers Confederation

(BEC) aimed at the analysis of energy consumption levels on construction
sites and establishing industry best practices to enable companies to
achieve appropriate cost savings in energy consumption on sites. This
project benefited from the Construction Industry Benchmarking Challenge
launched by the Department of the Environment in 1994. BEC formed a
benchmarking club comprising 57 of its members. Initial studies indicated
that energy costs were about 0.35 per cent of the annual turnover, with
electricity taking up the major cost share. The study identified that smaller
companies had a higher potential for energy savings. Bigger companies
with active energy policies had the lowest share of energy costs. The
study went further to identify and publish for its members best practice
leaflets for its members and target levels for energy consumption on sites
which would form the basis for future continuous improvement.
The above are just a few examples of the varied applications of

benchmarking in the construction industry discussed in the IMI/CPN
meeting. What emerged from the meeting was that, as yet there was no
clear broad framework for the application of benchmarking that identifies key
issues and relevant metrics. Indeed, one of the projects was aimed at
developing such a framework, although no results were presented. It is also
worth mentioning that there was no presentation which linked benchmarking
to TQM, as argued in this work, although most presentations indicated some
links with continuous improvement and customer focus - some of the key
aspects of TQM.
3.6.3 Other construction benchmarking initiatives
Some other construction industry benchmarking efforts include:
European Construction Instutite - Measurement and Benchmarking

Task Force
This is a task force set up by the European Construction Institute (ECI) with
the following mission:
86
To develop and implement a system for measurement and benchmarking
of construction projects that will provide objective evaluation of projects
performance to performance to identify best practices which, through the
sharing of this knowledge, will promote improvement within the European
Construction Industry.
The task force set the following objectives in order to realise the mission:
" design and implement a system for performance benchmarking

construction projects;
" develop and implement process benchmarking among ECI members;
" identify and establish a management and administrative structure abd
system to support bith performance and process benchmarking; and
" develop and manage a benchmarkingclub among ECI members.
The task force, which was set up in 1994, had achieved several objectives
August 1996.
Development of ECI benchmarking model
The model describes key processes that run through the following phases of
a construction project:
" feasibility;
" scheme design;
" detailed design;
" site construction; and
" site commissioning.
The main objectives of the ECI benchmarking project are to identify best
practices in the key processes, together with measures of process
effectiveness and efficiency. The model developed generic methodologies
for presentation of benchmarking results.
Benchmarking productivity in the UK construction industry - BOMEL
This is a benchmarking study aimed at establishing productivity benchmarks

for those activities and trades with the largest time and cost contributions to
the total project costs and those areas those areas which consistently cause
87
largest time and cost overruns. The following sub-objectives were identified
in the project:
" to develop a detailed methodology for the measurement of productivity

on building construction sites;
" to determine the major on-site cost and schedule factors, particularly
those which contribute to overruns;
" to collect baseline data in order to benchmark currently achived
productivity levels;
" to identifythe priority areas where productivityimprovementsshould be
made; and
" to provide the methodology for on-going measurement, review and
improvement of productivity.
The first phase of this work is completed and the second phase is in
progress. Benchmarking study results have not yet been published.
3.6.4 A framework for TOM and benchmarking applications in

construction
In this research, a conceptual framework that identifies a clear role of TQM

in benchmarking has been developed. The review of applications of
benchmarking in construction indicated that a generic framework has not
been developed. While some priorities have been identified, there is no
clear vision of how these priorities can be achieved. It is necessary to
embark on a continuous improvement journey in a sustainable manner.
Figure 3.11 represents a conceptual model for TQM implementation
developed in this work in order to appreciate the role of benchmarking in
kick-starting and sustaining the TQM implementation process - continuous
performance improvement programme.
88
Benchmarking II Total Quality Management
Entry tool in TOM

. (gap identification)
Metrics
Continuous
comparison
with external
Practices performance
Leadership
Best Continuous Teamwork
Practices Improvement
Training
Process
improvement
Customer
Satisfaction Customer
involvement
Figure 3.11: Benchmarking for TQM implementation (Lemaet at. 1994,pp.493)
3.7 Relevance of the conceptual framework in this research
Benchmarking is widely accepted as a tool for quality and productivity

improvement opportunity identification through comparison with recognised
leaders (Zairi 1992; Watson 1993; Karlof and Ostblom 1993). The search for
best practice is initiated through performance comparisons with leaders on
the basis of performance indicators or metrics. These indicators can be
business results such as profitability, turn-over, market share, share prices,
productivity etc. The identified performance gap provides the motivation for
embarking on a performance improvement programme. TQM is one such
viable programme. Once the programme has been initiated, the initial
quality and productivity improvement steps are achieved through the
incorporation of best practices throughout the organisation. As the practice
of TOM becomes part of the organisation's management philosophy,
benchmarking, both internally and externally, becomes part of the process
as a means of setting targets, identifying best practices and incorporating
them into the organisation's processes.
There is extensive scope for benchmarking applications for performance

improvement throughout the construction industry, some of these were
89
discussed in Section 3.6.2. Figure 3.12 illustrates a conceptual framework
which spans from industry to crew level. Performance indicators can be
identified at different levels and associated with various practices. The
framework can be used as the basis for identification of the main or critical
indicators, which in turn highlight the improvement potential. Focusing on
productivity as a central performance issue, specific indicators of productivity
at various levels can be isolated and prioritised. This research is based on
this logic in that it justifies labour productivity as a key performance indicator
and identifies best practices associated with best performance. In an
environment where competition is gradually becoming a key performance
driver, performance gap quantification becomes a main motivator as the
potential for improvement is unlocked. It is also important to identify the
potential where it can best be influenced. The reasons for the choice of
labour productivity at crew level are discussed in detail in Chapter 4.
Benchmarking
Practices Performance Indicators (metrics)

II
'9 Industry
Industry
0
a
c
F-
II Corporate -- Corporate
. 4c
3
Project Project d
N
C
fD
N
I Crew
Crew
Labour output
Work quality
luence Safety
Motivation
Labour cost
Teamwork
Skill
Figure 3.12: Benchmarking metrics and construction productivity

measures (Lema et al. 1994, pp. 494)
With the acceptance of TQM as a viable performance improvement tool in the

construction industry, a wider view of performance has to be adopted and
even redefining some of the common terminology is necessary. However, in
90
view of the widespread use of the traditional indicators, historical performance
can only be evaluated utilising these measures. Some of these indicators
have been used in Chapter 7 to evaluate construction industry performance at
industry level in Tanzania. In any case, the industry is under continuous
challenge to develop and adopt performance indicators that conform to
modern performance improvement philosophies such as TQM. The adoption
of modern performance improvement concepts requires a clear understanding
of the scope and characteristics and structure of the construction industry and
constraints to the implementations of these, especially in a developing
economy.
3.8 Summary
This chapter has reviewed the traditional approaches to construction

performance measurement and improvement. Their limitations were
explored. Total Quality Management (TQM) was identified as a viable
performance improvement tool which is gradually finding widespread
acceptance in the construction industry. The TQM concept, a management
philosophy widely applied in the manufacturing sector, offers a promising
future for the construction industry which faces major productivity
improvement challenges. Benchmarking has been identified as a powerful
tool not only for kick-starting the TQM process, but also for sustaining
continuous improvement efforts. Benchmarking was seen within the context
of TOM as an accelerator towards TQM achievement by learning from others
who have demonstrated excellence. This is the general philosophical
approach to this research. Furthermore, this chapter argued that it is a
necessary tool that exposes internal activities of competitive organisations to
external organisations, a competitive potential currently not fully utilised.
Several cases of the application of benchmarking in the construction industry
in the United Kingdom were cited.
Productivity improvement was identified as a major challenge facing the

construction industry. Conceptual relationships between TQM, productivity
and benchmarking were explored with the objective of developing a clear
understanding of the role of benchmarking in addressing the productivity
improvement challenge within the TQM framework. The specific role of
benchmarking in this research has been stated. In Chapter 4, construction
productivity will be examined in detail from the point of view of the
construction process performance measurement.
91
CHAPTER FOUR
CONSTRUCTION PRODUCTIVITY
92
CHAPTER FOUR
CONSTRUCTION PRODUCTIVITY
4.1 Introduction
Chapter 2 and 3 examined the need and techniques for performance

improvement. Productivity was identified as one of the key measures of
construction performance. The significance of construction industry
productivity in the economy was discussed in Chapter 2. Although there is a
general consensus on the need to achieve higher productivity, there is wide
agreement that there is no precise definition of what productivity actually
means (Lowe 1987; Abdel-Razek 1987; Olomolaiye et al. 1989; Thomas et al.
1990 and Lim 1995). Methodologies for measuring and quantifying
productivity are issues of considerable debate. Factors influencing
productivity, although widely researched, are not yet fully explored even in
developed countries. This chapter: examines some of the historical
developments leading to a definition of the term productivity; develops the
concept of productivity into the various definitions; discusses the accepted
definitions in general and in construction at industry, project and site
production level; and identifies appropriate productivity measures for use in
this research. Measurement of productivity and associated problems are also
discussed. Factors that influence productivity are also examined in this
chapter.
4.2 Productivity - origin, concept and basic definition

4.2.1 Origin of the word 'productivity'
In a formal sense, probably the first time the word 'productivity' was mentioned
was in an article by Quesnay in 1766 (Deurinck 1955). More than a century
later, in 1883, Littre defined productivity as the 'faculty to produce', that is the
desire to produce. It was not until the early twentieth century that the term
acquired a more precise meaning as the relationship between output and
means employed to produce that output. In 1950, the Organisation for the
93
European Economic Co-operation (OEEC) offered a more formal definition of
productivity as follows: (Deurinck 1951, pp. 22).
'Productivity is a quotient obtained by dividing output by one of the factors of

production. In this way it is possible to speak of the productivity of capital,
investment, or raw materials according to whether output is being
considered in relation to capital, investment or raw materials etc. '
In the 1950s, the OEEC actively promoted several productivity issues. During
this time, many European and Asian countries established Productivity
Centres and Councils. This led to a general widespread interest in productivity
especially in the manufacturing industry. Other definitions of productivity
emerged in the 1950s, 1960s, and 1970s, however, landmark definitions of
productivity were few. Some of these are listed in chronological order in Table
4.1.
Table 4.1: Chronology of important definitions of productivity (Sumnth1985,

pp.4)
Author Year Definition

18th century Quesnay 1766 The word productivity appears for the first time
19th century Littre 1883 'Faculty to produce'
20th century Early 1900s 'Relationship between output and means
employed to produce this output'
OEEC 1950 'Quotient obtained by dividing output by one of
the factors of production'
Davis 1955 'Change in product obtained for the resources used'
Fabricant 1962 'Always a ratio of output to input'
Kendrick and 1965 'Functional definition of partial, total factor and total
Creamer productivity'
Siegel 1976 'A family of ratios of output to input'
Sumanth 1979 'Total productivity - the ratio of tangible output to
tangible input'
94
4.2.2 The search for a' comprehensive definition of productivity
Despite attempts by several researchers to provide comprehensive definitions

of productivity, there is not yet a universally accepted definition. In the 1950s
and 1960s, when productivity research by economists was more prevalent (at
both national and industrial levels), there seems to have been less confusion
about the basic definition of productivity (Sumanth 1985). In recent years, with
the topic of productivity being addressed almost by anyone, other definitions
have emerged. The following section discusses some concepts and
definitions currently in use.
4.2.3 Dictionary definition
The Concise Oxford dictionary (1990) defines productivity as:
" the capacity to produce;

" the state of being productive; or
" the effectiveness of productive effort or production per unit effort.
The dictionary definition of productivity serves to illustrate some of the

difficulties of attaining a precise definition. The first and second part of this
definition are quite vague and may lead to confusion by regarding productivity
as synonymous to volume of production or the production process. The
second part is also misleading because it relates to production effort only,
which is only a small part of the inputs in the production process.
4.2.4 Other definitions
Some productivity definitions originate from special interest groups such as

economists, industrialists, trade unions and politicians. Individuals or groups
have meanings that fit their situation. Management and union leaders, for
example, may have different views on the definition of productivity. While
managers may argue for a definition that considers all inputs including
improved technology, unions may argue for labour input alone and attribute
any increase in productivity to labour alone. Riggs and Glenn (1983)
95
summarised slogans suggested by various groups at a productivity awareness
program as follows:
" Engineer - doing more with less;

" Scholar - work wisely;
" Manager - getting it all together;
" Philosopher - knowing you have done your best;
" Pragmatist - performing effectively;
" Psychologist - striving, yet at ease; and
" Optimist - making tomorrow better than today.
The message of the slogans is that productivity is closely related to efficiency

of the production process. Indeed, some of the messages in these slogans are
reflected in the mission statement of the Japanese National Productivity Centre
that describe productivity as follows (Williams 1991 pp. 102).
'Productivity is above all else, an attitude of mind. It is a mentality of

progress, of constant improvement of that which exists. It is the certainty of
being able to do better today than yesterday and less well than tomorrow. It
is the will to improve the present situation, no matter how good it may seem,
no matter how good it may really be. It is a constant adaptation of economic
and social life to changing conditions; it is the continual effort to apply new
techniques to new methods; it is faith in human progress. '
This statement embodies a number of aspects very much in line with the Total
Quality Management (TQM) philosophy. While it adequately covers the need
for the continued effort to improve productivity, it serves little purpose for a
researcher or a practitioner who has to work with precise definitions. It is
necessary to examine the concept of productivity which can then be used to
explain the commonly used productivity definitions.
4.2.5 Productivity concepts
Productivity is generally defined as the ratio of outputs to inputs in a production

process. Productivity results from the efficiency of the productive process, in
particular, the efficiency with which labour and capital (men and machines) are
96
used in transforming inert materials into socially usable end products. This
definition can be illustrated in a conceptual production process as shown in
Figure 4.1. The process illustrated comprises of three sub-processes A, B, and
C. Each sub-process has both inputs and outputs. Sub-process A, for
example, has inputs 1,2,3,4, and 5 and output 'i' (for simplicity only one
is
output considered here, although in practice there may be several outputs).
The inputs may be raw materials, labour, machines, energy etc. while output 'i',
is a semi-processed input into the subsequent sub-process B along with other
inputs which may also be semi-processed products of other sub-processes.
The same would apply for sub-process C. The final product is the output of
sub-process C.
The above illustration further identifies process boundaries, denoted as

boundaries I- IV, between inputs and outputs. Using the basic definition of
productivity as the ratio between output and input, it is therefore possible to
denote productivity at any of the defined boundaries.
Output iii
Boundary I: Productivity =
Inputs 1+2+... +5
Likewise for boundaries II and III, productivity can be expressed as:
Final product
Boundary III: Productivity =
Inputs a++x
However, for boundary IV, the expression of productivity becomes complex

due to multiplicity of both inputs and sub-processes. Whereas is an output of
sub-process B, and likewise a and x may be outputs of other sub-processes.
Sub-process B also has several inputs, some of which are output of sub-
process A. To achieve a final product, there are streams of inputs at each sub-
process and the task of determining productivity becomes complex. Such a
system of sub-processes can be illustrated by an example of a manufacture of
a car as illustrated in Figure 4.2. Figure 4.3 illustrates a construction process
example of concreting activity where such sub-processes are shown.
97
c.
0
LL
'
(D
U cz
0
L C
0
co
C
0
>
N
cd
C N
(D
O 0
m 0
c
0
V
V
U) co 0
C a
0
2
0
0 co
CL
4)
V
c
0
C)
0N
. Q.
C rn
U-
4)
U cd
D
O
CL 0
n
N
0.
98
CD-
0
lL
a)
U c
2
L C
a-
O L-
w
m
C.%
U)
N c fr 0)
c >. :.i
CL W rn 0
a- E
c ac O.
0 .5 0)
v
0
Ci
> 0
L-
Cl) CL
CO ctj ct U)
a)
U
2 C
a
a- 0 0 a)
co co
C
0
CO U)
d
ci
C)
0 U-
a)
U C13
0 C
I
IL
0
m
Cl)
I-
D.
99
4-
n
dE
0 - ccaCD
E
LL
0/
2!1
Cl) (Ti
U)
C)
U
O
L
c1 0
co
0

C
0
ev
N
U)
4- o0 c N
CL o a E d
c EcN v
0
Lo. Vg
a
a)
>
a,
C co
0 C 0
c
co
CO 0
U)
0
C ca
a)
U CL
O
O m
0 m CL
a,
c0 Z
0
U
M
x
: Y E -w a)
CL U i I-
0
c
co
C)
0 '0 LL
dddd
E
Q Q)
1
(0
100
While the representation in Figures 4.1 - 4.3 seems comprehensive, there are
number of difficulties in their practical applications. These include:
" Demarcation of process boundaries is difficult and sometimes vague. This

has already been demonstrated in Chapter 2 where the problem of defining
the construction industry boundaries was discussed. Similar problems are
associated with defining a construction operation at site level. For
example, a concreting operation was defined by Price (1986) as involving
pouring, vibrating, shovelling, tamping, troweling and covering. This was
mainly because concrete was delivered to site ready mixed. In this work,
the concreting operation included batching constituent materials, mixing
and placing, as all these activities were integrated in one place at the same
time, by the same crew. Similar demarcation differences can be
experienced in numerous other construction processes.
" Identification and quantification of all inputs and outputs is not easy as
some are tangible and others intangible. Further, even when they are
tangible, it is difficult to isolate the extent of each input's role as they are
inter-related in a complex way. One way out of this is to represent
productivity in terms of a single key input. This form of representation,
although a simplification, can be utilised to compare productivity of similar
processes where other inputs are more or less the same across the
processes, or they are unimportant.
4.2.6 Basic definitions of productivity
Despite these difficulties, it is possible to identify a number of basic definitions

of productivity. Sumanth (1985) proposed three different meanings of
productivity which can be easily associated with Figure 4.1.
Partial productivity
Partial productivity is the ratio of output to one type of input. For example,
labour productivity (the ratio of'output to labour input) is a partial productivity
measure. Similarly, capital productivity (the ratio of output to capital input)
101
and material productivity (the ratio of output to materials input) are other
examples of partial productivity.
Total-factor productivity
Total-factor productivity is the ratio of net output to the sum of associated

labour and capital (factors) inputs. Net output is the total output minus the
intermediate goods and services purchased. In this case, the denominator
of this ratio is made up of labour and capital inputs factors.
Total productivity
Total productivity is the ratio of total output to the sum of all inputs factors.
Thus, a total productivity measure reflects the joint impacts of all inputs in
producing the output.
In all the above definitions, both output and input(s) are expressed in 'real' or
'physical' terms being reduced to constant monetary currency of a reference
period (referred to as base period). If a company produces output of value Vt
over a certain period t, using human input Ht, material input Mt, capital input Ct,
energy input Et and other expenses input Ot, then partial, total-factor and total
productivity values can be computed as follows:
Partial productivity over period T.
roductivit, HV
Human PYP-
Hr
Materials productivity, Mp=V

Mr
V'
Capital productivity, Cp =
Ct
Vf
Energy productivity, EP =
102
V
Other expenses productivity, Op =V
r
Total-factor productivity:
net output
TFPP = where TFPt is total-factor productivity over period t.
Hr + Ct
Vt - materials and services purchased

Thus, TFP, =
Ht + Cr
Total productivity:
total output
TPt = where TPt is total productivity over period t.
total input
Vt
Thus, TP, =
H, +CC+Mt +EE+OO
Each of the above basic types of productivity measure has advantages and
disadvantages. Partial productivity measures are easy to understand, easy to
obtain data for, and easy to use to compute productivity indices. They are thus
widely used and industry wide data are available, but, they can be misleading
when used in isolation. Total-factor productivity, has the advantage that data
are available mainly at corporate level and are easy to compute. However, as
the factor does not capture all inputs, a full picture is still lacking leading to the
possibility of misguided decision making. The value added approach to
defining output is not common at corporate or project level. Total productivity
has the advantage of considering all inputs although quantification of these
remain a major disadvantage of this measure which makes it impractical.
Some of these measures may use different values in different situations
depending on the purpose of the analysis, type of process, and ease with
which data and information can be obtained.
103
4.3 Construction industry definitions and measures of productivity
4.3.1 The need for construction industry specific definitions
The need for a construction industry specific definition of productivity has been
long recognised. The Business Roundtable (1982) established that there is no
common definition for construction productivity. Even when definitions were
consistent, approaches to measure input and output vary greatly making it
difficult to compare results of different studies. Unless efforts are made to
harmonise these differences, research efforts will continue to yield results of
little value to the industry at large. It is also well recognised that the definitions
of productivity in the construction industry depend on the boundary of the
production system considered (Lowe 1987). These boundaries may include:
the industry as a whole (this may be equated to boundary IV in Figure 4.1);
company (firm); and project or activity (crews). - Productivity may be defined
within each of these boundaries for unique purposes. - In contrast with the
classic economic definition, construction industry has sometimes used
input/output as a measure (definition) of productivity. The reason given for this
is that, in the construction industry, cost is an important performance factor in
the estimating and project delivery stages (Ireland 1992). The use of this
measure is historical and difficult to justify.
4.3.2 Labour productivity -a partial definition of productivity
Construction labour productivity is usually defined as the gross product output

per person employed or man-hour worked. This measure of productivity takes
into account only one factor of input (i.e. labour). It is, therefore, a partial or
single factor productivity measure. The capacity to produce is a combination of
several factors or inputs as demonstrated in Figure 4.1, and further shown in
the different measures of productivity. It is difficult to measure the contribution
of each factor in the production process. When a labourer is, for example,
shovelling soil, it is difficult to differentiate between the contribution of the
labourer and shovel separately. It is also clear that the output will be a function
of the shape, weight and even size of the shovel. The output is only a measure
of the joint power of input surrounded by a complexity of other factors. Neither
input data nor output data can provide an analyst with conclusive evidence on
104
the productivity of any type of input: i.e., labour, capital, land, organisation and
methods, management skills, or enterprise. Focusing on labour productivity
only can be misleading because a number of other factors affect the output per
unit labour. These include managerial efficiency, economies of scales, use of
plant and equipment, and the introduction of new technologies (Ireland 1992).
In spite of the above limitations, labour productivity has been widely accepted
as a performance measure in the construction industry (Lowe 1987; Handa &
Abdalla 1989; Olomolaiye and Ogunlana 1989; Emsley et al. 1990 and Horner
1992). The emphasis on labour derives from several reasons:
" labour is the most important factor and most easily quantifiable;
" it is the only factor that has conscious control over its contribution to output;
" labour is a resource which can appreciably be influenced by the quality of
management; and
" labour productivity is a key potential issue of contention between
management and employees as regards performance, and both parties are
normally well equipped with relevant data.
4.3.3 Multi-factor productivity (total factor productivity)
Multi-factor productivity (or total factor productivity) is the ratio of weighted sum
of several inputs to gross product output. If all inputs have been taken into
account it is referred to as 'total productivity'. It is a comprehensive indicator of
production efficiency and it was initially introduced by Stigler in 1947 and later
developed by Kendrick in1956 (Chau and Walker 1988). This definition differs
slightly from that proposed by Sumanth (1985) who proposed that net output or
value added should be considered as the output and the inputs should only be
labour and capital. The purpose of developing total factor productivity was to
overcome the limitations of the single factor approach. This approach takes
into account all major inputs. The TQM philosophy was partially reflected by
the Business Roundtable (Business Roundtable 1983) productivity definition
which is very much in line with the total factor productivity approach although
this approach has so far found little usefulness in the construction industry due
to difficulties associated with this measure as discussed earlier. These
difficulties are further reinforced by the temporal nature of construction
activities.
105
4.3.4 Industry level productivity
Industry level productivity is usually of interest to economists who look for the
most appropriate production function and measure the required resources
inputs for this function (Burton, 1991). Literature suggests that the total factor
productivity (TFP) approach described above is popular. This is demonstrated
by work done by Webber and Lippiatt (1983), Ruddock (1991), Burton (1991),
Lowe (1987), and Chau and Walker (1988). These approaches vary in their
degrees of complexity and are more useful in economic studies. They are not
of great value in this work.
Construction industry productivity is simply expressed as a relationship

between construction industry gross product and the key inputs at that level.
The definition can be in the form of a single factor productivity in the case of a
single input (such as labour productivity), or multi-factor productivity for several
inputs depending on the inputs under consideration. Labour is the main input
used. Labour productivity in the construction industry can be obtained by
dividing the gross industry output by the number of persons engaged in the
industry over a period of time (Briscoe 1988). As statistics are normally
available on an annual basis, productivity is expressed as follows.
Gross industry output per annum

Construction industry productivity =
Labour engaged per annum
Problems associated with this form of definition have been examined in

Chapter 2 in which the very definition of 'construction industry' was found to be
imprecise due to difficulties in determining its boundary. The definition of
productivity at this level has to be flexible enough to accommodate the
inevitable flexibility of the industry's boundary. Value added per person
engaged is sometimes used and Kumari (1993) stated that it is actually
preferred by economists as it expresses the effectiveness of the industry in
adding value to the economy as discussed in Chapter 2. Both of these
definitions have been used in Chapter 7 to evaluate and compare time series
construction industry productivity in Tanzania, Kenya and Britain.
106
4.3.5 Project level productivity
Productivity at project level has received most attention because project

boundaries are fairly easy to define, and the benefits of improved productivity
are easily quantified. Indicators of project performance have been the subject
of considerable research (Sink 1985; Freeman and Beale 1992; Bresnen et al.
1990; Picard 1991and Lim 1995). At project level, the construction industry
determines productivity in two ways: by tracking project progress over a given
time against the associated costs or by evaluating at the end of the project the
inputs in comparison to the value. Again, the definition of productivity depends
on both the purpose and the party evaluating it. Productivity to a construction
contractor may be determined by examining the gross value in relation to costs
(or value in relation to labour or capital employed), whereas a building
construction client may evaluate productivity in terms of costs per square metre
of floor space or for a road construction client, cost per kilometre (or cost per
mile). Productivity comparisons across projects utilise both total as well as
partial productivity measures. Measures such as cost per square metre of floor
area, or labour-hours per square metre of floor area for building projects are
total and partial productivity measures respectively. The importance of and
methodology for determining project productivity and progress tracking by
construction contractors has been dealt with in some detail by Williams (1991).
The method is based on a comparison between planned and achieved input in
terms of labour-hours and planned and achieved output in terms of work
progress. Similar concepts and methods could be extended to productivity
evaluation by clients by comparing costs and value created in similar projects.
4.3.6 Crew level productivity
The definition of productivity at the crew level takes into consideration output of
individual activities. Alfeld (1988) and Thomas (1992) defined it as the ratio of
labour-hours to the quantity of work in place. For example, masonry wall
productivity would be expressed in terms of the number of labour hours
required per square metre. This partial productivity measure is contrary to the
measure used by economists, and Ireland (1992) associates it to traditional
practice other than anything else. The reason for using the measure in this
107
form, in construction, is probably because the product is normally well defined
in contract documents and the problem is actually in determining the labour
resources required. A more common measure of productivity is the ratio of
quantity of work in place to the labour-hours expended. Thus:
Quantity of work
Labour productivity .
Labour hours
This measure has been adopted in this research both for its popularity and for
the reason that the quotient increases as productivity increases. Even at this
level, productivity definitions differ depending on the purpose. Thomas et al.
(1990) argue that there is considerable difference of interest in the type of
labour productivity required for different groups at different times. Horner and
Talhouni (1990) differentiate between productivity calculated on the basis of:
" total or paid time;

" available time; and
" productive time.
This form of definition is particularly important for researchers because it offers

an opportunity to evaluate productivity over a relatively short interval of time
when particular sets of conditions exist. This can lead to the understanding of
what causes productivity to change and by how much. This research has also
utilised labour productivity based on productive time to evaluate the extent to
which labour productivity can be increased. This has been referred to in this
work as productivity rate and it is expressed as:
Quantity of work
Productivity rate, P, =
Productive labour hours
where productive labour hours are based on the productive time evaluation of
working crews. This measure can be used to analyse factors that affect
productivity when delaying factors are eliminated.
108
4.4 Productivity measurement
4.4.1 Significance of productivity measurement
Productivity measurement in the construction industry at various levels serves

several purposes. It forms a basis for:
" strategy and policy formulation at industry and corporate levels;

" performance evaluation at industry, corporate, project and crew levels;
" product costing and operational cost control at project and site levels;
" comparison with competitors at corporate, project and site level;
" operatives remuneration and motivation;
" performance target setting at corporate, project and crew levels;
" motivator for performance improvement; and
" planning, performance forecasting and schedule control at all levels.
Productivity measurement in a competitive construction industry has to be

seen within the broad context of performance measurement for performance
improvement. At this level, productivity data are important for the formulation of
the strategies for, and the evaluation of the effectiveness of productivity
improvement policies and programmes (Chau and Walker 1988). This may
also be used as a measure of investment effectiveness because the higher the
industry productivity the higher the return on investment at national level.
International competitive performance comparisons depend on industry
competitiveness as quantified by among other factors, productivity.
Measurement of productivity at industry level is, however, clouded by so many

factors some of which are beyond the control of the industry. Measurement at
this level serves the wider purpose strategy and policy formulation.
Measurement at corporate, project or crew level enables management to focus
on specific actions which are within reach such as streamlining company
procedures, effective planning and control of project costs and schedule.
Productivity measurement at corporate, project and crew level forms a
powerful tool for performance comparison in a competitive environment.
However, even at corporate level, there are still areas for which management
has little control over market conditions such as, available work, number of
109
other competitors and the risk exposure. Other factors include government
policies and regulations, available technology and environmental factors.
Measurement at corporate level serves a fundamental purpose either for
strategy formulation or as part of a strategic action (Edum-Fotwe 1995).
Measurement at this level is aimed at long term-action.
Productivity measurements at project and crew levels have both immediate

and long-term objectives. Schedule control, cost control, target setting, and
motivating the workforce are some of the short term objectives. Employees
need productivity data as a feedback on their performance and may be used
for pay bargaining. The same actions will lead to the provision of a
performance database for planning and evaluating performance at this level.
Performance at this level provides management with information that can
influence their strategic actions. This research concentrated on productivity
measurement at crew level mainly because:
" production takes place at crew level;

" crew level productivity is easier to study;
" it is less clouded with other uncontrollable factors in comparison to other
levels;
" it can lead to both short and long-term productivity improvement actions;
and
" modern performance improvement philosophies such as TQM recognise
the importance of understanding the basic operational processes in a
production system.
A wider discussion of these concepts has been presented in Chapter 2.

Productivity performance at industry level has been included in the study to
gain an appreciation of a wider picture.
4.4.2 General obstacles to effective measurement
In spite of the general understanding of the importance of measuring

productivity, the task has always been a difficult one (Pilcher 1992). Before
discussing the techniques that are currently in use, it is perhaps better to
examine some of the barriers to effective measurement. Ireland (1990),
110
provided a summary of the major causes of difficulties in measuring
productivity in the construction industry as:
Fragmented nature of the industry

The diversity of the skills and resources required by the industry (or to
successfully complete a project) is so great. For this reason, most
construction projects recruit many specialist organisations to meet the level
of the skill required. This makes information flow poor and performance
measurement difficult.
Cyclical nature of the industry

It is commonly acknowledged that construction firms are more highly
geared than their counterparts in other industries. This made them
particularly vulnerable to the boom-burst cycle. Fluctuations in the market
cause construction firms to reduce margins and trade with inadequate
amounts of working capital. Both boom-bust and fluctuation of the market
considerably affect productivity levels.
Diversity in organisational sizes and structures

Construction industry comprises organisations of varying sizes, these
include the biggest multi-million firms to the one person trade
subcontractor. There is a wide variation of the way in which data are both
interpreted and reported..
Ambiguous nature of inputs and outputs

Standard measures of productivity are insufficient to capture the complex
nature of construction industry's inputs and outputs. For example
determining contribution of hired plants on the inputs. Outputs of
construction industry are in most cases not standard. For example, there is
nothing like a standard house. Even similar houses may have different
types of materials and quality.
Lack of agreed methodology

There are a wide variety of productivity definitions and measurement
techniques in use. This is because no single technique can measure
productivity effectively. Productivity has already been described as a ratio
111
of some measure of input to some measure of output. By this definition it
means that productivity measurement will involve determining inputs and
output in a given production process. Constraints exist in measuring both,
inputs and outputs.
Problems of measuring input
Total productivity has been found to be the best measure of productivity

because it takes into account all inputs. Determination of the contribution of
each input is a difficult task. The main problem is the measurement of
contributions of inputs such as labour effort, skill, time, machine, equipment,
technology change, organisation and managerial expertise in such a way that
their contribution can be aggregated. As a result of these difficulties,
productivity measurement has been performed by using a simple measure of
labour productivity. Using a single measure like labour productivity, ignores
the contribution of other inputs and could be misleading if there is a variation in
the contribution of different inputs. For example, if there is an increase in the
level of mechanisation in a production process, labour productivity is likely to
improve as a result of decreased manpower input to achieve the same output.
Research into increased labour productivity as a result of capital substitution
(or technology substitution) is still at its infancy in the construction industry.
The change in construction technology has enabled the increased use of
semi-finished inputs in construction processes such as prefabricated units (or
modular construction). Their use reduces construction time and labour
requirements. This results in increased output per person. The measurement
of site crew productivity is likely to increase not because it has improved but
because some of the activities have been shifted elsewhere due to a change
in technology. This makes it difficult to determine the actual contribution of all
inputs. Apportionment of inputs which are shared by several processes
present some difficulties as well. Contributions of activities such as research
and development in an organisation are also difficult to quantify.
Problems of measuring outputs
Products of several industries are non standard. For example, in the

construction industry, products range from a small building to very huge
112
structures such as bridges or dams. Even a well defined basic unit, such as
placing a cubic metre, is influenced by numerous factors, which cannot all be
quantified in a standard manner. Quantifying quality or level of workmanship
is another serious problem. For example, production times for laying a square
metre of brickwork varies with the level of workmanship.
Despite the problems of both definition and measurement, attempts have been
made to develop some measurement techniques, some of which have evolved
from work study techniques. These are discussed in the following section
4.5 Work study
4.5.1 Brief history
Manufacturing became the early centre of attention for productivity

improvement in Europe and America during the 18th and the 19th centuries as
a result of rapid industrialisation. Currie (1977) and Drewin (1985) provided a
comprehensive account of the general historical developments in the field of
scientific management and especially productivity improvement through work
study. It is clear that the emergence of this field was closely associated with
the process of industrialisation, except in a few cases where great thinkers like
Leonardo da Vinci, a leader of thought in many fields got involved surprisingly
in the field of construction (Brown 1969). Brown described a precise account
of both work method and output in shovelling loose earth as written by
Leonardo da Vinci. Serious scientific studies to productivity improvement have
only been conducted over the last 100 years.
Frederick Winslow Taylor (1856-1915), who is often referred to as the father of

scientific management is widely recognised as the pioneer of scientific
management. Taylor (1929) promoted the basic management responsibilities of
planning, organising, and controlling with the objective of standardisation to
improve human efficiency in order to achieve the maximum output with minimum
input. He proposed a number of techniques of scientific management, including
time study, standardisation of tools and methods, bonus payments and
differential payments, in recognition of their potential impacts on worker
productivity. Most of Taylor's work was based in the manufacturing industry,
113
although some of the activities studied were very similar to some construction
activities, for example, he made a detailed time and motion study of 600 workers
shovelling pig iron onto railroad cars at Midvale Steel Inc., Pennsylvania in the
United States in 1911. He showed that by varying the load and the size of the
shovel, daily output per worker could be increased (Christian and Hachey
1995). As a result of this classic study, the cost of labour was reduced by 50 per
cent (Drewin 1985). Taylor's studies established that worker output standard
could be increased by between 50 - 300 per cent justifying a 30 - 60 per cent
wage increase. Indeed this could be considered as an early example of
establishing a worker productivity benchmark, that is the best output that could
be achieved by adopting better methods identified at the time. The success of
Taylor's work inspired many others to study and develop new techniques of
monitoring productivity by scientific means. An early example of productivity
studies in the field of construction was pioneered by Frank Bunker Gilbreth
(1868-1924) who together with his wife Lilian Moller Gilbreth (1878-1972)
concentrated on the application of motion study first on bricklaying (Gilbreth
1911). Gilbreth studied bricklaying methods and developed an improved
method that reduced the number of motions from 18 to 4.5 per brick. As a result,
350 bricks could be laid in one labour-hour instead of the then customary 120,
an increase of nearly 200 per cent (Drewin 1985). This illustrated the early
recognition of the method of work as a significant factor that influences
productivity. In effect, Gilberth's studies were also attempts to establish
productivity benchmarks for the specific activity.
Taylor and Gilbreth laid an important foundation on what was then known as
time and motion study which has been widely applied in manufacturing. The
applications of these techniques have expanded to what is basically now
known as work study (ILO 1979). Despite Gilberth's pioneering work in
construction, most of his important findings found ready application in the
manufacturing industry. Construction received attention much later and in fact
serious attention to productivity studies was only seen in the second half of the
twentieth century (Parker and Oglesby 1972).,
114
4.5.2 Objectives of work study
Work study is aimed at developing the most economical method of performing

a specified task (Price 1986). The present notion of work study is defined in
the 'British Standard Glossary of Terms' used in work study and organisation
and methods (BS 3138: 1992 pp. 3), as:
'The systematic examination of activities in order to improve the effective

use of human and other resources'
Work study is subdivided into method study which is the technique used to
record work procedures, to provide systems of analysis, to develop
improvements and work measurement, or time study which is the
measurement of the time required to perform a task so that an output of
production for a worker or machine may be established (Harris and McCaffer
1995). ILO (1979) stressed that method study is the principle technique for
reducing the work involved primarily by eliminating unnecessary movement on
the part of materials or operatives. Work measurement, on the other hand, is
aimed at investigating, reducing and subsequently eliminating ineffective work,
(i.e. the time during which no effective work is being performed). Both aim at
increased productivity. Figure 4.4 illustrates the concept of work study.
Work Study
Method study Work measurement

To improve methods of production To assess human effectiveness
Giving improved planning and Providing output data

control and better use of: giving improved:
Materials Estimating;
Plant and equipment Production planning;
Manpower Incentive scheme
Higher productivity
Figure 4.4: Work study (Harris and McCaffer 1995, pp. 45)
115
Details for performing method study have been described by various authors
including the British Institute of Management (1956), Larkin (1969), Currie
(1977), ILO (1979), Barnes (1980), Harris and McCaffer (1995). Various
recording techniques such as flow diagrams, process charts, string diagrams,
multiple activity charts, foreman delay surveys, time-lapse recording, and video
recording techniques are well described in Harris and McCaffer (1995). Of
these techniques, only the foreman delay survey technique used by
Borcherding and Tucker (1977) and Easton and Woodhead (1981) and the
video recording technique developed by Harris and McCaffer have their
origins in construction. In this research, emphasis was placed on establishing
construction worker productivity on construction sites utilising the work
measurement technique. A detailed examination of this technique as a
research tool and in construction practice is made below.
4.5.3 measurement (time study)

-Work
The British Standard BS3138 (1992, pp. 3) defines work measurement as:
'The application of techniques designed to establish the time for a

qualified worker to carry out a specified job at a defined rate of working.
This definition raises two main problems:
" the definition of a qualified worker is not clear leading to subjectivity in

judgement; and
" the definition of rate of working although expressed in quantitative terms, it
is also clouded with subjectivity.
The concept of standard rating was introduced to reduce subjectivity. This is

the average rate at which qualified workers will naturally work at a job
provided they adhere to the specified method and are motivated to apply
themselves to their work. In conducting time study, the practitioner must have
a concept of standard rating in mind. This can only come from experience in
judging different speeds of movement, consistency and effort. The standard
rating leads to a number of other ratios such as: basic time; standard time;
standard performance; relaxation allowance; and contingency.
116
Rating and allowances continue to be controversial aspects in productivity
studies (Thomas et at. 1984) as productivity studies are not exact sciences.
Rating and the allowances to be given for recovery from fatigue and other
purposes are still largely matters of judgement and therefore fairly subjective
despite the various studies and guidelines that exist. Where such subjectivity
exists, it is likely to have significant influence on the study results. It is then
best to utilise other methods. The main difficulty in using time study
techniques for labour productivity studies in developing countries is the lack of
work study experiences not only in construction but also in manufacturing.
Such studies would call for a method which requires little experience,
excludes subjective judgements and can be applied widely with a minimum of
training and resources. Activity sampling is one such study procedure.
4.6 Activity sampling
4.6.1 Background
Activity sampling is a method of determining the percentage occurrence of a

certain activity by statistical sampling and random observations (ILO 1979).
This method was first introduced by Tippett (1935) who used it to make
observations in the British textile industry. Activity sampling is also known as:
" ratio survey;

" observation ratio;
" snap reading method;
" random observation method; and
" work sampling.
The activity sampling procedure, in its simplest form, involves making

observations at random intervals of one or more operators or machines and
noting whether they are working or idle. The percentage of the day that the
worker spends idle is the ratio of the number of idle tally marks to the total
number of idle and working tally marks. If the sample taken is large enough, its
characteristics will have an insignificant difference from the characteristics of
the group. The percentage of time that the operatives or machines are
117
engaged in a productive activity is known as activity rate. This method can be
used to establish crew and equipment activity rates for the whole project, a
segment of a project or an activity within a project. A physical measure of
output achieved during the observation period would enable the determination
of a crew's productivity.
4.6.2 Theory
The activity sampling concept is based on the notion that a working day can be
subdivided into two major parts: productive and unproductive time. To make
an estimation of the percentage productive time for the crew based on a
sample, a systematic statistical estimation procedure is required. First, the
outcome of an observation may be either 'productive' denoted by p, or
'unproductive' denoted by q. Further, the proportions p and q add to unity
(i.e. q=1- p). A frequency distribution diagram corresponding to p and q
based on a large number of observations is approximately normal. In order to
be confident that the observed proportions are within specified limits of
accuracy (i. e. L per cent of the true proportions p per cent of the time), a
minimum number of observations has to be taken. The following procedure is
adopted.
The standard deviation a of the distribution is expressed by the relationship:
P( 1n ')
=p= where n is the size of the sample.
If the proportion 'p, the acceptable error size V, and the confidence is set at
say, 95 per cent, the corresponding sample size can be calculated by using the
standard normal variable corresponding to the selected level of confidence.
The standard normal variable 'z', at 95 per cent confidence level, Z95= 1.96 as
illustrated in Figure 4.5.
(i + L) - PI,
At 95 per
er cent confidence level, Z = 1.96
a
but (pN,
+ L) - p= L, Iimit of accuracy.
118
2 5%
Figure 4.5: Normal distribution for activity sampling application
The limit of accuracy is the percentage within which the calculated proportion
'p' is required to be. If, for example, the accuracy is to be within 5 per cent,
then, L=5 per cent.
Therefore 1.96 a=5; thus, a=2.5 (approximately).
If we go back to the standard deviation equation to obtain n:
2.5- 35x65
n
6.25n = 35 x 65
n=364
A sample size N of at least 364 is required if the specified confidence level and
accuracy criteria are to be achieved. In general, the following formula can be
used to calculate the required sample size and achieve the specified
119
confidence level and accuracy criteria (Larkin 1969; Currie 1978 and Harris
and McCaffer 1995).
N= z2xPx(1-P)
L2
Where: N= number of observations required;
P= activity rate observed (usually assessed from pilot study;
L= limit (in percentage) of accuracy required; and
Z= standard normal variable depending on the level of confidence.
In practice, other time classifications may be of interest in an activity sampling

study. An example of how the construction working day could be broken down
along with illustrating some of the classifications is shown in Figure 4.6.
WORKING DAY
ABSENCE TIME ATTENDANCE TIME
OFFICIAL RELAXATION TIME SUPERVISION TIME
INSTRUCTION
BREAK TIME II RESTING
UNPRODUCTIVETIME PRODUCTIVETIME
I UNRELATED
IIWIGII
IDLE TIME WORKING TIME
INTERNAL II EXTERNAL APPORTIONED II

ANCILLARY
DELAYS DELAYS WORK
EXTRA EARLY QUITS 8

RELAXATION LATE STARTS
Figure 4.6: Classification of construction working day (Olomolaiye

1988, pp. 87)
120
4.6.3 Activity sampling procedure
Before making actual observations, it is important that the objective is clearly

known as this will influence the level of detail of observations. The actual
procedure can be summarised in the following steps:
" determine the objective of the study and select the activity to be observed;
" make a preliminary observation to determine the approximate values of p
and q;
" in terms of chosen confidence level and accuracy range, determine n (the
number of observations needed);
" determine the frequency of observations, using random tables;
" design recording sheets to meet the objectives of the study;
" conduct the observations; and
" summarise and analyse the results.
Most authors recommend that the operatives being observed should be

informed of the study. Earlier work study observations based on observing
secretly led to hostile reactions from the workers and is now considered
unethical (Currie 1977). The procedure adopted in this research is described
in Chapter 6.
4.6.4 Activity sampling in construction
Work study has most of its roots in the building construction. Early
observations made by Leonardo da Vinci and extensive work done by the
Gilbreths (1911) were all in construction operations. Despite these early
successes, construction has only shown tentative interest in work study. Wide
applications of work study in construction have been hindered by:
" temporary nature of construction projects and associated tasks which make
data collection for planning purposes difficult;
" many variables affect construction operations which make the data
collection difficult; and
121
" construction operations environment differ so much, such that the value of
data collected is low.
A full discussion of these was made by Price (1986). Despite the difficulties,
Harris and McCaffer (1995) argue that work study can play a valuable role in
construction and,` in particular, the method study technique provides a
systematic procedure for recording information and its subsequent analysis.
Investigations conducted by Price established that only a few contractors
actively used work study. He further investigated 14 organisations, which used
work study. These included contractors, local authorities, and others whose
work was of similar nature to that of construction. Five out of nine
organisations investigated in detail, had utilised activity sampling technique in
their data collection. He concluded that:
" similar work study techniques were being used by various organisations to
collect data; and
" work study was mainly used to determine output rates for planning,
estimating, and bonus schemes and thus involved work measurement as
opposed to method study. This was contrary to the Harris and McCaffer's
(1995) views.
Literature on construction work study is dominated by the use of activity

sampling as a research tool under the generic terms of activity and delay
models especially in 1980s. Activity sampling was used by Gregerman (1981),
Thomas (1981), Drewin (1982), Thomas and Daily (1983), Thomas et al.
(1984) Loui and Borcherding (1986), Price (1986) Baxendale (1987),
Olomolaiye (1987), Horner et al. (1989), Handa and Talhouni (1989). The
technique has been used to perform productivity studies, establish labour
utilisation on sites, establish delays and identify their sources in construction
operations, and as a surrogate measure of productivity. The use of activity
sampling to establish the worker activity rate is based on the assumption that
reduction in delays will result in higher productivity. Although this assumption
had been verified by a number of researchers, its validity is questioned by
others. This dilemma is discussed below.
122
4.6.5 Activity sampling as a work measurement tool - the dilemma
Strictly speaking, activity sampling is not a labour productivity measurement

technique, but a labour activity measurement technique as suggested in the
definition. It is a technique for quantifying the time spent by construction or
maintenance craftsmen in certain predetermined categories and provides
valuable information on areas of low activity (Parker and Oglesby 1972). The
objective of establishing the level of activity is to enable identification of
opportunities for increasing productivity through better utilisation of time.
Activity rate is not synonymous with productivity and a person who looks busy
may not necessarily be productive (Parker and Oglesby 1972). When activity
sampling is used to quantify output, it is a productivity measurement tool.
Several studies have been performed on the basic assumption that a

decrease in unproductive time in the construction process has an influence on
productivity (Thomas et al. 1982; Thomas et al. 1984; Liou and Borcherding
1986; Price 1986 and Baxendale 1987). Some models suggested that by
using labour activity rate, labour productivity predictions could be made
through work sampling (Handa and Abdalla 1989; Liou 1984 and Liou and
Borcherding 1986). This relationship seems logical although arguments
against it have been pointed out in recent studies (Thomas 1991). Thomas
(1991) examined data from 159 work sampling studies from more than 30
nuclear power plants constructed between 1973 and 1985 in USA. These
studies were conducted in the form of general site wide tours. Direct work was
established on the basis of these tours. The study concluded that there was a
very weak correlation between direct work and labour productivity. This can
be taken to imply that:
" productivity cannot be predicted by activity rate; and

" improved labour utilisation, through better productive time, would not
necessarily result in increased productivity.
These two arguments, first advanced by Thomas (1991) and later adopted by
various other researchers including Horner and Talhouni (1991), have
probably diminished the value of activity sampling as a work measurement tool
123
in construction. A careful examination of work done by Thomas (1991)
revealed the following.
" The productivity definitions adopted were unconventional. Productivity was

expressed in form of a performance factor (PF) defined as the ratio between
the estimated and the actual unit rate (workhours per unit). In other words,
estimated output in workhours per unit was divided by achieved output to
obtain productivity. Firstly, uniformity in the estimated output across all the
thirty projects has to be ensured because two similar outputs from different
projects could result in two different performance factors depending on the
estimate. For example, an actual output of say one hour per unit will result
in a performance factor of 120 per cent and 80 per cent for an estimated
output of 1.2 hours per unit and 0.8 hour per unit respectively. It is unlikely
that all the thirty projects had the same estimate given that they were
constructed over a period of 13 years and most likely by different
contractors. Secondly, uniformity has to be maintained across all the
trades studied for the same reason for the activity sampling results to be
combined. This is particularly difficult. An evaluation of an estimating
manual for building work in the UK revealed considerable differences even
for the same trade. In general, estimated output is very unreliable for use in
this form of analysis. In the same paper, the concept of earned work hours
is used. This has limitations similar to the estimated output discussed
above.
" Thomas (1991) argued that direct work percentages on nuclear projects
decline as projects progress from start to finish. This argument was based
on an equation with a correlation determination, r2 of only 17 per cent. This
relationship was used as a basis for comparisons between productivity
(measured on the basis described above), and direct work. The
relationship between direct work and productivity was therefore dismissed
because the trends suggested in the equation were not supported. This
conclusion is based on a very weak relationship as shown by the r2value.
The two methodological factors on this work cast doubt on its value: firstly,
because it is illogical; and secondly, the long established relationship between
productivity and motivation is brought into disrepute because increased
124
motivation primarily leads to increased operative effort which has a direct
relationship with productive time (Olomolaiye 1988; Maloney 1991). To
conclude that there should be no correlation at all between productive time
and productivity raises a fundamental issue related to how construction
processes are executed, that is construction effort is generally not directed
towards productive activities.
It is fundamentally acceptable that direct work alone cannot produce a strong

correlation with productivity, other factors such as skill, ability, tools and
knowledge specifically related to work being executed have some influence on
productivity. A number of other productivity enhancing factors unrelated to
productive time have to be incorporated into the relationship for a strong
correlation. This aspect is further discussed in Chapter 9, where it has been
demonstrated that under some circumstances, productive time is a significant
factor that influences productivity.
Horner (1992) argued that because activity sampling represents a snapshot of

the work process, it fails to distinguish between unproductive time arising out
of long and short delays. He asserted that productivity was adversely affected
by interruptions longer than 15 minutes, while interruptions shorter than 15
minutes have no such effect. He concluded that the percentage of time spent
at work determined from activity sampling studies cannot therefore be used as
a surrogate for productivity. Whereas this statement of conclusion is
agreeable, the argument advanced does not necessarily lead to this
conclusion. If delays of less than fifteen minutes were encountered in the
study, productive time would not necessarily be a surrogate measure of
productivity for reason given above. Indeed, if these findings are to be upheld,
then an activity sampling study, which would be able to distinguish between
long and short delays, could be applied to counter this problem. An
observation interval of not more than fifteen minutes (as is the case in this
research) would largely solve this problem.
However, it should be mentioned that contradictions are prevalent at project

level productivity forecasting through activity sampling as is evident from the
studies Thomas et al. (1991). The approach to work sampling in these studies
is therefore different from the approach in other studies (Price 1986;
125
Olomolaiye 1988 and Parker et al. 1987). Relationships established in earlier
studies where the scope of work is narrowly defined, definitions of direct work
are restrictive, the process of producing output is very elementary, and detailed
measurement of output is possible, are still valid (Thomas et al. 1984). Despite
the dilemma, activity sampling is still a powerful tool in establishing labour
utilisation on sites when the scope of work is narrowly defined. This research
utilises activity sampling to observe crew level operations where the activities
are well defined and output achieved during the elapsed study time is
quantified. Further justification for the use of activity sampling is presented in
Chapter 5 which presents the methodology adopted in this research. A
detailed activity sampling study procedure is discussed in Chapter 6.
4.7 Factors that affect construction, labour productivity
4.7.1 Significance of the study of productivity factors in

construction
The factors that influence construction labour productivity have been the
subject of inquiry by many researchers. Justifications for continuous
productivity improvement have been discussed in this research. In order to
improve productivity, a study of the factors that affect it, whether positively or
negatively is necessary. Making use of the factors that have a positive effect,
and eliminating (or controlling) factors that have a negative effect will improve
productivity. If all factors influencing productivity are known, it would also be
possible to forecast productivity as demonstrated by Smith and Thomas
(1990). Several factors have been identified. However, research findings
encountered in the literature refer to studies performed in developed
economies where construction conditions may not be the same as in
developing countries where this research was based. Nevertheless,
experience in these studies provides a valuable resource from which this study
can tap. The approaches used to classify these factors and the identification of
relevant factors in the Tanzanian construction industry will be discussed. It will
be demonstrated that, in spite of the many years of research, there is no
agreement as yet on both a standardised approach towards the study of these
factors to enable wider comparisons and transferability of research results, nor
126
is there a universal agreement as to the key factors and their impacts on labour
productivity. Studies have tended to address the diversity of problems in
specific situations limiting the generality of research findings. This research is
specific to the building construction industry in Tanzania where economic
conditions, technology and the general level of construction performance and
expectations may be different from developed countries. Research findings
elsewhere provide a general set of factors that could be investigated.
Literature reviewed in this section serves this purpose. The factors identified
were tested for their validity by soliciting the views of key practitioners, that is
contractors and contractors' operatives, in the Tanzanian construction industry.
Methods used and the results of the survey are reported in Chapters 6 and 9
respectively.
There is no evidence in the literature of credible studies with regard to

construction productivity factors in sub-Saharan Africa. Construction industries
in the economies of post-colonial African states have had no significant
efficiency improvements, despite efforts at national level (Wells 1986; Ofori
1990). Construction productivity, in Tanzania and Kenya for example, has
continued to decline in spite of these efforts, as illustrated in Chapter 7 of this
research. Implementation of broad based improvement policies without an in-
depth analysis of root causes of poor performance have been partly to blame
(Ofori 1991). The application of scientific approaches to the study of
productivity at crew level is an approach aimed at analysing some of the root
causes of poor performance. A cautious approach was adopted in this work, in
view of the differences in the economic and technological advancements in
Tanzania and in the developed economies where most previous studies of
labour productivity factors have been based. The following methodological
steps were thus adopted in this work.
"A general literature review was performed to identify factors that have a
significant influence on construction productivity.
" These factors were classified within a framework that can be understood and
applied within the Tanzanian context.
Key practitioners in the Tanzanian construction industry were consulted of the
relevance of these factors and the extent to which in the practitioners' opinions
127
they had an influence on worker productivity. The consultations addressed
major actors in the industry (i.e. contractors, and operatives).
The extent of their influence on productivity was then explored through actual
field studies.
The first two steps are described in the following sections. The third and fourth
steps are described in Chapters 5,6 and 9.
4.7.2 Classifications of productivity factors
In spite of extensive research into factors that affect construction labour

productivity over the last 30 years, there appears to be no consensus with regard
to both factors, their classifications, or the extent of their influence on labour
productivity. A general literature review suggests that most studies have been
tailored for specific situations that make comparisons of findings difficult. In spite
of this, there are some common factors which emerge from the studies some of
which could apply across most construction situations. These are investigated in
the following sections.
Several researchers have investigated the factors that influence labour

productivity. These include United Nations (1965), Gates and Scarpa (1972),
Paulson (1975), Robson (1978), Burch (1979), Borcherding et. al. (1980),
Borcherding et al. (1981), Drewin (1982), The Business Roundtable (1982),
Maloney (1983), O'Connor (1985), Thomas et. al. (1986), Lemna et al. (1986),
Price (1986), Olomolaiye et al. (1987), Thomas and Yiakoumis (1987), Briscoe
(1988), Harris and McCaffer (1989), Ireland (1990), Homer and Talhouni (1990)
Osman and Griffith (1990), Smith and Thomas (1990), Price (1992) Halligan and
Demsetz (1994). In spite of such intensive investigations, researchers have not
agreed on a universal set of factors with significant influence on productivity nor
has there been agreement on the classifications of these factors. Several
approaches to the classification of the factors affecting construction productivity
can be identified.
A United Nations (1965) report stated that, in ordinary situations, there are two
major factors affecting site labour-productivity requirements: organisational
continuity and executional continuity. Organisational continuity encompasses
128
physical components of work, specification requirements, design details, and so
forth. It is sometimes called work content. Executional continuity relates to work
environment and how well a job is organised and managed. Management
aspects include weather, material and equipment availability, congestion, and
out-of-sequence work. A conceptual representation of the findings of the UN
report (United Nations 1965) is shown in Figure 4.7. The flow of inputs and
outputs is comparable to a pipeline. The work to be done and work environment
categories are analogous to catalysts in the form of resources and conditions
needed to efficiently convert inputs (work-hours) to outputs (quantities). Where
these are not present, as can be the case when materials, equipment or
information are unavailable or where specifications are beyond the norm in the
industry, inputs cannot be converted efficiently to outputs.
Work environment
Q
Input Output
Work to be done
Figure 4.7: Conceptual representation of UN Report (United Nations

1965).
The model classifies factors into those related to work environment and those
related to the work to be done. The model has been adopted and extended by
some researchers to form the basis of an extensive research into factors that
influence productivity (Thomas et al. 1990; Gary and Thomas 1990; Sanders and
Thomas 1993 and Thomas and Sakarcan 1994). The model has been referred to
as the factor model and is shown in Figure 4.8.
129
Congestion Sequencing Weather Supervision Plant status
1 Work
Environment
Jr
Information Equipment Tools Materials Rework
Conversion Outputs
Inputs techonology (Quantities)
(work hours) (Workmethod)
} Work
Size o pecification Work content Design features Work scope
components and quality r to be done
Figure 4.8: Factor model (Thomas and Sakarcan 1994, pp. 230)
There have been several approaches to the classifications of productivity. The

United States Centre for Productivity classified the productivity at industry level
under the following headings (Kellog 1981):
" management;
" construction process;
" manpower;
" government regulations; and
" economy.
On the other hand, the Business Roundtable (1984) classified these factors under:
" construction technology;

" labour supply and training;
" regulations and codes and management; and
" labour effectiveness.
Briscoe (1988) attributed low construction industry productivity to a lack of long

and continuous production runs and inapplicability of economies of scale. He
cited these as the main factors at both site and corporate level. Other factors
include frequent changes in specifications, inefficient methods of building,
overmanning on sites, and poor management on and off-site.
130
Most previous research has focused at site level productivity. Hazelman (1981),
for example, grouped the factors at site level into:
" direct factors such as site, weather, supervision; size of the project; and
" indirect factors such as attitude of workers, density of workers, phase of the
project and work regulations.
Other forms of classifications at this level have been suggested by Abdel-Razek

(1987), Emsley et al. (1990) and Halligan et al. (1994). The criteria for different
forms of classifications is not clear although it seems to depend on the scope and
level of the study ranging from industry-wide at industry level to site specific at
micro-level or crew level. One of the more popular classification methods is for
factors that are:
" Internal - those within the control of the contracting firms management; and
" External - those outside the control of the construction firm.
This form of classification is easier to understand and it enables the

concentration of productivity efforts to those factors which are within the
influence of the management. This classification concept was adopted by
Drewin (1985) and Pilcher (1992) in describing the process and process
variables that influence construction productivity. Drewin's open conversion
process model for construction is illustrated in Figure 4.9. This form of
classification has been widely adopted by various researchers including
Olomolaiye et al. (1987), Omar and Griff ith (1990), and Pilcher (1992). This
general form of classification was also adopted by Olomolaiye (1987) who
categorises labour productivity factors into:
" Internal factors: originate in or around the firm, and

" External factors: are outside the control of a firms management.
The construction process model at craftsman level proposed by Sanvido (1988)

recognises that even when such a classification is adopted, it is difficult to isolate
between the various factors due to their interrelationships.
131
Feedback Feedback
Internal environment:
Input Organisationalstructure, supervision,
factors management, span of control, Output
1Labour work rules, etc. factors
Rework
Capital Products
Conversion Output and
n rgy Input 10-
projects
mechanism technology mechanism
Materials
Dis urbance Dice
Equipment Dis rbance
Exogeneousfactors such as:
DisturbaN
Weather, economic conditions, unions,
government regulations, public, etc.
Figure 4.9: Construction as an open conversion process (Dre ih 1985,pp. 4)
A refinement to this form of classification has been suggested by Duff et al. (1987)
as follows for application at site level studies:
" factors within site management control;

" factors within site management influence; and
" factors that site management has little control or no control over.
Classification of factors on the basis of the level of application in the construction

process has also been used. The hierarchy model developed by Kellog (1981)
provides a comprehensive general framework of developing construction factors
relationships at all levels in the construction process. This form of classification
enables factors at different level of the construction process to be tackled
appropriately. The common levels have been:
" industry - factors that affect productivity across the industry for example:
available level of technology; skill base; economic factors etc.;
132
" corporate - factors that are inherent in an organisation due to
management practices (i. e. company policy and procedures,
top management style etc);
" project - physical, management and environmental factors surrounding
a project environment; and
" crew - factors that relate directly to productivity at this level.
In order to appreciate that factors influencing productivity at different levels are

different, and that within each level there are processes that are affected by
internal and external factors. A combination of the last two classifications was
considered appropriate in this work. Productivity studies were performed at crew
level with the objective of identifying factors that are:
" within company and site management control;

" within company and site management influence; and
" factors outside the control and influence of company and site management.
This approach enables linkages of factors with both performance indicators and
responsibility for control or influence of the factor at various levels. This research
draws its motivation from concern for poor productivity performance in the
construction industry in Tanzania. The conceptual model linking benchmarking
and construction productivity at various levels in the construction industry, as
developed by Lema and Price (1994), maps the linkages.
4.7.3 Factors affecting construction labour productivity
The diversity of methodologies of quantification or evaluation of the factors makes

it difficult to compare findings. The following examples illustrate the diversity of
factors.
In his paper on construction labour cost, Paulson (1975) identified the following
factors that affect construction productivity in the construction industry:
" location and regional variations;

the learning curve;
" skills;
133
" work schedule;
" work rules;
" weather and environmental effects;
" experience of the craftsmen; and
" management factors such as job morale, safety and motivation.
Borcherding et al. (1980) listed the following factors that influence motivation and
productivity on large projects at crew level:
" materials availability;

" tools availability;
" rework;
" crew interference; and
" inspection delays.
Leonard et al. (1988) listed another set of factors by investigating effects of

change orders on 90 case studies. The identified causes of productivity loss
include:
" stop and go operations;

" out of sequence work;
" demotivation of workforce;
" loss in productive rhythm;
" loss in learning curve;
" unbalanced crews; and
" excessive labour fluctuations and unbalancing of successive operations.
To identify many of the influences that can affect worker productivity, Harris and
McCaffer (1989) utilised video recorded time studies to help recognise particular
problems on site. They identified these as:
" flow of operatives and materials;

" equipment utilisation and balance; and
" safety and working conditions.
134
During the 1990 CIB conference held in Sydney, Australia, a number of papers
dedicated to labour productivity were presented. There was a clear diversity of
views with respect to factors that affect productivity, their methodology of
evaluation and impact on productivity. The factors considered significant in a
selection of these papers have been presented below.
Horner and Talhouni (1990) identified the following factors affecting the
productivity of bricklaying gangs based on a study at seven different sites:
" delays;
" gang composition;
" length of a working day;
" day of the week; and
" temperature and humidity.
They concluded that productivity of a gang of bricklayers may vary by 200 per cent
from one day to the next and that the productivity of one gang may be 65 per cent
higher than that of another gang carrying out a similar activity under identical
situations. Factors identified in their research accounted for only 54 per cent of
the productivity variability observed.
Thomas et al.- (1990) studied masonry productivity similarities in seven countries

using a standardised procedure. Their study utilised the following productivity
disruptive factors that were considered to be significant:
" materials delay or suspension of work;

" size and organisation of materials storage area;
" materials handling and distribution;
" interference with other crews or congestion;
" out of sequence work or prerequisite work;
" rework or fabrication errors;
" accidents;
" equipment availability; and
" supervision and staffing.
135
The evaluation of the different factors under different conditions, even when the
procedure is standardised was reported as one of the major problems in this
study.
Osman and Griffiths (1990) studied the factors that affect labour utilisation on
construction sites on the assumption that an increase in labour utilisation levels
would result in increased productivity. The concentration on productive time was
in recognition of its influence on labour productivity. The authors listed a number
of research publications which had validated this assumption. These included
work done by Aird (1963), Howeinstein (1975), Sebastian and Borcherding
(1979), The Business Roundtable (1982), Thomas and Daily (1983), Stevens and
Grant (1984) and Baxendale (1987).
In her paper presented at CIB 90 conference, Emsley et al. (1990) suggested that
factors that affect masonry productivity are related to decisions made during
design and others are related to site organisations. The authors demonstrated
that masonry productivity was influenced by the element geometry by using a
simulation model.
Other productivity factors have been identified by Thomas et al. (1992) in their
study of masonry labour productivity comparison for various projects in seven
countries. These included:
" manpower and labour pool;

" total quantity of a task;
" design features;
" environmental conditions and weather;
" construction methods;
" project organisation;
" project features;
" management practices and control; and
" daily diary.
Sanders and Thomas (1993) identified the following major factors in their study of
masonry labour productivity:
136
" work type;
" physical elements;
" construction method; and
" design requirements.
Besides identification of productivity factors, few have attempted to quantify them

for practical applications. Price (1986) developed a site factor model aimed at
building up planning times from basic operation times by adjustment of basic
operation times to cater for various inefficiencies observed on site through activity
sampling. These factors were analysed with respect to pay. He established that
there was a high correlation between total site factor (a measure of site efficiency)
and hourly pay with a clear inverse relationship between site factors and
remuneration independent of trade. His work was based on an activity model
which was used to establish site factors dependent on the crew time utilisation as
quantified by the activity rate. This approach was based on the assumption that
increased labour utilisation results in increased productivity and is supported by
several researchers as shown above. This assumption was however challenged
by Thomas et al. (1990) who suggested that there was little evidence to suggest
that increased labour utilisation results in increased productivity in spite of his own
earlier finding suggesting otherwise (Thomas 1983). Price (1986) did not
however investigate the relationship between performance and other factors such
as site conditions, skill, other forms of incentives and level of direct supervision. It
would have been interesting to look at these relationships, a view also supported
by Olomolaiye (1987).
The non-standard approach to construction productivity evaluation led Thomas et

al. (1990) to develop a standardised approach in the form of a manual for both
quantifying productivity and evaluating factors. The manual provides detailed
guidelines for evaluating productivity in the form of eight data forms, seven of
which have to be filled everyday for each project being studied. The applicability
of the manual for this research was considered inappropriate due to its complexity
and relevance to the Tanzanian environment.
137
4.7.4 Summary of factors affecting productivity
The study of productivity factors in construction has demonstrated a diversity of

factors. The following is a summary of factors that have been identified.
Industry level factors:

" management;
" construction process;
" labour effectiveness;
" government regulations;
" economy;
" construction technology; and
" labour supply and training.
Project level factors:

" geographical location;
" project size;
" building type;
" expected construction activities in the construction area;
" climate;
" complexity of the job (i.e.-maintenance compared to new construction);
" material handling and construction methods;
" level of capital substitution;,
" management effectiveness;
" extent of overtime;
" safety;
" buildability (constructability);
" technology usage; and
" changes in specifications.
Crew level factors:

" number of disruptions;
" interruptions to the job;
" length of working day;
" gang composition;
" extent of training;
138
" crew motivation;
" acceleration of the work;
" payment;
" methods of employment (direct work compared to subcontracting);
" work rules and procedures;
" experience of operatives;
" safety;
" inspection delays;
" day of the week;
" supervision;
" unbalanced crews;
" design requirements; and
" physical elements.
It is clear from the above that the factors are both numerous and diverse. It is
nevertheless not exhaustive. It would be unreasonable to assume that it is
possible to quantify the effect of each on productivity in a predictable manner
since what may be significant in one environment, may be insignificant in another
situation. Nevertheless, the list provides a pool of factors that may be considered
for productivity studies. In this research, a list drawn from the crew productivity
factors was tested for validity in the Tanzanian construction environment. The
procedure adopted is described in Chapter 6 and the results are presented in
Chapter 9.
4.8 Summary
This chapter has examined the concept of productivity, its general definition,
the need for construction industry specific definitions and the significance of
labour productivity in the construction industry. Methodologies for construction
productivity measurement at industry, project and operative level and
associated problems have been discussed. Operative level productivity
evaluation has utilised work study methods. Work study has been extensively
139
reviewed focusing on activity sampling. Applicability of this technique in this
research has been analysed and justified.
This chapter has presented results of extensive literature review with respect to
factors that affect labour productivity. However, these refer to studies
performed in developed economies where construction conditions may not be
the same as in developing countries where this study is based. Nevertheless,
these results provide a valuable resources from which this study can tap. The
investigation demonstrated that in spite of many years of research, there is no
consensus on:
" standardised approach towards the study of labour productivity factors;

" key factors that influence productivity; and
" extent of productivity influence of the factors.
Studies have tended to address the diversity of problems in specific situations

limiting the generality of research findings. This has resulted in a long list of
productivity factors some of which may have no relevance in the Tanzanian
construction environment. Selected factors were tested for validity by soliciting
the views of key construction process owners, that is construction operatives
and contractors, in the Tanzanian construction industry. The methodology
used and the results of the survey are reported in Chapters 5,6 and 9.
140
CHAPTER FIVE
RESEARCH METHODOLOGY AND BENCHMARKING

FRAMEWORK
141
\
CHAPTER FIVE
RESEARCH METHODOLOGY AND BENCHMARKING FRAMEWORK
5.1. Introduction
This chapter is aimed at describing the methodology used in this research

and to formulate a benchmarking framework within which the research
process can be fitted. The need to set a clear methodology for the research
has come about as a result of the variety of approaches available. There is
indeed a wealth of information on the subject of research methodology,
which, if not evaluated carefully may be confusing (Buckley et al. 1976). The
methodology described in this chapter includes the description and
justification of processes used in identifying and solving the research
problem.
Review of literature on performance improvement concepts, benchmarking

and productivity was performed in Chapters 2,3 and 4. These issues were
discussed with respect to their application in construction. Benchmarking
was identified as a powerful performance improvement tool in a competitive
environment which has found acceptance in the construction industry. In this
chapter, the research problem is revisited with the objective of describing
how it fits within the benchmarking concept discussed in Chapter 3. This is
followed by the formulation of the labour productivity benchmarking
based on the generic benchmarking procedure developed by
-framework
Xerox (Camp 1989; Karsnia 1991). Various steps in the benchmarking
framework are then identified, described and linked. An evaluation of labour
productivity models is performed to identify suitable models in the key
benchmarking process steps. Activity models are identified as suitable
models in this work. The concept of benchmarking gap is then described as
applied in this work.
142
5.2 Research methodology
5.2.1 Literature on research theory and methodology
Several references were consulted with the objective of developing a

framework for the research methodology. These included Saslow (1982),
Weisberg et al. (1989), Pedhazur and Schmelkin (1992) and Buckley et al.
(1976). Guidelines suggested by Saslow and Weisberg were oriented
towards social science studies. Pedhazur and Buckley provide a general
and comprehensive approach to research design of a more general nature.
These form the main references for the research design in this work.
Because the research is based in a developing country, it was found
necessary to consult other publications specific on conducting studies in
developing countries. Casley and Lury (1981) concentrate on practical
problems of data collection in developing countries based on their
experiences in East Africa. This publication was valuable in this work.
5.2.2 General framework
Dictionary definitions of research include expressions such as: a careful

search; a systematic investigation towards increasing the sum of knowledge.
Luck et al. (1961) offered an operational definition which requires the
satisfaction of the following conditions:
" it is an orderly investigation of a defined problem;

" appropriate scientific methods be used;
" adequate and representative evidence be gathered;
" that unbiased logical reasoning be used in drawing conclusions on the
basis of evidence;
" demonstrate or prove the validity of reasonableness of the conclusions;
and
" cumulative results of the research in a given area yield general principles
or laws that may be applied with confidence under similar conditions in
future.
The need to conduct research in a scientific method does not necessarily

mean following a set of formal procedures, a detailed roadmap or an infallible
prescription for research. Rather, the scientific method refers to an attitude
and philosophy which guide the researcher to abstract essential elements
143
from the overwhelming phenomena and sensations in the environment, and
encourages an examination of the relationships among these elements.
Nevertheless, a useful general research methodology framework suggested
by Buckley et al. (1976) was found comprehensive and useful. This is
illustrated in Figure 5.1. The framework essentially involves problem
identification and problem solving through six processes, namely:
" problem genesis;

" problem definition;
" determination of mode of research;
" identification of the strategy for the research;
" selection of the domain of the research; and
" identification and application of the appropriate technique.
These processes have been discussed in the following sections using the
framework as they apply to this research. Discussions have been limited to
those parts of the processes which were considered relevant in this work as
illustrated in Figure 5.1 (shaded areas).
144
C
a)
E
m C C C C
J C O O 0 m O C)
Q E c m m
Er c
2 Z
cn N Cl) 4) a) O U
O 21 C In N (n U (Z (U
C D Co U
Q.
Z m -0
0 -0
0 0 O O

u) a
w
Z)
O
U
d
m 'a
w c
~ -0 ca
C. '
U E
Z)
N 'fn
C 0 t
T
0 0 N
Q E
U) c 0 O c 0) E
:c C) Cl)
2 a m E m
E) c E 5 z v
Er 21
O C) E E cz U)
2 .ro -0 c0) ro Nm
2 (IS
O (n OE 4)
7
cn N E U w L
CL
O i Em
>
o0
u3
aO) T

M. +
Q
O ca
N m
a ro m o
Z _
E
2 ro o > ro
j
> 04)

c
o
QU
c
a
OM
C: _ C
CD N V
W
to
E
o ar
C
cf: 0 cc O
w Q Q 00
,
cca
w
0 T
-0
-2 L
a,
LL
2
w J
m
0
a
-0
oc -Z:
3
r
w
J
mw E
OZ
cLw O
a LL
145
5.3 Research processes
5.3.1 Problem genesis
This research was performed due to concern for poor labour productivity in
the building construction industry in Tanzania. Informal procedures were
used to arrive at the research problem. Consensus and experience amongst
practitioners in the construction industry, as suggested by publications in
Tanzania, were used (Ministry of Works 1977; Ministry of Works 1992; and
Bjorklof et al. 1992). The process of arriving at the research problem has
been extensively discussed in Chapter 1. This problem was further
confirmed in the research process by an analysis of construction industry
performance at macro level which has been presented in Chapter 7.
5.3.2 Problem definition
The main objective of this research was to benchmark labour productivity in

the building construction industry in Tanzania. An extensive literature review
was performed in relation to the concept of productivity and previous studies
in the construction industry as shown in Chapter 4. A literature review led to
the identification of benchmarking as a technique that can be used to solve
the identified research problem. The objectives of the research have been
set out in Chapter 1. An extensive literature review on benchmarking has
been summarised in Chapter 3.
5.3.3 Research strategy
The following two main approaches to research were suggested by Buckley

et al. (1976).
" Deductive research guided by hypotheses which are either proved or

disproved during the course of the research. The use of deductive
research requires some prior knowledge upon which to construct the
hypothesis.
" Inductive research is guided by scientific inquiry and the researchers'

efforts are concentrated on procedures for obtaining and analysing data
and focus the analysis on the generation of a new theory.
146
This research is primarily deductive in the sense that it is not essentially
aimed at generating a new theory. Known theories have been tested in a
new environment.
Several research strategies are available as indicated in Figure 5.1 whether

the research is deductive or inductive. These include:
Opinion research: seeking of views, judgements or appraisals of other

persons (individuals or groups) with respect to a
research problem.
Empirical research: based on observation or experience by the

researcher through experimentation or field work.
Archival research: based on examination and analysis of recorded facts

in primary, secondary or physical form. Primary
sources are original documents while secondary
sources refer to published data based on summaries
or analyses of primary data. Official statistical data
published by, say the United nations is considered as
secondary data. Physical data include physical
observations such as in archaeology.
Analytical data: analytical research does not rely on any of the above
modes. The researcher relies on the use of internal
logic to solve the research problem.
This research has necessitatedthe use of several of the above methodsat

different stages. An intricate approach to the research was used in view of
the limited prior information and research reports in the Tanzanian
construction industry in comparison to the construction industries say in the
UK or the USA. These comprised different survey processes which together
constitute the methodology for the research. These are indicated in Figure
5.2.
147
Research problem: low
E labour productivity in
C)
M
0 construction industry in
aI- Tanzania
Opinion survey Empirical study I Archival study I
w.
c
C)
E Contractors
0. 51
0 EO Constructionsites Published statistics
0 eratives
C) 0 L-----i
>
C)
a
ID Formal observation
C) Compilation
a Interview technique - activity
0 and analysis
2 sampling
3 Labour productivity
factors
Labourproductivitydata
O
Labour productivity
benchmarks and
y
d factors
Opinion survey Empirical study
II Construction sites
Operatives
c
N
C
0
Ii Contractors
O Formal observation
U) Prominentpersons technique - activity
'D
sam lin
I
Interview and questionnaires

c
.c
V
Labour productivity data
Validated benchmarking results
Figure 5.2: Labour productivity benchmarking research methodology
148
5.3.4 Discussion of strategies used in the research
Opinion survey on labour productivity factors
The literature review, presented in Chapter 4 on factors effecting

construction labour productivity, led to the identification of a long list of
factors. Most of the literature was based on studies performed in developed
countries. These factors were then used as a basis for the identification of
relevant factors in the Tanzanian construction industry through an opinion
survey of contractors and their operatives. A structured interview technique
was used. The survey process has been described in detail in Chapter 6.
Empirical study - labour productivity observations
Labour productivity data collection was based on the use of empirical

strategy, that is, field observations using formal techniques. Activity
sampling is a formal procedure used for this purpose. A detailed discussion
of the justification of this technique and the general procedure has been
discussed in Chapter 4. The actual process of data collection has been
discussed in Chapter 6.
Archival strategy - construction industry performance
The analysis of the Tanzanian construction industry's performance was

aimed at confirming the concern for low productivity and relied heavily on
primary and secondary data. Data were obtained from files at the Central
Bureau of Statistics in Dar es Salaam in Tanzania. Other data were obtained
from United Nations, International Labour Organisation, World Bank, and
International Monetary Fund publications. These are used in Chapter 7 for
the construction industry performance analysis.
5.3.5 Reliability and validity of survey results
Buckley et al. (1976) further suggested that two aspects of research quality
have to be ensured:
Reliability: this refers to the ability to achieve the same results under
replication of the same research methodology in identical
circumstances; and
149
Validity: which refers to the criterion of representativeness, that is, the
findings should mirror the reality which they purport to represent.
These are important aspects of this research which have been achieved by
the use of an independent field study data as shown in Figure 5.2. The
actual procedure used is discussed in Chapter 8 and 9.
5.4 Special aspects of research in developing countries
There are a few references that address problems of research that rely on
field data collection in developing countries. Casley and Lury (1981) had
specific experience in East Africa where this research is based. Although
their experience was based on research in the field of agriculture, their work
was published for research applications across various fields. Some of their
findings were of a general nature that can be applied to this work. They
established that mailed questionnaires had limited success in developing
countries, and that direct objective observation was found to be more
successful coupled with intimate knowledge of the local language. They
cautioned against using sophisticated techniques which may be
unsuccessful due to logistical problems as well as staffing difficulties inherent
in developing countries. This research has mainly relied on actual labour
productivity observations in the field as suggested by Casley and Lury. In
some cases structured questionnaires were used for face-to-face interviews.
The main advantages of such interviews were found to be:
" the interviewer can:
" explain in detail the objectives of the research;

" elaborate any queries that the interviewee might have; and
" query the interviewee where an unclear response has been given;
" sometimes interviews can provide lines of further inquiry which might
have not been considered by the researcher.
" interviews are more detailed than questionnaires; and
" sometimes confidential data can be obtained through interviews.
The main disadvantage is that it is time consuming on the part of the

researcher who has to be physically available for the interviews. In
Tanzania, where telephone systems are unreliable, several physical visits
150
were made to some offices just to make an appointment, let alone conduct
the interviews. There is also a possibility of being side-tracked.
Casley and Lury (1981) emphasised the use of smaller samples to increase
the accuracy of the information obtained. It is better to stick to a small
sample and aim for thorough, detailed and accurate information or
measurement rather than survey a large sample and slacken on the
accuracy of the information collected. In particular, they stressed the
objective of reducing bias which may outweigh the sampling error. This is
important in developing countries where resources to achieve both large
samples and accurate information is not possible. Where data have to be
collected by others, training of surveyors, careful supervision coupled with
careful supervisory inspection and control have to be exercised. In this
research, a detailed training programme was devised for all those involved in
data collection, daily supervisory visits were made to all observation sites
and at the end of each week all data collected were jointly quality checked by
the observers and the researcher before handing it in. A more detailed
discussion of this is presented in Chapter 6.
5.5 The formulation of benchmarking framework
5.5.1 The research problem
The preceding section has set out a methodological framework for this
research. This section develops a logical sequence of the research
processes within the benchmarking framework. This study was motivated by
the need to improve the performance of the Tanzanian construction industry
as discussed in Chapter 1. National performance improvement efforts
focused on policy and organisation of the industry have been applied without
much success (Wells 1986 and Ofori 1990). Indicators of performance at
this level have been identified in Chapter 2. However, these are clouded
with many variables, some of which emanate from corporate, project, site,
and operative levels. Possible isolation of these variables can be achieved
through the analysis of the processes starting from operative level. This
research has focused on operative level process performance improvement.
However, industry level performance has to be analysed to confirm and
quantify the extent of the problem. Issues which needed to be addressed,
with respect to this, include:
151
" identification of key performance indicators at this level;
" evaluation of performance at macro level; and
" comparison of performance with other construction industries.
Labour productivity has been identified as a key construction performance

indicator. Its improvement has to focus at operative level because that is
where production occurs. Understanding of the construction process at
operative level will enable establishing productivity improvement potential.
Possible causes of low productivity can also be uncovered. Realising the
operative productivity potential can be achieved by adopting productivity
enhancing practices. The adoption of these practices can be motivated by
the need for competition driven performance improvement. In situations
where competition is not yet seen as the main performance driver, national
initiatives have to be taken to accelerate the performance improvement
process by creating a mechanism which facilitates learning from better
performers even in circumstances where the overall performance is poor.
This is typically the situation in Tanzania as argued in Chapter 1.
5.5.2 Approach to formulating a benchmarking framework
The research problem discussed in the preceding section fits closely within
the benchmarking philosophy where the objective is to improve, by adopting
a systematic procedure of learning from those with better performance as
discussed in Chapter 3. This philosophy has been expressed in a generic
model shown in Figure 5.3 (Lema and Price, 1995b).
Assessment of strengths and Assessment of competitor

weaknesses of internal operations strengths and weaknesses
I Analyse to isolate best practices and I

competitive qap
Incorporate best practices by copying, I

modifying /adapting and improving
Gainsuperiority
Figure 5.3: Generic benchmarking model (LemaandPrice1995b,pp.29)
152
Benchmarking programmes were initially implemented in companies facing
considerable competition. These have been known as competitive
benchmarking programmes. In circumstances where individual companies
cannot implement these programmes, benchmarking clubs have been
formed for the mutual benefit of the competing participants as stated in
Chapter 3. In some cases, governments have assisted in their formations
and also provided resources and some expertise to implement these
programmes (Construction Monitor 1994).
Benchmarking efforts, in this research, aim at enhancing the performance of

the national construction industry in Tanzania. Most construction companies
in the country would not be able to implement benchmarking programmes,
let alone be aware of it. The philosophy adopted in this work is divergent
from that adopted in the mainstream benchmarking practitioners in which the
initiative and the implementation of benchmarking emanate from within the
company. This research argues that it is more efficient to execute the bulk of
the benchmarking work centrally and disseminate results for implementation
by individual organisations in circumstances such as in Tanzania. The
research focuses on efforts aimed at identifying practices within the industry
that can be adopted to enhance productivity. Sustenance of the
benchmarking programme can be maintained through a monitoring
programme. This has to be reflected in the benchmarking framework
developed.
5.5.3 Description of labour productivity benchmarking framework
The framework formulation is based on the Xerox generic benchmarking

framework (Camp 1989 and Karsnia 1991) and incorporates some of the
ideas from the benchmarking process proposed by Shetty (1993). These
models have been discussed in Chapter 3. The framework for the labour
productivity benchmarking used in this research, as illustrated in Figure 5.4,
consists of three main stages: planning; analysis; and implementation. It
should be noted that, while the framework illustrates the implementation of
the benchmarking exercise, the scope of this research extends as far as
identifying the labour productivity benchmarks and productivity enhancing
practices. Some aspects of the implementation are proposed as a follow-up
to this work.
153
1: Confirmthe needfor benchmarking
Identify benchmarking subject

.c2:
_ N
U
2 C .C
Q)
C.)
EL 3: Identify benchmarking partners E
(potentially competitive) E
I-
W
Cl) 0
0
W
It 4: Determine data collection methodology
N and collect data z
0
F- Q
LL D
0
T3
W
5: Analyse to pinpoint gaps in performance c
c13
IL U) z
c
0
0 0
E z
6a: Determine best 6b: Determine best
performance practices a cc
O
Q
z
O
7: Set performance targets for continuous
improvement
CC
w
F-
CL
c
z
Cv
0 18: Communicate
Q)
0
0
J
and develop joint action plans r-.
.C
0
c 2
9: Implement plan to bridge the gap and

continuously monitor results
E
Figure 5.4: Labour productivity benchmarking framework
The above benchmarking framework describes the research process

achievable through a number of steps subdivided into three major stages,
namely planning, analysis, and implementation. These are described below.
The research extended as far as Step 7 of the framework in which the
performance targets are set and continuous sustenance and improvement
mechanism is proposed. The implementation stage of the benchmarking
framework involves actual adoption of the research findings in the industry
and is outside the scope of this research. Nevertheless, the general
procedure for implementing Steps 8 and 9 are briefly described. The
linkages indicate the iterative nature of the benchmarking process and also
the need for continuous monitoring aimed at sustained improvement.
154
5.5.3.1 Planning
The main objectives of this stage were to:
" identify the benchmarking subject;

" select benchmarking partners;
" identify information requirements; and
" collection of data.
These objectives were achieved in the following ways in this research:
Step 1: Confirmation of the need for benchmarking
Concern for low construction industry performance in Tanzania cannot

be taken for granted. This need has been verified by analysis of
output and related statistics over a considerable length of time to
establish trends. Performance comparison with selected national
construction industries has been be performed. The performance
evaluation framework developed in Chapter 2 was utilised. The
results of this step are presented in Chapter 7.
Step 2: Identify benchmarking subject
The main objective in this step was to identify the main performance
indicators, which would provide a significant performance
improvement when addressed. Labour productivity has been
identified, mainly through literature, as a key indicator in the
construction process. An extensive literature review, related to this
aspect, has been conducted in both Chapters 2 and 4. Confirmation
of the concern for labour productivity has been achieved through
interviews with construction industry executives Tanzania. Interview
guide and results have been reported in Chapters 6 and 10
respectively.
Step 3: Identify benchmarking partners
In a conventional competitive benchmarking exercise, willing

competitors in an industry agree to co-operate. In this research,
building contractors allowed access to their sites to contribute to the
155
pool of data used in the benchmarking exercise. Assistance was
obtained from the National Construction Council in Tanzania which
has legal obligation to control the construction industry.
Step 4: Determine data collection method and collect data
Key data in this research relate to labour output and variables or

practices that influence this output. The choice of techniques for data
collection was influenced by the objective of the research and the
productivity model to be utilised. This warranted an in-depth
examination of the available labour productivity models in order to
either select, adapt or develop an appropriate model. One of the
critical qualities of the model requirements was its ability to identify
and quantify labour productivity performance gaps. Labour
productivity models are evaluated in Section 5.6 of this chapter for this
purpose. The data collection procedure is reported in Chapter 6.
5.5.3.2 Analysis
The main objectives in this stage were to:
" determine performance gaps;

" identify best performance;
" identify best practices; and
" propose mechanism for sustenance and continuous improvement.
These objectives were achieved through Steps 5,6a, 6b and 7 included in

Figure 5.4 and described below:
Step 5: Analyse to pinpoint gaps in performance
Identification of a performance gap provides an opportunity for

improvement. The choice of the labour productivity model in Step 4
had a profound influence on this step. The quantification of the
performance gap has been achieved by analysing the process
performance variability. Thus, by evaluating the statistical distribution
of productivity, it is possible to identify and quantify opportunities for
improvement. This is the main thrust of this research and it warrants
further elaboration. The main philosophy is discussed in Section 5.8
156
of this chapter. The results of the performance gap analysis have
been validated by using opinion survey data which are presented in
Chapter 8.
Step 6a: Determination of best performance
Analysis of the performance gap described in Step 5 identified

performance variability. A procedure was devised to identify a
productivity level that would be considered as the best performance.
The procedure is described in Section 5.7. It is essentially based on
identifying a cut-off percentile level in the productivity distribution, the
basis of which has been justified. The validity of this model is
examined in Chapter 8.
Step 6b: Determination of practices that lead to best performance
In this research, best practices are defined as those significantly

contributing to improvement of performance, that is practices likely to
improve productivity. This task results in identifying factors that
enhance productivity. The complexity of such a task has been
discussed in Chapter 4. Several approaches have been employed, as
discussed below.
"A survey was conducted involving key practitioners in the

Tanzanian construction industry as to their perception of these
factors. The conduct of this survey is presented in Chapter 6. The
results of the survey are presented, validated and discussed in
Chapter 9.
" Labour productivity surveys conducted on construction sites

involved an assessment of the extent of productivity factors in each
site. A rating procedure was used. The conduct of this
assessment is presented in Chapter 6 and the results of the survey
are presented in Chapter 9. A further regression model was used
to investigate the relationship between the extent of these factors
and productivity.
157
Step 7: Set performance targets for continuous improvement
This step identified some of the performance improvement

programmes and describe a mechanism for its implementation. This
task was based on the argument presented in Chapter 3 in which
benchmarking was identified in a conceptual framework as a kick-
starting tool for the implementation of TQM. The labour productivity
targets were identified and validated in Chapter 8. The model
illustrated in Figure 8.8 - 8.10 represent a continuous productivity
improvement framework.
5.5.3.3 Implementation
The main objectives in this stage are to implement the results of the
benchmarking study to effect productivity improvement. This involved:
" communicating the results;

" developing action plans; and
" implementation of plans to bridge performance gap.
The approach for their implementation is described in Steps 8 and 9.
Step 8: Communicate and developed joint action plans
In practice, benchmarking results in the identification of practices that

have to be adopted in an organisation to improve performance. In this
research context, the responsibility for communicating and developing
ways for implementation of the results rests with the government by
involving the construction industry. Actual mechanisms for achieving
this are beyond the scope of this research, and would form a basis for
further research.
Step 9: Implementation of plans to bridge the gap and continuously

monitor results
Adoption of the identified best practices in the construction process is

essential for productivity improvement. This is the main objective of
benchmarking. This is not a one-off exercise. It requires continuous
monitoring for sustenance of continuous improvement. This could be
158
through a TQM programme as stated earlier. A conceptual model for
its implementation has been proposed in this work as a basis for
further research.
Some of the steps described above call for the use of procedures and
techniques that require further analysis and justification. The most important
of which are:
" selection of appropriatelabour productivitymodel; and

" discussion of the performance gap identification concept.
The remainder of this chapter discusses these issues.
5.6 Review of labour productivity models
5.6.1 Significance of productivity models
The use of models in productivity studies is common. Models have been

used to illustrate complex productivity relationships, explain certain
phenomena and forecast productivity. Much attention has been paid to
productivity models at national and industry level. Most of these have been
of interest to economists as mentioned in Chapter 4. Construction
researchers have been interested in productivity models at project, site and
operative levels focusing mainly on labour productivity for reasons mentioned
in Chapter 4. The objective here is to review these models so as to analyse
their suitability in this research. The review will be limited to operative level
labour productivity models.
5.6.2 Review and evaluation of labour productivity models
Several labour productivity models have been developed by various

researchers. These include Kellogg (1981), Maloney (1981), Drewin (1985),
Price and Harris (1985), Thomas and Yiakoumis (1987), and Abdul-Razek
(1987), Sanvido (1988), Pilcher (1992), Sanders and Thomas (1993), and
Halligan et al. (1994). Thomas et al. (1991) classified the models into three
groups: conceptual productivity models; work study based models; and
factor-based models. These are illustrated in Figure 5.5.
159
LABOUR PRODUCTIVITY MODELS
CONCEPTUAL WORK STUDY FACTOR BASED

MODELS BASED MODELS MODELS
Construction Expectancy
Delay model
process model model
Hierarchical
Activity model Factor model
model
Task model
Figure 5.5: Labour productivity models
Conceptual Models
These models illustrate the conceptual relationship between input and output
in the construction process. One of the simplest construction process
models is the closed conversion process model proposed by Drewin (1985),
in which all factors affecting the work are held constant except for the known
input and output as illustrated in Figure 5.6. Because of the steady state
nature of this model, the system more closely represents the industrial
production process than the construction process. Drewin (1985) proposed
a more appropriate construction process model at site level which he called
the 'open conversion process model', in which both internal and external
factors as well as undefined disturbances affecting the construction process
were illustrated. This model was illustrated in Figure 4.9 in Chapter 4.
Similar construction process models at operative level have been illustrated
by Sanvido (1988) and Pilcher (1992). All the three models illustrate the
complexity of the environment in which the construction process at operative
level takes place.
KNOWN CONTROLLED CONVERSION WITH KNOWN

INPUT ISOLATED ENVIRONMENT OUTPUT
Figure 5.6: Closed conversion construction process (Drewin,1985, pp. 5)
160
The hierarchical model proposed by Kellogg et al. (1981) identified
construction boundaries and identified discrete level of productivity, decision
making and measurement. This model serves little purpose for labour
productivity studies at operative level besides identifying relationships with
productivity at higher levels. Conceptual models play a significant role in
simplifying complex construction processes relationships, and provide a
basis for developing other types of models which can then be used for
evaluation and analysis.
Work study based models
Work study based models can be classified into the following groups as
indicated in Figure 5.5 and discussed below.
" delay models;

" activity models; and
" task model.
Delay models
These models originate from work measurement and are based on dividing
the total available work time into two major parts:
" net available work time; and

" major delay time.
The net available is further subdivided into productive time and minor delays
as shown in Figure 5.7. The time spent in each mode is recorded with a stop
watch. This model is best suited for closed system production processes
which have few external influences. They have been applied to equipment
intensive construction process such as hauling earth (Clemmens and
Willenbrock 1977). This limits the model applicability to typical construction
processes which are normally labour intensive. However, they serve as
useful diagnostic tools for further in-depth analysis using other models.
161
TOTAL AVAILABLE WORK TIME
NET AVAILABLE WORK TIME

MAJOR DELAY
TIME MINOR DELAY PRODUCTIVE
TIME TIME
Figure 5.7: Delay model time breakdown (Bernardet al., 1973)
Activity models
The models are based on work measurement using an activity sampling

approach. They measure the time spent in different pre-defined activities and
can be applied to most construction activities. The main objective is to
identify the proportion of unproductive time on site. The assumption is that
the reduction of unproductive time will make more time available for direct
work. The proportion of time spent productively, known as activity rate as
mentioned in Chapter 4, is assumed to be related to productivity. A typical
work sampling model of productivity is shown in Figure 5.8.
Breaks
Late start and 7%
early quits -xl
3%
Direct work
32%
Waiting and idle

28%
Instructions and
read drawings
8%
Travelling and Tools and
transporting materials 7%
15%
Figure 5.8: Graphical representation of the activity model
162
Activity models have been used extensively in the construction industry
mainly due to their simplicity. Their extensive use has been demonstrated in
Chapter 4. Recent research performed by Picard (1991), Christian and
Hachey (1995) were based on the activity model. This model identifies
possible scope for performance improvement through better labour
utilisation. This model is well suited for labour productivity benchmarking
because of its ability to identify a performance gap in labour utilisation.
However, this would be based on the assumption that improvement in labour
utilisation leads to increased productivity. This assumption would require
verification.
Task model
The task model extends both delay and activity models into a concept of site
factors (Price and Harris 1985). The concept of site factors recognises that
in the construction process time utilisation, there are certain activities that are
basic or necessary, others are additional but necessary, while there are
activities which are unnecessary. The concept of site factors was developed
by combining the assessment of time utilisation, work rate and allowing for
relaxation. This model is illustrated in Figure 5.9.
BASIC TIME INDEYL

A T. B.
DELAY
L .
FI_
W. B. W. B.
WORKTIME
TOTAL BASIC TIME RATE W. T.
F1=
T. B. T. B.
EXTERNAL A. T.
WORKING TIME DELAY RELAXATION F2 = "
W. T. W. T.
EXTRA
ATTENDANCE TIME BREAKS W. H.
F3 =
A. T.
A. T.
.
I OFFICIAL WD
WORKINGHOURS BREAKS F4=
W. H. W. H.
WORKING DAY
W. D.
Figure 5.9: Application of site factors to basic times (Price, 1992, pp. 188)
The working day is related to basic time, with the site factors, (F) for elements
of construction work in which the working day is a function of the various
factors as follows:
163
F (Total)
Working day = Total basic times
F2 F F1
where factors F1, F2, F3, and F4 are factors as illustrated Figure 5.9.
This model allows for the assessment of performance gap on a construction

site through the quantification of site factors. However, rating and
allowances to be given for recovery from fatigue and other purposes are still
largely matters of judgement and therefore fairly subjective despite the
various studies and guidelines that exist. Where such subjectivity is likely to
have significant influence on the study results, it is best to utilise other
methods less likely to incorporate subjectivity. This model is likely to
produce subjective results in this research due to a complete lack of work
study experiences not only in construction but also in manufacturing. A
model that requires little experience, excludes subjective judgements and
can be applied widely with minimum of training and resources would be
more suitable.
Factor models
The factor model is a multivariant approach to modelling crew level

productivity (Thomas and Yiakoumis 1987). Factors are quantified through a
statistical analysis of crew productivity and related factors expressed in a
mathematical relationship as follows:
mn
AUR, = IUR(q)+a, x, +f(y)1
where AUR = the actual (or predicted) crew productivity for the time period t
IUR = the ideal productivity for broad classifications of work
performed under standard conditions;
q= number of quantities installed that have an influence
productivity improvement;
a, =a constant representing an increase or decrease in
productivity caused by factor i;
Xi =a zero-one variable denoting the presence of the factor;
and
164
f(y)j = continuous variable sub models representing yfactors in
sub model j. As many as n sub models may be included.
There are two major setbacks in the application of this model in this research:
" The model does not state the ideal productivity level nor does it describe
the standard conditions under which ideal productivity is achieved. Even
when these have been defined, their in
applicability the research
environment would involve subjective judgements. This model relies on
related research findings in the same environment. With no productivity
records of any form in Tanzania, this approach would be unsuitable.
" The model requires a large database for a general application. This
model was developed for use in conjunction with the standard labour
productivity manual (Thomas et al. 1990) in which a number of records
have to be filled on seven different forms based on daily site visits. The
degree of detail achieved in this procedure is not necessary in this more
basic work. Besides, it would demand resources beyond those available.
Expectancy model
This model is based on expectancy theory which explains productivity

variations in terms of the effort a worker is willing to exert to complete a task.
Incentives can enhance or diminish worker effort. These are a function of job
conditions, management actions, relevance of the work, rewards and so
forth. Performance will be high if effort is exerted, the worker has adequate
knowledge, skills, and abilities, proper direction is given and constraints are
removed. Figure 5.10 illustrates this model by Maloney (1991).
KNOWLE
EFFORT VORGANISATION
PERFORMANCE
(MOTIVATION) CONSTRAINTS
ABILITY
Figure 5.10: Performance determinants (Maloney,1991,pp.405)
165
Maloney and McFillen have presented a model of worker performance and
reported research that validated the model within construction context
(Maloney and McFillen 1983,1986). The model identified four variables that
influence worker performance:
" worker motivation as evident from worker's effort;

" the degree to which the worker possesses the requisite job-specific
knowledge and skills;
9 the degree to which the worker possesses the innate mental and physical
abilities; and
" the effectiveness of management in organising the work and providing
the necessary resources.
The greater the incentive to the operatives, the greater their effort to perform
the task. A crew with a strong incentive to perform a task will spend less idle
time and take fewer breaks. During the time they work, they will work more
intensely than a crew with less incentive. The conceptual relationships
expressed in the model above had been partially validated. Research
conducted by Olomolaiye (1988) established a relationship between
motivation and productivity by analysing bricklayers' output. He established
that motivation is not directly related to productivity, it first influences the
percentage productive time and then the output. He established that
motivation accounts for 25.3 per cent of the variation in percentage
productive time. Good supervision was found to be the only significant
motivator. This finding has some significance in this research, as discussed
in Chapter 9.
He asserted that production output was further influenced by skill, that is the
ability to combine all the necessary productive motions to achieve a standard
output and it accounted for the main difference in output between bricklayers
over the same period of productive time.
The performance determinants model in Figure 5.9 provides a good

conceptual basis for analysing worker productivity with respect to factors that
enhance productivity. These determinants have been taken into account in
the search for productivity factors. An expanded version of this model is
contained in Thomas et al. (1990).
166
Other models
Several other types of models have been used for productivity studies.
Survey models described in the literature include the craftsman
questionnaire and the foreman delay surveys.
The craftsman questionnaire was used by Borcherding (1976) to establish

productive time on a construction site. The procedure, although simple, can
yield biased information depending on operatives perception of the purpose
of the survey. The limitation of the operatives to determine actual causes of
productivity loss may lead to incorrect information. If direct and simple
questions are asked, this method can prove very useful. Operative
questionnaires were extended to include own productivity assessment by
Olomolaiye (1987).
Foreman delay survey is a study in which a foreman is required to estimate

and document time losses at the end of each working day in the form of a
checklist. Causes of delays are also identified. Foremen are used in this
survey primarily because of their close contact with both operatives and
management and is best placed to identify a delay, its cause, and its
duration. This would enable management to focus on issues that have a
significant reduction on lost time. This method was used by Tucker et al.
(1982).
Other models include action-response model proposed by Halligan et al.

(1994). The model analyses cause and effect relationships that may lead to
a loss of productivity deriving its name from the notion that productivity or
lack of it is a crew response to external events and actions of others. The
model is essentially based on the factors model and does not advance new
ideas besides assigning productivity factors to different sources.
5.7 Choice of a suitable model
A number of labour productivity models have been described in the

preceding section. Suitability of models for this research focuses on the
following criteria, some of which were suggested by Lim (1996):
167
" measure accurately both input and output parameters used in the
productivity calculations which in this case are labour hours input and
work quantity completed;
" represent calculated productivity in a form that is useful to the researcher

so that results can be linked to the progress of an activity and reflect site
productivity;
" use simple and effective procedures without the need for special skills or
training;
" identify the major factors influencing productivity in an objective manner

and be capable of establishing cause and effect relationships between
individual factors and calculated productivity; and
0 relate productivity to special causes.
Additionally, for this research, the model should be able to:
" identify and quantify productivity improvement opportunity;

" focus on primary factors of labour productivity at operative level.
Table 5.1 evaluates, in general terms, the suitability of the various models for
application in this research.
One of the main criteria used to test the suitability of the model is its ability to
represent the construction process effectively. Delay, activity and task
models are able to represent the construction process at operative level in
quantitative terms. Of these, activity model is detailed enough to enable a
study of the construction production process by analysing the sub-activities
unlike the delay model which is too general. The task model requires
specific knowledge and experience not available in the research
environment.
168
Table 5.1: An evaluation of productivity models for benchmarking
Productivity Main objective Suitability for Ease of application

model benchmarking In this research
Conceptual Understanding the Not directly applicable. Applicable as
models construction production conceptual model
process. only.
Delay model Identification and Suitable but too general Can be applied
reduction of delays. and does not focus directly with ease.
on productivity.
Activity model Identification and Suitable and detailed Can be applied
reduction of enough but also does not effectively.
unproductivetime. focus on directlyon
productivity.
Task model Identification of site This is essentially a Specific
factors (site production benchmarking model, knowledge and
efficiency factors). requires specific training experience
for effective utilisation. required for
effective stud V.
Expectancy Identifies influence of Not suitable for Can be used as
model motivation on benchmarking but forms a support
productivity. good basis for conceptual model
identification of practices. in the study.
Factors model Quantification of Robust, requires a Resources
influence of factors on substantial data base, and required for
productivity. a number of sub-models. effective
Suitable for identifying application of the
practices. model prohibitive.
Foreman delay Identifies delays Suitable for preliminary Can be used as a
surveys through foremen. productivity survey. support model in
this study.
Craftsman Identifies productive Establishes a good basis Can be used to
questionnaire time and productivity for self assessment of verify
through operatives. operatives. benchmarking
process results.
An activity model is well suited for benchmarking with respect to the input
side of the productivity equation as well as the key aspects of the process.
However, the model does not capture the output. For its effective utilisation,
output achieved during the study period has to be quantified by direct
measurement. A record of productivity factors inherent in the process during
the study period would enable the analysis of the practices in relation to
output. An evaluation of the relationships can then be utilised to benchmark
labour productivity. This approach has been adopted in this work. The
hybrid, approach as shown in Figure 5.11, was therefore adopted.
169
Direct output
measurement
Activity model factors

evaluation
Relationships
Benchmarking
model
Figure 5.11: Hybrid labour productivity study model
5.8 Productivity performance gap concept
Productivity variability is one of the key characteristics of a production

process which can be utilised to identify opportunities for improvement.
Construction process performance variability is often used in process
simulations. Normally, this variability refers to a known process in a defined
environment. For example, a study of equipment productivity variability in an
earth moving operation. Such a study would reveal a performance gap
which could be utilised for improvement within the process and is limited to
the practices employed in the operation. If a study of several such
operations is performed, a wider performance gap is likely to be encountered
and may lead to uncovering other better practices. This is the essence of
benchmarking. In this study, productivity variability is utilised to identify the
performance gap. The basic model assumes that:
" construction processes resulting in the same type of output have some
form of productivity distribution;
" the distribution which defines the productivity variability provides an
opportunity for its improvement. This can be quantified by determining
the difference between the current mean productivity (CMP) and target
mean productivity (TMP), and
170
" to achieve the target mean productivity, it is necessary to adopt identified
practices.
Within these assumptions, labour productivity benchmarking is referred to as

the 'target mean labour productivity (TMP)' as illustrated in Figure 5.12.
TMP is an achievable productivity level based on current evaluation defined
at a percentile level of current distribution. This is necessary within the
benchmarking concept in order to provide an opportunity for operatives to
learn from best performers. For simplicity, Figure 5.12 is represented by an
assumed normal distribution, an assumption which has been tested in
Chapter 8. The proportion of operatives' productivity which is below TMP on
the current distribution curve can improve upon lessons drawn from a small
proportion whose productivity is above TMP (see shaded area on Figure
5.12).
current productivitydistribution
1- CMP 2= TMP
Labour productivity (output/labour hour)
Figure 5.12: Labour productivity benchmarking concept
The benchmark in this context is to serve as current asymptotic upper labour

productivity figure for which a productivity gap analysis can be made for
individual construction companies. Conceptually, the productivity gap is
equal to the difference between TMP and the CMP. Ideally, TMP cannot be
achieved because it is dynamic, that is, as productivity improves, TMP also
171
moves further to the right thereby providing a continuously moving
productivity target - an essential characteristic of benchmarking. This
research utilises this basic concept to identify productivity benchmarks in the
following procedure which is implemented in Chapter 8.
" The normality of distribution of productivity data is tested.

" The benchmarking gap is quantified on the basis of the concept
discussed above.
Benchmarking concept validity is tested by comparison with operatives
opinion data.
5.9 Summary
This chapter has set out the methodology upon which this research was
based. Several research strategies have been discussed followed by a
justification of those adopted for this work. Special aspects of conducting
such research in developing countries have been identified from literature.
The research problem has been revisited in the light of literature review
performed in Chapter 2,3, and 4. The need to evaluate productivity at
industry level has been identified. A labour productivity benchmarking
framework was formulated around the research problem utilising literature
review findings. The various steps in the benchmarking framework have
been identified and described. The procedures for accomplishing each of
the steps have been outlined and fitted within the thesis framework. Labour
productivity models have been reviewed with the objective of identifying a
suitable benchmarking model for this research. Activity models have been
justified. -A productivity gap quantification model has been proposed as a
basis for identifying benchmarks. Applicability of this model is tested and
validated in Chapter 8. In the following chapter, the procedures used for
data collection are described.
172
CHAPTER SIX
DATA COLLECTION
173
CHAPTER SIX
DATA COLLECTION
6.1 Introduction
The methodology described in Chapter 5 outlined the research strategy in

which several domains for data sources were set out. Field studies were
identified as the main source of data. Other sources of data included
individual interviews and secondary archival information. Chapter 5 further
outlined the labour productivity model that best met the research objectives as
set out in the benchmarking framework developed. Activity models combined
with direct output measurement and productivity factors offered a rational
approach to achieving the research objectives of labour productivity
benchmarking. The actual process of data collection was conducted in two
separate phases:
Phase I: Productivity data were collected between 1985 and 1990. This
essentially constituted productivity data collection based on site
observations with the main objective of establishing a Schedule of
Rates for Building Work in Tanzania, a separate project sponsored
by the National Construction Council (NCC) in Tanzania.
Phase II: Data collected during the period of this research. The data was
essentially collected in late 1993 and in 1995. This comprised:
0 operatives' and contractors' opinion research on labour

productivityfactors;
" labour productivity observations on construction sites;
" operatives opinion research on labour productivity levels;
" construction industry performance data based on archival
survey; and
0 opinion survey of prominent persons in the Tanzanian
construction industry.
This chapter describes data collection procedures in both phases and

discusses some of the practical problems encountered.
174
6.2 Labour productivity data collection process - Phase 1
6.2.1 Operative level productivity
This phase was primarily aimed at establishing a labour productivity data

base in Tanzania, as a separate project sponsored by NCC, as described in
Chapter 1. Current research benefited from the wealth of data collected in
this project, a task that would have been difficult to achieve in the research
in
period view of both time and resources limitations. Nevertheless, since
most of the data collected during this period forms an important input in this
work, a full description of the procedure has been described.
As mentioned in Chapter 1, it was clear that the only source of reliable labour
productivity data would have to be obtained through field observations on
construction sites. This was inevitably a slow process for a reasonable
quantity of data to be collected. This required a number of site observers. An
early decision was made to train groups of third year civil engineering
students at the University of Dar es Salaam. Site observations were
performed during breaks which were between three and four weeks long.
Groups of 15 - 20 students were engaged at any one time for the purpose. A
decision was made then to base initial studies in Dar es Salaam for a number
of reasons:
" Dar es Salaam is the main business centre in with a larger proportion of
construction activity in comparison to any other location in the country;
" it was easier to obtain sites where productivity observation was conducted
due to personal contacts, and also because most companies had their
headquartersin Dar es Salaam;
" the researcher as well as the observers were based in Dar es Salaam,
hence travel and accommodation costs were not incurred; and
9 data collection in one centre would facilitate comparison by limiting the

number of variables.
Before deploying the students on sites, they underwent intensive training

which not only equipped them with the procedures for activity sampling, but
also acted as a source of motivation to collect the data as accurately as
175
possible. The actual activity sampling activity is not particularly complicated,
and can actually be boring especially when the observation period is long. A
thorough knowledge of the background knowledge and objectives of the
study, was judged as a motivator for the observers.
6.2.2 Training and deployment of site observers
Observers were trained for a period of one week before they started the
actual site data collections. The training period was designed so as to equip
the observers with a background knowledge to productivity and the
significance of: labour productivity; methods for labour productivity
measurement; and theoretical and practical aspects work study in general
and of activity sampling in particular. Intensive classroom training took about
three days. This was followed by two trial activity sampling studies, one
hypothetical in the classroom, and the second an actual study on a variety of
general activities within the Faculty of Engineering, in Dar es Salaam
University. During the fifth day of the training week, the students observed
actual construction activities on a large construction site under close
supervision. This enabled them to identify and discuss problems related to
the actual site observations. In the last day of the week, general discussions
were held and final preparations were made including the issuing of
observation sheets, measuring tapes and stop watches.
During the training week, a parallel exercise of identifying active sites was
performed. Common knowledge, personal contacts, as well as visits to
known major contractors and major clients played an important role in this.
Activities were then identified from the relevant sites. This was a continuous
exercise during the observation period which normally lasted for about four
weeks. An allocation plan for observers was drawn up at the end of each
week.
Observers were deployed on construction sites early in the morning before

activities began so that they could identify operatives responsible for specific
tasks on site before the actual work began. Observers were introduced to the
site engineer and/or site foreman who was to be the contact person on site
during the period of the study. Observers were informed of the tasks for the
day and labour deployment on site. This enabled the observers to perform
their observations smoothly. It was agreed that the workers involved were to
be informed of this study by the site engineer or site foreman. Actual
176
recording started as soon as a task related activity was performed. The
actual recording procedure is discussed in the next section.
6.2.3 Activity descriptions
In order to maintain consistency in performing the site productivity

observations and interpreting the results, a standard for describing site
activities was adopted. An initial survey indicated that the most common
standard method of measurement for building work in Tanzania was The East
African Standard Method of Measurement for Building Works, published by
the East African Architects Association (1977), referred to simply as the SMM
in this study. The SMM is an adaptation of an earlier version of the Standard
Method of Measurement of Building of Works of the Royal Institute of
Chartered Surveyors (1968). Data were collected for a number of activities,
some of which was very limited. For example, brickwork was observed on
two sites only. In this research, only those activities with more than thirty
observations were included. This was based on the assumption of normality
of data, in which case, the sample sizes should be equal to or more than 30.
The following are the descriptions of the activities which met this criteria.
Blocklaying: 230 mm thick sand-cement block vertical wall with 10

mm thick sand-cement mortar.
Blockmaking: manual making of sand-cement blocks of general size
450 x 230 x 150 mm in steel moulds.
Excavation: manual excavation of loose soil to reduced levels < 2.00 m.
Formwork fixing: timber formwork fixing to horizontal soffits of slab
including connected beams.
Concreting: concreting in ground beams, 150 - 450 mm slabs and
connected beams. Concreting in this case involves
mixing ingredient materials, carrying and placing in the
final shape.
Plastering: 12 mm thick sand-cement plastering to vertical walls.
Painting: one coat of emulsion paint to vertical plastered walls.
Ceiling fixing: 10 - 20 mm thick ceiling boards 120 x 240 mm fixing to
horizontal frame.
Studies were conducted on active sites during the period of the study in Dar
es Salaam centre. The information collected using this form of convenience
sampling was intended to reflect the general state of affairs in the building
177
construction industry. It should be noted that blockmaking as such is not a
building item but forms a very important activity for most building sites and is
therefore included in this study.
6.2.4 Data collection procedure
The study aimed at collecting information on: output per labour hour for
different activities; labour utilisation; identification of crew sizes for different
operations; and general factors within the production environment.
Techniques such as time study, activity sampling, craftsman questionnaires
and foreman delay surveys were considered for this research. The
justification for using activity sampling has been discussed in detail in Chapter
4 and 5. Activity sampling has been used by Thomas (1981), Thomas and
Daily (1983), Thomas et al. (1984), Grant and Stevens (1982), Drewin (1985),
Liou and Borcherding (1986), Price (1986), Baxendale (1987) and Olomolaiye
et al. (1987). The advantages for activity sampling at crew level include:
" observation skills required in this technique are not as demanding as

those required to accomplish similar objectives using other techniques
such as time study;
" it provides information on the extent of labour utilisation on construction
sites;
" it is possible to identify in detail the sub-activities involved in the
productionprocess;and
0 it is possible to attach statistically quantifiable inferences to the results
of the study.
In order to be sure that all possible proportions of productive and

unproductive time were within a 5 per cent accuracy at a level of confidence
of 95 per cent, about 400 observations were required for each work sampling
study, obtained by maximising the equation for N, which was developed in
Chapter 4, as follows:
N=Z2(P-P2)
L2
dp Z2
For maximum N, = (1- 2P) =0
178
P=0.5
Substituting P=0.5, L=0.05 and Z= 2 into the equation gives:
N=22(0.5-0.52)_400
0.052
Thus, if N= 400 is used, all possible proportions of P would meet the

accuracy criteria. This approach was also used by Handa and Abdalla
(1989), and recommended by Harris and McCaffer (1995). Further to activity
sampling, quantification of work accomplished during the period of the study
was recorded through actual physical measurement. The following procedure
was adopted for conducting the site study to meet the study objectives.
1. Construction sites were visited and on-going activities that could be

studied were identified in consultation with site staff.
2. The main activities were broken down into sub-activities and
categorised into two groups namely, productive and unproductive
activity.
3. An appropriate study sheet with random observation intervals of
between one and five minutes was designed. A sample of the study
sheet is presented in Table 6.1.
4. Before starting the actual observation, the following had to be ensured:
" all persons involved in the activity were clearly identified and
their specific roles defined; and
" work completed prior to the start of the study noted as
accurately as possible. This enabled an accurate assessment
of work accomplished during the observation interval.
5. Observations were then made at pre-identified times. At any
instantaneous observation, a record was made of the sub-activity being
performed by each of the workers. A sample of observation record is
included in Table 6.1.
6. In order to meet the accuracy requirement criteria, 400 observations
were made for each activity study period. A typical observation period
of a concreting team of ten members took approximately two hours
assuming an average observation interval of about three minutes.
7. An accurate physical measure of the work accomplished during the
observation interval was recorded.
179
Table 6.1: Sample activity sampling study sheet
UNIVERSITY OF DAR ES SALAAM NATIONAL CONSTRUCTION COUNCIL

Labour Productivity Project
Data Collection sheet
1. Activity: 2. Location:
........................................:......... .................................................
3. Name of Firm: 4. Day and Time:
........................................ .........................................
Observation Productive activi Unproductive activit

Time (iii) Others
i (ii) iii) Others
9: 15
9: 16
9: 19
9: 20
9:24
9:27
9:29
9: 32
9: 37
9:40
9: 45
9:49
9:51
9:53
9:55
10:00
10:02
10:04
10:05
10: 10
10: 12
10: 13
10: 16
10:20
10:24
Sub-total
Sheet of
............ ................
180
Table 6.2: Sample activity sampling study summary sheet
UNIVERSITY OF DAR ES SALAAM NATIONAL CONSTRUCTION COUNCIL

Data Summary Sheet
1. Activity: 6. No of observations:.............
.........................................
2. Location: 7. Labour-hours committed:...........
.......................................
3. Name of Firm: 8. Volume of work done:..........
.................................
4. Period of Observation: 9. Observer:..........................
.....................
5. Day and Time:
.................................
Observation Productive act' 't Unproductive activi!X

Time
i ii iii Others i ii (iii) Others
Number of
observations
Percentage
observations
Number of labour
hours
Productivity Computation:
Volume of work done

1. Actual productivity:
Number productive manhours
Volume of work done

2. Overall productivity:
Number labourhours committed
Productive activities:
(i)
........................................................................................
(ii)
........................................................................................
........................................................................................
(iv) Others
............................................................................
Unproductive activities:
(I)
........................................................................................
(ii>
........................................................................................
(iii)
........................................................................................
(iv) Others
..............................................................................
181
8. Additional information such as weather conditions, method of payment,
general level of supervision, and general site conditions were recorded.
9. The data were then summarised in a separate sheet at the end of each
observation period. A sample summary sheet is presented in Table 6.2.
Further to the above data, some other factors were recorded during the data
collection. These included:
" payment;
" basis for payment, that is whether on piecework, daily, weekly, or monthly;
" other forms of payments and benefits;
" weather conditions;
" tools and equipment used;
"' gang sizes where appropriate; and
" level of supervision.
Since the emphasis in this phase was not on labour productivity factors, there
were no clear guidelines of how these should be evaluated and recorded.
The records of factors were largely both incomplete and inconsistent.
6.2.5 Data quality control
Besides observer training, a close monitoring and supervision was maintained

throughout the study. Random visits were made on all sites everyday to
ensure that the observers were in place. Random checks on data collected
were made during the site visits. Measurement of work performed was
sometimes cross-checked with site records or by talking to the site foreman.
Data collected during the week were discussed with the main researcher and
summarised every Saturday of that week. Data were checked for both
accuracy and consistency for each observer on an individual basis. Data
were then handed in, both in raw and summarised form as shown in Tables
6.1 and 6.2. The whole exercise took over 4000 observation labour hours in
Dar es Salaam centre alone. The data analysed in this work are only part of
the data collected. Other data could not be included because they were
mostly incomplete for full analysis to be possible. Summaries and analyses of
data collected in this phase are presented in Chapter 8.
182
6.3 Data collection - Phase 11
6.3.1 Scope for data collection
This phase of data collection was done during the period 1993-1995. Data
collection in this phase involved:
" construction industry performance data based on archival survey;

" opinion survey on labour productivity factors;
" labour productivity observations on construction sites;
" opinion research on labour productivity; and
" opinion survey of prominent persons in the Tanzanian construction
industry
The procedures for the collection of the data is discussed in the following
sections.
6.3.2 Tanzanian construction industry performance data
Improvement of macro economic performance of an industry in any country is

very important. Most actions at national level are normally geared towards
improving the performance at this level. This is also true for the construction
industry. Although the main focus for this research was at production level,
an analysis of performance at macro level was considered important because
this is the level at which policies can be influenced. The measures of
performance at this level have been identified in Chapter 2. The key
measures of performance include:
" contribution to the Gross Domestic Product (GDP);

" contribution to the Gross Fixed Capital Formation (GFCF);
" creation of employment; and
" labour productivity.
Various sources of data were used to compile relevant information for the
evaluation of construction performance. The sources of data are listed below:
" Yearbook of National Accounts Statistics published by the United

-
Nations although about three years late. The most recent copy in 1995
was the 1992 Yearbook.
183
" Central Bureau of Statistics - This is the custodian office for national
statistical data in Tanzania. During the period of the data collection
(February - April, 1995) the office was compiling national accounts
statistics for 1993.
" World Development Report - Published by the World Bank annually listing
economic and social indicators of most countries.
" International Financial Statistics - This is a monthly publication by the

International Monetary Fund (IMF) on financial statistics including
exchange rates of IMF member countries.
" Yearbook of Labour Statistics - published by International Labour

Organisation listing detailed employment statistics for most countries.
Data for some developing countries may be outdated. For example,
Tanzanian construction employment data is missing altogether in the last
10 series.
" Construction Industry Statistics Yearbook - published by United Nations,

but was discontinued in 1985.
Data from the above sources were compiled to enable an analysis of

construction industry performance in Tanzania and comparison with other
countries world-wide. The results and the data analysis are presented in
Chapter 7.
6.3.3 Opinion survey - labour productivity factors
A literature review of factors that affect labour productivity revealed a long list
of factors as shown in Chapter 4. The list included the following crew level
factors:
" number of disruptions;

" interruptions to the job;
" length of working day;
" gang composition;
" extent of training;
" crew motivation;
184
" skill;
" acceleration of the work;
" payment;
" methods of employment (direct work versus subcontracting);
" work rules and procedures;
" experience of operatives;
" safety;
" inspection delays;
" day of the week;
" supervision;
" unbalanced crews;
" design requirements; and
" physical elements.
These factors were related to studies performed mostly in the UK and USA and
their applicability in the Tanzanian environment could not be taken for granted.
Indeed, some of the factors that may not be considered significant in the western
world may have a profound influence on worker productivity in a developing
country. Such factors may include things like free meals on site, provision of
transport to and from site, provision of uniforms, medical facilities and basic tools.
Such facilities, when not provided by the employer, may have a significant impact
on the net income of the workers. This aspect is discussed later in this chapter.
Identification of the relevant factors in the Tanzanian construction industry was
therefore performed in two stages:
Stage I: Selection of relevant factors from a list of labour productivity factors

identified from literature.
Stage II: Rating of selected factors to obtain significant factors.
Stage I: Selection of relevant factors
A long list of factors identified from literature, as shown above, were discussed
with a selection of practitioners with the objective of shortening the list for a more
extensive survey. The selection process involved key process owners in the
185
building construction industry in Tanzania, that is, contractors, consultants and
operatives. The researcher was of the opinion that consultants were not
necessarily conversant with productivity factors at crew level. Three building
contractors were asked to select from the list obtained from literature sources
those factors which in their view influenced labour output at crew level in the
Tanzanian environment. They were also given the liberty of suggesting factors
which in their opinion should be included. The same procedure was used with
six operatives working for the three contractors. This led to the reduction of the
list of factors for the contractors and operatives as indicated below.
Contractors' views
In evaluating their operatives, contractors were of the opinion that the following
factors were important in influencing productivity:
" level of skill;

" wage level;
" level of mechanisation;
" monetary incentives;
" other forms of incentives;
" level of interaction and relationship between workers and site
management;
" future prospects and job security;
" weather and climatic conditions;
" time of the day;
" geographic location;
" work environment - whether worker feels intimidated by site supervisors;
" level of bureaucracy; and
" workers' physical abilities - level of fitness.
Operatives' views
When operatives were asked to propose a list of factors that would influence their
productivity, the following list was the result:
" wage level;

" monetary incentives;
" leadership;
" other forms of incentives;
186
" level of skill;
" level of bureaucracy;
" social environment
" weather, climatic conditions and location; and
" level of supervision.
The above factors then formed the basis of questionnaires for the two groups
(Appendix 1 (a), and (b)) with the objective of rating the factors on the basis of
the extent of influence on site labour productivity.
It should be noted that operatives' remuneration are included in the list of factors
in three different forms. These include wage level, monetary incentives and other
types of benefits. It is worth explaining here why the operatives' payments and
benefits are sub-divided into different categories in the questionnaires.
Wages - these are taxable wages which are generally very low. Minimum legal
wages are not taxed. Contractors therefore normally declare to the
authorities that they are paying their workers the minimum wage in
order to avoid paying taxes.
Monetary incentives - these include allowances for lunch, transport, housing,

tools, etc. These are normally not taxed because in principle they are
considered as compensations at cost, although they form a substantial
part of the wage accounting for as much as 70 - 80 per cent of the total
income. It is not unusual, therefore, for a contractors to pay their
operatives both the allowances and provide the benefits for which the
allowances are meant. This form of payment structure applies to most
ranks of employees both in the private and public sector (taxable wage
for a university lecturer for example is only about 15 - 20 per cent of the
total payment).
Other benefits - these may include free transport, accommodation, uniforms,

lunch allowances, basic medical facilities, tools allowance etc.
These were separated in this study because their influences on labour

productivity may be different.
187
Stage II - Ranking of productivity factors
The factors selected by the two groups formed the basis for an opinion survey
for individuals in each group. Two types of structured questionnaires were
used as a basis for interviews with contractors and contractors' employees
(see Appendix 1(a), and (b)). Individuals in each group were contacted either
physically or by telephone to request for their willingness to be interviewed. In
spite of a concentration of building contractors in Dar es Salaam where this
study was based, only a few contractors agreed to participate in the study.
Twenty seven contractors were contacted but only seven were finally
interviewed.
Fifty questionnaires were completed on the basis of structured interviews with

contractors' operatives. It would have been possible to interview more
operatives except for the fact that interviews with operatives took much
longer, and had to be conducted in Swahili - the national language (see
operatives questionnaire - Appendix 1(a)). A further 40 questionnaires were
completed by operatives for purposes of validation as discussed in Section
6.4.6.
The main objective of the survey was to uncover the perceptions of the main
actors on the extent to which pre-identified factors influence worker
productivity. The interviewees were asked to rate the factors from one,
(meaning the factor has no influence at all on productivity), to six (meaning
that the factor has a very high influence on productivity). A rating of 3-4
indicated a moderate influence. A typical rating matrix extracted from the
operatives questionnaire is shown in Table 6.3.
Similar rating matrices were used for contractors with factors drawn up by the
contractors' group. The results of this survey are summarised and analysed
in Chapter 9. The summary of the results include an evaluation of the total
factor scores which was then used to rank the importance of each of the
identified factors. Ranking of the factors has a significance in the assessment
of factors on an actual construction process. This is discussed in the
following section.
188
Table 6.3: Operatives' productivity factors ranking matrix
Note that: 1 means that the factor does not have any effect or it is irrelevant;
6 means that the factor an extremely high influence on labour productivity; and
3-4 means that the factor has a moderate influence.
Score
Factor 1 2 3 4 5 6
Wages
Monetary incentives
Other benefits
Leadership
Level of mechanisation
Level of skill
Level of bureaucracy
Social environment
Weather, climate and location
Level of supervision
6.3.4 Site labour productivity studies and productivity factors assessment
In Phase II of data collection, productivity studies were performed on various

construction sites focusing on concreting activity only. Concreting was
selected due to its relatively high contribution to the total cost of typical
building in Dar es Salaam (see Appendix 2 for analysis of cost breakdown of
buildings). Activity sampling, combined with the physical measurement of
work performed, was used as in Phase I utilising similar study forms (see
Table 6.1 and 6.2). A more systematic procedure for the evaluation of
productivity factors was introduced. This was possible after a survey of
productivity factors described in the previous section. The survey enabled a
more focused evaluation during actual construction operations. Again, a
nominal scale of one to six was used. A factor assessment of one on site
meant that the factor was evaluated extremely low. For example, if there
were no financial incentives provided, this was rated as'1'. On the other
hand, if wages were regarded as very high on the same site, this would be
rated '5' or even V. However, assessments were based on discussions with
189
the operatives and opinions of the observers. The forms to assess the
productivity factors are illustrated in Table 6.4.
Table 6.4: Site productivity factors assessment form

Productivity factors assessment form
1. Activity: 2. Location:...............................
....................................
3. Name of Firm: 4. Day and Time:.........................
...........................
Score
Factor 1 2 3 4 5 6
Wages
Monetary incentives
Other benefits
Level of skill
Leadership
Work organisation and layout
This form of observer rating has been frequently used before (Pedhazur and
Schmelkin 1991) and relies on the assumption that the human observer is a
good instrument of observation capable of some degree of precision and
some degree of objectivity. Validity and objectivity of the ratings may vary
depending on who does the ratings, within what framework of reference and
in what settings. This has raised some concerns, the most common of which
is a bias known as the'halo effect' (Pedhazur and Schmelkin 1991), in which
a constant error occurs when the raters general impression creates a bias.
Ratings may be consistently too low or too high (leniency or severity errors).
In all these, the strategies aimed at minimising raters bias are training in the
application of a specific scales used. In this research, bias was reduced
through training and using the same raters in the identification, discussion and
rating of factors.
The summary and analysis of the factors is presented in Chapter 9. An

evaluation of the relationship between the productivity observed and the
190
assessment of factors is also performed in Chapter 9 through a multiple
regression modelling process.
6.3.5 Opinion survey on labour productivity
One of the ways used to validate Phase I findings (that is labour output
observed) and in particular the established benchmarks, was to conduct an
opinion survey of those involved in the construction process. The survey
involved both operatives and contractors separately. Both surveys were
conducted concurrently using two separate techniques due to time and other
resources limitations. These were:
" postal questionnaires to 150 building contractors; and

" direct site interviews with construction operatives.
Since the studies were concurrent, a postal questionnaire to contractors was

judged as more efficient than interviews. This approach for the more elite
group had been recommended based on experiences in East Africa (Casley
and Lury 1981). The use of questionnaires was not practical for the
operatives, some of whom may not have been able to read or write. A
structured interview approach was therefore considered more practical.
Detailed procedures for the data collection are discussed in the following
sections.
Postal questionnaires
In the study of productivity factors involving contractors, a direct approach

was used. A very poor response rate was obtained (only seven out of 27
agreed to be interviewed). A different approach was used in this survey.
Instead of interviewing, postal questionnaires were used. Also, instead of
approaching contractors directly, an indirect approach using an authority
recognised by contractors was used. All building contractors are registered
by the National Board of Architects, Quantity Surveyors and Building
Contractors of Tanzania (NBAQSBC) board before they can operate. This
board is also responsible for annual evaluations of contractors for their
eligibility for promotion and continued registration. Permission was obtained
from the chairman of this board to use both their office and their letter head to
send out questionnaires in anticipation of a better response.
191
A list of registered building contractors was obtained from NBASQBC. A
selection of 150 contractors, either based in or known to be operating in Dar
es Salaam was selected. The list included contractors from different classes
of registration as shown in Table 6.5.
Table 6.5: Contractors' opinion survey list and class of registration
Registration I II III IV V VI VII

Class
Number of 22 11 17 38 35 15 11
contractors
Total registered 48 20 53 85 105 78 632
as of 1993
Percentage 46 55 32 45 33 19 2J-l
contacted
A questionnaire which was primarily aimed at obtaining information relating to

contractors' perception of worker productivity was designed. The design was
partly based on a previous questionnaire used by Olomolaiye (1987) for a
similar exercise in Nigeria.
Questionnaires were then attached to a letter from NBAQSBC to the

contractors and sent to 150 selected contractors. A copy of the questionnaire
with a cover letter to the contractors is included in Appendix 3. A stamped
envelope addressed to the NBAQSBC was enclosed with instructions for the
return of questionnaire. Contractors were asked to fill and return the
questionnaires within a period of one month. Contractors were asked to
indicate whether they would like to have a feedback of the results of the
study. In spite of these efforts, only ten questionnaires were returned two
weeks after the expiry of the date of return. The results of the survey are
analysed in Chapter 8.
6.3.6 Operatives structured interviews
As discussed in Section 6.3.1 above, this form of survey was inevitable for
this group. A questionnaire on which the interview was based is shown in
Appendix 4. The main objective was first to confirm the factors identified as
significant for productivity at crew level, and secondly to validate labour
192
productivity benchmarks observed on construction sites through direct
interviews with the operatives. Thirteen building sites were visited and 40
operatives were interviewed. The summary of the results of the interviews
and analysis are presented in Chapter 9.
6.3.7 Interviews with executives in the Tanzanian construction industry
Structured interviews were conducted with top executives in the Tanzanian

construction industry. These included:
" Current President of Institution of Engineers Tanzania and Chairman of

National Board of Architects, Quantity Surveyors and Building
Contractors;
" Chairman of National Construction Council and past president of the
Institution of Engineers Tanzania;
" Executive Secretary of the National Construction Council;
" Special consultant to the National Construction Council;
" Chairman, Architectural Association of Tanzania and Managing
Director, National Estates and Design Company;
" Managing Director of a leading local construction consultant;
" Managing Director of a leading local contractor;
" Former Chief Engineer, National Bank of Commerce; and
" Managing director of one of the oldest Architectural firms in Tanzania.
" Director, Consulting firm and past president of the Institution of
Engineers, Tanzania.
The interview was guided by a set of questions focusing on a number of

issues including:
" general perceptions as regards the scope of the construction industry;

" the general perception of the general performance and constraints;
" current policy and operational efforts for its improvement; and
" general discussions with regards to modern management concepts and
their applicability in the Tanzanian construction industry.
The interview guide is included in Appendix 5. In general, the interviews were

very successful lasting between two to six hours. The results of the
interviews have been incorporated in various sections of this work. In
particular, they were included in Chapter 2 and also used to justify the
193
direction and emphasis of this research in Chapter 1. They also facilitated the
identification of the future direction of the Tanzanian construction industry
development and research. This aspect is discussed in Chapter 10.
6.4 Summary
This chapter has presented the procedure used for the data collection in this
research. A number of strategies, domains and techniques have been used.
Most of the data were obtain through actual field observations. Activity
sampling coupled with direct output measurement was used. Some
information related to the performance of the Tanzanian construction industry
was obtained from archival and published sources. A list of possible labour
productivity factors was drawn on the basis of studies mainly in UK and USA.
The relevance of these factors was tested in Tanzania through a two-stage
process: confirmation and/or modification of factors through opinion survey;
and ranking of factors identified in the first stage through a second opinion
survey. Further field observations were performed to obtain data to confirm
the results of the initial study. This was followed by a second opinion survey
of productivity and productivity factors. The analysis of archival and primary
data related to the macro performance of the Tanzanian construction are
analysed in Chapter 7. Site operative productivity data analysis and
establishing productivity benchmarks have been performed in Chapter 8.
Chapter 9 presents the results of the opinion survey and relates them to
actual productivity.
194
CHAPTER SEVEN
TANZANIAN CONSTRUCTION INDUSTRY

PERFORMANCE
195
CHAPTER SEVEN
TANZANIAN CONSTRUCTION INDUSTRY PERFORMANCE
7.1 Introduction
This chapter aims at quantifying and evaluating the concern for the poor
performance of the Tanzanian construction industry. Macro performance of
the industry is analysed by using national accounts statistics on the basis of
the framework discussed in Chapter 2. The analysis is based on the
hypothesis that performance has generally been deteriorating. Construction
industry statistics over a 25 year period, from 1969 to 1993, have been used
to analyse the extent of deterioration of performance. A brief history of the
construction industry in Tanzania is examined followed by performance trends
of key measures. A performance comparison with selected national
industries has also been performed, followed by wider international
comparison. Emphasis has been placed on productivity comparisons with the
objective of obtaining the relative performance of the Tanzanian construction
industry.
7.2 Measures
The significance of the construction industry in the national economy has

been examined in some detail in Chapter 2. It was established that there is a
close relationship between construction activity and economic growth.
Consequently, the construction industry's performance has a significant effect
on the performance of the economy as a whole. The framework for the
evaluation of the construction industry's performance at national level has
been examined in some detail in Chapter 2, where the following were
identified as reliable performance indicators:
" contribution to the Gross Domestic Product (GDP);

" contribution to the Gross Fixed Capital Formation (GFCF);
" creation of employment; and
" labour productivity in terms of value added per person engaged, and gross
output per person engaged.
196
These indicators have been used to analyse the Tanzanian industry's macro
performance on the basis of both published data, and data obtained from the
Central Statistics Bureau (CSB) in Dar es Salaam, Tanzania. Trend analysis
has been based upon detailed output data for a 25 year period between 1969
to 1993. The indicators have also been used to compare the Tanzanian
construction industry performance with that of Kenya and the UK.
International performance comparison was made on the basis of value added
per person engaged in the construction industry.
7.3 Tanzanian construction industry performance
7.3.1 History
Construction activity is as old as mankind. It first appeared to address the

basic need for shelter followed by other needs such as water supply,
communication systems, defence structures, etc. Traditionally, the provision
of shelter in a community has generally been a shared community
responsibility in a typical Tanzanian village. Individuals requiring a traditional
house, normally young men, would be encouraged by elders to organise the
collection of the necessary materials. These would normally be wooden
poles, thatch, sometimes clay and stone. The would be owner of the house
would provide food and local beer to the those involved in the collection of
the materials. Once the materials are all assembled, a special day for
building the house would be set. The house would normally be completed
within a day involving both men and women. This had been the practice for
many generations before the commercialisation of the construction process.
No records are available of the productivity which was achieved. In general,
no payments are involved in the whole process except for the provision of
food and beer all of which would be consumed on the day. Houses in the
village would, therefore, be of similar type and quality. Other construction
related activities in the village like water supply and irrigation systems,
footpaths and bridges, etc. were also undertaken by the community under
the supervision of the elders. Masao (1974) record that, during the past 400
years, the Wachagga people of the Mount Kilimanjaro region, in northern
Tanzania, have developed an extensive irrigation system which allows them
to farm large areas of otherwise barren land. They dug furrows manually by
using small sticks to plot the course. These practices continue even today in
some rural areas where 79 per cent of the population (22.5 million out of a
total population of 28.5 million) still live (United Nations 1993).
197
Commercialisation of the building process emerged when the community
started having wealthier individuals who wanted to have better houses than
the ones normally provided by the community. They would be willing to pay
- initially in the form of food, goats, sheep or cattle for the building services
provided (Ninatubu 1995). Later on, as money became a medium of
exchange, it played a central role in the house building process. Gradually,
the community responsibility for the shelter provision diminished. In some
rural areas, it is no longer possible to build any form of housing without
monetary payment for both the materials and labour. The provision of
shelter in both the rural and urban areas is still mostly catered for through
the informal construction industry.
7.3.2 Emergence of formal construction industry
Colonisation brought into existence the formal construction industry. Early

major construction projects such as railways, government office buildings,
hospitals, churches and later roads are evidence of construction
achievements in the country. Iliffe (1980) record that towns, construction
industry and plantations were the Tanganyika's chief market for wage labour
from early 1900s to the time of independence. In 1913, wage labour stood
at 172,000, and in 1937 was 244,000 (compared with 333,755 in 1964 and
871,215 in 1993 (Central Statistics Bureau 1995)).
Tanganyikan construction industry statistics, for the period before

independence in 1961, are scarce. Serious efforts towards the development
of the industry by the government came much later. Chohan et al. (1982)
recorded that, after the second World War, the volume of construction as
well as productivity were at a low level, partly as a result of the war as well
as the introduction of the Economic Control Board, through which most
materials were imported including cement. Chohan et al. record that during
the period between 1949 - 1954 artisans migrated from overseas (mainly
India) to the then Tanganyika and familiarised themselves with the local
conditions and within the period realised that productivity could be
significantly increased. During the period 1954 - 1960, efficiency of
supervision was heightened and task work basis of construction was
generally practised. Productivity improved significantly during this period.
The establishment of trade schools in the country at the time contributed to
this increase in productivity.
198
7.3.3 Post independence construction industry
Tanganyika acquired its independence from the British in 1961, after a

period of more than 60 years of both German rule and British protectorate.
Literature on the performance of the construction industry, just after
independence is limited. There is, however, a consensus that there was a
slump in the construction industry following uncertainty in the political
atmosphere of the newly independent nation and its overall effect on the
economy (Bienefeld 1968; Chohan et al. 1982; Wells 1986). This is also
clearly demonstrated by output statistics of the industry over the first few
years of independence (1961 - 1964) in which there was no increase, even
in current prices, of the construction output. The slump in the construction
demand led to the general exodus of foreign companies along with its
experts (Wells 1986).
A fairly detailed study on the performance of the Tanzanian construction

industry between 1961 and 1968 was performed by Bienefeld (1968). He
observed that the market conditions were such that differences amounting to
as much as 30 per cent were observed in the actual value of buildings for
which the same price had been paid by the client. He concluded that such
substantial differences overwhelmed the small percentage gains which could
be obtained through technical improvements. The main implication of this
observation was that, before technical innovations could be utilised for
efficiency improvements in the building industry, it was necessary first to
establish some control mechanisms especially in pricing to bring it to a
realistic level. Bienefeld (1968) further described: the structure of the
industry based on published statistics; a discussion of the shortcomings of
these data; and, a discussion of the various separate effective markets and
their performance within the industry.
7.3.4 Construction output quantification
The responsibility for collection, analysis, and dissemination of construction

statistics in Tanzania lies with the CSB, as part of their wider responsibility
as custodians of the National Accounts Statistics. Bienefeld (1968)
estimated construction activity, using Gross Fixed Capital Formation (GFCF)
in current prices and the consumption of cement in the country. He
demonstrated that the construction industry had already undergone two
major phases of development between 1960 and 1968.
199
" Between 1961 -1964, there was a general slump represented by a near
zero growth over the entire period. This is exhibited by the GFCF. The
growth in cement consumption increased only marginally.
" There was a steady growth rate of about 11 - 14 per cent per annum
between 1964 and 1968. The growth started in 1963 when the
construction industry output was at its lowest over the 1960 - 1968
period. Again, both indicators of output show similar trends.
These trends are shown in Figure 7.1 below.
120
100
O
II
80
ei
c
60
40
O -' N !9 Ih lo OD
o 10
Year
Figure 7.1: Construction industry output in Tanzania - 1960 - 68

(Bienefeld 1968, pp. C iii )
The second major source of information regarding the size and growth of the
industry stems from the data on employment and earnings output
quantification. Bienefeld (1968) demonstrated that, in spite of the
incompleteness of this data, it still provides fairly good representative trends
due to the concentration of the industry, i.e. a survey of a small number of
large firms can give reasonably good indications.
200
One of the major conclusions arising out of the study was that the benefits of
competition were only felt in a few isolated areas of the market, with the best
performance coming in the building of warehouses and such simple
commercial structures, and in the erection of a large number of simple
repetitive structures. All the other areas were characterised by a lack of
competition, combined with high and fluctuating profit margins. Bienefeld
(1968) pointed out that, in spite of the lack of competition and high profit
margins, construction labour utilisation was still good mainly due to the
introduction of Minimum Wage Legislation in 1962 which led to a decrease in
employment figures resulting in better labour utilisation and hence higher
productivity. This observation was echoed by Chohan et al. (1982), who
observed that during the period between 1960 and 1967, the level of
productivity was not affected, in spite of an increase in the volume of work to
the contractors who decided to remain and continue to operate in
Tanganyika.
7.3.5 Construction output between 1969 and 1993.
Construction output statistics for the period between 1969 up to 1993 were
obtained from CSB in Dar es Salaam. This period of analysis was selected
because statistics prior to this period were incomplete. Partial analysis of
the data prior to 1969 was done by Bienefield (1968) as discussed in the
preceding paragraph. There is no reason to believe that the data for the
period 1969-1993 are not subject to the same problems of both accuracy
and completeness as those observed by Bienefeld (1968). Information
presented in this section was compiled from various sources within the CSB.
It should, therefore, be treated with caution. In spite of this, the data can be
used to evaluate trends in the industry due to the same reasons observed by
Bienefeld (1968), that is, data are normally obtained from a small number of
large firms who undertake most of the construction work. Output data are
likely to be under estimated. Discussions with the staff at CSB indicated that
not all questionnaires sent out were returned. However, since the main
focus in this work is labour productivity, organisations making annual
statistical returns on their output would include labour engaged in the work
done for the year. It could, therefore, be argued that productivity trends
calculated from this data are representative of the industry trends. Several
indicators of construction performance have been computed from this data.
Since there has been a high inflation rate in Tanzania over a large proportion
of the period of analysis, as well as several devaluations of the national
201
currency, the indicators have been computed on the basis of 1987 pounds
sterling. This approach enabled comparisons to be made later in this
chapter, with selected national construction industries. A summary of
construction output both in current and constant national prices is included in
Appendix 6 (a).
7.3.6 Performance trends analysis 1969 -1993
7.3.6.1 Analysis framework
The performance of the Tanzanian construction industry, at macro level,

focuses on the size of the industry as measured by the output, its
contribution to the economic growth, and to the investment in the country as
measured by the GFCF. An analysis of value added per capita indicates
whether the construction industry's contribution to the economy is in
consonant with the population growth. -The industry's contribution to the
creation of employment is also considered an important indicator of
performance. Analysis of both value added per person engaged in the
industry and output per person engaged are important indicators of
productivity trends. These indicators have been examined over the period
between 1969 and 1993 for which complete data was available.
7.3.6.2 Output trends
Value of construction
Construction output can be measured by quantifying the value of total output

of the construction industry in purchasers' prices. Total output of
construction is inclusive of all new work and maintenance. The limitations of
this form of quantification have been examined by Briscoe (1988). Output
can be quantified by using current prices or can be expressed in constant
prices. Quantification of output in constant prices is particularly necessary
when dealing with economies with high inflation rates as is the case in
Tanzania.
Value added by the construction industry is a better measure of the true

output of the industry as it quantifies the contribution of the industry in
creating value. Value added is quantified at purchasers' prices and is
therefore inclusive of profit. Value added by the construction industry in a
202
competitive environment would, therefore, seem lower than in an
uncompetitive environment where excessive profits are permissible. This is
indeed a setback of this measure.
Both total output and value added trends in this analysis are expressed in
constant sterling pounds using relevant exchange rates as published by both
the World Bank (1991) and the International Monetary Fund (1995). Figures
7.2 and 7.3 quantify the construction output in Tanzania over the period of
analysis. Output trends, both in terms of total value and value added by the
industry, expressed in current Tanzanian shillings (Tshs) price are shown in
Figure 7.2. It should be noted that due to excessive inflation, the output
expressed in current prices would seem to be increasing at an almost
exponential rate. A logarithmic scale has therefore been used. When the
value added figures are expressed in constant prices (Figure 7.3), a more
realistic picture emerges. When output is expressed in 1987 Tshs,
construction output would seem to have peaked in 1973. This observation is
consistent with observations of the Ministry of Works (1992). The output is
thereafter fairly constant, fluctuating between Tshs 6000 mil - 8000 mil, up to
1982 when a downtrend started. The trend beyond 1986 is confusing
because it indicates a rise up to about 1990. This trend is considered
unrealistic when compared to the contribution to the GDP and the actual
GDP growth over the same period. An analysis of productivity, expressed in
value added per construction employee, performed later in this chapter
would also indicate an unrealistic trend. It is not clear for the moment as to
the reasons for this. A simplistic explanation for this is the decrease in the
value of the national currency from 14 Tshs to the US $ in 1984 to
approximately 550 Tshs in 1993. A fuller analysis of reasons for the trends
is beyond the scope of this work. An analysis of output in 1987 pounds
sterling shows a more realistic picture. It is, however, difficult to ascertain
the trends between 1986 and 1993 although it is consistent with the
observations in the country. It is mainly for this reason that subsequent
trends analysis are based on constant sterling pounds.
203
1000000
-
100000 w
N - . 01
- "I
c) -.
0 .e
C_ 10000
. 4p
- 01
1--
C) 1000
0
0
0
5 loo
c
value added
C 10
--- total value
0
1
G' c - NM LA %O r- co aO-NM In -.O r- OD O' O- (V M
DD 1- f- r- N r- r- r- r- r- r- OO CO CO CD CO CO 00 CO OO CO C O' O' a'
a' a' a' O' O' a' a+ 0% a' O% O' a+ 0% O' C% O1 G% O' ON OA O% O' O' O' O'
-------------------------
Year
Figure 7.2: Construction output in Tanzania

(in current prices)
20000
\ Output in 987 pounds x10000
\
18000 . -1 %
\\ Outputin 1987Tshs x1000000
16000 A
-_\
14000
1
12000
1
10000
CL
0 8000
6000 _` , -- ""%
i
v
4000
2000
0
O. O r- N N) '7 Lf) %D N CO 0% O .-NM 'e V) %O N CO 0% O .-NM
%D 1- N 1- NNN 1N NNN 00 OO 00 CO QJ CO 00 CO CO 0TT 0' O%
O' OA 0% 0' 0% 0% T 0% 0% O% 0% C% C% O' G% 0' OA G% O% 0' c% T 0' 0' C%
Year
Figure 7.3: Construction output in 1987 constant prices
204
7.3.6.3 Contribution to the GDP and GFCF
The construction industry is an important contributor to the economic growth

of any nation. It was stated earlier in Chapter 2 that the percentage
contribution of the construction industry is closely related the GDP per capita
of a national economy. Typical values of the contributions for a country like
Tanzania with per capita income of between US $ 120 and $350 over the
period of analysis should be of the order of 3.6 per cent (Wells, 1986). The
average contribution over the period of analysis was about 4 per, cent with a
maximum of 5.6 per cent in 1971 and 1990, and a minimum of 2 per cent in
1983. These contributions are within the expected levels. The trends
between 1969 to 1993 are shown in Figure 7.4.
Similarly, the contribution of the industry to the GFCF for a Tanzania's

income group is expected to be around 56 per cent (Wells, 1986). Actual
contribution was about 60 per cent between 1971 and 1973, and fell to
below 25 per cent in 1985 and rose again to about 30 per cent in 1993.
Trends over the period of analysis are also shown in Figure 7.4. The
contribution of the industry to the GFCF has been below the expected level
since 1974.
70
G
60
13 GF:
CF
50
C
O
40
v 30
C
O
20
C
8 10
cM-a
U-U-. -. -. .. '
0 ^rNM -qt In (O n 00 MOrN CI) It U) CO N 00 OO T- NM

(O N P- ^NNnP. co MM co CO CO MMMM CD rn o) O
OO 0) C) O C) C) O) C) G) C) O O) O) p) C) C) CD C) C) p) (3) W O) 0)
rrrrrrrrrrrrrrrrrrrr.. rrr -.
Year
Figure 7.4: Construction contribution to the GDP and GFCF
205
7.3.6.4 Creation of employment
Construction typically employs about 3-8 per cent of total employment

(Wells, 1986). Tanzania's average of 8.9 per cent employment is higher
than countries in its income group. Construction employment in Tanzania
peaked at about 22.19 per cent in 1973, when construction output was at its
highest (see Figure 7.5). However, in terms of employment contribution to
the overall population, the industry is still far behind a number of countries.
Employment contribution in relation to the population is also shown in Figure
7.5. This is observed to be consistently below 5 per 1000, except for 1973
when it rose to about 7 per 1000. A comparative analysis of the trends with
Kenya and United Kingdom has been presented in the next section of this
chapter. Total employment of approximately 30 per thousand in Tanzania is
about the same as the employment rate of construction industry alone in UK,
the construction industry in UK employs 3 per cent of the population,
whereas in Tanzania only 3 per cent of the population is in paid employment
(see Appendix 6 (a)).
35
N'--E-M-m
W-, Ell-, -.. m

30
M-M / M-M.
-
.01
0
0 25
No
0
20 total employment
c
E 15 construction
0 employment
n
E 10
w
0 0) OrNM Lf) Cp n co 0) OrNM In (D f co OrN Ch

cf
1*- I- I- I- I- I-- I- r- OD OD 00 CO co 00 co co co 00 c) 0) 0)
0) 0) 0) 0) p) 0) 0) 0) 0) 0) 0) 0) O) 0) 0) 0) 0) 0) 0) a) 0M 0) O)
rrrrrrrrrr r' rrrrrrrrr r' rrrr
Year
Figure 7.5: Total and construction employment trends
206
7.3.6.5 Productivity trends
In Chapter 2, it was establishedthat constructionindustry productivityis

computed in two ways:
" output per person engaged per annum (OPE), that is value of total output
divided by the total number of persons engaged in the industry; and
" value added per person engaged per annum (VAPE), that is value added
by the construction industry in the national economy divided by the total
number of persons engaged in the construction industry.
Productivity in the Tanzanian construction industry has been expressed in

both measures over the period of analysis using constant prices. Again,
these are both expressed in 1987 Tshs and 1987 pounds sterling. Both
should provide a consistent trend, unless there is a radical change in the
construction technology that significantly affects the proportion of the
construction value added. Figure 7.6 shows the productivity trends using
both measures based on 1969 productivity.
500
Value added (pounds)
0 p Value added (Tshs)
- 400 Output (Tshs)
.
n -0 - Output (pounds)
o,
co
-' 300
d
200
r
4-
100
0
L.
IL
0
OONM to CO f 00 OO
tt V-- NM d' In Co 1* CO OONM
tD NN f*l NN f t 1- I- f- 00 CO co 00 co 00 co co CO CO rn (M O) O)
OOO O) O O) 00000000000 O) Q) O) C) C) OO O)
TTTrrTTT1T P" T T- T T- T- TTTrrr P" T- T
Year
Figure 7.6: Trends for output and value added per person
(in 1987 prices)
207
In 1973, the industry experienced low productivity during peak output. This
is consistent with similar observations in the UK (Olomolaiye, 1987). An
increase of more than 200 per cent between 1973 and 1977 may seem very
high, but it will be shown in Section 7.4 that actual output throughout the
analysis period is almost consistently lower than that of Kenya (see also
Appendix 6 (b)). The picture beyond 1986 is again confusing and a
productivity increase of up to about 500 per cent, based on 1987 Tshs prices
is unrealistic. The trends expressed in 1987 pounds is more consistent with
general observations (Ministry of Works, 1992). A comparative analysis of
performance trends with other national construction industries is performed
in Section 7.4.
7.4. Construction industry performance comparison
7.4.1 Scope of comparison
International construction industry performance comparison is not a straight

forward exercise, in spite of the existence of apparently comparative data.
While output may be expressed in national currencies convertible to a base
currency, actual products may differ significantly. Indeed, technology
differences significantly affect the construction process and therefore
performance. It is for this reason that performance comparisons should be
made for comparable products as well as construction technologies. This is
difficult, especially for developing countries where data are generally scarce.
In this research, the Tanzanian construction industry is compared to that of
Kenya, whose construction industry is closely comparable to that of
Tanzania both in its products and technology. A further comparative
dimension is added by including the UK mainly to add more meaning to the
indicators. Figures used for these comparisons have been extracted from
various sources and are summarised in Appendices 6 (a), 6 (b) and 6 (c).
7.4.2 Output comparison - value added comparison
Construction industry output can be expressed both in terms of total output

or in terms of its contribution to the economic growth - value added. The
later is used here as the basis of comparison because figures for the total
output for the Kenyan construction industry could not be obtained in
published sources. Value added is expressed in constant 1987 pounds
sterling for the three construction industries. Figure 7.7 shows the
208
construction industry percentage contributions to the GDP. These
percentages are converted into actual figures and trends presented in Figure
7.8, which show the construction industry value added for Tanzania, Kenya
and UK for the period between 1969 and 1993. UK output is generally about
ten times the output in both Kenya and Tanzania up to about 1985. Beyond
this time, output in UK increased to a maximum of sixty times the output in
Tanzania in 1989 and about fifteen times that of Kenya. Tanzanian
construction output has generally been going down after peaking in 1973.
Output in 1989 was only about a quarter of the output in 1973. Kenyan
construction output has also been decreasing but at a slower rate than that
of Tanzania. Output peaked in 1978. At its lowest level in 1993, output was
about 60 per cent of the 1978. One of the characteristics of the output is its
erratic nature which has had an adverse effect on the performance of the
industry. The erratic nature of the Tanzanian construction industry output is
particularly evident when total output trends for Tanzania and UK are
compared (Figure 7.9).
10
c8
zr_
a7
a
06
0
` 5
c
0
3
12
w3
C
0
v2
1
0
OOrN CM tt tl) CO I. OD OOrN CO c!' LA CO f- 00 O) O T- N CO
co r" I- r- r- N- n r. r. ti N. (0 oo co co co co co co co co 0) 0) 0) 0)
rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn rn
rrrrrrrrrrrrrrrrrrrrrrrrr
Year
Figure 7.7: Comparison of construction industry contribution to GDP
209
35000 I
Tanzania 0 Kenya -"- UK x 10
co 30000
25000
l0
20000

15000
0 10000
V
5000
Cl)
O) r Cl) U) f- d) T M U) f- Q) r
c0 f f f f N- CD Co OD OD 00 C) Q)
O) G) O) O) Q) Q) C) C) C) C) C) C) O)
r T T r T T T T r r T T
Year
Figure 7.8: Construction value added in 1987 constant prices
10000c
C
9000c
80000
7000C
C
6000C
c'. 5000C
4000C
3000C
2000C
v 1000C
OOTN (') %T LO (D I*- 00 OOTNM et to CO (- CO OOTN c0

CD NN fl. NNNN I_ NN 00 Co CO Co CD OD Co CD Co 00 O C7 OO
000000000 (3) 000000000 0) 0) 0) 0) 0) 0)
TTTTTTTTTTTTTTTTTTTTTTTTT
Year 1 -Tanzania - value added x 10000

2- Tanzania - total output x 10000
3- UK - value added x 500000
4- UK - total output x 500000
Figure 7.9: Value added and total output in 1987 constant prices
210
The size of the construction industry in relation to the population can be
quantified by analysing output per capita. When this is expressed in value
added per capita, it signifies the additional wealth created by the industry per
capita - its contribution to GDP per capita growth. Indeed, this is a good
measure of the strength of the construction industry in any economy.
Comparative trends of the construction industry contribution to per capita
income in constant prices for Tanzania, Kenya and UK are shown in Figure
7.10. First, it should be noted the UK per capita value added is considerably
higher than that of both Kenya and Tanzania. This ranged from a minimum
of about 293 in 1969 to 546 in 1989. Throughout the period of analysis,
UK per capita construction contribution to GDP is higher than the total GDP
per capita in both Kenya and Tanzania. Secondly, the contribution of the
construction industry to the GDP per capita in Kenya has been consistently
higher than that of Tanzania. Kenya's contribution peaked at 21 in 1973
while that of Tanzania was only about 12 in the same year. GDP per capita
contribution for the Tanzanian construction industry has been consistently
below five pounds per annum since 1983 with a minimum of only 1.72 in
1993. While these figures'may seem alarming, it should be emphasised that
they should be treated with caution as not all construction output is captured
in national accounts statistics. However, the observed trends are generally
representative of actual trends and it can be concluded that the construction
industry contribution to the GDP per capita was at its lowest in 1993.
211
U) 70
c
0 0
-'-Tanzania Kenya " 0.1 x UK
60
ti
co
T 50
40
"
30
CL
20
as
10
0
> te
OOrNM IRf LA CO f- 00 OO V- NM et lA (O I 40 OO T- NM
co 1. fl_ fl f- I- I- I- 1- 1- f- co co OD co 00 00 00 00 co a0 rn rn rn rn
C) C) Q) O) C) O) Q) d) 0) 0) C) C) O) 0) 0) 0) 0) O) 0) Q) C) C) O) 0) O)
T T- T P" rT T- T T- P" T T- r1Tr r' TTT T- r' TrT
Year
Figure 7.10: Construction value added per capita comparison
7.4.3 Creation of employment
One of the important measures of the construction industry's performance is

its contribution to the creation of employment. This indicator is particularly
useful in developing countries where wage labour employment is very low.
Since construction processes are largely labour intensive, the relative ability
of the industry to create employment is high. Employment trends in the
construction industry for the Tanzanian, Kenyan and UK construction
industries have been analysed. Figure 7.11 indicates the contribution of
construction employment in relation to the population for the three
economies. Figure 7.12 indicates construction industry contribution to the
total employment. Construction industry employment was as much as 22
per cent of all employment in Tanzania in 1973, while that of both Kenya and
the UK has been below 8 per cent. The contributions of the Tanzanian
construction industry to employment have been fairly erratic, UK and Kenyan
figures have been fairly consistent between 5.6 and 7.9 per cent and 4.6 and
6.3 per cent respectively over the period of analysis. Construction
employment in Tanzania peaked at 22 per cent in 1973, and was at its
lowest at 3.3 per cent in 1993. The employment contribution of the
construction industry has been consistently decreasing. It should be
212
mentioned, however, that informal construction industry contribution, whose
data are not represented in the above analysis, may be considerable and
should not be ignored in the national accounts data (Nuru, 1990). A
methodology for estimating its contribution has to be worked out. This is
beyond the scope of current work.
40
Tanzania 13 Kenya " lK
35
0
0
30
..
Q, 25
20
E
a
0 15
c.
E
w 10
o+ T CY) U) t- 0) r- t) :A I- O) Cr)
C) t- CO CO CO CO CO
T
( 0)
CD a) O) O) 0) () Q) Q) Q) (7) 0) (7) O)
C) TTTTTTTTTTTT
Year
Figure 7.11: Construction employment and population comparison
213
25
Tanzania
20
c 0 Kenya
C
2 15 LK
C\
0 10
-.. "-
_. _. -.
>. -
5
0 0
E
w
c) U) 1 O) M
(0 r- r- t- t- CC) 00 00 Co CD (3) (
C) 0) 0) rn 0) 0) CD 0) C) O) C) C) C)
TTTTTTT TTTTTT
Year
Figure 7.12: Construction industry contribution to employment
7.4.4 Construction industry productivity
Construction industry productivity between Tanzania, Kenya and UK were

compared on the basis of output per person engaged per annum (OPE), and
value added per person engaged per annum (VAPE) in spite of the
limitations of productivity measurement at industry level (see Chapter 2).
OPE comparison was made for Tanzania and the UK only. Figure 7.13
presents OPE in constant prices for both Tanzania and the UK. Generally,
construction OPE in Tanzania was approximately one third of that in the UK
for most of the period of analysis. Minimum output for the Tanzanian
construction industry was approximately 4000 per person, in 1973 during
peak output. This observation is consistent with observations elsewhere
(Olomolaiye, 1987). It is not clear as to the reasons for this, although this
could possibly be associated with major civil engineering projects in progress
at that time under the East African Community. Productivity increased to a
maximum of 13,817 in Tanzania in 1977. Productivity had consistently
dropped back to the 1973 level by 1993, which was only about one fifth of
that of UK at the time.
214
30000
0 Tanzania UC
25000
Co
rn
a 20000
C
15000
a..
10000
/ i
Q.
0 5000
"
__
\'a
0) r- Ce) lf) N- 0) i-- Cr) lf) N- 0) TM

CD N- N- I- N- fl- 00 00 DO 00 CO C) Q)
0) 0) Q) Q) 0) 0) 0) 0) Q) Q) 0) O) a)
TT
rr r- r'rrrrr
Year
Figure 7.13: Productivity in output per person engaged (OPE)
22000
20000 Tanzania 13 Kenya " lK

c
Q 18000
CD 16000
14000
c
12000
10000
8000
o6000
Co
4000
2000
0-F--
cm T C) UI I'- C) 9- CC) U) I- (3) T Cr)
(0 N r- r- NN 00 00 OD Co OD cm 0)
Q) Q) 0) (3) ( O) 0) C) Q) C) (7) C) C)
TTTTTT1TTTTTT
Year
Figure 7.14: Productivity - value added per person engaged (VAPE)
The second measure of productivity is value added per person engaged

(VAPE). This is computed by dividing the contribution of the industry to the
GDP by the number of persons engaged in the industry in the year. Again,
comparison was made by converting all contributions to 1987 pounds
215
sterling. Figure 7.14 presents the trends for the three national construction
industries. UK construction productivity, as expected, was consistently
higher than those of both Kenya and Tanzania. The Kenyan construction
industry was also consistently higher than that of Tanzania, except for 1977
when they are almost equal. Both the Kenyan and Tanzanian construction
industry's productivity levels have consistently been decreasing with the
Kenyan industry decreasing even faster. When comparisons were made
based on 1969 productivity, UK productivity was observed to have increased
to more that 200 per cent by 1987, and then decreased to about 175 per
cent of the 1969 level by 1993. Productivity in the Tanzanian construction
industry increased to about 250 per cent by 1977, but decreased
consistently to only about 72 per cent of the 1969 productivity level by 1993.
The situation was worse in Kenya where productivity increased to a
maximum of only about 118 per cent in 1973 and then consistently dropped
to about 38 per cent of the 1969 productivity by 1992 although the actual
value added per employee was still higher than that of Tanzania. These
comparisons are illustrated in Figure 7.15.
300
250
rn
rn
200
x
c
C
150
100
(s
ru 50
0 (3) U) h a) CV) Ui r- (3) CV)

T fr) T T
Cfl f I N- N- f- Co 00 CO Co 00 ( C)
O) O) C) C) C) (D C) C) C) C) (M C) C)
TTTTTTTTTTTTT
Year
Figure 7.15: Productivity trends in 1987 constant prices
216
Productivity trends were compared to 1969 productivity levels and expressed
in constant prices. An analysis of productivity trends, on the basis of OPE,
did not reveal dramatic changes except for Tanzania. Productivity in the
Tanzania construction industry was about 290 per cent of 1969 level in 1977,
otherwise dropped to only about 95 and 105 per cent in 1992 and 1993
respectively. These trends are presented in Figure 7.16. Productivity in the
UK increased to a maximum of about 149 per cent of the 1969 level in 1993.
Kenyan construction industry productivity was analysed between 1969 to
1985 for which data was available. Productivity increased to 145 per cent in
1973 and later dropped to only about 87 per cent in 1983.
Kenyan and Tanzanian construction industries experienced considerable

productivity decreases over the period of analysis. Kenyan construction
industry VAPE decreased more that that of Tanzania although in absolute
measures Kenyan construction industry productivity was still higher than that
of Tanzania by 1992. As such there would be little to learn from Kenya as
far as productivity improvement is concerned.
400
0
0 350 Tanzania 0 Kenya " UK
T
CD
co
C)
300
/\
x
250
C_
C
0
200
\
N
d) woe
CL
I-
150
u)
Q.
d
100 c"`" .-/ =,.,
7
0 50
0
rn C) U) 1- Q) r CC) U) F- 0) rM
CD 1' f,- DO CO 00 OD Co O) (3)
O) O) C) G) C) (3) C) (3) O) 0) Q)
CD (7)
rrrrrrrrrrrr
Year
Figure 7.16: Productivity trends in output per person engaged

(OPE)
217
7.5 Wider international productivity analysis
7.5.1 - Significance of relationship between productivity and GDP per

capita
In order to avoid misinterpreting the results of construction industry

productivity comparisons, it is necessary to investigate the relationship
between productivity and the performance of the economy expressed in
income per capita (GDP per capita). It is expected that the level of economic
performance would have a significance influence on the productivity due to
higher levels of capital investment and use of more advanced technology.
Higher costs of inputs such as labour would also have an influence although
this is indeed a distortion. The following section investigate this relationship.
7.5.2 Productivity and GDP relationship for 1979 data
The relationship between VAPE in the construction industry and income per
capita is firstly performed using data for 1979 extracted from Annex 2 of
Edmonds and Miles (1984, pp. 17-20) to compute the value of output per
employee and later repeated using 1992 data compiled by the researcher
from various sources. Table 7.1 lists computed figures for VAPE in the
national construction industry for 69 countries along with their income per
capita.
218
Table 7.1: Construction value added per person engaged and GDP for
selected countries (based on 1979 statistics)
Country GDP per VAPE Country GDP per VAPE

capita capita
Algeria 980 12800 Luxembourg 7842 31360
Austria 9107 37300 Malawi 105 615
Bahamas 4220 4850 Mauritius 171 8700
Barbados 2659 8974 Mexico 1749 8357
Belgium 11260 61250 Netherlands 10624 21242
Bolivia 924 14411 New Zealand 6896 15500
Botswana 830 2275 Nicaragua 889 8666
Brazil 1809 11447 Norway 11486 28000
Burundi 3690 8500 Panama 1343 7500
Canada 9586 15332 Papua New Guinea 540 12400
Rep. of Cent. Africa 240 15000 Paraguay 1150 3647
Columbia 1064 6285 Peru 716 1636
Costa Rica 1843 5619 Philippines 646 11500
Cyprus 2828 13846 Poland 3830 14000
Czechoslovakia 5290 15675 Portugal 1610 2709
Denmark 12925 33125 Puerto Rico 4081 7571
Egypt 435 2090 Rwanda 206 10000
El Salvador 755 18500 Senegal 436 11000
Fiji 1609 9000 Seychelles 940 2666
Finland 8701 17687 Sierra Leone 283 3400
France 10720 32273 Singapore 3829 11850
Fed. Rep. of Germany 12419 45641 Spain 5300 3400
Haiti 180 2432 Sri Lanka 240 2133
Honduras 608 3414 Swaziland 680 1071
Hong Kong 3140 8888 Sweden 12831 26903
Hungary 3850 14375 Syria 1030 1555
India 205 4500 Thailand 590 4428
Ireland 4412 16350 Turkey 1136 5181
Italy 5686 16800 USSR 4110 10666
Ivory coast 1014 13692 UK 7192 16956
Jamaica 1554 6013 Cameroon 612 8000
Japan 8627 21805 Tanzania 254 1750
Kenya 394 5000 USA 10777 21666
Rep. of Korea 1613 9375 Yugoslavia 2620 10357
Liberia 427 29000
Notes: 1. All figures are in current US $.

2. VAPE refers to total value added by construction divided by total number of persons
engaged in construction in the year
Source: Edmonds and Miles, (1984, pp. 17- 20)
The correlation relationship between GDP per capita and VAPE is

expressed in Figure 7.18 - below where it is shown to be positive between
GDP per capital and productivity with a coefficient of correlation, r2 = 60 per
cent, that is 60 per cent of the variability in productivity can be explained by
the GDP.
219
80000-
y=2.246x + 4724.726
r2= 60%
co
D
t Belgium
a) 13
60000
a
C)
ca
at
a
a, 13
C
0 40000
aO
MpO
O D
p
2 20000
O pO
pU
ppO
O
Malewr O Cb
pO
0
0 Tanzania 5000 10000 15000
GDP per capita
Figure 7.17: Relationship between VAPE and GDP per capita

(in current US $ based on 1979 statistics)
The relationship between VAPE and GDP per capita is expressed in the
following relationship:
VAPE = 2.246 GDP + 4724 (in US $); R2 = 60%
Figure 7.17 illustrates that in general the higher the GDP, the higher the
value added per construction worker. As in investigations discussed earlier,
there is no consensus amongst researchers as to which one causes the
other. Indeed, higher income countries are able to invest more in the
construction process thereby increasing its productivity which, on the other
hand, enables the creation of more wealth that in turn enhances the ability of
the country to invest even more in the industry. The data incorporated both
civil and building works. It is also generally accepted that civil engineering
work is normally more capital intensive than building works and, therefore,
220
would have a higher apparent labour productivity figures measured in terms
of value added per employee. It follows that if a particular country has an
unusually higher proportion of civil engineering work in any one year, this is
likely to be reflected in the apparently higher value added per employee in
that year. This could partly explain some of the deviations from the general
relationship.
As for the Tanzanian construction industry, the computed VAPE is only

about US $ 1750 (see Table 7.1). With a GDP of about US $ 254, the
expected VAPE is about US $ 5300 computed from the relationship.
Tanzanian construction industry productivity was therefore less than half of
the expected value. This, although not a reliable indicator, still demonstrates
that productivity in the industry was very low for its per capita income in
comparison with other countries.
7.5.3 Productivity and GDP relationship for 1992 data
The relationship between GDP and VAPE as a measure of productivity is

further explored using more recent statistics published by the United Nations
(1994), ILO (1989 International Monetary Fund (1994) and World
-1994),
Bank (1992 1994). Table 7.2 lists the VAPE and the income per capita for
-
countries extracted from the above publications (refer to Appendix 7 for more
detailed data). Although most of the data refer to 1992 statistics, data
between 1988 and 1992 has also been incorporated especially with respect
to a number of lower income countries whose most recent data is missing.
221
Table 7.2: Construction employment, GDP and VAPE for selected countries
(based on 1992 statistics)
Country No. engaged GDP/capita Constr. Constr. VAPE Value added
in (in US $) GDP GDP per capita
construction (in %) (Mill US $) (in US $) (in US $)
Antigua and Barbuda 3109 3568.42 12.83 36.62 11778.42 457.74

Australia 517000 16530.31 6.70 19403.86 37531.65 1106.89
Austria 303900 26059.04 7.44 15276.97 50269.73 1938.70
Bahrain 30206 7505.88 6.41 250.27 8285.58 481.30
Barbados 8300 6064.27 3.55 55.98 6744.89 215.32
Belgium 245500 24575.38 5.49 13275.66 54076.01 1349.15
Benin 51655 335.64 3.12 48.11 931.39 10.48
Bolivia 55505 680.74 3.20 152.45 2746.68 21.81
Botswana 58048 2634.16 5.89 215.80 3717.53 155.25
Brazil 3588651 2262.36 8.39 27979.17 7796.57 189.82
Burundi 5810 216.20 3.35 39.55 6806.37 7.24
Canada 778000 21614.71 6.19 35540.92 45682.41 1337.13
Chile 321800 1888.84 5.14 1237.10 3844.31 97.03
Colombia 248777 1257.43 5.53 2337.35 9395.37 69.54
Costa Rica 69197 2077.73 2.65 170.48 2463.67 54.99
Cyprus 24000 10504.10 9.97 743.30 30970.75 1046.90
Denmark 200000 30333.85 4.62 7249.22 36246.12 1402.17
Equador 182060 1230.39 4.06 536.00 2944.06 49.91
Egypt 864004 750.58 4.46 1847.10 2137.84 33.49
El Salvador 53292 1318.37 2.84 201.33 3777.87 37.42
Equatorial Guinea 1929 455.89 2.80 4.59 2380.41 12.76
Ethiopia 45769 118.36 2.56 167.08 3650.47 3.03
Fiji 8480 1615.98 3.55 42.41 5001.14 57.31
Finland 190000 20922.17 5.45 5739.87 30209.85 1141.13
France 1572600 25702.73 5.26 77118.65 49038.95 1351.77
Gambia 3500 361.72 5.03 16.01 4575.63 18.20
Fed. Germany 2061000 30920.26 5.83 115584.68 56081.84 1802.63
Greece 246300 8158.03 5.72 4693.02 19054.06 466.50
Grenada 3531 2628.49 8.48 17.83 5050.18 222.90
Haiti 28001 150.06 4.52 45.86 1637.79 6.78
Honduras 81491 661.15 5.65 196.56 2412.03 37.37
Hong Kong 228400 12438.99 5.34 3825.13 16747.50 664.08
Hungary 317000 3567.68 5.19 1910.88 6028.02 185.16
Iceland 12400 25948.18 6.21 402.73 32478.35 1610.93
India 5543205 278.45 5.14 12161.94 2194.03 14.31
Indonesia 2414083 1008.20 6.16 11669.89 4834.09 62.15
Iran 1207459 3428.02 3.56 6610.34 5474.58 121.96
Ireland 79800 12659.66 4.83 2170.47 27198.86 611.40
Israel 107600 9733.95 8.02 3879.22 36052.24 780.53
Italy 1824000 24302.58 5.81 80575.43 44175.13 1412.37
Jamaica 64600 1379.90 12.93 422.97 6547.49 178.47
Japan 6190000 31395.86 10.12 395206.94 63846.03 3178.44
Kenya 72400 285.22 4.95 381.07 5263.35 14.12
Korea Rep. 1652000 6750.38 15.16 44676.21 27043.71 1023.28
Kuwait 114534 10425.11 1.99 436.69 3812.72 207.95
Luxembourg 18800 23405.72 6.74 615.15 32720.80 1577.31
Malawi 52101 141.94 4.09 53.03 1017.91 5.81
Malaysia 423900 3042.52 3.88 2199.24 5188.12 118.18
222
Table 7.2: Construction employment, GDP and value added per
employee for sel ected co untries (continued)
Country No. engaged GDP/capita Constr. Constr. VAPE Value added
in (in US $) GDP GDP per capita)
construction (in %) Mill US $ (in US $) (US $)
Maldives 3151 421.06 9.01 8.73 2769.13 37.94

Mauritius 21000 2494.12 6.06 161.68 7699.19 151.11
Mexico 1879231 3653.00 4.76 15562.77 8281.46 173.81
Morocco 308888 1196.70 4.97 1518.19 4915.03 59.42
Myanmar 174000 936.22 1.82 726.52 4175.42 17.00
Namibia 19000 1633.84 1.95 48.74 2565.14 31.85
Netherlands 389000 23362.53 5.35 18969.46 48764.69 1249.64
New Zealand 80900 13293.10 4.21 1909.23 23599.90 559.89
Nigeria 545600 284.75 1.51 496.86 910.68 4.30
Norway 126000 24606.27 3.92 4136.74 32831.30 964.28
Pakistan 1975000 469.17 3.77 2045.27 1035.58 17.70
Panama 32385 2390.84 5.01 300.90 9291.34 119.88
Paraguay 39816 923.76 5.63 216.19 5429.66 51.97
Peru 124371 1705.17 9.29 3555.74 28589.78 158.38
Philippines 1100000 1393.18 4.98 2662.57 2420.52 69.41
Portugal 380000 6097.06 6.88 4144.94 10907.74 419.53
Romania 583143 616.20 4.36 612.69 1050.67 26.88
St Vin. & Grenadines 3535 1767.86 8.06 15.67 4433.17 142.47
Seychelles 1651 5352.26 5.14 19.24 11655.25 274.90
Sierra Leone 7300 208.28 1.25 10.50 1438.86 2.59
Singapore 101000 16698.04 7.84 3691.17 36546.23 1308.93
South Africa 373700 2919.81 2.70 3134.29 8387.18 78.71
Spain 1530000 14454.03 9.05 51065.82 33376.35 1308.71
Sri Lanka 258630 546.64 6.54 621.63 2403.55 35.73
Sweden 273000 31325.59 6.58 17914.09 65619.37 2061.46
Switzerland 340000 35043.16 8.35 20194.61 59395.91 2926.76
Syrian Arab Rep. 357000 2550.52 3.81 1259.14 3527.01 97.16
Thailand 701500 1753.89 6.81 6798.55 9691.44 119.44
Trinidad & Tobago 75000 4183.23 9.63 503.80 6717.29 403.04
Tunisia 247000 127125 4.70 472.23 1911.87 59.70
Turkey 1100000 2541.87 6.49 9691.77 8810.70 164.88
United Kingdom 1679000 18373.20 5.39 57048.07 33977.41 989.56
Un. Rep. of Tanzania 30080 107.86 2.09 62.68 2083.72 2.25
United States 8471000 22172.38 4.01 226600.00 26750.09 888.56
Uruguay 83200 2734.85 3.79 324.78 3903.57 103.76
Venezuela 647301 2943.65 6.52 3886.27 6003.81 191.91
Zambia 23000 360.17 4.65 144.79 6295.15 16.76
Zimbabwe 51000 427.22 2.29 100.35 1967.62 9.77
Sources: United Nations (1994),

ILO (1989-1994)
World Bank (1990-1994)
World Bank (1992)
Table 7.2 presents construction employment in terms of total number of

people engaged in construction activities, GDP per capita, construction
industry contribution to the GDP, VAPE, and the construction industry's value-
223
added per capita expressed in 1992 US dollars for 87 countries. The data
were compiled from various United Nations, International Labour Organisation
and World Bank statistical publications. Relevant exchange rates were used
to convert national currencies into US dollars. While it would be interesting to
study the relationships between the various indicators listed, the data have
only been used to explore the relationship between construction industry
productivity expressed in VAPE and the GDP per capita for 87 countries.
Productivity varied from as low as US $910 per employee per annum for
Nigeria to US $65,619 for Sweden followed by Japan at US $63,846 per
employee. Switzerland, with the highest GDP of US $ 35,043, had a
construction industry productivity of US $ 59,395. Tanzanian construction
productivity for 1992 was about US $2,083 per employee and that of UK was
US $33,977 per employee. Figure 7.18 shows the relationship between the
strength of the national economy, expressed in GDP per capita, and VAPE.
The analysis was sub-divided into three income groups in accordance with
the World Bank (1994) catagories' as follows:
" all 87 countries with GDP varying from as low as US $107 for Tanzania
to US $ 35,043 for Switzerland;
0 sixty-six low and middle income countries with GDP of up to US $
10,000 i. e. Tanzania to Israel; and
" forty-one low and lower middle income countries with GDP of less than
US $ 2,500.
1 Strict
classification of the World Bank (1994) is as follows:
Low income countries < US $ 675 GNP per capita in 1992;
Middle income countries
Lower-middle US $ 675 2695 in 1992;
-
Upper-middle US $ 2695 8356 in 1992; and
Higher income countries US $ 8456 -
- or more.
224
12
Y =1.665x + 2710 = 0.877
Fs"
to Sweden
v
60000

E
u

1
1191o
4000
10
o

13 =3
O
10000- O
C3C3
Kuwait
0
o0
00
o 0
C. CO
GDP per capita (US $)

Figure 7.18 (a): All countries
40000 38
y"2150x+1778.314 r2m0.497
30000-
v 13
Peru p
, 20000
.
13 13
10000
13
n 13
-oC. 00
0000
NoN
NN l p
ti
GDP per capita (us $)
Figure 7.18 (b): Low and middle income countries
See next page
Figure 7.18: Value added per employee (VAPE) and GDP

per capita (based on 1992 national accounts statistics)
225
Y =1.946x + 2061.341 r2=0.304
10000
re
Co
d
0
CL
E
a,
a,
a
a
to
m
O C. OOC.
OOOC. O
NaN, O Vs
NN
Figure 7.18(c): Low income countries
Figure 7.18: Value added per employee (VAPE) and GDP per capita
(continued)
In general, there is a positive correlation between construction productivity

and the strength of the economy. One explanation for higher productivity for
higher income economies is the level of investment in the construction
process. This explanation could be verified by relating productivity and
investment per workplace. However, as mentioned earlier, the quantification
of the level of investment in relationship to labour and the overall effect on
productivity in the construction industry is not a straight forward issue and is
currently attracting research interest (Norman and Flanagan 1993)
The relationship between VAPE and the strength of the economy can be
expressed as follows from Figure 7.18(a):
VAPE = 1.665 GDP + 2710; R2=87.7%;

n= 87 countries.
This is a strong relationship with a regression coefficient r2 of 87.7 per cent

when all the 87 countries for which information was available are considered.
The strength of the
relationship diminishes with a regression coefficient of 50
per cent when 66 countries with GDP per capita of less than US $ 10,000
226
are considered (Figure 7.18 (b)). The relationship is further explored for 41
lower middle and low income countries with GDP of less than US $2500 and
the relationship becomes very weak with a regression coefficient of only 30
per cent (Figure 7.18(c)). Using the three relationships generated,
Tanzania's construction industry productivity would be expected to be US
$2888, $2008 and $2208 respectively. The calculated actual productivity is
US $2038 per person engaged implying that the Tanzanian construction
industry productivity is not very different from what would be expected for its
economic strength. However, given that there are 19 countries in lower-
middle and low income group with construction productivity higher than that
of Tanzania, (see Figure 7.18(c)), there is a strong justification for exploring
strategies for increasing Tanzania's construction productivity. Productivity
benchmarking for performance improvement is therefore very much in order.
7.5.4 Construction contribution to GDP
The relationship between the contribution of the construction industry to

GDP per capita and actual GDP for all the 87 countries is shown in Figure
7.19. Again, as established using the 1979 data, the 1992 data shows a
stronger correlation with a coefficient, r2 of 89.4 per cent The relationship for
lower income countries in not as strong, with a correlation coefficient, r2 of
69.4 per cent as shown in Figure 7.19 (b).
227
4000
U)
3000

v 2000
m
Co
11000
co 00
Cl 00
CD
H
Figure 7.19 (a): Eighty seven countries
Z 200
CO
150
a
cc
loo
C
0
5o
O
U
Tanzania 00000
C3, N
N
LC,
"
" CV 44
Figure 7.19 (b): Forty-one countries
Figure 7.19: GDP and

construction value added per capita
228
It is worth noting that Tanzanian construction industry's contribution to the
GDP per capita in 1992 was only about US $ 2.25 per capita, the lowest
amongst the 87 countries investigated. Indeed, there is a strong case for
strengthening the construction industry for its effective contribution to the
countries development.
7.5.5 Productivity comparison with selected developed economies
Further analysis of productivity was made by comparing time series

productivity with selected construction industries whose data was readily
available. A trend comparison, based on 1980 productivity in relative terms,
is presented in Table 7.3 and Figure 7.20. The productivity indices are
computed by dividing value added by the construction industry in constant
national prices to the number of people engaged in the industry. Indices for
Kenya and Tanzania are based on value added expressed in 1987 pounds.
Table 7.3: International construction Industry productivity comparison

(Index based on 1980 = 100)
Year UK USA Japan France West Italy Kenya Tanzania

Germany
1980 100 100 100 100 100 100 100 100

1981 99 94 105 102 97 98 100 118
1982 104 94 104 104 98 95 106 150
1983 111 99 97 106 102 96 96 99
1984 121 96 100 110 103 98 116 113
1985 131 94 104 113 102 100 120 125
1986 142 95 107 116 105 102 95 85
1987 162 94 120 117 105 104 90 76
1988 160 NA 126 124 107 107 84 78
1989 155 NA 128 127 110 112 70 58
1990 136 NA 133 127 109 112 65 80
1991 127 NA 132 128 112 110 61 76
1992 129 NA 131 132 117 108 55 66
Notes:
1: Productivity indices based on VAPE in national constant prices

2: Indices for Tanzania
and Kenya based on VAPE in constant British pounds
Sources: Department
of Environment (1994); United Nations (1990,1994)
229
200
180 -m- LK
0
0 a USA
16O
" Japan
co 140
-O- France
x 120
W.
c Germany
100
-. - Italy
80
V
-0- Kenya
v
60
a - Tanzania
40
ONO" to (C I- CO C1 O r- N
CO CO CO Co Co Co CO Co Co Co C) C) 0)
0) C) C) C) C) C) C) O) O) O C) C) C)
r1r. r. .-r .-rrr
Year
Figure 7.20: International construction productivity trends comparison
Relative to the productivity of a selection of national construction industries,

UK productivity attained its peak in 1987 and 1989 when productivity was
about 160 per cent of the 1980 productivity. Decrease in productivity for the
Kenyan and Tanzanian construction industries is a major distinguishing
feature when compared with the other industries (complete data for USA are
not available). During the period, productivity in Kenya nearly halved and
that of Tanzania decreased by about one third, a trend which is indeed very
disturbing. This research is an attempt to uncover ways of reversing this
trend.
7.6 Summary
This chapter has examined, in some detail, the macro performance of the
Tanzania construction industry within the framework developed in Chapter 2.
Analysis of output, contribution to the economic growth, productivity, and
creation of employment has been made over a considerable period of 25
years from 1969 to 1993. The analysis was based on data collected from
Tanzania and data extracted from various international statistical
230
publications. A comparison with the Kenyan construction industry over the
25 years showed that the Tanzanian construction industry has lagged behind
that of Kenya when the following performance indicators were examined:
" output as quantified by value added by the industry;

" contribution to the GDP;
" actual productivity as quantified by value added per person engaged
VAPE); and
9 contribution to employment.
Kenyan construction industry's productivity in 1992 was found to have

dropped to nearly one third of the 1969 level although in absolute terms it
was still higher than that of Tanzania. On this basis, the Tanzanian
construction industry has little to learn from its Kenyan counterpart on the
aspect of productivity improvement. Comparisons with the British
construction industry were generally aimed at providing transparency to the
analysis. Of significance in the comparison over the period 1969 -1993 are:
" peak UK output in 1989 was about 60 times that of Tanzania;

" OPE in UK was about three to five times that of Tanzania; and
" VAPE in UK was between four and nine times that of the Tanzania.
It has been established that there is a strong relationship between

productivity as measured by VAPE and the GDP per capita based on
international level comparison of 87 world economies. The Tanzanian
construction industry was observed to be less productive than 19 other
countries within its income group. This is a good enough reason to want to
improve the industry's productivity. Further, the Tanzanian construction
industry's contribution to the GDP per capita was found to be the lowest of all
the 87 countries investigated. A wider international comparison indicated
that while both Tanzanian and Kenyan productivity decreased, productivity of
five other industrialised countries increased. It is also noted that Kenyan and
Tanzanian GDP per capita continued to decrease. This adds more meaning
to the term used by Wells (1995) when she referred to the sub-Saharan
countries as 'countries that are not developing'. Construction labour
productivity will be investigated in the following chapter with the objective of
determining productivity potential.
231
CHAPTER EIGHT
ANALYSIS OF LABOUR PRODUCTIVITY AT SITE LEVEL
232
CHAPTER EIGHT
ANALYSIS * OF LABOUR PRODUCTIVITY AT SITE LEVEL
8.1 Introduction ''
In Chapter 7, Tanzanian construction productivity trends, examined over a

25 year period between 1969 - 1993, indicated a downward trend. Several
factors may have accounted for this trend, one of which was actual
productivity at operative level. Several reports have expressed concern
over the decline of productivity although they lacked quantifiers and
explanations (Ministry of Works 1977, Ministry of Works 1992 and Bjorklof
et al. 1992). It was therefore necessary to collect extensive field data to
quantify operative productivity. This chapter examines labour productivity
variability for eight building activities on 46 different sites belonging to 23
contractors with the objective of determining labour productivity
improvement potential. The procedure used to collect and summarise the
data was described in Chapter 6. The results of variability analysis have
been compared with similar studies in Britain and the United States. A
systematic methodology developed for data analysis to determine an
appropriate statistical distribution model is described in this chapter. Basic
statistical data were first obtained and histogram plots for the distributions
produced. Data were then tested for outliers using the Box and Whiskers
method. Normality tests using chi-square tests for goodness of fit,
comparison of skewness with standard error of skewness and Wilk-Shapiro
tests were used. These were not conclusive. The data were then fitted on
general empirical distributions using the estimates of the first four moments.
Johnson SB distributions were found to be the best approximation of the
labour productivity data distribution. Chi-squared tests for goodness of fit
were performed to the transformed data to validate the statistical model.
Unlike most other studies, in which the variability of productivity data was
shown to fall within a specific distributional shape, the Johnson SB is
flexible and represents in a simple form most productivity distributions as
shown in the contrasting shapes of concreting and excavation productivity
distributions. The productivity benchmarking model developed in Chapter
5 has been modified on the basis of the Johnson SB distribution, tested and
validated.
233
8.2 Site productivity data analysis
8.2.1 Objectives of the analysis
The first objective of this chapter is to investigate the variability of

construction labour productivity data of different activities collected from
different construction sites in Tanzania. The variability is compared with
findings of similar studies in the United Kingdom and the United States.
Eight different activities have been studied from 46 sites belonging to 23
different contractors working in and around Dar es Salaam, the main
business centre in Tanzania as described in Chapter 6. The activities
studied included blockmaking, blocklaying, excavation in loose soil,
formwork fixing, plastering, painting and fixing ceiling boards. The sources
of the data, with respect to contractors, and sites is shown in Table 8.1. The
criteria for the choice of the activities was mainly on the perceived
contribution to the total cost of the building, labour intensity, ease of study
and sufficiency of data for analysis.
Table 8.1: Productivity data sources - contractors and sites

Activities description Sample size Contractors involved Sites involved
1 (2) (3) (4)
Blockla in 91 16 20
Blockmakin 56 13 18
Plasterin 119 14 21
Concretin 56 15 18
Paintin 62 12 15
Formwork fixing 102 12 18
Ceilin boards 38 4 11
Excavation - loose soil 46 9 11
As shown from Table 8.1, the data sources were fairly widespread.
Thirteen of the contractors were Class I contractors (or top contractors in the
country), of which in 1990, there were only forty in the country (National
Board of Architects, Quantity Surveyors and Building Contractors, 1993).
The rest of the data came from Class II -IV contractors. The data were
therefore very much representative of the formal construction industry.
The second objective is to determine the statistical distribution model(s)

that best describe the variability of labour productivity. This provides a
basis for further research into the analysis of improvement potential and
would also be useful in construction process simulation analyses.
234
In the process of the analysis of the data, this chapter develops a
systematic methodology in the form of a flow chart for the determination of a
statistical distribution model for productivity data. The flow chart provides
general guidelines for analyses of this form.
8.2.2 Review of previous studies
The study of construction labour productivity variability is not new.

However, in view of the difficulties associated not only with the
measurement of productivity but also its definition, it is difficult to draw
conclusions based on the comparisons of the different research findings.
This problem is further compounded by the fact that conditions under which
measurements are made differ considerably. A construction contractor may
be interested in broad conditions under which a project is to be executed.
A researcher may be interested in a relatively shorter period of time under
which specific conditions may exist. This may be as short as a working day
or even a five minute period. On the other hand, a broader industry wide
perspective may be used to guide policy issues. Examples of various
measures of productivity that may be associated with each level of detail
have been discussed by Thomas et al. (1990).
At micro level, there have been studies investigating the variability of labour
productivity over a working day. A typical variation profile has been
presented by Harris and McCaffer (1995). They suggested that productivity
gradually increases to a peak during the morning session and decreases
as lunch time approaches. The afternoon session is also characterised by
the same variability profile.
The shape of the distributions for productivity variability over a longer

duration was studied in some detail by Noor (1992) for four different trades
in a project. He concluded that the normal distribution can be used to
model the variability on the basis of skewness of the data which is between
-1 and +1 and provides cumulative linear plots on the normal probability
paper. However, this approach is deficient for the following reasons:
" the skewness described above is only a necessary condition but not a
sufficient one in that there are a number of other distributions that exhibit
a skewness of zero. All distributions that are symmetrical about the
235
mean have a skewness of zero. For example, uniform distribution,
Cache distribution, extreme value distribution and some special cases
of beta distribution; and
" the subjectivity that surrounds the evaluation of cumulative frequency
plots on the normal probability paper.
Goodness of fit tests provide more reliable results of the extent to which a
distribution can be approximated by the normal distribution and indeed by
any other continuous distribution. Chi-squared, Kolmogorov/Smirnov and
Wilk and Shapiro are some of the tests that can be performed (Gardiner
and Gardiner 1979; Shapiro and Gross 1981 and Hahn and Shapiro
1994). Price et al. (1985) performed goodness of fit tests using a number of
distributions on the duration of a concreting activity on a cyclic operation.
He concluded that the log normal distribution represents a good
distributional model. The distribution investigated by Price is equivalent to
the distribution of the inverse of productivity as elapsed time to complete a
unit task is inversely proportional to achieved productivity. It cannot,
therefore, be concluded that the same model can be used for productivity
distributions as defined in this research.
A study by Emsley et al. (1990) suggested that frequency distribution of

daily labour productivity on site are skewed to the left. She suggested that
such a frequency distribution indicated that productivity frequently falls
below the mode. This in turn indicates that factors which cause daily
productivity to fall below the norm, represented by the mode, are disruptive
rather than systematic. No specific distribution was suggested by the study.
In this research, productivity variability were investigated for a number of

activities across the building construction industry. This makes it different
from any of the studies described above. The above findings cannot be
directly adopted in this work. However, possible links with this work are
discussed. A summary of study results of the eight activities is presented in
Appendix 8.
236
8.3. Variability analysis
8.3.1 Significance of variability and statistical distribution
Declines in construction productivity have been traced back to a lack of

improvements in the construction process as a whole. The key to
improvement lies in understanding construction processes, process
variables and process performance indicators (Burati et al. 1992; Koskela
1992). Process performance improvement can be effected by changing the
variable or set of variables that most influence the performance indicator.
The initial stage in seeking performance improvement involves the analysis
of the extent of the process performance variability. This would reveal the
extent of the performance gap. Labour productivity, one of the key
performance indicators in the construction industry, can be analysed to
reveal the extent of the performance gap.
Construction is a labour intensive industry, especially in developing

countries where the level of mechanisation is low and labour is both
available and cheap. Labour often forms a key input in the construction
process along with materials and to a lesser extent equipment. Labour
productivity norms, usually represented by a single figure, are an important
input in the pre-construction process cost and time estimates as well as
during the actual construction stage. However, productivity is better
understood as a quantity that varies not only over the duration of the
activity, but also from one site to another. An understanding of its variability
would provide an opportunity for managers to attain a leading edge in a
competitive environment - the essence of benchmarking. This highlights
the importance of understanding the variability of this key input for
enhancement of the construction process performance.
8.3.2 Summary and analysis of productivity study results
The data analysis procedure is summarised in the flow diagram presented

in Figure 8.1. A summary of the productivity study results of the eight
activities is presented in Table 8.2, which also provides the basic statistics
for the productivity data. The sample sizes for the activities range from a
minimum of 38 for ceiling board fixing to a maximum of 119 for wall
plastering. The labour productivity mean expressed in output per labour
hour is indicated in Column 3 of Table 8.2.
237
Start
Plot frequency histograms - using k= Determine basic data i.e. minimum, maximum
n= sample size and k= number of classes mean, standard deviation, skewness, kurtosis
Test for outliers

Box and Whiskers Determine range: std.
method deviation ratio
are
Yes Data from skewed distribution
outliers onl Yes Different
(eg. lognormal):remove
ne sided rom David et al
extreme outliers only tatisti
No
No
Data from symetrical norm for
distribution (eg. normal):
normality
removeoutliers
Compare skewness with II Chi-squared test for II Wilk-Shapiro test

std. error of skewness goodness of fit for N_ 50
I comparison Are
Yes
expected an normality test re.
frequencies conclusive?
Stop No
Fit data onto empirical distribution
Plot square of skewness, bi, and kurtosis, b2, onto (13i, (32) plane for all data sets
Chi-squared test and Yes Can data set b?r-,,

plot expected and described by a single
observed freauencies distribution? /
Plot (bi, b2) onto (pi, (32) plane demarcating Johnson system of distributions
Are ail IN
Determine Johnson
No data sets with
parameters for each Johnson Se
system independently
i Yes
Chi-squared test and Determine Johnson SB

plot expected and parameters for each
observed frequencies data set
Figure 8.1: Flow chart for data analysis methodology
238
Table 8.2: Productivity results and basic statistics
Activities Description Sample can output tan ewness Kurtosis Ranoe Histogram
Size .o
Variation
per labour Deviation Std Dev. Classes
hour
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Blocklaying 91 0.69m2 0.48 70% 1.251 5.52 5.64 10
Blockmakin 56 11.29 no 6.85 61% 1.219 5.73 5.25 8
Plastering 119 1.32m2 0.86 65% 1.551 6.26 5.68 11
Concreting 56 0.17m3 0.18 105% 1.297 3.60 6.70 8
Painting 62 4.47m2 2.96 66% 1.133 4.03 4.56 8
Formwork fixing 102 0.75m2 0.55 73% 1.118 3.485 4.02 9
Ceiling Board Fixing 38 1.03m2 0.47 46% 0.340 2.18 3.72 6
Excavation - loose soil 46 0.37m3 0.25 68% 0.966 3.26 3.75 7
8.3.3 Analysis of the variability of the results
The summarised data were investigated for any outright outliers. The Box-
and-Whiskers method was used to investigate the extent of outliers. In
order to represent all the data in a uniform scale, all the data were
transformed into z-scores. The resulting box-and-whiskers plots for the
productivity data are shown in Figure 8.2. The method assesses the
spread of the data by plotting 50 per cent of the data in the form of a box
with a median and bounded by the 25th percentile on the lower end and
the 75th percentile on the upper end. The length of the box corresponds to
the interquartile range, which is the difference between the 75th and 25th
percentile. Cases with values that are between 1.5 and 3 box lengths from
the lower and upper edge of the box are called outliers. Cases outside this
range are known as extreme values. The lines drawn from the end of the
box to the start of outliers are sometimes called whiskers thus the name
box-and-whiskers.
It was observed that outliers are almost consistently on one side suggesting
that the data are skewed to the left. A secondary procedure was used to
determine the extent of skewness. Several tests exist for this purpose, one
of which is to determine the ratio between the range and the standard
deviation (sometimes referred to as David et al's statistic) (Gardiner and
Gardiner, 1982). The results of this test have been included in Table 8.2.
There was agreement between the skewness and the results of the David
et al's test. Outliers were not eliminated outright during this process in view
239
of the fact that the data were consistently skewed to one side suggesting
that the data may belong to a distribution which is itself skewed. However,
a few extreme values were eliminated based on the practicability of
achieving the recorded productivity which may have resulted from a
genuine error in the data recording.
0
painting
1311Oa
plastering
i1
DO O
formwork
excavation
vo 1313
concreting
ceiling boards
0
blockmaking
OO
blocklaying
012345
-4 -3 -2 -1
z-scores
Figure 8.2: Box-and-whiskers plot for productivity data.
8.3.4 Comparison of productivity variability with other studies
Standard deviation is a key measure of the variability of the data. It is a

measure of the reliability of the mean of the data. Standard deviation has
little meaning on its own and should be expressed in relation to the mean.
A more representative measure of the reliability of the mean is the
coefficient of variation which is the ratio of standard deviation, s to the mean
x expressed as a percentage.
240
Coefficient of variation c =Sx 100%
,
Indeed, this statistic is particularly important in this study as it provides a
relative measure of the spread of the data for the different activities. It is
also a relative measure that can be used for variability comparison with
findings in other similar studies as it is independent of the units used.
There have been a number of studies investigating the variability of

construction labour productivity. However, direct comparisons of the
findings are difficult because of the inconsistencies in both the definitions of
labour productivity and methodologies of data collection adopted in the
studies. Most previous research has been based on contractors' historical
data and on activity sampling results (Noor, 1992). Productivity has been
measured on the basis of total time, paid time and sometimes productive
time. Another problem is the quantification of output which has been
expressed in terms of: labour hours per house; labour hours per structural
element; and sometimes in labour hours per unit item of work (or units of
item per labour hour). In spite of these limitations, it is possible to compare
various research findings on the basis of the coefficient of variability which
is independent of the definition adopted and units used in quantifying
productivity. Noor (1992) summarised the results obtained by different
researchers who investigated productivity variability in different trades. This
has been presented in Table 8.3. The comparison is made on the basis of
the coefficient of variability and range ratio obtained by dividing the largest
with the smallest observed value.
The variability in productivity from other studies is far less than in this study
on the basis of both the range ratio and the coefficient of variation. This
was because of the large difference between lowest and highest observed
productivity in this study. The minimum range ratio in this study was five for
plastering and ceiling boards fixing. This figure is higher than the
maximum range ratio of 3.5 obtained for the other studies. The minimum
coefficient of variation in this study is 46 per cent compared to a maximum
of 35 per cent in the other studies. The coefficient of variation for concreting
productivity, for example, is in excess of 100 per cent and the range ratio is
66. The concreting productivity data are also extremely skewed to the left
resulting in the observed coefficient of variability. The large variability of
productivity data in this study is clearly one indicator of the improvement
potential.
241
Table 8.3: Productivity variability from various studies
Source Activity Productivity Range ratio Coefficient of

measure variation %
Shippam (1987) Masonry & Pt 3 -
Plastering
Bishop (1965) Masonry & Pt 3 -
Plastering
Walker (1971) Masonry internal Pr - 9.1
finishes 6.6
Oxley (1975) Dry lining and PP - 25
Plaster boarding
Forbes (1971) Masonry Pp - 18
McLeish (1981) Masonry Pp - 14
Grant & Stephens Dry lining & Pp - 30 - 35

(1982) partitioning
Logcher and Tiling PP 2.5 -
Collins (1978)
Noor (1992) Masonry Pt 3.50 34
Pp 2.58 27
Wall plastering Pt 2.58 28
PP 2.50 28
Floor screeding 1.41 9
t 1.16 5
P
Tiling 1.96 20
Pt
1.93 19
P
This study Blocklaying Pt 36 70
Blockmaking Pt 27 61
Plastering Pt 5 65
Concreting Pt 66 105
Wall painting Pt 40 66
Formwork
Erection pt 55 73
Ceiling board
fixing Pt 5 46
Loose soil
excavation Pt 14 68
* Note that Pt is productivitybasedon total time and Pp is productivitybasedon productivetime
8.4 Frequency distributions of the productivity data
8.4.1 Graphical method - histograms
The purpose of summarising the principal characteristics of a frequency

distribution is to ease interpretation and comparison with other
distributions. In this way the complexity of the information is considerably
reduced as individual values are generalised, whilst its usefulness is
enhanced and the general pattern of the data is retained. Frequency
242
distributions in the form of histograms affords a simple visual impression of
variability of the productivity for the different activities for purposes of basic
comparisons. The choice of the number of classes is critical in the
construction of the histograms. One drawback of a grouping may be that
too much information is lost. A careful balance has to exist between
generalisation and amount of information retained, and several authors
have provided some guidelines as to the number of classes, K, based on
sample size, n, instead of using subjective judgement. Brooks and
Carruthers (1953) suggested that:
K> 5 Log1o n
Later, Huntsberger (1961) proposed the following relationship:
K=1 + 3.31og10 n
This relationship is suitable for a limited set of data and for a large n value,
it is considered conservative (Gardiner and Gardiner, 1982). A further
approach was suggested by Norcliffe (1977) who proposed that:
K=I.
Histogram plots for the productivity data were therefore made on the basis
of the approach suggested by Norcliffe for its simplicity. Productivity data
under investigation for the eight activities vary from n= 38 to n= 119.
According to Norcliffe (1977), the appropriate number of histogram classes
range from a minimum of six for n= 38, to a maximum of 11 for n= 119.
The number of histogram classes for each activity are included in Table 8.1
along with the basic statistics. The resulting histogram plots have been
presented in Figure 8.3.
The histogram plots show that the data are skewed to the left as suggested
by the statistics included in Table 8.2. This generally suggested that there
was a considerable number of operatives with low productivity. The next
step is to determine whether the normal distribution is an appropriate
model for the productivity distributions. Normality tests were performed for
the data before either proceeding to other types of distributions or
transforming the data.
243
JV 30
n=62
20 20
T
0)
U
C
0-
a)
a) 7
10
U- 10- C)
0
LL
0 --F j
00 0.6 121.8 2.4 30 0 00 3 50 7.50 10 S0 14 00
Productivity - m2/Ihr
Productivity - m2/Ihr
e) Wall painting
a) Blocklaying
20 30
F--n=-5-61
In = 102
V 20
C a)
10
0
Q
N LL 10
U-
10 1S 23 25 30 35 40
Productivity - No/Ihr rroauctivity - m-nnr
b) Blockmaking f) Formwork erection
11
Fn-=i-1-91
T
10
U
C 8
4) 1)
Q
a
(1)
IL 4-
U- l
2
U. 0 1G2.0 30 4.0 50 00 04 08 2 16 20 24
Productivity - m2/Ihr Productivity - m2/Ihr
Wall plastering g) Ceiling boards fixing

c)
12
__
10
C C
Q)
Q
a)
LL
0 t--
JO 0 170 0340 0.510 000 010 00 0 45 01") O11.0 90 1 05
Productivity - m3/Ihr Productivity - m3/Ihr
d) Concreting h) Loose soil excavation
Figure 8.3: Frequency histograms for selected construction activities
244
8.4.2 Normality test
Fitting an empirical distribution to data is necessary when there is no

sufficient theoretical grounds for selecting a particular model. The normal
distribution is most commonly used although it is usually necessary to verify
this assumption. Several techniques exist for the verification of normality.
A graphical assessment of the frequency distribution or by using the normal
is
probability paper plot sometimes sufficient. It is often necessary to
quantify the extent of data fit by using specific tests. The productivity data
collected in Tanzania were first tested for normality by assessing the
skewness. The test was based on the comparison of skewness of normal
distribution, which is equal to zero, and the calculated data skewness.
An inspection of the frequency histogram plots in Figure 8.3 does not reveal
any obvious fit to the normal distribution. However, for more reliability in
this conclusion, the value of skewness was tested for significant difference
from zero when the distribution is normal. This is compared with the
standard error of skewness, ss given by (Tabachnick and Fidel, 1983):
SS=
where N is the sample size.
The probability of obtaining a large skew value for data obtained from a
normal distribution can then be evaluated using the z distribution, where:
z=S-0 where S is the skewness value.

SS
The productivity data distributions were tested for normality and the results
have been tabulated in Table 8.4 below. A value of z outside the range of
1.96 would lead to rejection of the assumption of normality of the
distribution at a probability, p 0.05.
_<
The normality assumption cannot be rejected for the distributions of ceiling

board productivity data. However, to further confirm this preliminary
conclusion, the chi-square test for goodness of fit was performed for the
ceiling boards productivity distribution.
245
Table 8.4: Normality tests using standard error of skewness
Std. Error of
Activity Skewness, S Skewness, S_p Conclusion:
6 Z= isz>1.96?
Ss
SS _, I
(1) (2) Vn (5)
3 (4)
Blockla in 1.251 0.258 4.45 Reject
Blockmakin 1.219 0.327 3.73 Reject
Plastering 1.551 0.224 6.92 Reject
Concreting 1.297 0.327 3.96 Reject
Paintin 1.133 0.311 3.64 Reject
Formwork 1.118 0.242 4.62 Reject
Ceilin boards 0.340 0.397 0.86 Accept
Excavation 0.996 0.361 2.75 Reject
The chi-squared test assesses the degree of correspondence between

observed and expected observations in each category. The null
hypothesis, Ho, is tested using the statistic:
(I r El )2
x2 = E1
1=O
where O;= the observednumber of cases in the 11hcategory;

E; = the expected number of cases in the 11hcategory when Ho is
true; and
k= the number of categories.
In general, the larger the value of xz, the less likely it is that the observed
frequencies came from the population on which the hypothesis (Ho) and
the expected frequencies are based. The test is performed at a5 per cent
level of significance for the ceiling boards fixing productivity distribution.
The normal distribution assumption for ceiling boards fixing cannot be
rejected at 5 per cent level of significance because the calculated x2 = 1.98
which is less than X2o. o5= 11-1-
A further alternative test was performed on the ceiling boards fixing data (N
< 50) using Wilk - Shapiro test for normality for ceiling board fixing. This
test is referred to as an exact test for normality which does not depend on
any population parameter estimates based on the sample and is
246
particularly suitable for small samples (i. e. sample size less than or equal
to fifty (Hahn and Shapiro, 1994 and Gardiner and Gardiner, 1979)). The
test showed that the probability of the ceiling board data having come from
a normal distribution is approximately 15 per cent. This probability is
considered low, therefore, there is no strong evidence to suggest that the
data can be modelled by the normal distribution assumption. It must be
pointed out that the acceptance of the normal distribution model in this
particular case is weakened by the fact that the sample analysed is the
smallest (N = 38) as opposed to say plastering for which the sample size is
119. It could be argued that the normality of the data is only by chance and
the result cannot be used to generalise for other cases.
In a situation where most of the data do not conform to the envisaged

model, it is necessary either to transform the data or to test the data for
other distributional assumptions. However, before testing for other
distributions by trial and error, an investigation of the moments of the data
can provide a clue to the best distributional model. The following section
examines the moment and the transformations that would enable the use of
the standard normal variables to evaluate the goodness of fit for the
appropriate distribution.
8.4.3 Fitting data on empirical distributions
In order to define a distribution, it was necessary to describe its spread,

symmetry and peakness. These characteristics may be summarised by the
moments of the distribution. The central moments, referred to in this paper
simply as moments, are normally used to describe a distribution (Hahn and
Shapiro, 1994). A kth moment of a distribution is denoted:
E(x-m1)k
where ml'is the expected value E(x) of the random variable'x.
Thus, the first moment of a distribution is zero. The second moment is a

measure of dispersion referred to as the variance. The third and the fourth
moments are known as skewness (denoted i) and kurtosis (denoted 02)
respectively. In general, if all the moments of a distribution are known, it is
equivalent to knowing the distribution itself. Normally, the first four
moments are sufficient to get a good idea about the distribution (Shapiro
247
and Gross, 1981). Visual inspection of the frequency histograms presented
in Figure 8.3 and data presented in Table 8.2, suggests that the
distributions are all skewed to the left (i.e. skewness > 0). This suggests a
log normal, beta or gamma distribution. A standardised moment plot ,
versus 2 (i.e. plot of the relationship of the square of the skewness and
the kurtosis as extracted from Table 8.2) provides a better idea of the best
distributional assumption for each activity. Figure 8.4 indicates the
suggested distributional assumption (Hahn and Shapiro, 1994). For any
given set of data, p, and 2 are estimated by the corresponding sample
square of skewness, (b1) and kurtosis, (b2)and therefore subject to
sampling fluctuations. The plots of sample b, and b2 on Figure 8.4 suggest
a wide range of distributional models that can be adopted.
Instead of testing the distributional assumptions for each, a unified

distributional model can be adopted in order to simplify the analysis and be
used to draw conclusions across all activity distributions based on the
same model. The Johnson's family of distributions adopted from Shapiro
and Gross (1981) is described in the following section.
8.4.4 Justification for Johnson SB distribution
The Johnson family of distributions incorporates a number of common

distributions and is capable of describing most of them in fairly simple
terms. This form of distribution has been used to both model processes
(Department of Statistics 1989) as well as fit data on distributions (Chou et
al. 1994). The main advantage of using the Johnson system of distributions
is that it is based on the transformation of the standard normal variable and
covers the entire allowable range of the (l, 2) plane. Another advantage
is that it can also adopt a variety of shapes depending on the shape
parameters y and rl, location parameter c, and scale parameter X.
Estimates of the percentiles of the fitted distribution using a table of areas
under a standard normal distribution can be obtained. The first step in
deciding on the model type is to plot on the (l, 2) plane, the
corresponding b1 and b2 for the different data sets. Figure 8.5 adopted
from Hahn and Shapiro (1994) and modified for the purpose shows the
relative positions for the eight activities being analysed.
248
1
,- (J Impossible area
2
dstrbvfi
3
o, O
das
eet
d'sfrbv/017
6Gf, hap
4
-Q nJ
re6'i
e
dsfibGoshdp
fd'
Q2 5 as , edJ
-
0
f,
6 Pg:
e
O cd
7
KEY
8
O O- Wallpainting
locklaying
2O - Blockmaking GO - Fonnwork fixing

Exponential
9O"
10- Plastering 7O - Coiling boards fixing
distribution
0- Excavation
Concreting
0 2 3 4
i t
Figure 8.4: Plots of (b1, b2) for productivity distributions on (

P2) plane for variaus distributions
249 1
i KEY
Oj
2 - I3locklaying
O2
- lockmaking
3
O3
Q2 - Plastcring
4
4O - Concrcting
5 SO - Wall painting
Formwork fixing
6
7O - Cciling boardsfixing
7.
O
- Excavation
8
0 1234
Figure 8.5: Plots of bi and b2 for various productivity distributions
All activity distributions fall within the Johnson SB region. This family of
distributions can be used to describe the productivity data distributions.
The next step is to calculate the corresponding distribution parameters.
The Johnson SB distribution is based the following general form of
transformation (Hahn and Shapiro, 1994):
Z= y- i7z(X; 6, A),
for
(1)
.......................................
where
X-C
Z(X; , A_In
and
eSxSe+,,
The actual distribution is described by the function
250
77 21 1 x-E l2
f_(x) ex, -2[y+rln(.
27c(x-e)(A-x+e) t -x+e
where """(3)
s<_X-<+A,, 17>0, -oo<7<00, A,>0, -co<<
where r is an arbitrary function, y, il, e and %are four Johnson SB

parameters and z is a standard normal variable. Theoretically, c and % that
define the population range of the productivity data are not known.
However, they can be estimated from practical considerations. In theory,
the minimum possible productivity is zero (i.e. when labour employed on
site does not produce anything over the observation period). No such
cases were encountered although cases of near zero productivity were
observed. In the estimation of the distribution parameters, the minimum
productivity level, c, has therefore been taken as zero. On the other hand,
the maximum possible productivity is difficult to estimate although the
sample data estimate is a good starting point. This approach was adopted
in this work. On these bases, Johnson SB distribution equation (1) above
becomes:
X-C
z -y+ri in (4)
+E -x) .........................................................................
,
since e=0, the relationship simplified to:
(xx
z =y+ll In
_x) .
In order to obtain the corresponding y and i values, a procedure of matching

percentiles was used. A set of two simultaneous equations were solved by
setting z= za and z= zl_a as follows:
Xl-ai%,
innXa(%, -Xai
-XI1 -aJ
and
251
Xi
In -a
- XI-cc
where za and zl are the ax 100th and (1- a)100th percentiles of a

standard normal distribution, and xa and xl are the corresponding
variables at the stated percentiles. Thus, xa is the value of the a(n+1)th
ranked value obtained from the data, using interpolation if necessary.
From the wall plastering data for example, the maximum output observed, ?._
4.95. It is decided to match 10th and 90th data percentiles. From the
histogram in Figure 8.3, the interpolated 10th percentile figure is equal to 0.4
and the 90th percentile is equal to 2.6. Likewise zo.i and z0 9 are -1.28 and
1.28 respectively. Substituting these into equation (6) above, a value of i=
1.1 is obtained. When this value is then substituted in equation (7), the
resulting y=1.18. When these calculations are repeated matching 5th and
95th percentiles, the values of il and y are 0.76 and 0.78 receptively. By
repeating the calculations using different matching percentiles, the values of
rj and y are found to be close to 1. The same procedure is repeated for all
the other distributions varying the matching percentiles. To simplify the
calculations, the relationships were built into a spreadsheet.
Several percentile levels were used until a stable set of shape parameters
were obtained. These were found to be converging towards one for most of
the productivity distributions. The low percentile levels were found to distort
the shape of the distribution in view of the extent of skewness. The value of
a was therefore set at between 10 and 20 per cent. Table 8.5 presents the
Johnson SB parameters for the productivity distribution for each activity, for
most activities -y= i=1 produces a fairly good fit for the productivity
distributions except for concreting and ceiling boards for which the
parameters were as indicated in Table 8.5.
The chi-squared goodness of fit was used to determine the extent to which
the distribution can be modelled using the Johnson SB distribution with the
above parameters. The results of the test are included in Table 8.5 from
which it is noted that the model can be used to represent the productivity
data distributions for most activities at a5 per cent level of significance. It is
noted that the distribution for concreting data has different shape
parameters. This is evident from the histogram in Figure 8.3 which
indicated that there was a larger proportion of low productivity in this
252
indicated that there was a larger proportion of low productivity in this
activity. Different shape parameters were also exhibited by ceiling boards
fixing activity. This could actually be attributed to the limited data. In
general, Johnson SB distribution shape parameters 'y= r=1 provide a
good model for productivity distribution except for concreting and ceiling
boards. A typical Johnson SB distribution for y= i=1 is shown in Figure
8.5 generated from equation (3).
Table 8.5: Johnson SB distribution parameters and chi-squared test
Activity Johnson SB Parameters Chi-squared est

Description n %
x2 x2o.o5 is (a1<)?
(1) (2) (3) (4) (5) (6) (7) 9)
8
Blockla in 91 1.0 1.0 0 2.82 7.4 9.5 accept
Blockmakin 56 1.0 1.0 0 37.00 9.9 9.5 accept
Plasterin 119 1.0 1.0 0 4.95 8.0 12.6 accept
Concretin 56 1.6 0.7 0 0.68 7.2 9.5 accept
Paintin 62 1.0 1.0 0 14.00 8.8 9.5 accept
Formwork fixing 102 1.0 1.0 0 2.24 11.9 11.1 accept
Ceilin boards 38 0.4 1.2 0 2.08 2.7 7.8 accept
Excavate loose soil 46 1.0 1.0 0 1.05 4.0 9.5 accept
2.5
2.0
1.5
1.0
0.5
0.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 8.6: Typical Johnson SB distribution with y= il = 1; E=0; X=1
253
40
- actual
30
- expected
N
n=119
20
LL
10
Q 05 10 15 70 75 on 3.5 4.0 4.5 5.0

Productivity
- m2/lhr
a) Wall plastering
40
Q- actual
T
U
C
30
20
- expected
n=56
H
LL
10
nn nut 0.170 n F n zen 0.510 0.595 0.680
Productivity - m3 /Ihr
b) Concreting
20
actual
- expected
U
n=46
10 } f.. ',..
D
Q
: jj'
n All 0 0n 075 n<n in
c) Loose soil excavation
Figure 8.7: Actual and expected frequencies using Johnson SB distribution
Figure 8.7 further presents the sample Johnson SB distribution models for
concreting, plastering and manual excavation of loose soil in comparison to
the histogram distributions for the same activities. There was insufficient
evidence to accept the same model for blocklaying and ceiling boards
fixing activities. Furthermore, the acceptance of the model for formwork
fixing and blockmaking was only marginal due to the small difference
between the respective X2o.o5 and x'. It is argued that this was caused by
the insufficiency of the data as observed in the histogram distributions in
Figure 8.3. The general distribution trend is quite clear, as shown in the
histogram plot and the superimposed ideal models for the selected
productivity distributions in Figure 8.7. The main reason for the rejection of
the model is because of the gaps in the histogram plots due to limitations in
data quantity.
254
8.5 Comparison with- other research findings
A systematic procedure has been developed for the determination of a

productivity data distribution model. Similar studies have based their work
on either a pre-conceived distribution or on a trial and error basis (Price
1985; Noor 1992). Assumptions relating to the shape of the distributions for
the productivity values vary significantly. Noor (1992) suggested a normal
distribution based on studies of four trades on one construction site. His
conclusions were on the basis of daily labour productivity based on
productive time. The value of his findings is reduced due to his simplistic
assumptions for normality. Other researchers such as Van Slyke (1963)
have suggested the log normal shape. Other commonly used shapes
include the beta and the triangular distributions (AbouRizk 1992). A
detailed description of fitting distributions to activity duration data was
performed by AbouRizk (1992) for use in simulation. He concluded that the
positive skewness with long right tails was common for construction
durations although there were occasional situations where symmetric or
negatively skewed distributions were encountered. He suggested that it
was necessary to use a flexible family of distributions capable of attaining
different shape to model the diversified nature of construction duration data.
A family of distribution like the beta family, the Johnson system and
Pearson system were recommended. It should be noted that his work was
based on activity duration as opposed to productivity. When activity
duration are transformed into output per unit time, a mirror image of the
distribution is obtained. Thus, it can be concluded that productivity
amongst operatives expressed in output per unit time (as opposed to time
required per unit) would be skewed to the right. This property is utilised in
the following section to quantify the improvement potential.
8.6 Quantification of productivity benchmarking gap
The productivity benchmarking gap model was presented in Chapter 5.

This was based on the normal distribution of productivity data. The
analysis performed in this chapter established that Johnson SB distribution
with shape parameters, il =1 and y=1 as illustrated in Table 8.4, are
reasonably good distribution models for labour productivity data in the
Tanzania construction industry. The concept is therefore modified to suit
the established model. As observed in both Table 8.3 and Figure 8.6,
255
is
productivity skewed to the left, which indicates that most of the time,
operatives' productivity is low. That is, most operatives' productivity is
below the median as conceptually illustrated in Figure 8.8. The area below
x=0.5 on Figure 8.8, represents the proportion of the productivity which is
below the median. On this basis, this proportion is approximately 85 per
cent. This is in contrast to other research findings elsewhere in which the
time required to produce a unit item of work is skewed to the left, which
means that most operatives' productivity in terms of output per unit time is
skewed to the right (Price 1986; Emsley 1990 and AbouRizk 1992). That is
most of the operatives' productivity is high and only a few have low
productivity. Conceptually, this can be illustrated as shown in Figure 8.9.
While productivity frequency distribution in Tanzania is as indicated in
Figure 8.8, the target scenario is similar to that illustrated in Figure 8.9.
Figures 8.8 and 8.9 are superimposed as illustrated in Figure 8.10 to
illustrate the productivity improvement potential. The target productivity
improvement gap is the difference between i and 92, that is the difference
between target (TMP)and current mean productivity (CMP).
256
85%
U
C
15%
Q'
N
LL
u l = 0.3A, 0.57 A.
Productivity in output per labour hour
Figure 8.8: Current labour productivity distribution
U
Q)
N
L-
oo
LL
D 0.5X, u = 0.7%
Figure 8.9: Target labour productivity distribution

CMP TMP
a) (b)
LL
n
l=0.37 0.5X 2=0.77, A.
Figure 8.10: Labour productivity benchmarking gap
257
This can be justified by revisiting the benchmarking concept. Both
distributions are bound by the minimum and the maximum productivity as
defined in Section 8.4.3, thus c=0 and ?.= maximum observed productivity.
In which case, both distributions are based on previously achieved
productivity levels. The only factor that needs attention is the distribution
pattern of productivity. In the first scenario, only about 15 per cent of the
productivity is above the median. To improve the 85 per cent productivity
which is below the median, there is a potential 15 per cent upon which
lessons can be drawn for improvement. The mean productivity based on
the current distribution from Figure 8.10 (a), gi = 0.3%while that of the target
distribution, 2 0.7? This represents a potential general increase in
= .
productivity of about 233 per cent. For blocklaying for example, the target
mean productivity would be about 2.33 x 0.69m2 per labour hour, that is,
current mean productivity times the potential increase. The same potential
would apply for all activities which have the same form of distribution as
shown in Table 8.5. Current mean productivity and target mean
productivities for the selected activities are as indicated in Table 8.6.
Table 8.6: Current and target mean labour productivity for selected
activities
Activity Johnson SB Parameters Productivityper

labourhour
Description Y I e CMP TO
(1) 3 (4) (5) (6)
Blocklaying 1.0 1.0 0 2.82 0.69 m 1.60 m2
Blockmakin 1.0 1.0 0 37.00 11.29 no 26.30 no
Plasterin 1.0 1.0 0 4.95 1.32 m2 3.00 m2
Concretin 1.6 0.7 0 0.68 0.17 m3 0.39 m3
Paintin 1.0 1.0 0 14.00 4.47 m2 10.40 m2
Formwork fixin 1.0 1.0 0 2.24 0.75 m2 1.75 m2
Ceilin boards 0.4 1.2 0 2.08 1.03 m2 2.40 m2
Excavate loose soil 1.0 1.0 0 1.05 0.37 m3 0.86 m3
While the above illustration is conceptual, its practicability has to be

assessed on the basis what can actually be achieved on site. In this
respect, a survey of what blocklaying operatives have previously been able
to achieve was conducted. The results of this survey are discussed in the
following section.
258
8.7 Validation of benchmarking gap
8.7.1 Operatives' perceived productivity potential
In the preceding section, a conceptual productivity benchmarking gap was

presented. In order to establish the validity of the gap, an opinion survey of
48 masons was performed with respect to their perceived current average
and previously achieved maximum daily blocklaying productivity. This
survey was performed as part of a wider opinion study based on interviews
(see Appendix 4). The summary of the productivity opinion is presented in
Table 8.7. Current mean blocklaying productivity and target mean
productivity are included in Table 8.6 for comparative purposes.
Table 8.7: Comparisons of operative productivity potential
Observed mean and target

Operative opinion survey
productivity (m hr)
Average roductivi m2/hr Maximum productivity (m2/hr)
Mean Ran e Mean Ran e Current mean Tar et mean
0.75 0.50-1.25 1.75 1.125 - 2.50 0.69 1.60
Results of the opinion survey in Table 8.7 indicate that the operatives'
range of their productivity is similar to that observed in the field. On the
basis of at- test, observed productivity of 0.69m2/hr is not significantly
different from 0.75m2/hr claimed by operatives. The maximum that the
operatives claimed to have achieved is very similar to the target mean
productivity presented in Figure 8.10. Again, on the basis of a t-test, there
was no significant difference between 1.60m2/hr obtained from the model
and 1.75m2/hr based on the opinion survey. This opinion survey has
therefore indicated that the target mean productivity based on the
benchmarking gap concept is realistic and achievable.
Similar studies conducted in Nigeria indicated that operatives over

estimated their productivity by as much as 45 per cent (Olomolaiye and
Ogunlana 1989). Results obtained in this research did not support the
findings of the Nigerian study. The difference could probably be explained
by the fact that in this research, the survey was based on what the
operatives had previously been able to achieve as opposed what they
thought they could achieve. It has to be mentioned however that observed
259
mean productivity for some activities were similar in the two studies as
illustrated in Table 8.8. Significant differences were observed in slab
formwork fixing between the two studies, otherwise differences in the other
three activities were not significant. For activities which are less technology
dependent such as blocklaying and plastering, the differences are
practically insignificant.
Table 8.8: Productivity comparisons - Nigeria and Tanzania
Mean productivity in units per labour hour

Activity Nigerian study Tanzanian study % difference
Blockla in 230 mm wall in m2 0.74 0.69 -7
Concreting- floor slabs in m3 0.22 0.17 23
Plasterin - walls in m2 1.16 1.32 -14
Slab formwork in m2 0.47 0.75 -59
Nigerian productivity - Tanzanian productivity

Note: % difference = x 100
Nigerian productivity
8.7.2 Contractors' quantification of productivity potential
Contractors were asked to give their opinion on labour productivity

potential (see Appendix 3). Contractors were asked to indicate the
minimum acceptable, average, and maximum possible operative
productivity over an 8-hour day. Only four contractors responded fully.
Table 8.9 presents a summary of the range of their responses.
Table 8.9: Contractors' quantification of productivity potential
Activity Productivity per labour-day for 8-hour day

Minimum Average Maximum nits
Manualexcavationin loose soil 1-5 2-7.5 3 -10 m3
Blocklain 230mmwall 2.7- 12 5- 16.2 7- 20.2 m2
Blockmaking - 150 x 230 x 450 mm 36 - 200 43 - 300 50 - 500 No.
Plasterin - 12 mm - verticalwalls 8-24 10 - 27 12 - 30 m2
Concreting
- slabsandbeams 0.20-0.30 0.30-1.5 0.40-2.5 m3
Formworkfixing - slabs and beams 4-6 6-16 8-20 m2
Ceilingboardsfixing 4-16 4-18 6-23 m2
Wall aintin 8-38 12 - 40 16 - 54 m2
260
The contractors' assessments of the operatives productivity potential is
disturbing. Table 8.9 indicates that contractors' perception of operatives
potential varied considerably. This is clear from the ranges of what they
perceived as maximum, average and minimum acceptable productivity.
For example, a maximum output per day for manual excavation varied from
as low as 3 m3 to a maximum of 10 m3. A similar trend was observed for
almost all activities. Three of the respondents were Class I contractors, the
fourth was a Class IV contractors. There was no obvious differences
between the level of consistencies in the two groups. There is no basis to
conclude that the responses were distorted by the respondents
classifications.
Within each organisation, productivity potential quantified by the maximum

to mean productivity ratio varied by an average of only 1.4, that is, the
maximum productivity that the contractors thought was achievable within
their organisation was only about 1.4 times their current mean productivity.
For example, while one of the contractors thought that the minimum output
for blocklaying per day was 2.7 m2 and the maximum was 7 m2, another
contractor thought that the minimum acceptable was 12 m2. On the other
hand, some contractors had over-estimated the operatives' productivity
potential. For blocklaying, for example, the maximum observed was 39
blocks per labour hour while the maximum potential claimed by contractors
was about 63 blocks. In some cases, contractors underestimated
operatives' productivity. The maximum observed painting productivity was
14 m2 per labour hour, while contractors' estimate was 7 m2 per labour
hour. These results, although limited, indicate that contractors are not fully
aware of their operatives productivity potential. This observation
emphasises the significance of benchmarking because it can provide the
contractors with a wider view of their operatives potential. The next step
was to determine the practices that will lead to the realisation of the
operative productivity potential. This is the main theme of Chapter 9.
8.8 Summary
This chapter has concentrated on the analysis of site level productivity data
with the objective of quantifying performance improvement potential. All
construction productivity data analysed were found to be skewed to the left.
The spread of the data, as measured by the coefficient of variations ranged
261
from 46 to over 100 per cent. This is one indicator of productivity
improvement potential if compared with similar studies in Britain and the
United States where the coefficient of variation was between 5 and 35 per
cent. A systematic methodology for the determination of a distribution
model was developed. While there is no obvious theoretical distribution
that explains the distribution of labour productivity data for construction
activities, empirical models were used to model the data. There was no
strong evidence from the data to assume the normal distribution model was
valid and where this has been observed, this has merely been attributed to
chance. This specifically refers to the ceiling board productivity distribution
which had the smallest sample size but exhibited a normal distribution.
There was insufficient evidence either to support adoption of any one of the
commonly used models, such as beta, log normal or triangular
distributions. A more general and flexible model was thus used. The
Johnson SB distribution was found to represent the data distribution fairly
well as shown by the chi-squared goodness of fit test. Where it failed, this
could be attributed to insufficient data.
Johnson SB distribution was used to quantify the benchmarking gap. This

gap was validated on the basis of opinion surveys of construction
operatives potential. An assessment of the contractors' perception of the
operatives productivity potential indicated that within individual
organisations, the productivity gap was relatively small. When the
potentials were compared between organisations, it was evident that
benchmarking would greatly widen the scope of the contractors'
perceptions. The next step is to determine some of the possible actions
that may lead to productivity improvements. The main theme of Chapter 9
is to determine the practices that will lead to the realisation of the operative
productivity potential.
262
CHAPTER NINE
CONSTRUCTION LABOUR PRODUCTIVITY FACTORS
263
CHAPTER NINE
CONSTRUCTION LABOUR PRODUCTIVITY FACTORS
9.1 Introduction
Chapter 8 examined the variability of labour productivity in Tanzania at

operative level. Comparisons with productivity in some developed
countries indicated a large difference, not only in magnitude, but also in the
variability. Johnson SB distribution was established as a good distribution
model for operative labour productivity data in Tanzania.
The next important step in this work was to explain the variability of
productivity by identifying its sources. This task was divided into three
stages summarised below, and is described in detail in Sections 9.2,9.3
and 9.4 respectively.
" Construction productivity factors opinion survey.
The results of a rating survey of fifty construction operatives have

been summarised and evaluated for consistence. These were
used to rank the factors in their order of perceived importance. The
ranking was then validated by results of a similar independent
survey involving forty operatives. The results of the operatives
survey were compared with results of a similar study for contractors.
" Concreting productivity observations and evaluationof productivity

factors.
Factors identified in the opinion survey were evaluated on thirty-six

different sites during floor slabs concreting operations. Productivity
achieved in each of these sites was also recorded along with other
operation specific factors. These include operatives gang sizes in
mixing and placing the concrete, sizes of slabs, types of slabs,
productive times and productivity achieved in each pour.
Relationships between these factors have been explored.
264
9' Evaluation of influence of factors on concreting productivity.
The results of site factor evaluations on actual construction process,

along with operation specific factors, were then regressed with
productivity in order to identify significant factors.
The relationships between these stages have been illustrated in Figure 9.1.
265
START
Cl)
}
W
Cc:
Validate Independent factors
Ranking of 0-1
productivityfactors rankingby forty operatives
z
0 by fifty operatives
Compare Rankingof factors by
z seven contractors
0-
O
Conductsite productivitystudies Evaluaterated

T factors on each site
Z Determine dependent and independent variables

O
D Independent variables
-j Dependentvariables
j Observedfactors Rated factors
W Labour productivity
Element Wage level
F- concreting; Gang size Monetaryincentives
cn mixing; and Pour sizes Benefits
placing Productive time Skill level
Supervision
Mechanisation
Investigate basic relationships amongst variables
PERFORM MULTIPLE REGRESSION ANALYSIS
0
z
Rated factors weighed on linear scale
W
8 Ratedfactors weighedequally
z0 Rated factors weighed on average score
W
tY
0F COMPARE RESULTS
W
cc
Use appropriate factor weights for multiple regression analysis
Investigate non-linear models to Improve R2
Determine critical model
Figure 9.1: Productivity factors evaluation framework and

regression modelling
266
9.2 Summary and evaluation of productivity factors rating
9.2.1 Summary of factors ratings
An opinion survey, leading to the discovery of the relevant labour

productivity factors in the Tanzanian building construction industry, was
described in Chapter 6. Construction operatives formed the main focal
group for the survey, mainly because they are the key construction process
owners at site level. The survey was based on a structured questionnaire
focusing on the rating of pre-identified productivity factors. The process of
identifying these factors has been described in Chapter 6. An ordinal
rating scale, as described by Siegel (1988), was used for the evaluation.
The interviewees were asked to rate the factors on a scale of 'one' to 'six':
one representing factors that have no influence at all on productivity; and
'six' representing factors that have a very high influence on productivity.
A total of ninety skilled operatives were interviewed. Unskilled operatives are

normally employed on casual basis. Their involvement in construction
activities is only on an ad hoc basis and were therefore not included in this
study. The results of the interviews were divided into two independent
groups of fifty and forty. The results of the two groups were summarised and
compared. The ratings of the first group has been summarised in Table 9.1.
These results are graphically presented in Figure 9.2 using the mean factor
scores.
Table 9.1: Rating of productivity factors by fifty operatives
Productivity factors Scores Total Mean Ranking

score scores
1 2 3 4 5 6 Max. =30 Max. =6
Wages 1 1 4 4 20 20 251 5.0 2
Incentives 1 0 0 6 24 19 259 5.2 1
Other benefits* 4 0 2 20 14 10 220 4.4 3
Leadership 3 4 16 12 12 3 185 3.7 6
Level of mechanisation 1 6 16 12 11 4 188 3.8 5
Level of skill 3 6 11 10 12 8 196 3.9 4
Level of bureaucracy 14 17 13 5 1 0 112 2.2 10
Social environment 10 11 15 7 7 0 146 2.9 8
Weather, climate and 10 16 19 2 3 0 122 2.4 9
location
Level of supervision 9 3 20 8 9 0 152 3.0 7
*These include lunch, transport, uniforms, tools and in some cases basic medical facilities.
267
Financial incentives
Wages
Other benefits
Level of skill
Leadership
Social environment
Weather
2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0

Average factor score
Figure 9.2: Ranking of productivity factors by operatives
Twenty operatives rated 'wages' with the highest score of '6' whereas,
fourteen rated 'level of bureaucracy' with the lowest score of '1'. The total
score was computed from the relationship:
fisi where, m= number of rating categories;

i=1
f; = score frequency in scale i and

s; = score value in scale i.
The mean score is calculated by dividing the total score by the total number
of respondents. The mean score forms the basis of ranking.
The operatives ranked 'financial incentives' highest in the list of factors that
enhance their productivity. This was followed closely by the 'wage level'.
The provision of non-financial benefits, (such as transport, meals, tools and
uniforms), was ranked third. All these are associated with remuneration.
Factors such as the weather conditions were not ranked as important. It
should also be noted that close supervision was not considered favourable
by the operatives as far as productivity is concerned.
9.2.2 Evaluation of degree of agreement amongst operatives
The consistence of the operatives' ratings were evaluated by testing the null
hypothesis "that there was no significant agreement among the group on the
rating of the factors". The analysis was aimed at establishing that the ratings
268
had not been agreed upon by chance, and that there was a true agreement
amongst the operatives.
To test the hypothesis, non-parametric statistical tests were used (Siegel,

1988). These tests do not rely on the distribution of data, unlike most other
parametric tests. The test was performed using the Kappa statistic denoted,
K. The statistic is used in a typical situation where a group of N objects or
subjects, each of which is to be assigned to m categories by a group of /
raters (Siegel, 1988). The test statistic was applied to both groups of
respondents, that is the operatives and contractors' senior management.
In the operatives' rating, there were N= 10 factors, evaluated by I= 50

operatives, each assigning each of the factors to m=6 rating scales. The
Kappa coefficient of agreement, K, is the ratio of the proportion of times that
the raters agree (corrected for chance agreement) to the maximum proportion
of the times that the raters could agree (again corrected for chance
agreement) (Siegel, 1988):
P(A) - P(E)
K= ; P(A) is the proportion of time that the raters agree; and
1- P(E)
P(E) is the proportion of time the raters would be expected
to agree by chance.
If there is complete agreement among the raters, then K= 1; and if there is no

agreement, (other than that which would be expected to occur by chance)
among the raters, then K= 0. The equations for the computations of P(A) and
P(E) are shown below:
m
P(E) = P12 where p =N and
j=1
Cj is the number of time an object is assigned to category j.
In Table 9.2, for example, Cj is the sum of the column
frequencies.
To obtain the total of proportion of agreement P(A), the following relationship

is used:
269
[NIJ_ N 'm
P(A) 21
__ 1)il -I-1
where, N is the number of factors being rated;

m is the rating categories;
/is the number of raters; and
nj are the scores in the matrix.
Using the above relations, the computed values for P(A) and P(E) for the
operatives' productivity factors rating matrix are 0.255 and 0.180 respectively.
The computed value of Kfrom the relation given above is equal to 0.091.
After determining the value of the Kappa statistic, K, it is necessary to

determine whether the observed value is greater than the value that would
be expected by chance. Siegel (1988) stated that the value of K has been
found to be normally distributed with mean zero and variance, var(K). Again,
the computation of var(K) is given by the relation:
2 P(E) - (21- 3)I P(E)]2 + 2(I - 2)E p3

var(K)= NI(I-1)X [1-P(E))2
therefore,
K
Z
var(K)
This statistic was used to test the null hypothesis, Ho: K= 0 against the
alternative hypothesis, H1:K #0.
The computed value var(K) is 0.000468 which was then used to calculate z.
The computed z, using the relationship, is equal to 4.205. At 5 per cent level
of significance, z=1.645. Since the computed value is less than z0 05, it can
be concluded that there was significant agreement among the operatives,
and that the degree of agreement is beyond that which could have occurred
by chance. The null hypothesis was therefore rejected. The ranking of factors
presented in Figure 9.2 represents a consensus amongst fifty operatives.
270
9.2.3 - Validation, of productivity factor ranking
The fact that there was consistence among operatives within the survey of the
fifty operatives, as demonstrated above, did not validate the ranking order of
the factors. All it demonstrated was that within a single survey, there was a
degree of agreement beyond that which would be expected by chance. To
validate the order of importance of the productivity factors, the results of the
second independent group of forty operatives was used. Table 9.2
summarises the scores. These results are presented graphically in Figure
9.3.
Table 9.2: Validation factors rating by forty operatives
Productivity factors Scores Total score Mean scores Rankin

1 2 3 4 5 6 Max.= 240 Max. = 6.0
Wages 0 0 0 1 33 6 205 5.12 1
Incentives 0 0 1 13 21 5 190 4.75 2
Other benefits 11 8 12 5 1 3 106 2.65 7
Leadership 5 6 20 7 1 1 116 2.90 5
Level of skill 0 2 5 22 9 2 164 4.10 4
Social environment 11 6 16 5 2 0 101 2.53 8
Weather, climate and 8 9 16 5 1 0 99 2.48 9
location
Figure 9.3: Ranking of productivity factors by operatives
Again, the degree of agreement amongst operatives was evaluated using the
Kappa statistic on the summary of factor ratings in Table 9.3. The null
271
hypothesis would be rejected if the computed z is greater that z0 05,that is,
there was no agreement amongst the operatives beyond that which would be
expected by chance.
The computed values of P(A) and P(E) were 0.349 and 0.198 respectively.
The computed value of Kappa statistic K, was 0.187. and var(K) = 0.0000955.
Substituting these values in
K
=19.201.64.
var(K)
The computed K was, therefore, significantly greater than zero. The above
result led to the rejection of the null hypothesis, leading to the conclusion that
there was a degree of agreement amongst the forty operatives on the factors
that influence productivity beyond that which would be expected to occur by
chance. In fact, even at 1 per cent level of significance for which zo.01 = 2.32,
the null hypothesis would still be rejected.
The factors ratings were then compared with previous factors ratings by the
fifty of operatives in order to confirm the ranking order. The Kendal's
coefficient of concordance, W, is the appropriate test statistic for such
comparison. The test is used to determine the association among k sets of
rankings of N objects (or subjects) (Siegel 1988). In this case, two sets of
rankings of ten factors by two independent groups of operatives were
compared on the basis of the test statistic W, which was computed from the
relationship (Siegel 1988):
N
121 R? -3k2N(N+1)2
w= i=1
k2N(N2 -1)
where, k= number of sets of rankings, for example, number of groups;

N= number of objects being ranked; and
R; = sum of the ranks assigned to the ilh object.
The ranking of the set of factors in the two surveys are summarised in Table
9.3. For ease of interpretation, the results of the two groups are refereed to
as Group 1 and Group 2, corresponding to initial and validation survey
respectively.
272
Table 9.3: Factor ranking comparison in two independent groups
Productivity factors Group 1 Grou 2

Mean score Ranking Mean scores Ranking Ri R2
Wages 5.0 3 5.12 1 4 16
Incentives '5.2 1 4.75 2 3 9
Other benefits 4.4 3 2.65 7 10 100
Leadership 3.7 6 2.90 5 11 121
Level of mechanisation 3.8 5 4.38 3 8 64
Level of skill 3.9 4 4.10 4 8 64
Level of bureaucracy 2.2 10 2.10 10 20 400
Social environment 2.9 8 2.53 8 16 256
Weather, climate and location 2.4 9 2.48 9 18 324
Level of supervision 3.0 7 2.78 6 13 169
Substitutingfor relevant parametersin the equation,the value of W is:
W=0.948
To test the significance of W, the chi-squared statistic, x2, is used. This is

computed from:
x2 = k(N - 1)W with (N - 1) degrees of freedom.
Thus, x2 =2x (10-1) x 0.948 = 17.06. with (10 - 1) degrees of freedom.
If the computed value of x2 equals or exceeds the critical value at a

particular level of significance, and a particular degrees of freedom, (N - 1),
then the null hypothesis that the two rankings are unrelated may be rejected.
In this case, the critical value of x2 was 16.92 at a5 per cent level of
significance. The null hypothesis can be rejected, leading to the conclusion
that the operatives agreed on the ranking of the factors that influence their
productivity.
The analysis performed above established not only the factors that
construction operatives consider important in enhancing their productivity,
but also the order in which these factors were considered important. The
analysis of possible influence of these factors on construction labour
productivity based on actual site studies during concreting operations on
thirty six different sites is reported later in this chapter.
273
9.2.4 Comparative factor ranking by contractors' senior
management.
The contractors' interviews on the influence of certain factors on operatives

productivity was based on the same procedure as that of operatives.
Contractors were first asked to identify critical factors on the basis of a pre-
identified list of factors. The contractors list was therefore not exactly the
same as that of operatives. The procedure used has been described in
Chapter 6. This list was then used as the basis of the opinion survey.
Out of 27 contractors contacted to provide their views on the influence of the

pre-identified factors on labour productivity, seven agreed to be interviewed.
Most contractors said they were too busy and some only agreed to allow a
junior staff to be interviewed who, in the opinion of the researcher, did not
have sufficient knowledge or experience to provide a well informed opinion.
A summary and ranking of factors influencing productivity, based on the
opinions of the contractors' senior management, has been presented in
Table 9.4. The rankings have been presented in Figure 9.4.
Table 9.4: Rating of productivity factors by contractors' senior

management
Productivit factor Ratin Total scores Mean scores Rankin

1 2 3 4 5 6 Max. = 42 Max. =6
Level of skill 0 0 1 0 1 5 38 5.4 2
Wages 0 0 0 0 5 2 37 5.3 3
Monetary incentives 0 0 0 0 4 3 38 5.4 2
Other benefits 0 0 3 2 2 0 27 3.8 7
Level of interaction -skilled 0 0 0 0 3 4 39 5.6 1
Level of interaction -unskille d0 0 0 3 3 1 33 4.7 5
Future prospects 0 1 2 2 2 0 26 3.7 8
Weather and climate 1 1 2 1 2 0 23 3.3 9
Time of the day 0 1 1 3 2 0 27 3.8 7
Location 0 1 2 3 1 0 23 3.3 9
Work environment 0 0 2 4 1 0 27 3.9 7
Physical features 0 0 1 2 2 2 33 4.7 5
274
Figure 9.4: Ranking of productivity factors by contractors' senior
management
This survey indicated that contractors perceived that their interaction with
skilled workers as the main factor enhancing operatives' productivity. This
was followed closely by the operatives' level of skill, level of supervision,
monetary incentives and wage level. The differences between the scores
are, however, very marginal as indicated in Figure 9.4, and the factors could
as well be considered as being equally important. It is, however, clear that
factors such as future prospects, location of the project and climatic
conditions are not considered important in influencing worker productivity.
The degree of agreement amongst contractors is evaluated using the

procedure used for the evaluation of the contractors' degree of agreement.
The summary of factor ratings in Table 9.4 is used to compute the Kappa
statistic.The computed values of P(A) and P(E) are 0.270 and 0.250
respectively. The computed value of Kappa statistic K=0.031 and the
variance of K, var(K) = 0.001426. Therefore, substituting these values in
K
-0.77<1.64.
var(K)
The computed K is, therefore, not significantly different from zero. The above
result leads to the acceptance of the null hypothesis, leading to the
275
conclusion that degree of agreement amongst contractors could have been
by chance. It is very possible that the small sample size used in this analysis
may have influenced the outcome as the computation of var(K) relies on the
number of raters.
In spite of the conclusion that contractors' ratings could have been agreed
upon by chance, and that there was a high degree of agreement amongst
operatives, it was still possible to compare the results of the two groups. The
top six factors, in accordance with operatives' perception, were compared
with similar factors as perceived by the contractors, as shown in Table 9.5. In
summary, contractors agreed that five out of the six factors were important.
The order of ranking is similar for three of them which represents a
considerable degree of agreement. The contractors further indicated that
close supervision was important for improved productivity, ranking this factors
as very important. The skilled operatives did not support this view and
ranked this factor as moderately important. Operatives ranked factors related
to remuneration highest. While contractors recognise these as important,
they also ranked operatives skills and close supervision as equally important.
Table 9.5: Comparisons of factors ranking between contractors

and operatives
Factor Operatives Contractor Comments

rankin rankin
Monetary incentives 1 2 These represent similar ranking
Wages 2 3 scores in statistical terms
Other benefits 3 7 There is probably a difference
between the two groups
Level of skill 4 2 The difference is also small
Leadership 5 1 There is difference in perception
Level of mechanisation 6 4 There is little difference
9.2.5 Ranking of factors
The above analysis gave credibility to the operatives' ratings for further
analysis. Table 9.3 lists the total rating scores and the average score for
each factor. This formed the basis for the ranking, that is, the order in which
the factors were perceived to be important in influencing productivity.
Before the factors were considered for the next step in the analysis, a
276
decision was made on the list of factors that are considered significant. The
first six factors in the initial survey were selected on the basis of being
important, very important and extremely important (see Table 9.1). Since
average factor scores are not in integers, it was necessary to interpret
scores to the nearest integer. An average score of 3.7 for example was
taken to be equal to four which is equivalent to being important. Likewise,
an average factor score of 5.2 means the factor was very important. Thus,
factors scoring an average of more than 3.5 were selected as being
significant. These include:
9 financial incentives;
" wages;
" other non financial benefits;
" level of skill;
" level of mechanisation; and.
" quality of leadership.
These were used for possible analysis of their influence on operative

productivity. The rest of this chapter reports on this aspect.
9.3 Productivity and factors evaluation on construction sites
9.3.1 Concreting process and productivity observation
The second phase of labour productivity studies focused on concreting

activities as discussed in Chapter 6. The main reasons for concreting
productivity studies were:
" the significance of the cost contribution of concreting in relation to other

building activities;
" concreting would provide an opportunity to study a variety of factors due
to its complexity; and
" in spite of the complexity it is a relatively easy activity to observe and
quantify.
Activity sampling along with the physical measurement of output was used.
An evaluation of productivity factors during the process was also
performed. All concreting operations were performed by specialist
concreting teams which have become very common in Dar es Salaam in
277
the last five to eight years. Teams assemble each morning (at about six
o'clock) at a known location and leaders allocate teams to different sub-
contracts for the day. The teams were also responsible for hiring all
equipment for the task. These comprised a concrete drum mixer, spades,
vibrators and headpans. All materials were provided by the project's main
contractor. Concreting work would normally start between eight and nine
depending on how far the site was from the point of assembly. The
operations were subdivided into the following sub-operations:
Mixing: Manual volume batching of cement, sand and

aggregates. These materials were loaded into headpans
(about 10 - 15 kg) and then loaded into a drum mixer in a
predetermined ratio. Water was added manually. In
cases where pour sizes were more than about 50 m3, two
mixers were used. The mixer operator was generally
responsible for the quality and consistency of the mix.
Placing: Mixed concrete was then poured onto a heap from which
another group would load the concrete onto headpans
using spades. These would then be carried on the head
to the point of placing the concrete, either on the ground,
first or second floor slabs. Spreading, vibrating and
tamping the concrete was done by a team of two to four
masons depending on the size of the pour.
In general, all concreting operations were completed within the same day.
It was, therefore, necessary for the team leaders to have enough men and
equipment to complete the pour. A diagrammatic representation of a
general layout is shown in Figure 9.5.
278
0
Sand
Mixer r. f f.
Second floor slab
r. r"r r. r
. r. f. f.
First floor slab
Cement Fresh concrete
Ground floor slab, -,'
Water
Aggregates
Transporting
CONCRETE MIXING I CONCRETE PLACING
CONCRETING OPERATION
Figure 9.5: Schematic representation of concreting activity
During these operations, a number of parameters were observed and

recorded. A short summary of the main parameters observed in relation to
the 36 concrete pours are presented in Table 9.6. Further details are
presented in Tables 9.7 and 9.8 in which concrete mixing and placing are
treated separately.
.,
279
Table- 9.6: Concreting productivity study results
Site Pour size Slab type Volume Duration Gang size Productivity Productive
no. (in m3) poured (hours) (No) (in m3/hr) time (%)
(in m3)
1' 23.20 GFS 4.69 1.62 34 0.09 0.64
2 30.90 GFS 8.70 2.25 27 0.14 0.57
3 21.40 GFS 8.53 3.20 36 0.07 0.58
4 34.70 GFS 19.31 4.45 24 0.18 0.60
5 28.00 GFS 13.12 3.75 23 0.15 0.58
6 22.00 GFS --" 7.48 2.72 18 0.15 0.68
7 26.15 GFS 11.22 3.43 16 0.20 0.67
8 21.60 GFS 11.98 4.43 20 0.14 0.77
9 17.30 GFS 6.80 3.14 17 0.13 0.75
10 34.25 GFS 10.20 2.38 19 0.23 0.69
11 64.95 GFS 19.62 2.42 32 0.26 0.60
12 28.00 GFS 11.67- 3.33 27 0.13 0.58
13 28.45 GFS 11.14 3.13 21 0.17 0.76
14 20.35 GFS 6.70 2.64 22 0.12 0.75
15 43.55 GFS 17.59 3.23 27 0.20 0.67
16 38.40 FFS 20.49 4.27 23 0.21 0.65
17 51.10 FFS 20.01 3.13 28 0.23 0.63
18 26.35 FFS 10.09 3.07 19 0.17 0.72
19 30.00 FFS 8.61 2.01 30 0.14 0.68
20 11.50 FFS 5.30 3.68 17 0.08 0.82
21 20.20 FFS 7.52 2.23 29 0.12 0.55
22 75.80 FFS 7.86 0.93 58 0.15 0.54
23 50.80 FFS 7.30 1.15 41 0.15 0.60
24 40.70 FFS 5.03 0.99 36 0.14 0.62
25 45.10 FFS 10.15 1.80 28 0.20 0.66
26 39.20 SFS 11.65 2.08 40 0.14 0.63
27 49.20 SFS 15.48 2.52 30 0.20 0.72
28 55.20 SFS 14.17 2.57 44 0.13 0.56
29 28.60 SFS 13.58 3.80 46 0.08 0.72
30 47.40 SFS 13.76 2.03 59 0.11 0.60
31 64.95 SFS 18.40 2.27 66 0.12 0.74
32 11.25 SFS 5.54 3.45 19 0.08 0.72
33 37.50 SFS 11.25 2.70 32 0.13 0.70
34 33.80 SFS 8.26 1.47 33 0.17 0.63
35 8.00 SFS 1.94 1.93 19 0.05 0.66
36 24.65 SFS 4.05 1.15 35 0.10 0.46
Notes: GFS - ground floor slab; FFS - first floor slab; SFS - second floor slab
On site number 1 for example, the pour size was about 23 m3 and 4.69 m3
of concrete was poured for the ground floor slab (GFS) by a group of 34
operatives over a duration of 1.62 hours, the period in which the activity
sampling study was conducted. During this period an average productivity
280
of 0.09 m3 per labour hour was achieved working 64 per cent of the time. A
more detailed breakdown of concrete mixing and placing productivity have
been summarised in Table 9.7.
Table 9.7: Concrete mixing and placing productivity
Site Slab CONCRETE MIXING CONCRET E PLACING

no. type (1) (2) (3) (4) (1) (2) (3) (4)
1 GFS 16 0.20 0.37 55 18 0.16 0.23 72
2 GFS 11 0.30 0.41 74 16 0.24 0.53 46
3 GFS 18 0.15 0.25 58 18 0.15 0.25 58
4 GFS 10 0.40 0.73 55 14 0.31 0.48 64
5 GFS 10 0.36 0.49 75 13 0.27 0.60 45
6 GFS 8 0.25 0.33 76 10 0.28 0.44 63
"'7 GFS 7 0.94 1.40 67 9 0.36 0.55 66
8 GFS 9 0.59 0.80 73 11 0.25 0.31 81
9 GFS 10 0.30 0.38 78 7 0.31 0.43 71
10 GFS 9 0.38 0.50 75 10 0.43 0.67 64
11 GFS 12 0.67 0.96 70 20 0.41 0.75 55
12 GFS 12 0.32 0.42 77 15 0.23 0.27 85
13 GFS 9 0.40 0.51 78 12 0.30 0.40 75
14 GFS 11 0.55 0.69 80 11 0.23 0.33 70
15 GFS 11 0.82 1.09 75 16 0.34 0.56 61
16 FFS 9 0.45 0.62 73 14 0.34 0.57 60
17 FFS 11 0.41 0.57 73 17 0.38 0.67 57
18 FFS 8 0.53 0.68 79 11 0.30 0.44 68
19 FFS 14 0.38 0.48 80 16 0.27 0.47 57
20 FFS 6 0.69 0.73 95 11 0.13 0.17 75
21 FFS 11 0.30 0.46 64 18 0.19 0.39 49
22 FFS 26 0.33 0.55 60 32 0.26 0.54 48
23 FFS 15 0.27 0.49 56 26 0.24 0.39 63
24 FFS 18 0.15 0.30 52 18 0.28 0.39 72
25 FFS 10 0.54 0.72 75 18 0.31 0.51 60
26 SFS 15 0.39 0.52 75 25 0.22 0.41 55
27 SFS 13 0.37 0.50 74 17 0.36 0.51 71
28 SFS 17 0.32 0.55 58 27 0.20 0.37 55
29 SFS 15 0.39 0.51 77 31 0.12 0.16 70
30 SFS 20 0.29 0.46 63 39 0.17 0.30 58
31 SFS 24 0.45 0.63 71 42 0.19 0.26 75
32 SFS 9 0.53 0.69 77 10 0.16 0.24 68
33 SFS 13 0.29, 0.38 76 19 0.22 0.34 65
34 SFS 15 0.29 0.51 57 18 0.31 0.47 67
35 SFS 9 0.52 0.70 74 10 0.10 0.17 58
36 SFS 16 0.20 0.41 49 19 0.19 0.43 43
Notes: (1) - Gang size; (2) - Productivity in m3/Ihr;

(3) - Productivity rate in m3/Ihr; (4) - Productive time in %.
281
9.3.2 Basic concreting productivity statistics
Basic statistics of the 36 concreting productivity studies are summarised in

Table 9.8. These statistics were based on data presented in Tables 9.6
and 9.7. Three measures of productivity were considered. They include
output expressed in cubic metres of concrete per labour hour, productive
time, that is the percentage of time spent working by the operatives and
productivity rate which is the output of the operatives during productive
time. Other measures of the spread of the data, such as standard deviation,
coefficient of variation, and skewness have been presented. The range to
standard deviation ratio is also a measure of the spread of data.
Table 9.8: Concreting operation productivity basic statistics
Main parameter Type of Mean Standard Coeff. of Skewness Range

operation Deviation Variation Std Dev.
(1) (2) (3) (4) (5) (6) (D
Output mixing & placing 0.15m3/hr 0.05m3/hr 33% 0.25 4.2
mixing 0.41m3/hr 0.18m3/hr 43% 1.04 4.4
placing 0.26m3/hr 0.08m3/hr 32% 0.11 4.1
Productive time mixin & placinq 65% 8% 12% -0.09 4.5
mixing 70% 10% 14% -0.31 4.6
placing 63% 10% 16% -0.07 4.2
Productivity rate mixing & placinq 0.28m3/hr 0.08m3/hr 34% 0.20 4.3
mixin 0.58m3/hr 0.23m3/hr 39% 1.58 5.0
placing 0.42m3/hr 0.15m3/hr 35% 0.13 4.0
Table 9.7 shows that all the productivity (output per labour hour)
distributions are skewed to the left. This was found to be a typical
characteristic for most activities in Tanzania, as established in Chapter 8.
This indicated that even for a narrowly defined construction operation, such
as concrete mixing or placing, most of the operatives are producing below
the median. The coefficient of variation of productivity varies from 32 to 42
per cent. This is still high compared to that observed in studies elsewhere
(Noor, 1992). The coefficient of variation of productive time is considerably
lower than that of output per labour hour. This may indicate that there is not
much to be gained by improving the productive time, since the performance
gap using this indicator is relatively small. However, the specific
relationship between productivity and productive time has to be
282
investigated first. This will be performed later in this chapter. Frequency
histograms were drawn for mixing and placing productivity to provide a
visual illustration of the frequency distributions. These have been
presented in Figure 9.6.
ZU-
'
U
C
T
U
C
:3 a) 10-
U 7
Q
_N N
IL
0. 01 -_ u ,., 1 00 1.25 1.50

.. . .
in 3/1hr productivity rate in m3/Ihr
mixing productivity m
(a) Mixing productivity (b) Mixing productivity rate
6-
>
2- U
N
T C
U Q
N 8- 2
7
Cr
2
4
0
0. 00 0.0 8 ut, u" 't u_32 0.40 0.4 8
concrete placing productivity in m llhr concreting placing rate in m'/Ihr
(c) Placingproductivity (d) Placingproductivityrate
Figure 9.6: Frequency histograms concreting productivity
9.3.3 Exploration of some productivity relationships
Several parameters were observed during the study. The first step in the
determination of factors that influence productivity is to explore some
relationships between the parameters if any. The main parameters for
which relationships were explored include:
" concrete pour size;

" concrete mixing gang size;
" concrete placing gang size;
" slab type, that is ground, first and second floor slabs;
" productivity indicators; and
" productive time.
283
The first factor considered was gang balancing. Placing of concrete could
only be performed if there was mixed concrete. The mixing gang could
therefore significantly affect the productivity of the placing gang. On the
other hand, if the placing gang was too slow, the mixing gang would be
forced to stop mixing and wait for the mixed concrete to be placed
(otherwise too much mixed concrete would set). It is necessary to establish
that, in general, there was no significant delay caused by either of the
gangs. First, the mean ratio of mixing and placing productivity rate was
examined. This ratio was found to be 1.38, that is mixing productivity rate
was 1.38 times higher than concrete placing productivity rate. In general,
the rate of mixing the concrete should be equal to the rate of placing if the
operation is to be performed smoothly. The concrete placing gang should
therefore, be about 1.38 larger than the mixing gang on the assumption that
the productive times for the two operations are the same.
From Table 9.9, the mean gang size for mixing operation was thirteen while
for the placing operation was eighteen representing a ratio of about 1.4.
This reflects a good crew balance between the two operations. It is
observed in Table 9.8 that the average productive times for the two crews
do not appear to be the same. The mean productive time for the mixing
gang was 70 per cent while that of the placing gang was 63 per cent. A t-
test to determine whether there was a significant difference between the
two means was performed. At value of 2.98 was obtained which at 70
degrees of freedom represented a level of significance of 0.004 < 0.05.
This leads to the conclusion that the placing crew had less productive time
than the mixing crew. This may not have been unexpected because it was
also clear from the field observations that placing operation was physically
more demanding. This may have led to the need for more relaxation time.
In order to draw more specific conclusions with respect to optimal gang

ratios, relationships between productivity and gang ratios for ground, first
and second floor slab were investigated. The mean gang ratios were 1.24,
1.48 and 1.49 respectively. Again a t-test indicated a significance
difference between ground floor slabs and the first and second floor slabs.
The relationship between productivity and gang ratios are shown in Figure
9.7 and 9.8. A linear relationship is observed for the ground floor slab as
shown in Figure 9.7. The higher the ratio, the higher the productivity, that
is, the smaller the mixing gang in comparison to the placing gang, the
284
1.8
1.6
1.4 .
ca .
I- .
.
rn . U
1.2
rn
tm
X1
E
0.8
C

A 0.6
0.4
0.2
0
0 0.05 0.1 0.15 0.2 0.25 0.3
productivity in m3/hr
Figure9.7: Gang size ratio and productivity for ground floor slab (GFS)
2.5
O
2 O
0 O+
+a
'- +
1.5 G
rn '0
rn
c 0+ 13
O 13
.1
C

0.5
13 FFS
+ SFS
0
0 0.05 0.1 0.15 0.2 0.25
.
productivity in m3/hr
Figure 9.8: Gang size ratio vs productivity for FFS and S FS
285
higher the productivity. Figure 9.8 shows no clear relationship between
productivity and gang ratios for first and second floor slabs.
In general, gang size should be proportional to the size of the pour since
most concrete pours have to be completed within one day. This
relationship was explored for mixing and placing activity for GFS, FFS and
SFS. The scatter diagrams for these relationships have been presented in
Figures 9.9 and 9.10, which in general, show that there is a linear trend. In
two separate pours of similar sizes on the ground floor slab, marked GFS3
(that is, ground floor slab on site Number 3- see Table 9.7 and 9.8) and
GFS6, the concrete mixing gang sizes were 18 and 8 respectively. This
was reflected not only in productive times, but also in productivity achieved.
Productivity was 66 per cent higher in GFS6 when compared to that of
GFS3 which has a larger gang size. Similar observations can be made
between FFS16 and FFS24 for which pour sizes were about 40 m3
whereas gang sizes were 9 and 18 respectively. Consequently, productive
times were 73 and 52 per cent and productivity of 0.45 and 0.15 m3/Ihr
respectively were achieved. Similar analysis could be made with respect
to the concrete placing activity shown in Figure 9.9, in which a concrete
pour of 47 m3 on the second floor slab had 39, operatives whereas a
similar slab of about 49 m3 was placed by 17 operatives.
In general, there was little rationale in the determination of the appropriate

gang size and the study may establish some guidelines in deciding on the
optimal gang size. However, relationships incorporating a number of
factors such as productivity, gang size, type of slab and pour size have to
be explored and will be discussed later on in this chapter.
286
80
13
70
I +
60
50 13 13
+
E t
U C3++a
.n FFS24
40 0 FFS16
A
G)
N
-f
0
30
0 "o
0. +_ GFS
GFS6' p GFS3
20
o FFS
O+
10
t t SFS
5 10 15 20 25 30
numberof operatives
Figure 9.9: Pour size and mixing gang size relationship
70
60
1-
50
Cl) t
SFS2 t
SFS3
7
E 0
U 40
4+
U
C
03 4-
N 30 1
N . p I-
13
L
0
a
20 .+ 0
GFS
o
v FFS
10 t
t t SFS
1
number of operatives
Figure 9.10: Pour size and placing gang size relationship
287
Figure 9.11 - 9.12 present some relationships between productivity and
gang sizes. In general, the larger the gang the lower the productivity. In
the previous paragraph, it was established that unnecessarily large gangs
were being used. There were indications that these may have resulted in
low productivity. Congestion could have been one of the possible
contributors of the low productivity. However, the extent of congestion was
not quantified in this work. As hinted in the preceding paragraph, the
relationship between gang size and productivity cannot be studied
exclusively because other factors such as pour size have some influence.
' GFS
0.9
1
0 o FFS
0.8
0 + SFS
0.7
0
0-
0.6
aD
E 1+o'
Z 0.5
13 +
0.4 0
+0
0
23 0.3 '13 ++++
0
a
0.2
13
0.1
0
0 10 20 30
Figure 9.11: Mixing productivity vs gang size
288
n nr,
V. -TIP
"
GFS
0.4
13 FFS
0 0.35 +
+SFS
0
.n 0.3 + 130
p
a- 13
ci
0.25
"U++
E
0.2 +
U C+ -F
++
0.15
>1 13 +
U
0.1
0
EL 0.05
0
0 10 20 30 40 50
Figure 9.12: Placing productivity vs gang size
One of the main objectives of this study was to investigate the relationship
between productivity and productive time. Indeed, the activity model used
in this research relies on the assumption that increasing productive time
has a positive influence on productivity. There have been a number of
studies examining the relationship between productivity and productive
time. These were extensively discussed in Chapter 4 and 5. In general,
the studies have focused on the effect of increased productive time on
productivity. Thomas (1991) concluded that there is no linear relationship
between productivity and productive time. His conclusions were based on
a site wide 'productive time - productivity' relationship unlike, in this study
which focused on the relationship at activity level. There is justification,
therefore, in the examination of 'productivity - productive time' relationships
in the concreting work for the 36 sites. Figures 9.13 - 9.14 depict
graphically relationships between productive time and productivity for
mixing and placing operations respectively. In this respect, correlation was
performed between 'productive time and productivity' for the mixing
289
operation for all the 36 concrete pours. The graphs depict GFS, FFS and
SFS separately for ease of visualisation. Investigations will be performed
later to identify other factors that influence productivity. In general, Figure
9.13 indicates that as productive time increases productivity also increases
(with R2 = 24 per cent). This does not represent a strong correlation,
nevertheless, a trend exists. I
An examination of Figure 9.14 of 'productive time and placing productivity'

does not suggest a linear relationship between the two, irrespective of slab
type. This suggests that there are other more significant factors that
influence productivity and the effect of productive time is overshadowed.
Other factors are considered later in this chapter.
100
13
90
80
.r+ '+ L3 n+U
70
60
v
0 'c' GFS
50
,
Q FFS
40
+ SFS
30
0 0.2 0.4 0.6 0.8
productivity in cubic metres per labour hour
Figure 9.13: Mixing productivity vs productive time
290
90
80'
C3 +
70
m -E- - 0+
+ C+
E +
60 C3 15
tU + "+ UO
++
0 50
DD" GFS
0
+0
40 FFS
+ SFS
30
0 0.1 0.2 0.3 0.4 0.5
productivity in cubic metres per labour hour
Figure 9.14: Placing productivity vs productive time
The next relationship involved productivity rate against productive time.

Productivity rate is a measure of net productivity, that is productivity
excluding any delay. This provides an indication of the effectiveness of the
methods and skills being applied. If these are the same in similar
production processes, then productivity rate will be the same and output
will depend on productive time. As in
shown the preceding paragraph,
productivity is only remotely dependent on productive time. A visual
inspection of Figure 9.15 shows that there is a positive correlation between
productivity rate and productive time for the mixing operation. Unlike the
mixing operation, there is a negative correlation between productive time
and productivity rate, that is the higher the productive time the lower the
rate as shown in Figure 9.16. This form of relationship would be expected
because as more time is spent working, the effort to accomplish work
cannot be sustained throughout the working period. The contrary is
observed for the mixing operation which supports an observation
mentioned earlier that the mixing operation seemed less physically
demanding. In spite of the mixing gang spending more time working, with a
mean productive time of 70 per cent in comparison to 63 per cent for the
placing gang, their productivity rate is not affected by the increase in
productive time.
291
100
E3
90
80
w IL , ,9
E
. +" +I 0
CD 70
+-13
++ 0
t
2 60
n
CL " L+ GFS
10
50 13
o FFS
+ SFS
40
30
0 0.5 1 1.5
productivity rate in cubic metres per labour hour
Figure 9.15: Mixing productivity rate vs productive time
90
U
GFS
U
80
o FFS
0) + SFS
+
70
m -I- 13+
f
w
U' 13
60 O
O
0
d ++
50 13 13
40
30 -{
0 0.2 0.4 0.6 0.8
productivity rate in cubic metres per labour hour
Figure 9.16: Placing productivity rate vs productive time
292
9.4 Site evaluation of factors that influence productivity
9.4.1 Summary of evaluation of factors
Several factors that may have some influence on productivity were

recorded during the concreting operation. Some of these have been
mentioned in the preceding section. Some relationships between them
have also been explored. These included:
" size of the pour;

" sizes of the gangs, both for mixing and placing;
" type of slab, ground, first and second floor slabs; and
" percentage productive time for both gangs.
Factors identified by the operatives as significant (discussed in Section 9.2)

were evaluated and rated for each site. These included:
" financial incentives;

" wages;
" other benefits;
" level of skill;
" leadership;
" level of supervision; and
" layout and site organisation.
It should be noted that an additional factor relating to the general layout

and organisation of the site was introduced in the evaluations. This was
found necessary because it was noted that on different sites, concreting
materials stockpiles were located in such a way that they could influence
productivity. The location of the mixing areas was selected on the basis of
the most convenient location with respect to batching and to a lesser extent
on transporting and the placing of concrete. The layout was therefore
evaluated with respect to its perceived influence on productivity on each
site. The results of the rating of these factors on each site are shown in
Table 9.9. It should be observed that there was little difference in the
incentives and benefits given to the workers on different sites. There was
also no significant difference in the level of mechanisation between the
sites. Differences were observed in wage level, skill level, level of
293
supervision and quality of leadership. These factors are therefore
considered later in the regression model. This is not to say that the other
factors are not important, but rather the differences observed on sites are so
little that they are unlikely to have any significant differences in influence if
any.
Table 9.10: Rated site productivity factor scores
FA CTOR SCORES
Site No Incentive Wages Other Mechanise Skill Supervise Leadersh Layout &
benefits ip organ.
1 3 4 4 3 2 3 3 2
2 3 4 3 3 5 4 4 5
3 3 3 3 4 4 6 4 4
4 3 3 3 3 5 3 4 5
5 3 4 3 3 5 4 4 5
6 3 4 3 3 4 4 4 3
7 3 3 3 3 5 4 3 5
8 3 3 3 3 4 3 3 4
9 3 4 3 4 4 4 4 5
10 3 4 3 3 6 5 4 4
11 3 3 1 3 4 5 3 4
12 3 5 3 3 4 3 3 4
13 2 4 2 3 5 5 4 4
14 2 3 2 .3 3 4 4 4
15 3 4 3 3 5 5 4 4
16 3 3 5 3 4 4 4 4
17 3 3 3 3 2 3 3 2
18 3 3 3 3 5 5 3 5
19 3 4 3 3 4 3 4 4
20 3 5 3 5 3 2 2 3
21 2 3 3 3 4 5 4 4
22 3 3 3 3 6 5 4 5
23 3 3 3 3 3 3 4 4
24 3 3 3 3 1 1 1 1
25 3 3 3 3 3 3 3 2
26 3 3 3 3 5 3 2 5
27 3 3 3 3 4 5 4 5
28 3 3 5 3 5 5 4 5
29 3 3 2 3 5 4 4 5
30 3 4 3 3 6 4 4 3
31 3 4 3 3 5 5 4 5
32 3 3 3 3 2 4 3 3
33 3 3 3 3 4 5 4 5
34 3 3 3 4 4 3 3 4
35 3 5 3 3 3 4 3 3
36 3 3 3 4 4 5 4 4
294
Some of the rated factors were included in regression modelling to identify
influential factors and their effects. The process of building relevant
regression models is described in the following section.
9.4.2 Productivity and productive time comparisons
Concreting productivity observations were made on three separate

elements, that is ground floor slabs (GFS), first floor slabs (FFS) and
second floor slabs (SFS) as shown in Table 9.6. Investigations into the
general trends between some key parameters were performed in Section
9.3. It was pointed out that a combination of parameters rather than single
parameter had some influence on productivity. This requires some
investigation of the combined effect of the parameters on productivity. This
is known as multiple regression modelling. However, in order to combine
the results of the three types of slabs in the regression model, it was
necessary to establish that there were no.statistically significant differences
in the productivity achieved among the groups. Productivity of both mixing
and placing operations were investigated focusing on two measures:
" productivity in cubic metres of concrete mixed or place; and

" productive time in the mixing and concreting operations.
These investigations were performed by testing the hypothesis that the

achieved measures of productivity were not the same for the three types of
elements. Analysis of variance tests (ANOVA) are appropriate for such test.
The technique examines the variability of the observations within each
group as well as variability between the group means. Based on the three
estimates of variability, conclusions can be drawn about the population
means. One-way ANOVA was appropriate here because only one variable
was used to classify cases into different groups. ANOVA test results at 5
per cent level of significance performed on Microsoft Excel 4 indicated that
there was significant differences in placing productivity for the three types of
slabs. However, there were no differences for the mixing activity. These
results are shown in Table 9.10.
295
Table 9.10: - ANOVA tests on concrete mixing and placing
productivity for different slab types
SS df MS F P-value F crit
Mixing productivity
Between 0.036 2 0.018 0.560 0.576 3.28
Groups
Within 1.050 33 0.0318
Groups
Total 1.086 35
Placing productivity
Between 0.044 2 0.022 3.776 0.033 3.28
Groups
Within 0.194 33 0.059
Groups
Total 0.239 35
Mixing productive time

Between 0.005 2 0.003 0.259 0.773 3.28
Groups
Within 0.346 33 0.011
Groups
Total 1.085 35
Placing productive time

Between 0.011 2 0.006 0.554 0.580 3.28
Groups
Within 0.339 33 0.010
Groups
Total 0.239 35
Regression modelling for the mixing operations can thus be performed

irrespective of the slab type. Placing productivity was found to be
dependent on slab type. This would require building up different multiple
regression models for all the three slab types. Since there was a limited set
of data for the different floor slabs (only 15,10 and 11 for the GFS, FFS and
SFS respectively), it was decided to proceed with multiple regression
modelling, for the mixing operation only, for all the concrete pours.
Regression modelling for the mixing operation is discussed in Section 9.5.
296
9.5 Regression modelling
9.5.1 Basis of modelling
Regression models have been used to analyse the variability of a

.
dependent variable, or criterion, by resorting to information on one or more
independent variables or predictors (Pedhazur and Schmelkin 1991). In
this work, ' a number of factors that could influence productivity have been
investigated. The objective was, therefore, to build up a regression model
aimed at explaining the variability of productivity for concreting mixing
operations. This is known as multiple regression analysis (regression
analysis with multiple independent variables). The variables used for
model building were subdivided into two groups as listed in Section 9.4.1:
" observed factors during concreting operation; and

" rated factors are those perceived by the operatives to influence their
productivity and ranked as shown in Figure 9.4.
Several methods of incorporating rated factors into the regression model

were considered. These relate to the weights attached to the factors as
follows:
" linear weighting in which factors ranked highest were allocated a weight
of six and those ranked lowest allocated a weight of one;
" factors are allocated equal weight irrespective of ranking; and
" factors are allocated weighting equivalent to the average factor score in
the rating survey.
The analysis process leading to the building of the regression model has
been summarised in Figure 9.1. Multiple regression analysis was
performed on SPSS version 6.1 using the stepwise method. The stepwise
method includes in the model, only those variables which have a significant
influence on the dependent variable in an interactive step-wise process.
The variables that are ultimately retained in the model are referred to as the
critical variables, or in this case critical factors.
Initial results of the multiple regression analysis using different weighting

criteria for the factors did not suggest differences in the critical factors. A
linear weighting of the factors was therefore pursued for further analysis.
297
The results of the regression analysis of the mixing operation are presented
and discussed in Section 9.5.2 with respect to their statistical significance
measured by the F statistic and the accuracy of the model as quantified by
the coefficient of determination R2. Analyses of residuals were investigated
as hints towards an exploration of possible non-linear relationships in
attempts to improve the accuracy of the model.
9.5.2 Concrete mixing regression models
General regression models were constructed using all variables, that is,
measured and rated factors. Two dependent variables were investigated
independently for the mixing operation. The variables were:
" concrete mixing productivity (in m3 per labour hour); and

" productive time (in percentage).
The following factors were included in the investigation of all three

variables.
Measured factors:
" pour size;
" type of slab; and
" size of the mixing gang.
Rated factors:
" wage score
" benefits;
" layout;
" supervision;
" skills;
" financial incentives; and
" level of mechanisation.
The first step involved generating general regression models with all the
factors included. Some analyses of variance results of the general models
are presented in Table 9.11.
298
Table 9.11: General regression models - analysis of variance
Dependent variable: concrete mixing productivity
Sum of Deg. of Mean

Source Squares Freedom Squares F-Ratio Prob>F R2
Model 0.529 21 0.044 1.901 0.089 0.498
Error 0.557 23 0.024
Total 1.086 35
Dependent variable: concrete mixing productive time
Model 0.201 11 0.018 2.907 0.014 0.571

Error 0.151 24 0.006
Total 0.352 35
The general models had R2 values of 49 and 57 per cent with F values of
1.83 and 2.91 for productivity rate and concrete mixing productivity
respectively. At a level of significance of 5 per cent, Fcrit= 2.2 for 11 and 24
degrees of freedom for the regression and residuals respectively. These
results are of little value because all factors are included, and the relative
importance of each factor is not clear. It was, therefore, necessary to
eliminate from the model variables that do not significantly contribute to the
value of R2. Stepwise regression analysis is designed to eliminate such
variables in a stepwise manner until a model which meets the set
significance level is achieved.
Concrete mixing productivity was defined as productivity per labour hour in

the mixing operation. The general regression model for this variable was
reduced to the following sub-critical model through stepwise regression.
All factors were not significant except productive time, pour size and gang
size.
mixing productivity = 0.193 + 0.510 x productive time

- 0.022 x gang size
+ 0.005 x pour size;
R2 = 40%
Similarly, for concrete mixing productive time the following sub-critical

regression model was obtained through stepwise regression analysis. All
299
other factors were not significant except the wage level on site and
concrete mixing gang size.
productive time = 0.65 + 0.010 x wage score - 0.011 x mixing gang size;
R2 = 43%
Since gang size had a negative correlation with both variables, then the
productivity model can be improved by transforming the gang size. Several
trials were made aimed at improving the coefficient of determination. An
inverse of the square of gang size was used improving R2 from 40 per cent
to about 50 per cent as shown in the concrete mixing critical model in Table
9.12. The F-ratio indicated that the model was significant.
Table 9.12: Critical concrete mixing productivity model
Variable Std. Err. t

Name . Coefficient Estimate Statistic Prob >t
Constant -0.165 0.187 -0.882 0.384

pour size 0.004 0.002 2.621 0.013
productive time 0.337 0.274 1.229 0.228
1/(gang size)A2 22.245 5.606 3.968 0.000
Sum of Deg. of Mean

Source 2
Squares Freedom Squares F-Ratio Prob>F R
Model 0.541 3 0.180 10.586 0.000 0.498
Error 0.545 32 0.017
Total 1.086 35
The critical concrete mixing model can be expressed in the following form:
22.245
Ptm = -0.165 + 0.004 x psm + 0.337 x prtm +
9sm2
R2,= 49.8%
where Pt,,, = productivity over total time for mixing operation;
psm = pour size expressed in cubic metres;
prtm = percentage productive time for the mixing operation; and
gsm = the size of the concrete mixing gang expressed in number of
operatives.
300
This model has the following general implications with respect to
productivity:
" the larger the pour, the higher the productivity, therefore construction
operations should be scheduled so as to maximise pour sizes;
" increased productive time has a positive effect on concrete mixing
productivity; and -
" larger gangs have a negative influence on concrete mixing productivity.
This model did not directly incorporate any of the factors identified in the
opinion survey. This may lead to rather simplistic conclusions with respect
to the influence of productivity factors. - The study environment has to be
analysed with respect to the observed parameter variability and the
possible influence that this may have had on the regression modelling.
This aspect has been discussed in Section 9.6.
9.6 Discussion of regression results and productivity factors
The main objective in this chapter was to identify factors that may have
some influence on operative productivity. In Section 9.2, results of
operatives' opinion survey were analysed. Several factors were identified.
Factors ranked highest were related to operative remuneration. Other
factors such as quality of supervision and leadership were also identified.
These factors have influence on motivation, a subject that is not new in

construction productivity research. These include wage level, financial
incentives, other benefits, supervision and leadership. These factors play a
significant role in motivating the operatives. Indeed, the operatives ranked
motivation factors as key factors that influence their productivity, as
established in section 9.2.
Contractors perceived that their interaction with skilled workers was the
main factor enhancing the operatives productivity. This was followed
closely by the operatives' level of skill, level of supervision, monetary
incentives and wages. Some of these factors have direct influence on
motivation. Skill level, identified by the contractors, is very significant in
influencing productivity. The expectancy productivity model suggested by
Maloney (1991) and discussed in Chapter 5 recognised that even when
operatives were highly motivated, if they do not possess the relevant skills,
301
their productivity would not reflect the level of motivation. This relationship
has been partially validated by Olomolaiye (1989).
Concreting operations in all the 36 sites were all executed using very
similar methods. There was little variations in the level of mechanisation for
the different sites as indicated in Table 9.10. Similarly, there were
insignificant variations in the benefits and incentives provided to the
operatives. There was some variability in the wages, level of skill, quality of
supervision and leadership. However, these factors did not have a
significant influence on productivity in comparison to other measured
factors. However, the fact that productive time was a significant factor
indicated that the motivating factors had an indirect influence. Olomolaiye
(1987) established that percentage productive time was influenced by
motivation which then had a direct influence on output. In any case, the
measured factors could explain about 50 per cent of the mixing productivity
variability. This is a step in the right direction. The variability caused by the
other factors remain unquantified. A different approach to the study of the
influence of these factors has to be used, possibly by adopting an
experimental study where the factors could be influenced at will. This
approach could form a basis for future work.
9.7 Summary
This chapter has presented the results of an opinion survey aimed at

identifying labour productivity factors at crew level based on factors rating.
Operatives factors ratings were found to be consistent. The results of the
survey resulted in the ranking of factors in the order of their influence on
operatives productivity. The ranking of factors was verified through an
independent survey. The factors were evaluated on thirty six concrete
pours on different sites with the objective of evaluating their influence on
productivity. Multiple regression modelling incorporating the evaluated
factors along with other physical factors measured on each site were used
to establish their relationship to concrete mixing productivity. Concrete
pour size, mixing gang size and productive time were found to be the
critical factors. The fact that productive time emerged as a significant factor
suggested that motivating factors are important. However, the influence of
individual factors could not be quantified due to the limited variation across
the study sites.
302
CHAPTER TEN
CONCLUSIONS AND RECOMMENDATIONS FOR

FURTHER RESEARCH AND ACTIONS
303
CHAPTER TEN
CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER RESEARCH

AND ACTIONS
10.1 Introduction
This research was motivated by the need to improve construction industry

performance in Tanzania, where there was concern about its poor
performance (Ministry of Works 1977; Wells 1987; Lema and Mutabazi
1988; Ministry of Works 1992; Maro et at. 1992 and Bjorklof et at. 1992).
Early initiatives to improve its performance failed because they aimed at
broad based policy issues which did not address root causes of the poor
performance. Only recently, the Tanzanian government published a policy
paper aimed at guiding the development of the industry, with emphasis on:
increased productivity, efficiency in design, construction and distribution of
resources; establish realistic construction costs and cost trends; and
promotion of research and application of research findings (Ministry of
Works 1992). This research was aimed at uncovering some strategies that
would help improve the performance of the construction industry in line with
some of the policy objectives. This research has achieved the following:
" macro-performance trends were analysed over a considerable period

and comparison with other national construction industries indicating
downward trends;
" Total Quality Management (TQM) was identified as an appropriate

performance improvement framework with benchmarking as a tool for
kick-starting and sustenance the programme;
"a site labour productivity benchmarking model was developed and

validated on the basis of actual observations and opinion survey results;
and
" site practices that influence productivity were identified mainly through
operatives opinion survey and actual site evaluations.
This chapter presents these findings in some detail together with identified
areas that require further research.
304
10.2 Conclusions
10.2.1 Continuous construction productivity improvement philosophy
This research has concluded that construction productivity is still very much
a subject for further research, especially within the continuous performance
improvement concepts such as TQM. TQM, a management philosophy,
widely applied in the manufacturing sector in developed countries, offers a
promising future for the construction industry. TQM focuses on process
performance improvement, whereas traditional indicators focused on
financial indicators which could not explain what happened in the value-
adding process. In spite of this recognition, historical performance
evaluation relies heavily on traditional indicators, mainly because these are
the only historical indicators available. This was the philosophy adopted in
this work.
TQM provides a general framework upon which to embark on a long-term

business performance improvement programme. In Chapter 3, it was
argued that the quantification of a performance gap provides a motivation
for an organisation to embark on a continuous improvement programme.
Benchmarking was identified as a powerful tool that enables organisations
not only quantify the performance gap with competitors but also uncovers
practices that can provide a competitive edge. Thereafter, benchmarking
can be used for continuous performance comparisons. Where competition
has not yet thrived, internal performance analysis can be used for gap
quantification. In this study, a benchmarking framework has been used
primarily as a tool for kick-starting the continuous productivity improvement
by:
" establishing the labour productivity gap; and

" uncovering some key practices that can be adopted to improve
performance.
In this respect, the study established a methodology for benchmarking

labour productivity, that is, establishing both labour productivity metrics and
proposing methodologies for both identifying and adopting best practices
within the TQM philosophy as a possible framework upon which to initiate
and sustain productivity improvement efforts.
305
10.2.2 Tanzanian construction industry performance
This research argued that the initiation of the performance improvement

efforts have to be made at national level for effectiveness, as discussed in
Chapter 5. Individual contractors have neither the resources nor the
expertise to implement a benchmarking programme. In order to create
performance awareness at the national level, an analysis of the
construction industry performance was performed. The analysis has been
presented in Chapter 7. Performance trends over a period of 25 years
were analysed using several indicators. Analysis of output, contribution to
the economic growth, productivity, and creation of employment were
evaluated from 1969 to 1993. Some limitations of these data were
discussed. The industry's performance declined on the basis of all of the
indicators examined. Productivity was evaluated using output per person
engaged (OPE) and value added per person engaged (VAPE). OPE was
erratic with no significant improvement in 1993 compared to 1969 figures.
A more bleak picture was observed when VAPE trends were analysed. By
1993, VAPE had decreased to about 72 per cent of 1969 levels. A
productivity comparison with the Kenyan construction industry over the 25
years showed that the Tanzanian construction industry had lagged behind
that of Kenya when the following performance indicators were examined:
" output as quantified by value added by the construction industry;

" contribution to the GDP;
" productivity as quantified by VAPE; and
" contribution to employment.
Kenya's construction industry productivity in 1993 was found to have

dropped to nearly one third of the 1969 level although in absolute terms it
was still higher than that of Tanzania. On this basis, it was concluded that
the Tanzania construction industry had little to learn from its Kenyan
counterpart in the aspect of productivity improvement. Comparisons with
British construction industry were generally aimed at providing
transparency to the analysis. Of significance in the comparison over the
period 1969 -1993 were:
" peak UK output in 1989 was about 60 times that of Tanzania;

" OPE in UK was generally about three to five times that of Tanzania; and
" VAPE in UK was between four and nine times that of the Tanzania.
306
The research established that there is a strong relationship between
productivity as measured by VAPE and the GDP per capita based on
international level comparisons. Construction productivity for 69 countries
based on 1979 statistics indicated a strong relationship. These observations
were confirmed using wider and more recent data. An evaluation of the
productivity of 87 world economies based on 1992 statistics, confirmed that
there was a strong relationship between productivity and GDP per capita.
Sweden had the highest VAPE of, US $65,619 in current prices. Others with
high VAPE were Japan, Switzerland, Federal Republic of Germany and
Belgium. Lowest productivity were observed in Nigeria, Benin, Malawi,
Pakistan and Romania. The UK construction industry productivity of US
$33,977 was slightly more than half that of Sweden and average for its
income, while USA productivity of US $ 26750 was below average for its
income. Tanzanian productivity was slightly below average for its GDP. The
Tanzanian construction industry was observed to be less productive than 19
other countries within the low income group (as defined by the World Bank
(1994)). This is a good enough reason to want to improve the industry's
productivity. These countries form a possible pool from which Tanzania can
learn. Further, the Tanzanian construction industry's contribution to the GDP
per capita was the lowest of all the 87 countries investigated. This is a
matter of serious concern. A wider international comparison indicated that
while both Tanzanian and Kenyan productivity decreased, productivity of
five other industrialised countries increased. It was also noted that Kenyan
and Tanzanian GDP per capita continued to decrease. The analysis
performed in Chapter 7 provided a firm justification for embarking on
performance improvement efforts.
10.2.3 Labour productivity at operative level
Determination of an appropriate methodology for data collection
This research argued that productivity improvement efforts should focus at

operative level because that is where actual production takes place. On the
basis of TQM, the construction worker was not only seen as an important
customer in the construction process chain, but also as the ultimate supplier
in the chain. Besides understanding what this customer wants, it is
important to understand the process performance at this level. Labour
productivity was identified as a key measure of process performance.
307
Initial surveys indicated that there was no reliable labour productivity
database in Tanzania. It was necessary to collect primary data from
construction sites. A practical methodology that could be used to quantify
the productivity gap had to be used. Labour productivity models were
reviewed with the objective of identifying a model suitable for
benchmarking. Activity models were selected because they can:
" describe the construction process and quantify the process performance
variables such as output, resources expended and therefore productivity;
" quantify labour unproductive time, a factor that could be utilised for
performance improvement;
" identify factors and practices that may have some influence on the
process; and
" be widely applied due to their simplicity especially in view of to lack of
expertise and resources that other models would demand in Tanzania.
A basic labour productivity benchmarking model was developed to identify

the current mean productivity (CMP) and target mean productivity (TMP).
TMP was used as the labour productivity benchmark. The benchmark, in
this context, served as a current asymptotic upper labour productivity figure
for which a productivity gap analysis can be made for individual
construction companies. Conceptually, the productivity gap was equal to
the difference between TMP and the CMP. Ideally, TMP cannot be
achieved because it is dynamic, that is, as productivity improves, TMP also
moves further to the right thereby providing a continuously moving
productivity target - an essential characteristic of benchmarking. The
model relied on the variability of the productivity which was assumed to be
normal for simplicity. This assumption had to be verified.
Productivity variability and improvement potential quantification
Chapter 8 examined labour productivity variability for eight building

activities on 46 different sites belonging to 23 contractors with the objective
of determining labour productivity performance improvement potential.
Labour productivity was measured in terms of work done per unit labour
time, that is, output per labour hour, unlike in some other studies in which
productivity is sometimes expressed in time required per unit of output
(Price 1985; Emsley et al. 1990 and Thomas et al. 1990).
308
All construction productivity data analysed were found to be skewed to the
left. These observations were consistent when both frequency histograms
and box and whiskers methods were used. The spread of the data, as
measured by the coefficient of variations ranged from 46 to over 100 per
cent, indicating a significant productivity improvement potential. Similar
studies in Britain and United States, indicated that the coefficient of
variation was between 5 and 35 per cent (Noor 1992).
A systematic methodology for the determination of a distribution model was

developed. There was no obvious theoretical model that could explain the
distribution of labour productivity data for construction activities, empirical
models were then used to model the data. Normality tests were performed
using visual checks, standard error of skewness, Wilk-Shapiro tests and
chi-squared tests for goodness of fit (Hanh and Shapiro 1994). There was
no strong evidence to suggest a normal distribution model and where this
was observed, it was merely attributed to chance. This specifically referred
to the ceiling board productivity distribution which had the smallest sample
size but exhibited a normal distribution. There was insufficient evidence
either to support adoption of any one of the commonly used models, such
as beta, log normal or triangular distributions.
A more general and flexible model was thus used. The Johnson SB
distribution was found to represent the data distribution fairly well (Hahn
and Shapiro 1994). Goodness of fit tests using the chi-squared test
provided sufficient grounds to adopt the Johnson SB model. Where it
failed, this could be attributed to insufficient data. The analysis established
that Johnson SB distribution with shape parameters, il =1 and y= 1 were
relatively good distribution models for labour productivity distribution in
Tanzania. On the basis of this model, it was concluded that 85 per cent of
the productivity was below the median, that is, only a small proportion of
operatives had a high productivity, most of them had very low productivity.
This was found to be contrary to findings elsewhere where most of the
productivity was high and the proportion of low productivity was
comparatively low (Van Slyke 1963; Price et al. 1985; Emsley et al. 1990
and AbouRizk 1992). This formed the basis for the modification of the
benchmarking concept.
The Johnson SB distribution was then used to quantify the benchmarking

gap. The model established that the CMP was only about 30 per cent of
309
the maximum observed productivity. On the basis of the model, the TMP or
labour productivity benchmark was about 70 per cent of the maximum
observed productivity. The potential for productivity improvement was
therefore more than twice the current productivity level. For blocklaying for
example, CMP was 0.69 m2 per labour hour. The TMP or productivity
benchmark using the model was about 1.60 m2 per labour hour,
representing a 133 per cent increase. CMP for plastering was found to be
1.32 m2 per labour hour, while the computed TMP was 3.08 m2 per labour
hour. This potential was the same for activities whose productivity
distribution could be modelled using Johnson SB distribution with shape
parameters il =1 and y=1. The same concept would apply for
distributions with other shape parameters.
This potential had to be as validated. An opinion survey amongst 48

mason from 14 sites belonging to 13 different contractors to evaluate their
productivity potential was performed. Their perceived hourly average
productivity for blocklaying was about 0.75 m2 per labour hour, and the
mean of their perceived maximum productivity was about 1.75 m2 per
labour hour. A t-test indicated that there was no significant difference
between the computed TMP and the perceived maximum productivity. The
model was therefore producing achievable productivity targets. An
assessment of the contractors' perception of the operatives productivity
potential indicated that within individual organisations, the productivity gap
was relatively small. When the potentials were compared between
organisations, the gap was much higher, and closer to the productivity
benchmark actually observed. This indicated that contractors did not have
a clear perception of what labour could achieve, nor did they know what
was being achieved in other companies. It was evident that benchmarking
would greatly widen the scope of the contractors' perceptions. The next
step was to determine some of the possible actions that may lead to
productivity improvements. This focused on factors that influence operative
productivity.
10.2.4 Factors that influence operative productivity
TQM philosophy advocates that construction process improvement efforts

should focus on key process owners. Therefore, improvements at
operative level need to focus at this level. A rating opinion survey amongst
Tanzanian construction operatives was performed on the influence of pre-
310
identified factors on their productivity. Two independent groups of fifty and
forty operatives were surveyed. An analysis of the rating using the Kappa
statistic, K, (Siegel 1981) indicated that there was a high degree of
agreement amongst operatives for both groups. The rating was used as
the basis for the ranking of factors in the order of importance. Again, factor
ranking for both groups was found to be consistent on the basis of
significance test on Kendall's coefficient of concordance, W. Six factors
were identified as most important. These included:
" financial incentives;

" wages;
" other non-financial benefits;
" level of skill;
" level of mechanisation; and
" quality of leadership.
Operatives' rankings were compared with that of the contractors. In spite of

the limited data from the contractors, there was agreement on four out of the
six factors identified by operatives. Contractors were of the opinion that
close supervision was an important factor that influences productivity.
Operatives did not agree. Contractors thought that interaction with skilled
operatives was an important factor. The same was not the case with
unskilled operatives. Operatives' ranking was used for on-site evaluation
for productivity influence.
10.2.5 Evaluation of productivity factors on sites
Activity sampling along with the physical measurement of output was used
to evaluate concreting productivity on 36 sites. An evaluation of
productivity factors during the process was also performed. An analysis of
productivity distribution indicated that it was skewed to the left -a typical
characteristic for most activities in Tanzania. Several factors that may have
some influence on productivity were recorded during the concreting
operation. These related to the physical conditions pertaining to the
operation such as: size of the pour; sizes of the gangs, both for mixing and
placing; type of slab, ground, first and second floor slabs; and percentage
productive time for both gangs. Other pre-identified factors in the
operatives opinion survey were evaluated and used to generate a concrete
mixing regression model. The following model was found to be critical.
311
22.245
Ptm = -0.165 + 0.004 x psm + 0.337 x prtm +
gsm 2
R2 = 49.8%
where Ptm = productivity over total time for mixing operation;

psm = pour size expressed in cubic metres;
prtm = percentage productive time for the mixing operation; and
gsm = the size of the concrete mixing gang, i. e. number of operatives.
Pour size, percentage productive time and gang sizes were found to be the
critical factors. In general, higher productivity was observed in larger pour
sizes. There was no attempt in this research to explain why this was the
case, although larger pours could influence productivity because all the
pour had to be completed in one day. The workers would therefore feel
obliged to put in more effort for larger pours. Productive time had some
influence on productivity, contrary to claims by some researchers (Thomas
et al. 1991, Horner 1992). Gang size had an influence on productivity, and
the larger the gang the lower the productivity. Again, there was no attempt
to explain this relationship in this research. It is possible however that
congestion, resulting from gangs of more than optimal size, had a negative
influence on productivity. On the other hand, it is also possible that due to
large gangs, the operatives commitment to work is diminished.
The measured critical factors could explain about 50 per cent of the mixing
productivity variability. The fact that productive time emerged as a
significant factor suggested that motivating factors were important.
However, the influence of individual factors could not be quantified due to
the limited variation across the study sites. This is a step in the right
direction. The variability caused by the individual factors remain
unquantified. A different approach to the study of the influence of these
factors has to be used, possibly by adopting an experimental study where
the factors could be influenced at will. This approach can form a basis for
future work.
312
10.3 Recommendations for further research and actions
10.3.1 Performance measurement in the construction industry
Indicators of performance
The industry still puts much reliance on historical financial performance

indicators whose value in influencing process performance is limited. This
research identified the need to adopt a wider view of construction
performance indicators following the acceptance of the TQM philosophy.
The identification of relevant indicators is largely unexplored. Issues such
as, customer satisfaction in the construction process chain, team working at
various levels, empowerment and continuous improvement, do not have an
immediate relevance in an industry which is fragmented. Tackling
fragmentation and aligning objectives amongst practitioners is currently an
area which requires serious research effort.
Productivity measurement
This research discussed the concept of productivity in some detail. The

advantages and disadvantages of different measures of productivity were
highlighted. Construction productivity continues to utilise partial measures
of productivity, the most common of which is labour productivity. As a
partial measure, it can only present partial results and therefore partial
explanations. Effects of mechanisation or off-site prefabrication cannot be
captured by this measure. Comparisons between countries with different
levels of technology are therefore of little value. There is a need to adopt a
wider measure such as total factor productivity so that the effects of key
inputs can be evaluated. Such research would be particularly useful in the
choice of technology options in developing countries.
Links between performance indicators at various levels
This research focused on construction productivity improvement at

operative level. The reasons for this have been discussed. Concentration
at this level assumed that improvement at operative level would be a
significant impact on the overall construction industry performance.
However, this impact could not be quantified in this research. Indeed, there
has been little research on the effect of operative level performance on
313
performance at project, corporate and industry levels. For example, the
extent to which operative productivity improvement will affect industry level
productivity is an area which requires further research, even in developed
countries. With the adoption of the TQM philosophy in the construction
industry, a wider scope of performance indicators has to be adopted. This
is an area which requires further work in order to ensure that performance
improvement efforts are well directed.
10.2.3 Framework for implementation of research findings
Communication and exchange of information in the construction industry is

a problem in developing countries due to poor infrastructure, lack of interest
amongst practitioners and weak trade associations. Dissemination of
research results is therefore difficult. The benchmarking framework
discussed in Chapter 5 identified three steps that would enable the
implementation of the research findings. These included:
" communicating the research results;

" developing action plans; and
" implementation of plans to bridge the performance gap.
In practice, benchmarking results are implemented within an organisation

to improve its performance. There have been limited studies targeting the
industry as a whole. Modalities of adoption of such results have therefore
not been investigated, even in developed countries. This research argued
that initiatives for such implementation have to be taken by the respective
government for effectiveness, especially in developing countries. It is clear
that such implementation has to adopt a long-term continuous improvement
view. This research argued that continuous improvement philosophy such
as TQM provides a general framework upon which such efforts can be
initiated and sustained. At the time of conducting this research, there was
little knowledge, if any of modern performance improvement programmes
in the Tanzania construction industry. Interviews conducted with twelve
executives in the industry indicated that only one had an in-depth
knowledge of TQM. For the success of continuous improvement
programmes in developing countries, initial efforts are necessary to inform
executives and decision makers of modern performance improvement
philosophies. The form and delivery of such knowledge is subject to further
investigation.
314
10.2.4 Widened scope of performance comparison
This research argued that meaningful construction performance

comparisons and possible benchmarking can be made on comparable
economies. There would be little to learn by the Tanzanian construction
industry if comparisons of performance and practices were made with say,
Japan, the USA or the UK. The technological, organisational and
environmental differences would make it difficult to implement findings.
There could be better chances to learn from similar construction industries
and adopt an incremental continuous improvement. Learning from one
another in Sub - Saharan African countries is one such option. However,
there have been few studies on construction labour productivity in these
countries. There is a need for similar studies in other countries in the
region in order to widen the pool from which productivity improvement
lessons can be drawn. Similar cross - country studies within a geographic
region are common. For example, there have been numerous comparative
construction performance studies in the European Union. Similar studies
in Eastern and Southern African countries would facilitate comparisons.
10.2.3 Detailed studies on labour productivity factors
This research identified productivity factors through an opinion survey.

These were evaluated on actual construction operations. Their variability
across the sites were very limited. Their influence on productivity through
regression modelling could not be quantified due to this factor. It is
generally difficult in practice to encounter wide variability because industry
practice tends to be uniform, that is workers in the industry tend to receive
similar treatment from contractors. In order to evaluate the influence of
radical actions, it is necessary to perform pilot studies where factors could
be influenced at will. Differential treatment of operatives on different sites
could provide some clues as to the actual influence of productivity factors.
The above areas of further research have prompted drawing up of a list of

actions and responsibilities as shown Table 10.1.
315
Table 10.1 Proposed list of actions and responsibility
Proposed action Responsibility

Compare productivity with other Eastern and National Construction Council
Southern African countries (NCC)
A comprehensive programme for improving the Ministry of Works and NCC
general performance of the construction industry
has to be drawn and implemented
Executivetrainingon modernmanagement NCC
approaches
Initiate specific productivity benchmarking scheme NCC
for building contractors
Evaluate own productivity and compare with Construction companies
established benchmarks
Focus attention on operative motivation and skills NCC and Construction
improvement companies
Monitor productivity improvement on continuous Construction companies/ NCC
basis
Improve own skill level through training Construction operatives
Launcha comprehensiveresearchprogrammeon NCC and researchinstitutions
constructionindustryperformancein general and Universityof Dar es
Salaam
It is hoped that the implementation of the actions proposed above will

contribute towards the improvement of the Tanzanian construction industry
productivity and performance in general.
316
REFERENCES
317
REFERENCES
Abdel-Razek, R. H., 1987. Estimating innacuracy and tender variability: causes,

evaluation and consequences. Unpublished PhD Thesis, Loughborough
University, UK.
AbouRizk, S. M., 1992. Statistical properties of construction duration data.

Journal of Construction Engineering and Management, ASCE, 118(3), pp. 525-
544.
Aird, D. C., 1963. Manpower utilisation in the Canadian construction industry. A

pilot study, National Research Council of Canada, Division of Building Research,
Ottawa, Canada.
Alfeld, L. E., 1988. Construction Productivity: On-Site Measurement and

Management. McGraw Hill, N. Y.
Allan, A. L., Hollwey, J. R. and Maynes, J. H. B., 1968. Practical Field Surveying
and Computation. Heinman, London.
Barnes, R. M., 1980. Motion and Time Study. Design and Measurement of Work.
7th Edition, John Wiley - Interscience, New York.
Baxendale, A. T., 1987. Measuring site productivity by work sampling.

Proceedings of the C/B W65,5th symposium on OMC, London, pp. 821-823.
Benchmarking -A Management Process, 1992. Du Pont Corporation, USA.
Bendell, T., Kelly, J., Merry, T. and Sims, F., 1993. Quality: Measuring and
Monitoring. Century Business, UK.
Bernard, D. A., Ferragut, T. R. and Neumann, D. L., 1973. Production Efficiency

Study on Large Capacity, Rubber Tired Front Loaders. Report No. FHWA-
RDDP-PC-520, Federal Highway Admonistration, Washington, DC, UAS.
318
Bienefeld, M. A., 1968. The Construction Industry in Tanzania. Paper No. 69.4,
Economic Research Bureau, University College, Dar es Salaam.
Bishop, D., 1966. Architects and productivity - 2. Building, 211(49), pp. 79-82.
Bjorklof, S., Klingberg, T., Muhegi, B. and Temba, A., 1992. Birches Don't Grow
in Serengeti - An Evaluation of Development Policy and Nordic Support to the
Construction Sector in Tanzania 1975 - 1990. S IDA, Stockholm.
-
Bon, R., 1990. The world building market, 1970 -1985. Proceedings, C/B 90
Joint Symposium on Building Economics and Construction Management,
Sydney, 14-21 March, 1, pp. 16-47.
Borcherding, J. D. and Gamer, D. F., 1981. Workforce motivation and productivity

on large jobs. Journal of the construction Division, ASCE, 107(CO3), pp. 443-453.
Borcherding, J. D. Sebastian, S. and Samuelson, N. M., 1980. Improving

motivation and productivity on large projects. Journal of the Construction Division,
ASCE, 106, pp. 73-89.
Borcherding, J. D., 1976. Improving productivity in industrial construction.

Proceedings of ASCE. 102(CO4), pp. 599-613.
Bresnen, M. J., Haslam, C. 0.. Beardworth A. D., Bryman, A. E. and Keil, E. T.,
1990. Performance on Site and the Building Client, CIOB, Ascot, UK.
Briscoe, G., 1988. The Economics Construction Industry. B. T. Batsford Ltd,

of
London.
British Institute Management, 1956. Outline Work Study, Part 1, Introduction.

of of
British Institute
of Management. London.
British Standards Institution, 1992. BS3138 Glossary Terms Used in

of
Management Services. British Standards Institution, London.
319
British Standards Institution, 1992. BS7850 Total Quality Management.. British
Standards Institution, London.
Brooks. C.E.P. and Carruthers. N.. 1953. Handbook of Statistical Methods in

Meteorology. London. H.M.S.O.
Brown, M., 1969. On the Theory of Measurement of Technological Change.

CambridgeUniversity Press, UK.
Buckley, J. W., Buckley, H. M. and Chiang, F. H.. 1976. Research Methodology

and Business Decisions. National Association of Accountants, New York, N. Y.
Burati, J. L. Mathews, M. F. and Kalindini, S. N., 1991. Quality management in

construction industry. Journal of Construction Engineering and Management,
ASCE, 117(2), pp. 341-359.
Burati, J. L. Mathews, M. F. and Kalindini, S. N., 1992. Quality management

organisations and technique journal of construction Engineering and
Management, ASCE, 118(1),
pp. 113-128.
Burch, T., 1979. An Causes of Delay Encountered on Construction

assessment of
Projects. MSc Thesis, University Salford.
of
Burton, F. M.. 1991. A for ACEE

method measuring construction productivity.
Transactions,pp. L3.1-4.
CaMP, R. C., 1989. Benchmarking: The Search for the Industry Best Practice
that Lead to Superior Performance. Quality Press, Milwaukee, USA.
Casfey, D. J. Lury, D. A.. 1981. Data Collection in Developing Countries.

and
Oxford University Press.
Central Bureau Statistics. Survey, 1994. Ministry of Finance

of 1995. Economic
and Planning, Dar
es Salaam. Tanzania.
320
Chandra, M., 1993. Total quality management in management development.
Journal of Management Development, 12(7), pp. 19-31.
Channon, D. F., 1978. The Service Industries. Macmillan, London.
Chase G. W. 1993. Effective total quality management in construction. Journal

of Management in Engineering, ASCE, 9(4), pp. 433-443.
Chau, K. W. and Walker, A., 1988. The measurment of total factor productivity of
the Hong Kong construction industry. Construction Management and Economics,
6(3), pp. 209-224.
Chohan, V. V., Ghor, F. M. and Hassuji, S. S., 1982. Productivity in the building
construction Industry in Tanzania since 1949. Conference proceedings, Annual
Seminar, Institution of Engineers, Tanzania, Arusha, pp. wc1 - wc13.
Chou, Y. M., Turner, S., Henson, S., Meyer, D. and Chen, K. S., 1994. On using
percentiles to fit data by a Johnson distribution. Communications in Statistics:
Simulation and Concepts, 23(2), pp. 341-354.
Christian, J. and Hachey, D.,1995. Effects of delay times on production rates in

construction. Journal of Construction Engineering and Management, ASCE,
121(1), pp. 20-26.
Clemmens, and Willenbrock, 1977. The SCRAPESIM computer simulation.

Journal of the Construction Division, ASCE, 104(4), pp. 419-435.
Codling, S. 1995. Best Practice Benchmarking: a Management Guide.

Aldershot, Gower.
Commission
of the European Communities, (1993). Strategies for the
Construction Sector. 1993.
Construction Industry Development Board, 1988. Construction Sector in a

Developed Singapore, CIDB, Singapore.
321
Construction Industry Institute, 1992. TOM -An Industry Necessity. CII Annual
Conference, 1992.
Construction Monitor, (1994). Building. The Builder, Ltd. London, No. 7852.
Construction Specifier, (1992). TOM, partnering lead to industry productivity

trends. August, pp. 17-1B.
Costanzo, L, 1993. Benchmarking: top of the class? Engineering, 233(8), pp.

27.
Cottrell, J., 1992. Favourable recipe. The Total Quality Management Magazine,
February, pp. 17-20.
Currie, R. M., 1977. Work Study. 4th Edition, Pitman Publishing, London.
Davies, P., 1993. Help or Hazard? The Total Quality Management Magazine,
April 1993, pp. 35-38.
Department of Employment, 1990. Employment Gazette. Department of

Employment, HMSO, London.
Department of Environment, 1994. Digest of Data for the Construction Industry.

(Second edition), Government Statistical Service, HMSO, London.
Department of Statistics, 1989. Modelling Input process with Johnson

Distributions. Purdue University, Purdue.
Deurinck, G.. 1955. Productivity Measurement. Organisation for the Economic

Cooperation
and Development, Paris.
Drewer, S., 1980. Construction Habitat

and development: a new perspective.
International, 5,
pp. 395-428.
Drewin, F. J., 1985.

Construction Productivity. American Elsevier, New York, N. Y.
322
Duff, A. R., Pilcher, R. and Leach, W. A., 1987. Factors affecting productivity
improve,ment through repetition. CIB Proceedings W65 - London, 2, pp. 634-645.
East African Association of Architects, 1977. The East African Standard Method
of Measurement for Building Works. Nairobi, Kenya.
Easton, G. R. and Woodhead, R. W., 1988. A survey of Motivation and Productivity

on Australian Construction Sites. Report prepared for the Australian Project
Managers Forum, School of Civil Engineering, University of New South Wales,
Australia.
Edmonds, G. A. and Miles, D. W. J., 1984. Foundation for Change: Aspects of

the Construction Industry in Developing Countries. Intermediate Technology
Publications, London.
Edum-fotwe F. T., 1995. A Framework for Strategic Management of Construction

Contractors. Unpublished PhD Thesis, Loughborough University
Emsley, M. W., Duff, A. R. and Zambetakis, D., 1990. Bricklayers productivity

measurement. C/B 90, Sydney, Australia, pp. 161-171.
European Construction Institute, 1996. Implementing TO in the Construction

Industry practrcal Guide. Thomas Telford, London.
-A
Eurpean Construction Institute, 1993. Total Quality in Construction. June 1993.
First find your bench, 1991. The Economist, 319(7706), pp.72.
Fisher, D.. Miertschin, S. Pollock, D. R., 1995. Benchmarking in the

and
construction industry. Journal of Management in Engineering, ASCE, 11(1),
PP. 50-57.
Fleming, M. C., 1980. Construction Related Professions. Royal

and the
Statistical Society Social Sience Research Council, Pergamon, 1980.
and the
323
Forbes, W. S., 1971. Dimensional disciplines and the output of bricklayers.
Building, August, 221(35), pp. 82.84.
Freeman, M. and Beale. P., 1992. Measuring project success. Project

Management Journal, 23(1), pp. 8-17.
Gardiner, V. and Gardiner, G., 1979. Analysis of Frequency Distributions. Geo

Abstracts Ltd, Norwich, U.K.
Garnett, N. and Pickrell, S., 1995. Construction industry benchmarking.

Unpublished Seminar Proceedings, School of Business and Industrial
Management, Kent, UK.
-
Gary, R. S. and Thomas, H. R., 1990. A conceptual model for estimating

productivity on change order work. C/B - 90, Sidney, Australia, pp. 498-509.
Gates, M. and Scarpa, A., 1972. Learning and experience curves. Journal of
Construction Division, ASCE, 98(CO1), pp. 79-101.
Gilbreth, F. B., 1911. Bricklaying Systems. McGraw Hill, New York.
Gilbreth, F. B., 1911. Motion Study. D. Van Nostrand Co. Princeton, N. J. 1911.
Grant, E. and Stevens, A. J., 1982. House building productivity: activity

sampling study at Crawley. Internal Note N73, BRE, 1982.
Gregerman, 1. B., 1981. Productivity of construction labour using

measurement
activity sampling. ASCE Transactions, 1981 pp. 181-185.
.
Groak, S., 1994. Is industry? Construction Management and

construction an
Economics, Vol. 12,
pp. 287-293.
Guevara, J. M. & Boyer, L T., 1981. Communication problems within

construction. Journal of Construction Division, ASCE, 107(CO4), pp. 551-557.
324
Hahn, G. J. and Shapiro, S. S., 1994. Statistical Models in Engineering. John
Wiley and Sons, New York.
Halligan, D. W., Demsetz, L. A., Brown, J. D. and Pace, C. B., 1994. Action
response model and loss of productivity in construction. Journal of
Construction Engineering and Management, ASCE, 120(1), pp. 47-64.
Hammer, M. and Champy, J., 1993. Re-engineering the corporation -A

manifesto for Business Revolution. Nicholas Breadley Publishing, London,
1993.
Hammer, M., 1990. Re-engineering work, don't automate, obliterate. Havard

Business Review, July August, pp. 104-122.
-
Handa, V. K. and Abdalla, 0., 1989. Forecasting productivity by work sampling.

Construction Management and Economics, 7, pp. 19-28.
Harris, F. and McCaffer, R., 1995. Modem Construction Management, 4th Edition,
Blackwell Science Ltd, Osney Mead, Oxford.
Harvey, R. C. and Ashworth. A., 1993. The Construction Industry of Great

Britain. Butterworth-Heinemann, Oxford.
Heizelman, J. R., 1981. Productivity vs closed shop. Transactions of AACE,

- open
June 28 July 1,1981, Toronto, Canada.
-
Hellard, R. B., 1993. Total Quality in Construction Projects. Thomas Telford,

London.
Hillebrandt, P. M., 1975. The

capacity of the industry. In Aspects of the
Economics of Construction, Turin, D. A. (ed. ) George Godwin.
Hillebrandt, P. M., 1984. Analysis British Construction Industry.

of the
Macmillan, London.
325
Hillebrandt, P. M., 1985. Economic Theory and the Construction Industry. 2nd
edn, Macmillan London.
Homer, R. M. W. and Talhouni, B. T., 1990. Causes of variability in bricklayers

productivity. C/B-90, Syndey, Australia, pp. 238-250.
Homer, R. M. W., 1992. Productivity, the key to control. The Practice of Site
Management, Vol. 4, The Chartered Institute of Building, pp. 166-175.
Housing and Construction Statistics, 1970 - 1994. Government Statistical

Services, HMSO, London.
Howeinstein, E. J., 1975. Productivity trends in the construction industry. AACE

review, December, pp. 189.
Huntsberger, D. V., 1961. Elements of Statistical Inference. Wm. C. Brown,

Dubuque, Iowa, USA.
Iliffe, J., 1980. Wage Labour and Urbanisation. In Tanzania Under Colonial
Rule. Kaniki (ed.), Longman Group, London.
ILOIDFK, 1983. Draft Report an 1LOMission to Tanzania to Advice on

of
Contractor Support. International Labour Office, Geneva.
International Labour Office, 1979. Introduction to Work Study. 3rd (Revised)

Edition, International Labour Office, Geneva.
International Labour Office, 1983. Building, Civil Engineering and Public Works
Committee, General Report, International Labour Office, Geneva, pp. 120-127.
International Monetary Fund, 1995. international Financial Statistics, XLVIII(8).
International Standard: Quality Assurance Standards, Guidelines for Selection

and Use, 1987. International Organisation for Standardisation, Geneva,
Swtzerland.
326
International Standards Organisation, 1987. Quality Assurance Standards,
Guidelines for Selection and Use. International Organisation for
Standardisation, Geneva, Switzerland.
Ireland, V. J., 1992. Productivity in the construction industry. W65-1992 Meeting,

Delft University of Technology, Belgium.
Juran, J. M., 1988. Juran for Planning for Quality. The ' Free Press, New York, N. Y.
Karlof, B. and Ostblom, S., 1993. Benchmarking: A Signpost to Excellence in

Quality and Productivity. John Wiley and Sons, 1993.
Karsnia, A. L., 1991. Toward world class development: benchmarking to

improve project management practices. Project Management Institute Seminar
Proceedings, Dallas Texas September/October, pp. 1-9.
Kearney, A. T. Inc., 1991 Creating the Environment for Total Quality

Management. Keamey AT Inc. USA.
Kearney, A. T. Inc., 1992. Total Quality Management: A business Perspective.

A. T. Kearney, USA.
Kellog, J. C., Howell, G. E. and Taylor, D. C., 1981. Hierarchy model of construction.
Journal of the Construction Division, ASCE, 107(CO1), pp. 137-152.
Kendrick, J. W., 1961. Productivity Trends in United States. Bureau of

Economic Research, Princeton University Press, Princeton, New Jersey.
Kishimbo, S. I., Lema, N. M., Mamiro, J., 1992. Standard formwork

development. Seminar Proceedings, Research and Development Issues in the
Construction Industry in Tanzania, National Construction Council, Dar es
Salaam, Paper No. 7.
Koskela, L., 1992. Application of New Production Philosophy to Construction.

Technical Report No. 72. Centre for Integrated Facility Engineering,
Department of Civil Engineering, Stanford University, California, September.
327
Kumari, A., 1993. Productivity in public sector - an analysis at industrial group
level. Economic and Political Weekly, Bombay, India, November.
Lake, G. D. and Ulrich, D., 1992. Introduction to exemplary practices. Human

Resources Management, 31(1) & (2), pp. 1-7.
Larkin, J. A., 1969. Work Study - Theory and Practice. - McGraw Hill, London.
Latham, M., 1994. Constrructing the Team. Joint Review of the Procurement and
Contractual Arrangements in the United Kingdom Construction Industry, Final
Report.
Laufer, A. and Bocherding, J. D., 1980. Financial Incentives to raise productivity.

Journal of Construction Division, ASCE, 107(CO4), pp. 745-756.
Lema N. M. and Mutabazi, M., 1988. The need for review of bid evaluation
framework in Tanzania. Proceedings of the 1988 Institution of Engineers
Tanzania Annual Conference, Arusha, Tanzania, pp. NML1- 13.
Lema, N. M. and Price, A. D. F., 1995a. A model for construction industry

performance stimulation for a developing economy. Proceedings of First
International Conference on Construction Project Management, 11 - 12
January, Nanyang Technical University, Singapore, pp. 373 - 382.
Lema, N. M. and Price, A. D. F., 1995b. Benchmarking: performance

improvement toward competitive advantage. Journal of Management in
Engineering, ASCE, 11(1), pp. 28-36.
Lema, N. M. and Shirima, A., 1992. Establishment of schedule of rates project -

Phase 3- 1992, unpublished report to the National Construction Council, Dar
es Salaam, Tanzania.
328
Lema, N. M. Price, A. D. F. and Abdul-Kadir, M. R. M., 1994. Conceptual
linkages between benchmarking, Total Quality Management (TQM) and
construction productivity. Proceedings of the 10th ARCOM Conference,
Loughborough, UK, pp. 485 - 495.
Lema, N. M., Maro, W. E., Moshi, H., Likumbo, S. S., 1992. Construction cost
trends study. Proceedings of seminar on research and development issues in
construction, National Construction Council, Dar es Salaam Tanzania, Paper
No. 6.
Lemna, G. J., Borcherding, J. D. and Tucker, R. L., 1986. Productive foremen in

industrial construction. Journal of Construction Engineering and Management,
ASCE, 112(2), pp. 192-210.
Leonard, C. A., Fazio, P. and Moselhi, 0., 1988. Construction productivity: major
causes of impacts. AACE Transactions, pp. D10.1-D10.7.
Liebfried, K. H. J. and McNair, C. J., 1992. Benchmarking for Continuous

Improvement Human Resources Management. Harper Business, New York.
Lim E. C., 1996. The Analysis of Productivity in Building Construction.

Unpublished PhD Thesis, Loughborough University.
Liou, F. and Bocherding, J. D., 1986. Work sampling can predict unit rate
productivity. Journal of Construction Engineering and Management, ASCE,
112(1), pp. 90-103.
Logcher, R.D. and Collins, W.W., 1978. Management impacts on labour

productivity. Journal of Construction Division, ASCE, 104(CO4), pp. 447-461.
Lothian, N., 1987. Measuring Corporate Performance. The Chartered Isntitute

of Management Accountants, UK.
Lowe, G., 1987. The measurement of productivity in the construction industry.

329
Luck; D. J. Wales, H. C. and Taylor, D. A., 1961. Marketing Research,
,
Englewood Cliffs. Prentice Hall, Inc.
Macdonald, J., 1993a. Demise of the gurus. The Total Quality Management
Magazine, December, 1993, pp.5-6.
IP
Macdonald, J., 1993b. TQM: Does it always Work? Technical Communications,

Letchworth, England.
Main, M., 1992. How to steal the best ideas around. Fortune International,
126(8), pp. 86-89.
Maloney, W. F. and Federle, M. 0., 1991. Organisational culture and management.

Journal of Management in Engineering, ASCE, 7(1).
Maloney, W. F. and McFillen, J. M., 1983. Research needs in construction worker

performance. Journal of Construction Engineering and Management, ASCE 109,
pp. 245-254.
Maloney, W. F. and McFillen, J. M., 1986. Motivation in unionised construction.

Journal of Construction Engineering and Management, ASCE 112, pp. 122-135.
Maloney, W. F., 1981. Motivation in construction -a review. Journal of

Maloney, W. F., 1983. Productivity improvement : the influence of labour. Journal

of Construction Engineering and Management, ASCE, 109(3), pp. 321-334.
Maloney. W. F., 1990. Framework for analysis of performance. Journal of

Construction Engineering and Management, ASCE, 116(3), pp.395 - 415.
Maro, W. E., Moshi. H. P., Lema, N. M., Likumbo, S. S. and Kagoro, L., 1992.
Construction Cost Trends and Control Mechanisms in Tanzania. Final Report,
National Construction Council, Dar es Salaam.
Masao, F. T., 1974. The irrigation systems in Uchagga: An ethno-historical

330
approach, Tanzania Notes and Records (Tanzania), pp. 1-8.
Mawenya, 'A. S., 1995. Private communication. Dar es Salaam, Tanzania,

March.
Mbwambo, W., 1988. Systems Analysis of Urban Low Cost Housing Construction
Methods. MSc Thesis, University of Dar es Salaam.
McLeish, D. C. W., 1981. Manhours and interruptions in traditional house

building. Building and Environment, 16(1), pp. 59-67.
Mefford, R. N., 1991. Quality and productivity - the linkage. International

Journal of Production and Economics, 24, November, pp. 137-145.
Merrow, E. W. and Crocker, W. H., 1994. Improved business results through

benchmarking. Proceedings, ECI Conference, Lisbon, Nov. 1994.
Miles, D. W. J. and Neale, R., 1992. Building for Tommorow -International

Experience in Construction Industry Development. International Labour Office,
Geneva.
Ministry of Works, 1977. Local Construction Industry Study: General Report.

Ministry of Works, Dar es Salaam, Tanzania.
Ministry of Works, 1992. National Construction Industry Development Strategy.

Ministry of Works, Dar es Salaam, Tanzania.
Mohsini, R. A. and Davidson, C. H., 1992. Determinants of performance in

traditional building process. Construction Management and Economics, 10, pp.
343 - 359.
National Board of Architects Quantity Surveyors and Building Contractors

(Tanzania), 1994. List of Registered Architects Quantity Surveyors and Building
Contractors as of 31st Dec. 1993. NBAQS&BC, Dar es Salaam, Tanzania.
Ninatubu, N., 1995. Private communication. Moshi, Tanzania, April.

331
Noor, L:, 1992. A Study of the Variability of Labour Productivity in Building
Trades. Unpublished PhD Thesis, Dundee University.
Norcliffe, G. B., 1977. Inferential Statistics for Geographers. London.
Norman, G. and Flanagan, R., 1993. Changing the Labour Capital Balance in
Construction. Unpublished Internal Report, University of Reading, UK.
Norman, M. and Stoker, B., 1991. Data Envelope Analysis: the Assessment of
Performance. John Wiley, Chichester, New York.
Now for quality comparisons, 1992. Management Today, August, pp. 80.
Nuru, K., 1990. African architecture between tradition and innovation.

Conference proceedings, Sustainable Habitat on Urbanized Planet. Berlin,
March.
O'Connor, J. T., 1985. Impacts of constructability improvement. Journal of

Construction Engineering and Management, 111(4), pp. 404-410.
Oakland, J. S., 1993. Total Quality Management - The Route to Improving

Performance. Butterworth/Heinemann
Ofori, G., 1990. The Construction Industry: Aspects of its Economics and
Management. Singapore University Press, Singapore.
Ofori, G., 1991. Programmes for improving the performance of contracting firms
in developing countries: A review of approaches and appropriate options.
Ofori, G., 1994a. Establishing construction economics as an academic

discipline. Construction Management and Economics, 12, pp. 295-306.
332
Ofori, G., 1994b. Formulating a long term strategy for developing the
construction industry in Singapore. Construction Management and Economics,
12, pp. 219-231.
Olomolaiye P. O. and Ogunlana, S. 0., 1989. An evaluation of production

outputs in key building trades in Nigeria. Construction Management and
Economics 7, pp. 75-89.
,.
Olomolaiye, P. O. 1988. An Evaluation of Bricklayers' Motivation and Productivity.

Unpublished PhD Thesis, Loughborough University, Loughborough.
Olomolaiye, P. 0., Wahab, K. A. and Price, A. D. F., 1987. Problems influencing

craftsman's productivity in Nigeria. Building and Environment, 22(3), pp.
Osman, O. and Griffith, A., 1990. Factors influencing productive time achievements
during on site production. Proceedings of CIB-90 Conference, Sidney. pp. 382-
392.
Oswald T. H. and Burati, J. L., 1992. Can Total Quality Management work in the
public construction arena? Transport Research Record No 1340, pp. 40-49,
Oxford Concise Dictionary, 1990). Oxford University Press, New York, 8th Edition.
Oxley, R., 1975. Optimising Productivity and Cost in Repetitive Construction.

MSc. Thesis, UMIST, UK.
Pall, G. A., 1987. Quality Process Management. Prentice-Hall New Jersey,

USA.
Parker, H. W. and Oglesby, C. H., 1972. Methods improvement for construction

managers. McGraw Hill, New York.
Parker, N. A. and Lema, N. M., 1986. Establishment of Schedule of Rates for

Building Works - Phase I. Unpublished Report, National Construction Council,
Dar es Salaam.
333
Parker, N. A., Lema, N. M. and Mlingwa, G., 1987. An analysis of labour
productivity in Tanzania. CIOB, C/B Managing Construction Worldwide,
London, Vol. 2,835-845.
Paulson, B. C., 1975. Estimation and control of construction labour costs. Journal
of the Construction Division, ASCE, 101 (C03), pp. 623-633.
Pedhazur, E. J. and Schmelkin, L., 1992. Measurement, Design, and Analysis:

an Integrated Approach. Hillsdale, N.J.; Hove: Lawrence Erlbaum Associates.
Pender, R., 1993. Partnering for prtofit. Total Quality Management Magazine,
Bradford, UK. pp. 13-16.
Pera International, 1991. Benchmarking for Excellence. Unpublished Internal

Report, British Telecom, UK.
Pheng, L. S., 1991a. World markets in construction: IA regional analysis.

Construction Management and Economics, 9,63-71.
Pheng, L. S., 1991b. World markets in construction: IIA country by country

analysis. Construction Management and Economics, 9,73-79.
Picard H. E., 1991. Productivity and value management on utility construction and
maintenance projects. Proceedings of the 1991, PM!, Annual
Seminar/Symposium, Dallas Texas, pp. 201-208.
Pike, J. R., 1994. TQM in Action -A Practical Approach to Continuous

Improvement. Chapman & Hall,. London.
Pitcher, R., 1992. Principles of Construction Management. McGraw Hill

International Series in Civil Engineering, 3rd Edition, London
Planning, Organising and Managing Benchmarking Activities: User's Guide,

1992. American Productivity and Quality Press, Huston, USA.
334
Price A. D. F., 1986. An Evaluation of Production Output for lnsitu Concrete Work.
PhD Thesis, Loughborough University.
Price, A. D. F. and Harris, F. C., 1985. Methods and Rates for Concrete and
Bituminous Road Paving Operations. Internal report published for S. E. R.C.,
Loughborough University.
Price, A. D. F., 1992a. Factors influencing construction productivity. Unpublished

Internal Report, Loughborough University of Technology.
Price, A. D. F., 1992b. Construction operative motivation and productivity. Building

Research and Information; 20(3).
Priestley, C., 1994. British Construction: in Pursuit for Excellence. Business

Roundtable, London.
Quality Systems, 1987. BS5750 British Standards Institution, London.
Ranbird, H. and Syben, G., 1991. Restructuring a Traditional Industry:

Construction Employment and Skills in Europe. Berg, New York.
Rasdorf, W. and Herbert, M., 1990. Automated identification systems - focus on

bar coding. Journal of Computing in Civil Engineering, 4(3), pp. 279-296.
Research Technology Management, 1990. Pushing to Improve Quality. 33(3),

May.
Riggs, J. L. and Glen, H. F., 1983. Productivity by Objectives. Prentice Hall, N. J.
Robson, P. E. B., 1978. Interaction of design and production. BT & M, CIOB, pp. 13-
15.
Royal Institute of Chartered Surveyors, 1968. Standard Method of

Measurement of Building Works. 5th Edition, London.
335
Ruddock, L., 1991. Productivity in the UK construction industry. AACE
Transactions, pp. L1.1 - L1.4.
Sander, S. R. and Thomas, H. R., 1993. Masonry productivity forecasting model.

Journal of Construction Engineering and Management, 119(1), pp. 163- 179.
Sanders, S. R. and Eskridge, W. F., 1993. Managing implementation of

performance. Journal of Management in Engineering; ASCE, 9(4), pp. 365 -
381.
Sanders, S. R. and Thomas, H. R., 1991. Factors affecting masonry productivity.

Journal of Construction Engineering and Management, ASCE, 117(4), pp. 626-
644.
Sanvido, V. E., 1988. A conceptual construction process model. Journal of

Construction Management and Engineering, ASCE, 114(2), pp. 294-310.
Saslow, C. A., 1982. Basic Research Methods. Reading, Mass.

London: Addison-Wesley.
Sebastian, S. J. and Borcherding, J. D., 1979. Exploratory study of the major

factors influencing craft productivity in nuclear power plants. Texas University,
Austin, USA.
SERC, 1994. Innovative Manufacturing Initiative: A New Way of Working.

London.
Shapiro, S. S. and Gross, A. J., 1981. Statistical Modelling Techniques. Marcel

Dekker, New York.
Shetty, Y. K., 1993. Aiming high: competitive benchmarking for superior

performance. Long Range Planning, 26(1), pp. 39-44.
Shippam, R. 1987. House-building productivity in USA. Building, 15, Dec.,

pp. 127-132.
336
Sidwell, A. C., Van Metzinger, W. A. and Tucker, R. L., 1988. Japanese, Korean
and U. S. Construction Industries. Construction Industry Institute, Source
Document No. 37, Austin, Texas.
Siegel S. and Castellan, N. J., 1988. Non-Parametric Statistics for Behavioral

Sciences. McGraw Hill International Editions, 2nd Edition
Singh, K. D. and Evans, R. P., 1993. Effective benchmarking: taking the

effective approach. Industrial Engineering, 25(2), pp. 22 and pp. 65-66.
Sink, D., 1985. Productivity management: planning, measurement and

evaluation, control and improvement. John Wiley and Sons, New York, N.Y.
Smith, G. R. and Thomas, H. R., 1990. A conceptual model for estimating

productivity on change order work. Proceedings of CIB-90 Conference, Sidney.
pp. 382 - 392.
Stanhope Properties Plc, 1993. Setting Standards in the Construction Industry.

An Internal Report, Stanhope Properties Plc.
Stevens, A. J., Grant, E. and Howe, M., 1984. Housebuilding productivity - activity
sampling on Phase 2 of Pitcoudie. BRE note 76/84, Unpublished.
Strategies for the ConstructionSector, 1993. Commissionof the European

Communities,Brussels, Belgium.
Summanth D. J., 1985. Productivity and Management. McGraw Hill Book

Company, New York.
Tabachnick, B. G. and Fidel, L. S., 1983. Using Multivariate Statistics. Harper

and Row, New York.
Taiichi, 0., 1990. Toyota Production System: Beyond Large Scale Production.
Productivity Press. Cambridge, Mass.
Taylor, F. W., 1929. The Principles of Scientific Management. Harper and

337
Brothers, New York.
Tenah, K. A., 1986. Construction personnel role and information needs. Journal
of Construction Engineering and Management, ASCE, 112(1), pp. 33-48.
Thamhain, H. J., 1991. Benchmarking of project management systems: how to

measure yourself against the best. Project Management Institute Seminar
Proceedings, Dallas Texas, pp. 471-476.
The Business Roundtable,1982. ConstructionIndustry Cost EffectivenessProject.

ReportsA, B, C, D, & E, New York.
Thomas H. R., 1981. Can work sampling lower construction costs. Journal of
Thomas H. R., 1992. Effect of scheduled overtime on labour productivity.

Journal of Construction Engineering and Management, ASCE, 118(), pp. 60-
76.
Thomas H. R., Guevara, J. M. and Gustenhoven, C. T., 1984. Improving productivity

estimates by work sampling. Journal of Construction Management and
Engineering, ASCE, 110(2), pp. 178-188.
Thomas, H. R. and Daily, J., 1983. Crew performance measurement via activity
sampling. Journal of Construction Management and Engineering, ASCE, 109(3),
pp. 309-320.
Thomas, H. R. and Sakarcan, A. S., 1994. Forecasting labour productivity using

factor model. Journal of Construction Engineering and Management, 120(1), pp.
228-239.
Thomas, H. R. and Yiakoumis, A., 1987. Factor model of construction productivity.

Journal of Construction Engineering and Mangement, ASCE, 113(4), pp. 623-639.
Thomas, H. R. Homer, R. M. W. and Smith, G., 1990. Procedures manual for

collecting productivity and related data of labour intensive activities on commercial
338
construction projects. 4th version, National Science Foundation, Washington, D. C.
Thomas, H. R. Mathews, C. T. and Ward, J. G., 1986. Learning curve models of

construction productivity. Journal of Construction Management and Engineering,
ASCE, 112(2), pp.245-258.
Thomas, H. R., 1991. Labour productivity and work sampling: the bottom line.
Journal of Construction Management and Engineering, ASCE, 117(3), pp. 423-
444.
Thomas, H. R., Handa, V. K. and Homer, R. M. W, 1990. Productivity similarities

among masonry crews in seven countries. Proceedings of CIB-90 Conference,
Sidney, pp. 543-553.
Thomas, R. H., Maloney, W. F., Homer, R. M., Smith, G. R., Handa, V. K. and
Saunders, S. R., 1990. Modelling construction labour productivity. Journal of
Construction Engineering and Management, ASCE, 116(4), pp. 705-726.
Tippett, L. H C., 1935. Statistical Methods in Textile Research. Shirley Institute

Memoirs, 26, pp. 51-55.
Total Quality Management, 1992. BS7850, British Standards Institution,

London, England.
Total Quality Management: A Business Process Perspective, 1992. Kearney, A.

T. Inc.
Tucker, R., Roggie, D., Hayes, W. and Hendrickson, F., 1982. Implementation of
foreman delay surveys, Journal of the Construction Division, ASCE, 108(4), pp.
577-591. !. 0
Turin, D. A. et al. 1972. Construction and development: a framework for

research and action. A paper prepared for IBRD, London, University College
Environmental Research Group.
Turin, D. A., 1969. The Construction Industry: Its Economic Significance, and its
339
Role in Development. UNIDO Monograph on Industrial Development, No. 2.
Turin; D. A., 1975. Introduction. In Aspects of Economics of Construction, Turin,

D. A. (ed.) George Godwin, London, pp. viii-xvi.
United Nations, 1965. Effect of Repetition on Building Operations and

Processes on Site. United Nations Committee on Housing, Building and
Planning (UNCHBP), ' Publication No. ST/ECE/HOU/14, New York.
United Nations, 1965. The effect of repetition on building operations and

processes on site. ST/ECEIHOU/14, Committee on Housing, Building and
Planning, New York, N. Y.
United Nations, 1971. Yearbook of National Accounts Statistics. Vol. I Part 1&
2, United Nations, New York.
United Nations, 1983. Yearbook of National Accounts Statistics. Vol. I Part 1&
2, United Nations, New York.
United Nations, 1988 - 1993. Demographic Yearbook. New York: United.

Nations.
United Nations, 1990. National Accounts Statistics - Main Aggregates and

Detailed Tables, 1988. Part I& II, United Nations, New York.
United Nations, 1994. National Accounts Statistics: Main Aggregates and

Detailed Tables, 1992. Part 1 and 2, New York.
Van Sykes, R. M., 1963. Monte Carlo methods and the PERT problem.
Operations Reserach, 11, pp.839-860.
Walker, A., 1971. A study of the variation in output in building operatives. The
Architect and Surveyor, 10, NovJDec.
Watson, G. H., 1993. Strategic Benchmarking: How to Rate your Company's

Performance Against the World's Best. John Wiley and Sons, Inc.
340
Webb, W. (ed.), 1988 - 1995. Business Periodicals Index. H. W. Wilson
Company.
Webber, S. F. and Lappiatt, B. C., 1983. Productivity Measurement for the

Construction Industry. National Bureau of Standards, Washington, D. C.
Weisberg, H. F. and Bowen, B. D., 1977. An Introduction to Survey Research

and Data Analysis. Freeman, San Francisco, Calif.
Wells, J., 1986. Construction Industry in Developing Countries: Alternative

Strategies for Development. Croom Helm, Kent, UK.
Wells, J., 1993. Appropriate building technologies: An appraisal on the case

studies of building projects in Senegal and Kenya. Construction Management
and Economics, 11, pp. 203-216.
Wells, J., 1995. Population, settlements and environment: the provision of

organic materials for shelter. Habitat International, 19(19) pp.73-90.
William, P., 1994. Measure of success. New Builder, July, 1994, pp. 40.
Williams, I., 1991. Improvement and measurement of productivity and progress.

in Guidelines for Management of Major Construction Projects, NEDO, London,
HMSO, pp. 102-110.
World Bank, 1984. The Construction Industry, Issues and Strategies in

Developing Countries. World Bank, Washington DC.
World Bank, 1991. World Tables 1991. The John Hopkins University Press,
Baltimore and London.
World Bank, 1994. World Development Report - 1994, -Infrastructure for

Development. Oxford University Press, New York.
341
Zairi, M., 1992. Competitive Benchmarking: An Executive Guide. Technical
Communications Ltd, Letchworth, England.
Zairi, M., 1994. Measuring Performance for Business Results. Chapman & Hall,
London.
342
Appendix 1
Construction operatives' and contractors' questionnaires

Phase I
343
UNIVERSITY OF DAR ES SALAAM DEPT OF CIVIL ENGINEERING
RESEARCH ON BENCHMARKING OF LABOUR PRODUCTIVITY
CONSTRUCTORS' QUESTIONNAIRE Quest 1
1. Date and Time of Interview: Date:

Time:
2. Company Name:
3. Address:
Year of establishment: Current class registration
4. Ownership: Private: Local:

State: Foreign:
5. Main Activities: 1.
2.
3.
6. Other Activities: 1.
2.
7. a) Volume of construction work in Tshs. value per annum:

b) Volume of other activities in Tshs. per annum:
8. Personnel
Number Qualification Education Experience (years)
Chief Executive(s)
1.
2.
Other Key Personnel
1.
2.
3.
4.
PermanentEmployees
1. Skilled
2: Unskilled
Temporary Employees,
1. Skilled
2. Unskilled
Part-time
344
9. Mode of work preferred by management:
Piecework/Hourly/Daily/Weekly/Monthly Payment
Reasons:
10. Mode of work preferredby workers:

Payment
Piecework/Hourly/Daily/Weekly/Monthly
Reasons:
10. Resources- Equipment:
Description/Type, Value Quantity Level of utilisation

(hours/dayover the year)
1.
2.
3.
4.
5.
11. Other Support Services:
Description/Type Use
1.Office
2. Workshop
3. General transport pool
4. Stores
5. Others
12. Welfare facilities:
1. Medical services
2. Canteen
3. Toilets & washroom
4. Housing
5. Transport
6. Others
13. Is there any staff training programme and development ? Yes/ no.
If no why?
If yes, what type of training has been offered in the last two years?
1.
2.
3.
4.
345
14. Productivity:
Indicate by assessingthe extent to which the following factors affect labour

productivity on site by ticking the appropriate box.
Note that: -1 means that the factor does not have any effect or it is irrelevant.
that it has an extremely high influence on labour productivity.

-6 means
it has a moderate impact on the labour productivity.
-3-4 means
Score
Factors
1 2 3 4 5
Level of skill
Wages
Monetary incentives
Other forms of incentives
Level of Skilled
interaction
Unskilled
Work environment
Prospects
_ Weather / climate
Time
Location
Physical fitness
Thank you for agreeing to be interviewed
N. M. Lema
Principal researcher
346
LOUGHBOROUGH UNIVERSITY OF TECHNOLOGY
CIVIL AND BUILDING ENGINEERING DEPARTMENT
CONSTRUCTION PRODUCTIVITY BENCHMARKING IN TANZANIA
CONSTRUCTION CREW QUESTIONAIRE
1. Date and time of-interview: Date: Time:
2. Place and Address:
3. Employer's Name:
Private: Local: Self:

Public: Foreign:
4. Type of Employment: (tick in the box)
Type Permanent Temporary Self Skill type

Skilled
Unskilled
7. Qualifications and Experience:
Trade qualification: Year:
Level of Education:
Experience: ears
6. Typical working day: Start: Finish: Break: min.
7. Main activities and productivity: Kindly list the main

activities that you normally undertake and state the maximum
average and minimum productivities that you have previously
achieved.
Activity Productivity
Maximum Average Minimum
1.
2.
3.
4.
5.
6.
347
8. Can you rank the main reasons you think made you produce at
your maximum capacity (work harder)?
Score
1 2 3 4 5 6
Payment
Other Incentives
Other benefits*
Leadership
Quality of tools
Level of skill
Ph sical stren th
Teamwork
Weather/climate
Level of
Su ervision
*Tick types of benefits
Transport: Bonus: Uniform:

Housing: Meals: Med. Care:
Others:
8. Payment achieved over the last two years:
Type (tick) Wage (Tshs)

Low Average High
Hourly
Daily
Weekly
Monthly
Contract
Other
*state basis
9. On which activities do you make most money:
a)
b)
10. What would be the main reasons for making most money? Please
tick.
a) The employer is generous:

b) 'You worked very hard:
c) The employer did not know
how to price the job:
d) Luck
348
11. What keeps you working for your present employer?
Score
1 2 3 4 5 6
Wages
Incentives
Other benefits*
Leadership
Prospects
Job security
No other
o]2]2ortunity
*Tick types of benefits that you receive

Housing: Meals: Med. Care:
Others:
11. Payment for welfare facilities on site

Free:
Little"
Normal:
High:
12. How would you describe the managerial capability of your

supervisors or management in general:
Very good:
Average:
Poor:
13. If you were the manager, what things would you improve:
1.
2.
3.
14. Have you been on any form of training in the last two years?
Describe:
Was it any good for your work? Yes /No
Thank you for your co-operation
N. M. Lema, Main researcher, July, 1993.
349
Appendix 2
Cost Breakdown of Selected Buildings in Dar es Salaam

(based on the winning tenders)
350
Mechanical and Machine Design Building
University of Dar es Salaam
Cost breakdown
Q Concrete - 36%
Finishing - 18%
Earthworks - 14%
Roofing - 7%
Walls - 6%
Windows - 5%
Decorations - 4%
Plumbing - 3%
Doors - 3%
Fittings - 3%
Electrical Inst. - 1%
Kawe Branch Building National

Bank of Commerce
Cost breakdown
Concrete - 44%
Finishing - 13%
Roofing - 11%
Windows - 8%
Walls - 8%
" Fittings - 5%
E] Decorations - 4%
Plumbing - 3%
Earthworks - 3%
Q Doors- 1%
Q Electrical - 0%
351
Kawe Managers' House
National Bank of Commerce Cost breakdown
Roofing - 23%
Walls - 18%
Concrete - 16%
Finishing - 14%
Plumbing - 6%
Windows - 5%
Q Fittings - 5%
Earthworks - 4%
Doors - 4%
Q Decorations - 3%
Housing No 19 - Phase
Cost breakdown
Q Concrete - 20%
Finishing - 17%
Roofing - 12%
Windows - 9%
Walls - 7%
Q Plumbing - 7%
Doors - 6%
Fittings - 6%
. Decorations - 4%
Earthworks - 3%
352
Housing No. 19 - Phase II
Cost breakdown
Q Concrete - 36%
Q Finishing - 13%
Plumbing - 11%
Roofing - 7%
Windows - 6%
Walls - 5%
Doors - 5%
Fittings - 4%
Decorations - 3%
Q Earthworks - 2%
Estates Workshop
Cost breakdown
0 Roofing - 25%
Concrete - 22%
Earthworks - 17%
Doors and Windows - 12%
Walls - 8%
Finishing - 7%
Plumbing - 5%
Decorations - 4%
Electrical Inst. - O%
Q Fittings - 0%
353
Road Engineering Laboratory
Cost breakdown
Q Concrete - 44%,
M Finishing - 13%
Windows - 10%
Plumbing - 5%
Earthworks - 4%
Q Fittings - 4%
Doors - 3%
Paving - 3%
Roofing - 1%
Decorations - 0%
New Estates Centre

Cost breakdown
Q Concrete - 37%
Finishing - 30%
Carpentry - 7%
Earthworks - 6%
Decorations - 5%
Walls - 4%
Plumbing - 4%
Windows - 3%
Doors - 3%
Q Metalwork - I%
354
NPF Building - Dar es Salaam
Cost breakdown
Q Concrete - 57%
Q Finishing - 15%
Plumbing - 8%
Joinery - 5%
Painting - 4%
Roofing - 3%
Q Earthworks - 3%
Carpentry - 2%
Drainage - 2%
Paving - 2%
Staff Village - National Bank of Commerce

Cost breakdown
Q Concrete - 39%
Electrical Inst - 12%
Walls-11%
Finishing - 8%
Decorations - 7%
Roofing - 6%
Doors - 6%
Windows - 6%
Fittings - 3%
Q Plumbing - 2%
355
EDP - Staff Housing - Dar es Salaam
Cost breakdown
Concrete - 25%
Q Finishing - 20%
Windows-14%
Walls - 13%
Roofing - 12%
Decorating - 6%
Fittings - 2%
Plumbing - 0%
Earthworks - 0%
Q Electrical Inst. - 0%
Staff Housing - National Bank of Commerce

Cost breakdown
Q Concrete - 37%
Walls- 12%
Plumbing - 8%
Finishing - 7%
Decorations - 6%
Q Roofing - 5%
Doors - 5%
Windows - 5%
Fittings - 2%>
Q Earthworks
- 1%
356
Semi - Detached Houses
National Bank of Commerce
Cost breakdown
Q Roofing - 27%
Finishing - 16%
Walls - 13%
Concrete - 11 %
Plumbing - 8%
Fittings - 7%
Decorations - 5%
Windows - 5%
Doors - 4%
Q Earthworks - 2%
Mechanical Engineering Building

University of Dar es Salaam Cost breakdown
Concrete - 47%
Q Finishing - 17%
Roofing - 8%
Walls - 7%
Windows - 6%
Decorations - 3%
Plumbing - 4%
Fittings - 3%
Earthworks - 3%
Q Electrical Inst. - 2%
357
Appendix 3
Contractors' Questionnaire - Phase II
358
UNIVERSITY OF DAR ES SALAAM
DEPARTMENT OF CIVIL ENGINEERING
DAR ES SALAAM, TANZANIA.
LOUGHBOROUGH UNIVERSITY OF TECHNOLOGY

CIVIL AND BUILDING ENGINEERING DEPARTMENT
LOUGHBOROUGH, UNITED KINGDOM.
CONSTRUCTION PRODUCTIVITY BENCHMARKING IN TANZANIA
CONTRACTOR'S QUESTIONAIRE Quest 3
A: GENERAL INFORMATION
1. Date :
2. Company name and address :
Year of establishment: Class:
3. Place: (HQ/Site)
4. Ownership: Private: Local: Foreign: State: (pleasetick)
5. Main activities: (please tick) 1. Building and civil:

2. Building:
3. Civil engineering:
4. Sub-contractor:
6. (a) Other activities: (not necessarily construction related)
1.
2.
3.
B: CONSTRUCTION VOLUME AND OVERALL OUTPUT

7. a) Volume of construction work in Tshs. value per annum for last
year:
b) Volume of other activities in Tshs. per annum for last
year:
8: (a) Approximate construction work load at present:
Job No. Type of work Contract Contract Duration % completion

Sum from to
1.
2.
3.
4.
5.
6.
359
(b) What is your estimated turnover this year?
At what capacity do you think you are now operating: %
Can you briefly state the reasons?
9. What has been your work load (expressed in annual value of work done) over the last
fifteen years in comparison to the number of your employees?
YEAR Value in Tshs. Total employees Annual

productivity
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
10. If you have had fluctuations in your work load in Tanzania briefly state the main reasons
for the fluctuations:
(a)
(b)
(c)
(d)
11. What approximate percentageof your expenditure in a year goes into the following:
(a) Maintenance of company overheads %

............... %
(b) Labour ...............
(c) Management %
............... %
(d) Equipment
...............
(e) Materials %
............... %
(e) Others:
...............
360
C: PERSONNEL AND CREW PRODUCTIVITY
12. Average number of personnel

Number Qualification Education Experience (yrs)
Chief executive(s) (titles only)
a)
b)
Key personnel (titles only)
a)
b)
c)
Administrative and
support personnel
Permanent employees
a) Skilled
b) Unskilled
Temporary employees
a) Skilled
b) Unskilled
Part-time
Others
13. What modeof work do you prefer for your workers/operatives:

Payment
Piecework/Hourly/Daily/Weekly/Monthly
L O. 'JIIJ.
14. What mode of work is preferred by workers/operatives:
Piecework/Hourly/Daily/Weekly/Monthly Payment
Reasons:
15. Typical working day: Start: Finish: Break: min.
361
16. Main activities and productivity: On the basis of the main activities listed below state
the maximum, average and minimum productivity (output per man per day -8 hours)that
you have previously observed.
AMY PRODUCTIVITY
MAXIMUM AVERAGE MINIMUM UNITS
Manual excavation in cu. m.
loose soil
Blockla in 230 mm wall sq. m.
Blockla in 150 mm wall sq. m.
Blockmaking no.
150x230x450mm blocks
Plastering 12mmwall s. M.
Plastering 12mmceiling sq. m.
Concreteslabs&beams Cu.m.
Concretewalls & columns Cu.M.
Concreting - bulk cu. M.
Formwork fixing (timber) sq. m.
slabs sq. M.
beams sq. m.
columns sq. m.
Ceiling board fixing sq. m.
Wall aintin sq. m.
Ceiling painting sq. m.
Floor painting sq. m.
17. What conditions are necessaryfor labour to achieve maximum productivity mention in
Indicate by the to
assessing extent which the following factors affect
section 16. above.
productivity by ticking the appropriate box.
Note that: meansthat the factor doesnot haveany effect or it is irrelevant.
-1 it has high influence labour
-6 means that an extremely on
productivity.
FACTOR SCORE
1 2 3 4 5
Wages
Other financial bonuses
Other benefits*
Leadershi /mana ement
Teamwork
Level of skill
Physical strength
Experience
Training
Quality of tools
Safety
Materials availability
Weather/climate
* Other benefits include housing, meals, transport, uniforms etc.
362
18. How much can your craftsmen make in one day in Tshs?
Low: Average: High:
Can you rank the potential for income for craftsmen for the following activities? Use the
scale as in 13 above.
ACT 'y SCORE

1 2 3 4 5 6
Manual excavation in
loose soil
Blockla in 230 mm wall
Blockla in 150 mm wall
Blockmaking
150x230x450mm blocks
Plastering 12mm wall
Plastering 12mm ceiling
Concreting slabs and
beams
Concreting walls and
columns
Concreting - bulk
Formwork fixing (timber)
slabs
beams
columns
Ceiling board fixing
Wall painting
Ceiling painting
Floor painting
Briefly explain the reasonswhy the potential is higher for other activities:
19.(a) To what extentdo you usesubcontractors:

Labour only: Very frequently Regularly Only occasionally
(b) For what activities do you use them?
i)
ii)
iii)
iv)
V)
363
(c) Can you briefly justify your answer to 19(a).
(d) If you had concreting work, would you use concreting gangs, why?
D: CONSTRUCTION COMPETITIVENESS
20. a) How would you judge the competitionin the following areas:
Building work:
Very high High Medium Low Very low
Civil engineering work:

Very high High Medium Low Very low
Briefly justify your choice.
b) Indicate the key factors that are important for the success in your business. Rank them in
their order of importance.
1st
2nd
3rd
4th
5th
6th
21. a) Are there areas in your business which you think you can learn from others in the same
business? e.g. quality, productivity, organisation etc.
a)
b)
c)
d)
e)
b) If your company was to be comparedin performancewith the rest of companiesin the

country, wherewould you placeyourself?Pleasetick.
Top 5% Top 15% Top 30% Top 50% Below the top 50%_
Explain briefly why you think you are in the above position indicating your what you look
at in judging performance.
22. Are there specific areas which you think they can best be learnt from others not necessarily
in the contracting area.
a)
b)
364
C)
d)
e)
Do you intend to do this in the near future? Yes: No:
Why?
E. GENERAL STATE OF THE INDUSTRY
23. There have been a number of claims that productivity and quality has declined over the last
25 years. If today's productivity and quality is say given 100% what would you give
these performance indicators for the following years?
FACTOR SCORE
1970 1975 1980 1985 1990 1995
Quality of construction
Productivity (industry)
Productivity (site)
Other:
24. What do you think are the main reasonsfor the change?
Quality
a)
b)
c)
d)
e)
Productivity
a)
b)
c)
d)
e)
25. Have you had any of your staff on a training programme?
Yes/ no.
If yes, what sort of programmes?
a)
b)
c)
26. Have you heard of Total Quality Management? Yes: No:

Dohy
? think it is applicable in your business? Yes: No:
Y
27. Have you heardof Benchmarkingin the businessworld? Yes: No:

Do you think it is applicablein your business? Yes: No:
why?
28. What do you think of the construction industry in Tanzania in general with respect to the
365
following:
" Indicate by assessingthe extent to which the following issues by ticking the appropriate
box.
Note that: 0 means that the aspectis extremely poor or extremely low.
6 means that it is has an extremely high or extremely good
ISSUE SCORE
1 2 3 4 5
Productive - site
Productivity - industry
Quali of work
Competitiveness
Skills quality
Management competence
Teamwork in a project
Ethics
29. Given the state of the industry in Tanzania, can you kindly suggest specific actions by
various people that could improve performance:
Government:
Consultants:
Contractors:
Your company:
30. Kindly indicate whether you would like to be informed of the outcome of this research.
Yes" No:
Thank you very much for participating in this work

Name of the respondent (optional):
Position:
N. M. Lema
Principal researcher
366
Appendix 4
Operatives' site questionnaire - Phase II
367
DEPT. OF CIVIL AND BUILDING ENGINEERING, LOUGHBOROUGH
UNIVERSITY
Labour productivity benchmarking in the construction industry - The case of

Tanzania
CRAFTSMEN QUESTIONAIRE GUIDELINES Quest 4
A: GENERAL INFORMATION
1. DateandTime of Interview: Date: Time:
2. Placeand Address:
3. Employer'sName: Type:
Private: Local: Subcontractor (labour only)

Public: Foreign: Self.
4. Type of Employment:(tick in the box)
5. QualificationsandExperience:
Trade qualification: Year:
Level of Education:
Age: Experience: ears
6. (a) How long have you been with the current employer: years
(b) How long have you been on this site: months.
(c) How manysiteshaveyou beenon in the last two years?
Which onedid you considerthe bestand why?

Site:
Reason:
368
B: PRODUCTIVITY
7. Typical working day: Start: Finish: Break:
-min.
8. Main activities and productivity: Kindly list the main activities that you normally undertake
and state the maximum averageand minimum productivities that you have previously
in
achieved say one day (one day =8 hours)
10. Can you rank the main reasonsyou think made you produce at your maximum capacity
(work harder)?
Indicateby assessingthe extentto which the following factorsaffect productivity by ticking
the appropriatebox.
Note that: meansthat the factor doesnot haveany effect or it is irrelevant.
-1 it has high influence on labour
-6 means that an extremely
productivity.
FACTOR SCORE
1 2 3 4 5
Wages
Other financial bonuses
Other benefits*
Leadership/management
Teamwork
Level of skill
Physical strength
Experience
Training
Quality of tools
Safety
Weather/climate
*Tick types of benefits given to you

Housing: Meals: Med. Care: Others:
369
11.(a) What is your normal output per day in your current task?
(b) Are you satisfiedwith the currentoutput level? Yes: No:
Why?
(c) What is the maximum output that you can achieve in this task?
12. How would you classify the following on your site:(tick appropriatecolumn)
SUBJECTS Excellent V.Good Good Fair Bad Very bad
Work organisation
Supervision
Level of pay
Work Environment
Teamwork
13. How much money can you make in one day in Tshs?
Employed: Low: Average: High:

Self employed:Low: Average: High:
14. Can you rank the potential for income for following activities? (If you were working in the
following activities, in which you think you would make most money?) Use the scale as in
9. above.
Briefly explain the reasons:
370
15. Payment achieved over the last one years:
Type (tick) Wage (Tshs)
Low Average High
Hourly
Daily
Weekly
Monthly
Contract
Other
*statebasis
16. On which activitiesdo you makemostmoney:

a)
b)
c)
17. What would be the main reasonsfor making mostmoney?Pleaselist

a)
b)
c)
d)
18. How would you describe the managerial/leadership capability of your supervisors or
managementin general:
Very good:
Average:
Poor:
19. If you were the manager, what things would you improve:
1.
2.
3.
20. Have you been on any form of training in the last two years?
Describe:
Was it any good for your work? Yes

___/No
21. In a typical working day what percentageof your time would you estimate is spent in the
following classification:
Classes % of typical working day
371
Working
Official breaks
Waiting (e.g. materials)
Recovery
Unaccountable
C: MOTIVATION
22. List in orderof importancethe first threethingsthat give youjob satisfactionin your work:
Ist
2nd
3rd
23. List in order of importance the first three things that give you job dissatisfaction in your
work:
Ist
2nd
3rd
24. Which problems if solved would give yield improvement on your speedand quality of
work in order of importance:
Ist
2nd
3rd
25. Pleasegive suggestions on how best to improve the productivity of workers and quality of
work produced in your trade:
Thank you very much for participating in this work
Name of the interviewee: (optional):

Position:
Nameof interviewer:
N. M. Lema
Principal reseracher
April/ May, 1995
372
Appendix 5
Construction industry executives' interview guide
373
DEPT OF CIVIL AND BUILDING ENGINEERING
LOUGHBOROUGH UNIVERSITY - UK
Labour Productivity Benchmarking in'the Construction Industry

The Case of Tanzania.
A: Economic Significance and Performance
1. What is generallyconsideredthe Tanzanianconstructionindustryin Tanzania?
2. How significantis the role of the constructionindustryin the Tanzanianeconomy?
3. Would you necessarily draw up differences in this respect between say building
and civil engineering construction? Why?
4. Has the industrys capacity in comparisonto other sectors been a bottleneckin the
developmentprocess? If yes, in what particularway?
5. What have been the common Tanzanian construction industry performance

indicators in your experience? Which one would you consider the single most
important?
6. What is your observation with regard to the performance of the construction

industry over the last 20 - 25 years with respect to the single most Important
indicator?
7. What about the other performance indicators over the same period?
8. If there are any trends in performance,would you attribute these to any specific
factors/conditions/events?
9. Is there a single most Important factor resulting in the performance trends

observed?
374
10. With regard to current changes in in the country and possibly outside the country
(political process and orientation, internal economic policies and global economic
policies), what are the likely significant effects on the performance of the Industry?
How soon will these effects be felt?
B: Performance Improvement.
1. What do you consider are'the critical success factors for a sustainedgrowth in the
constructionindustry?
Externalto the industry;

- Internalto the industry.
2. Is there competitionin the Tanzanianconstructionindustry?
3. How do you see the role of competition in the Tanzanian construction industry as a
performance improvement basis?
4. Do you think there are significant performance gaps between individual

organisationsin the constructionindustryin Tanzania?
- Consulting organisations;
- Construction companies.
5. Do you see a role for a performance improvement philosophy such as Total

Quality Management(TQM) in an industry such as Tanzanian?Why? [TQM aims
at continous process improvementthrough customer (both internal and external)
satisfaction.]
6. What about the role of benchmarking? What about nternal benchmarking?
(Benchmarking is the continuous search for industry best practices that leads to
superior performance, by measuring against and learning from best companies
recognised as industry leaders.)
375
7. Given the poor performance state of the industry at the moment, where would you
think initial efforts should focus for the creation of the critical success criteria?
8. The fragmentation in the construction industry in developed countries has often

been blamed for the its poor performance in relation to manufacturing. What is your
commenton the same in Tanzania?
9. In the customer-supplier chain of the construction process, (arising out of the TOM
philosophy), the construction worker is not only seen as the ultimate supplier, but
also as the main internal customer. Efforts for performance improvement therefore
should initially centre on the worker. Do you think this approach would apply in
Tanzania?.
10. Again, on the basis of your wide experience, what factors would you consider
significantin influencingworkerproductivity?
THANK YOU VERY MUCH

N. M. Lema
February/March, 1995.
376
Appendix 6
Construction industries performance statistics

-
Tanzania, Kenya and United Kingdom
377
O a Cl t0 -e of t ei O Co N h N
O O h e N aD U) h 41
M Y) 0 N h a O N ID O N h V' . .
O h h O Co
Co 0 i 0 (D 40 N
O Co le u N N ^ M CO t` M
a0 N h IA f c4 07
N M CO
N N 10
O
h N
a a M M 1A 01 O M O N
0 0 0 h Co Co M 0 '" M
R N 10 01 t0 h h 01 0
U) N
0 N aT O h h N
r- N
e CY CO N N Co O
O a ^ 00 , h e M O
' h N h 7 Co p M
V
N
O M h M N r O 0 '? ^ h
O O - Co M M V h O N
to n Cl 0 1q CD 0 UI N C) Of 10 .-
O M Co t0
C/) O O a0 v M O> M Co OI CO M
O ,p t0 M tp O)
N M N h N O N - NM
C) t0 h N p
... 0) h N
C) e h
0) -
! M
10 N to CO CO N "r h U) O O N M O
O O h U) h Co t0 h OD N
c7) Cl Cl V M h O t0
10 in
O Of M; le OD N h
to r_ 0 (3) (9) 0 le O O
N M m M N cli "t M - - w
O IA
Qi r r Co M N
'a - -e Co
T-
M
N
L Ol O 10 O M O Co C
O) N CO M N to an N
^ O O r Co O O 0 tl ? M Co M
N N In O O M O aD -
O O ri j
O N h Co u) N N m ti N
tu 0) , r4 It Ip l'9
N r M ' h -e M N r Co Co
O h in
t0 N
h
"
N a0 Of OD In N r M n r h
O O l'9 V' h 10 0/ Cl
(V N V O O 1q h 1p to
O h r t0 V' U) ^
ei Co
r
h
a
N M 00 0 Co
M N Of 111 to j
O
_ M N
M
Si
N
E
N h M U) h N V -. 0 OD M O1 N
0 0 in e Co N in N W
yO .- N O) at h - t0 OI U) N Co a0 Cl -.
to N p
O M
a% p O CO M N
01 N h M O N r
OD p O N r r N M
CL t0 a to thD
Co QI M
M - N
O O V M h Co to O O h N 1[)
: Si h (0 N V N V 0
O O O h - - 09 9 Mn o! o n
0 r Co rn Co .- v
m CD h h ^ N t N
IA " CO t0 N
M O 40 f0 N t
4- 10 h - le r,
_ CI) ^ Cl
cu M
M N
m M
h N O M of O OD 1ff 0 N Co N 0
L
"Fr
p> O
p
O
.
1p "
M
-co
r
CO
r
CO
O
O
O Co .- O C9 M O 1 10 r 00 O
m N V h N
(ii O N N O 00 f Of '7
a N N h to
Co Co N O Co
In 10 O
M
-
c z
d
V N
C
L
Co 0
V a
N
q 1 p
_
C S
p N '
N ~ tll
y Q ii
N t C a m
C o $ }t
$
I e "- h
2 i
1 0 y
m e a C q -
m .- 1
1 1 1
1 1
0)

iu 1:
U
1
5
378
Ol O N a O O O 07 Co
to o O N a0 e - n 0 Co CO to
M r U) U) r N 00 Co
O N O OI UA 4
Co -e M r- cy
Of In U) M
OMi 4
O
CY) M /0 O
N O Of () M .- N ^ N
O/ M U ^ r
r Q N
N r
N
N
0 Co O M O n O CO 4') r to
O O N Ol ' M O N M O) n CO
N 0) M t0 aD at 40 at OI N et U)
Co O O) ca O N 0 O
10 O Of O 1
O aD
N M r 1(1 'Q M LO
r
N Co (O V M N O e O
M M Omi
a0 to O r N
to Co
Z
N
-Co M t0 CO t9 CD n O tO Cf M
y"' O O U) n N M a r U)
C') N U) U) n n OD Co t0 n N Co
' 0 r ^ N (C CO t0
V 0 M N "- c M r it M ^ O N
0 M too N r .d
Co Co N
(i CO M cl
Nn n r to

N
y
U) V a O c) O) O f. aD n V
N O O CD EO N N CD O
0)
a
U)
n
M 1n Of r O o r M n m r
M 6 U) U) r N
0 0 Co r)
gn N to c2
cu Co 0 M M to CCo
) (
4.0
C) M M V CD N
U CO O t0 - cli M
M V N
U) U) Co r Y
à
0
C N
"
M 0
cu N 0) N
O O Oi to O C9 M t0 ^ P. 0
r- U) n M O N n
Co 0 p N fu Co N N M O O
0) N n n a M CCO r
r_ 0) N M M - Q N f
M OJ 4O ; M O M
N Of Y
- 0) ', Co - O r O
M r r
E
L
Of
r
O
0) Co Of 0 M it) m c*
Co N le O O O -
Q N -e V b 0 r U) N 0 U -v
U) b N In fq
O N N
CL C> Co p
N
O- 40
CD
r
n
N N Co N
Co
:2
r ,r
n N n r)
Y 0 O -
- N V' l
'(i - r N
N M M n N M O w n a 'p
00 0 M 0 0) n ^ C N
t r 04 O n 'C Nv ^ 0) O
N n N M P.
O aD ^ N
Co M N Co rO ^
Ri C) c4 0
n N
N O Ol M n Co
0)) to 40 -i N m O
CO n a r
- e M
L
M M O n O M N f O1
r r' M Oh
Co O O n N Y a Y ^ CO O Co 40
p n O tM0 OD p 10
( M N n Of C-1 m im
0 r M , N 0
T% r t0 Co
Co N Co o
C r' Co M Co
M r N
" n
C r
0
i
yam
Co _
Co
.
p
O
..
Op
q
O W4
= O G "'
O y m
C
L g n ~ ,t
i 3
C r CO c a
N Fh
r_ p p 'G a n
r
Co y1
$ O
3$ = t t; ao
S [a
"' c c S
s r.
a
0
m 0) 0 r G
po_ t o >, U
p n " m m s)
C c
n bn 10
Co
S
Gl " r Q. 1p
Yl n
$ $ E
n
E
c
v
0 0,
n
n c
Co
= Pi va
3
V
,g
V
a
E
ID
m

E
E
2
i
i
W

'U
)
g i
C

is ` 1 l
i i 11
Q' 0 o m
c n E >
Q
9. G C7 0 (7 > > > Z5 F- W U j O O
379
O Of M N N h O 0 a0 CA
O N h O) h O N P M0 M M
M O N O O M 1 M M O
O 0) In (0 N
` M p O N N M
O Co , u1 .O cm ' 4n
e Co co ^ Co N Or O
O N M N r O
, t+ N V'
M P M ^
O M
O (0 N ".
O M
o N
`/ P V O Co h Co Co
M - p c M N W _ r
N O) 0 b 0 t0 r O 40 00 C 0
C') 1^w N (7)
p. Co m N N -
0 n N O r C' ' r'
n C') - c (n '
N n M O U) CY)
O O On N N N p
". + N Co N M C>
c M
N
L:
`t n M t0 O N Co N O
O M O N 'a O 'Q
CD NN r P M 0 N
o 0 M n O M 00
M O p N
O M N c e
t0 Y o r 0
C) p N N M M M M
N O M N m
r Co r M Co
V O
(0 N N N
r%
N
N OV ^ N Y 0)
C cl 0 0 N M U, N
p . P. O! i CN+) O
i0 CY r_ N
G M9 O ce e4 N
M O " 00) Co w
O) a0 Nf1 N Co
O M Co O N N
r M O n -
N N
[I N r
M
U)
L N
w
a M n N o in O CO U`
pD O o 01 -co
N
O
N
-Co
O N
h M 0
N
0
O N O N m m 10
O N N N M N
p M ^ N., m et n N
o r e4 y
O M., M ^
C. M M 'p r
'- O - V
Co V N n N
Co -
N
M (0 M O N 0 M O M (3)
0 0 U, e f1 e c,) h e0 - N M (0
M eD N of t Q
Co Co o a0 v) r Y Of t
a0
O
P
M
to r4 O M c4 10 Co
'(f O e of N p r N
M O W N N
p r Co M 07
0 M N ^
^ Co M
M N
L N
,Q
-0.- N
U) M N N O O N O Of
O L1 O N N O Co N (0
O O O M t0 CO N 0 O U) m ^ O!
a N - N 'n 01 1)
'a O V 0 l') N eD
0 e N O
V ' 0
( h N
o N V N
r p Co O
to Ip N
W0 N '
to 0
M N N r- to
N
N
0
N
C
1
O
C
v
d)
y
H to d
49
C
Cu f" y ' tP-
o
n N y' m
o O (7 My
0 0
CO) C,
Co
N
= 0
o
F-
o ' o

2
CCo
=
X
3
c a
n Co
o
O C N at . Co 0 Cp
m
o j, C C r CO
g N "rs r
v O1
r o n V S y N Y:
e
m C

q n a a rn
z 0
!C 0 aEi is E
CL E 0 C
d a ' m 7i 'i
15 u a 10 0 n 2 2
0
rn m
`m m m
p. m Co E > > > E - c c S
- a a a a C
c > E i a > i LL
Q 0 0 a a z u7 t c c7
N N (0 - P O to e0 e N
tO O O O) M N O 00 N - -e N
0) Co to ll e0 C') M O to eD N O)
0 to e N P M Co
0 r M ^ m h m
toe m
(0 C N N N ' M Co
0
Cl
N
'f
O Co ..
- M P M
M 2
N (O N M
O .-
M
N
380
n O O M M e r o n N M 1 '7
t- o o Co s!
N O O h Co M O M LO f- O CO) t0 Of N
N GO st e !- (3) (0 N N M tl1 N M
O) O a p Co pi Co N m N r
p CV e 0 M r
r h
.) ,a 0 N U) F- u)
M n
M M A M
Cl M O)
Co
M to O) O CO M CO N
(0 O O N U) O lit -e p O p O O
c,
r
d
lit
c, M u) fZ r M h O M
N O f- p; tl CO OD O (0 r
CO CO CO r M CO ' Cl
01 O O O) O) N O) t0 O ^
M N IA 0 1A r r
r h Cf O N M O) r M u) O O) 1n
N N M u0 ^
Co Co N M Co
n
0 0 O) 0) N 0 V' N M r r
M (1 C') r (0 Co 0 r
U) O O -
O M N h ui O N tt N O Co O> CO r M
N CO IA r O CO N (0 N CO (0 M
C) O p r- F) O) in N CD r
Q) st p p O U) CO U)
r r M p Co O r
N tt) U) lir M O 0 M
N M aNp M M O
ei r,
r2 N -
C) r
Qi
z C7 CO N O to r to -* O tD f0 M ! 0
O O CO 00 IA M d'
CO N C) to O O O et 0) r c
N O Co in U, O I- O) C7 P. N to M M
N to N O -ir
2
M O N ,r d" (D 0 LO D - U) N (0 CO
N r p O et p 1[)
M (O CO
N cr) d tD N N o
to N
Q)
r
T N
O in O r 0 O a0 r m Co
Of 0 f0 O t0 e -t
M O O h t0 p N
Co N M u) O 1n O N Un
fl o h r r N C;
e
L. O u3
0) ui aD M U') N CO - N 0 CO
W N M N N r r Co 0 O) IA (0 O r
O r O h O> r r r
r r M Co M M Co
y,,, n Co Co ^ Co
r r O N r
0
n p r
:. N
t0 <O Of 01 O1 O M M 0 9 <O
C O M N h N O1 O r N N
N O t M Co p N ei N
" t- o n N N
h Co ll of r
r O N a1 M CO r r aT Cf M
M O O N r0 h r In M c r r to
Co le st -
r N A Of r N r M)
r O) 0 r
(i N r N t (C 1 r N r
= N o
p 0 N O N O 0)
O O ' e O Co Cl r ^ nt "
lle
N 6 cD N In i! N r .- O!
O P O O i (6 N M et r_ p)
I[) r (0 Cl r - Of 0
M O h N (0 Co 0 0) r_ p
N C) 1n er f' In r
- O N U) Co r C') p
a- p r
N st 0 M Co
O ^
N
Q 1l) U)
00 O " P CO r O ! M O Ol O
t e9 10 00 N O
O O o D r
P O In Co a t N R
Co r u)
cm O of
e Co e r M p Co Co LO Ln 0 N v- e M
Co r O C M h Co)
Cu r M at N le 00 V
r' to N to
Cu
'a r
CD O r M 'et <D f0 O CO r r CO O
O O O Co f- N M O
N M c6 N r r r (
V Co O N O; Vi c; O O `1!
r -f pr
p LO N M 1 O
r 0 -e Co P
(A O
N p p) 0 e
CO M N N r r
tY N N C%j CO {f
CO r
N 1[) Co t0 M Co
r ~ r
M
C
O
16-
CI)
C o o "- a
O o o O ff
m
+
CD r
E
m C w
'=v tt G
C O - N O l: 1
Y p
r r- V
O
u
"a
op ...
y Co y
o 9
o 5 a '"S
i
P P
= p p *0
j 0 r-
t 2
_Y4 00 00
Q O O 0 O C C v
m CO O C 0)
Y 0 0 m o
C) o . w o m m , s .- r "-
Cx. r m II
5 S g a m m
CL I- II . m> m
5 CO E m C"
_ E o ' to
0' E " m
(D
0 ' 0 m
" w - E .. E E
c , ". m
U o ,r c d a o a m m a o E E
( a r
C
CL 0 v v Cl- E E o c w m
C o 2 m .. ..
O '0 -0 t E E * t b
E CO 0) C) C) 2=
C)
O is v ' m co - 12 a y W S" $'
CL E o m 15
ca 5
> ; > > > >
a n. a
a E d
cl C) 0 12 > > > a
-a 0 z >
no. (D (D ( 0 > > > 1- u7 U U 0 0
381'
O O O In h O d O
N N N r O) O
Co O O O O
Co C) Co !O to Co N d O
N d d CO O) d
Ol O d N 0) Co 0) ^ O
O 10
r Co tN Co 0) N
r 1f>
CO r to M C N
NM r
O
N
0 0 d 0 N N O Co) CO O N tD
O O Co d t_ Co O N "f 0
O O Co IZ 1A T O
Co O l) ^ M CG M O to Co d CO N d N
O r d M Co In r N M Co d CO O)
O Co M r CO N d 0
co oO r 01 p T 0)
r p d (3) r r d r
0 d Co O) - t
1[) U) Of
0 Co N
r O)
O M
CV
In d 0 0 0 e 0 0
O d d O O O CO r
d O O O N 'ct O
O O ^ N N O N CD N N M
r d
Co O 10 d o N r' d 0 0
O O N Co Lii
Co ci
M u)
9) Co - Co N O O U) T O
Co
Co
(O
N
C) M T
CO Oi Co r- T M T
O M p M d ^ U r O '
d d N N O)
O
r r U)
r M
r
0 0 Co N to to 0 N in C') NN 0
0 Co CO N O) Co *'
M O 0 to O N M ~
h Co C''I r 0 00 CO "
4i Co O - M 1A h to 1 d CI) 1 C0
C r 0 N A
O 0 (0 n Co co r Co 01 a M ll cri Co
O e2 Co
M M C) Co O) 0 M
v f, uA Cl) 0) O CI) Co Co
D
Co r M
N T O Co O 1 Of iL'1
L- N O N. O O O Co d CO Co
N O O Co r Co c4 IA
O p f M IA d O M In t0 tC 'f
Co O T d to CO t0 M O O h 0 rNN (0 (0 O
M IA M N U) T
Co Ol O p M r N N 0 Co N0 0M M T
*" r p Co 0 M T t0 r Co M Co U) - d
h c0 d O O Co Co 0
01
0
O *' O)
C T M
O O M t0 O in O d LA N O! 'lf
N Of N N N N
ci O CID O O M O r CD O
O N a Ip Co M O) O M 1- .e M O CO 1A r'
pj O
CI)) CD 0 Co 0 Co O to T d fN 1N Co v-
C N in Ict N T
N t0 N O Co tY
Co Co d Ip d O M (0 O
n O O N
r' r
L
fN
0 0 0 U) - - O - M - Co - -
Co -O -' - Co - - O - O - 1A - CD M U) O

M Co N to
iC.. O O 0 0 N Of r.: 1: (0
i' Co O O Cy
M M O U-) d O r N Co Co r Ln O
Qi 01 Co O U T Co r CO d
yn Co O CO M 1A M 4Y d r
Q. N O d O Co O
Co d O T
M M h LO d O O Co U[) 0
Ln Co T
,a O T
T O M
C r.
M
O O 1A Co M 0) O O 1 C1 t0
N O, O N O N O
O O O "
h O O 0 Co to N O O
M
0 ei tO (0 c; Co M N ~
O O
Cp h In
M h
Ln
N O
M
N
Co v ' d' to Of
y,,,
O Co
Co
(O
d
Co c)
O1 d O r' h
M Co 0
N
O
d
r Co
p <O N YY
d M d d O 0) N) Co
Zi Co
O OI T
O N
r
7x; O O N 1D r ( O M M N u!
Co O O r d w CO Kn
CO) Of M t0 N O
h O Of c to d O O M CD f N U) M r r
O O O d CO Co O O
d r d In O N Co (0
'C r N O p N 0) IO LO C) M d d
d Co d d Co M M Co
C
Co Co
C r
O
16.
U) C Q
y
C
V 0 CL
O
a
p
o
p m
U m
m II C O. 'C N 0
O 0
O y "a
T O a
>% O 0 C A
A r
C t
C> y Q
0
O O N
C m ^ C C M S
0 v
Y O p d Q)
o_ O m pC m S T T
o
r r
( p
g i111S 11
S C 1 11 y m II C C
C
' rj- Of
T
0 0
o 0) t (Y) , m CL
E E d to
.:
CD H .a T r' to ' m a
G) 0 c m a o o d a E E E
Ci. CL
C El
'O Y
n
v v C u ai
c 43
m CO 0 m d 4m3
ma
E T V (Cl ..
CL
CL KL CL IL CL
a)
CL 92 c3 0 13
o >tu
> >cu >C O C.
)
0 Co> Z P- W
CL D, (D,( CD>
382
O 19 N O O Co O) O e
N O A <D T O 1n C7 (3)
d p i
Of O d' O h t Co h N v CO
; 0 CD Of
O h Co in CO CO N
p h CO er N p
O U) (0 N r' h Co
0 C9 h N 0 lt
h N to
(Si
-O T - CO - N - N - fO - h - N - - - Co - O - - O -O f0 Co O O
0 V
O 10 4 M N T c to N to h
.i ui c9 r
Q h OI "cT (0 n 02 N
0 at r de
(i r a0 Co h
v - N N h N Co
0 -f
r
T
` N
-co O I h Op CO O r O
O h CO Co CO O
o M
O M . e Co Co to t
(3) Of .. a 1f) e p
O) O 1 Il) CO Of N O U)
0 N ,a r '-t .,
O 1n (Si CO C) T h O
T CO
o h N st
U) r
" N
N f- O) O 0 O N Of c! N
m N N Co T p
Ci Ol 07
00 CO N KA Co CO
a O N h t0 -e N CO c0 01 le Ln
C) O tt f0 h 1f) 1n
h O r CO
Cu U) (Si CO zr CO U)
-
CO N cy ei
st
E
L'
ep
r
r
N
0 O) O Ol r'
i CD 1 O N O
CO O O O C9 O r CD O)
Q1 a) N 4)
Co O O
aD M U) d' ad h c+) h C p p
a 01 O t IC) CO CO C9 Lf)
f0 p C er r 00
in r
C p) r ...
r (0
Co CD V) CI)
m e - C7 r
N T
Cu N
O O O lt h 9 <e O
h O O O O 't O CI) N le
Cu C) 0 N O) 0
CO O h p t0 N O C Co Co -e
0I O V"
p '. t G> M N
Cu O h CO h CO Co
O O T T r"
.- h CD to (0
-
*0 le C) O in Co) N
T
T
/1
I N
.a.
'C
C
C
O
- 0
m V H
o p r 0
90
O 0 E 0 y Y
o ayi
Ny d II c
0
p d
C Y O `
y o y
0 h `j 2 cu
U
"C
Co 0 CL
Co C) Co
a
C - 9 : :
c h h
0 C: )
C N h
. O ^S
Y v ' ob
Co 0 8
o (D (P r
O
0" r
C m
5 O n o
Z a
C u E m uu 0 o c o i ui rn
CL)
5 r
u) r C) rn 9) rn c n a n
c rn
. o Cn A
E E `m ca E E
cC)
% c
r- m . o E n d 0 n 0 E
K _ 0 T T n d c
a v v 0 0
a v Ci
n 0 43
' '
C i aho 'c "o v c 9 `m o E i ..
i 5 2
CL
CL

a n. Cl. n. >m m
> m
> > an
> d
7 E n 0+ O t
'Co j E >
Q Co (11 >0 0
> z' Io ui 0 0 0 0
d0 C07 (C)7 Cta7(07 > >
383
o 9 o o v v in M v r 9 o ci n
n
n o Co M rn un rn Co
M N cd d =
n O co
M N O) 1q O cd N
a of rn Z OD
of
0
co
Co .
0
^ M 0 r
o 2 e M 1p r N N .- ^ r N M N N '? 0 r
01 co Co n n F F N
CO r r 1A C9
C9 N
Lil
n O f Co CO N Of r O ( O N 1
O M CO r e r 80 (D O CD O N
t0 n OD Of 'a lY n
n O CO M GD r M N t N t0 u3 Of CO
cj (0 0 C') N C' ) 0) M
O 0 N (7) IA
M M n M n M 0 C') N - C) M
2 to O
n N r -' r 1-t r-
0 n n to
O r
M r d' r r N M r'
N N r r
M N
O n N 00 't r r 1A O Ci 1 O
O N <O N O) O h
90 r F M N sr
n O fit) M p r r r r r M c3)
M
n `f
C9
0)
n
O
M M M 10 0 Cl
(3) o Co 0)
0
r M r M M
(0 cO r le CI) 0) CO
r (D In n r r r r p O N Co vi
N 0 p p) r n I-
C M r
r r r r
F r a}
M N
to
_UT n M O CO O 0/ tf Cf O C7 f-
N 00 t0 00 M 00 M 1-f
ei er O in Co M N M N h n n N V'
n Co .4
n O to N aj
N t0 C,j p O Co O> Co
In
d M M N 0 N Co Ir W
C c3) 0 CK) O N N F CO
(0 M r r t00 an 0)
r tt) U) 0N O) OOi Co 0 O) 0 1p
Q1 c0 a 0 0
N s! r N 't r
T r4 N N r r' "ef
Co M N
1 ul
0 0 O M r Iq f f t O a0 li
Q1 c0 1n M n r a0 aD ee r O O
o it r N 4 O N n M
CO n O N r O ,4 U n n r r
N n N U) Il) O M M t to N N CO M O
T C) O N V) (0 F t0 d' 0
N n r t! ) t0 M Co of r
M N N r 't
r 'l n O
n Co N F N Co M r
N N r
Co O N
C')
0 N O e (C Of C1 CO Of O to c!
10 O r O r V' r
Co N O r N le CO Co M Co N O c16
t
M Ol Oi
n p u) M M N N aD n u;
r O N O N N M M I 'p N M W) U() O
V pi O C) - F le C#) r c3) V)
N p e Co tn - r
n `t r F N
r ) (0 C) ul Co
(0 F r r M r
r_ 0 "- r "p
M N
N Co fD i0 10 O 10 O O 10 O
to -e C) st 01 Co
() O O t0 CD O 'V' O
r n Co O) r 1lf
r n O e , Co c) c; et c; N r r
pi Co (0 F- o
_ 0) 0 n "4> n Co Co N M C') N F r N ^
O) IG r
r Of n r M t0 (0 r O r M V'
t0 Co N N
N Y9 n O) O) 0
N r N
0
Co O !7 N 10 Co Of !7 t0 1f O
O 01 M f 01 h N
O O O le O M 0 c;
00 n 1 t 00
rn a n n n F n o N 0 M n rn 0) vat (0 0
N p r r' M M IA r r i0 r r
N M N r N r O '
,a r)
_ r
O U) n r C9 n
IA p) n 0) M M "7 r
r_ CD OD r N
O N
Co - -r - - - - - - - `! -O - a0 et le w o o Of N [1 le O O O
-n
N N n Of I 'C t Ol to
'~ O n F O In O r of r
t0 ^ ei In I[) 0 M O
Cy) O 1A O
n Co Co O M O cy, CO f r) N Kn
M
r
CO
r
N' *n QO
N 'et
10 r n p N c{ r r n 1[) f0 N e CO
t0 t0 Co Co i.
e
i CD r er
N N
a n
N
0 c^

n
0 a g ed
p
N Co
v a

m
0 I
CL
CL i
n o g
y o o o CL
r_ a o a, ym o r 2 a r p a
0 0
O r r x
V p p C 5 p
dE
2Q5
p
Co
0 X m p Q
m it
NT Op j Na yr
co m S
o C E. C
-. r r r r v 's 4 n
0 m
il y II II m m
V C E p owl `
vl C n aEi Q r a i j
m m Of O m D S
(Q N m <O . . W
c m o (M C m co CO , m m E E a a r
m ) : a m m E E
CL o
o
v
r- a r a 9 9 9 9
C Y m
t; M vv vv c c - c c e C b
0
Cp v m y uEj r r
ano ,O m
t3 b is ` g
o
Q o Cu cis m a ,
'5
E r V 0 19
m m m m m m v
CL, a n a a a a > > > > > > a E m
L in 0 > > > > > > > Z F ui r o 0
I (510 c9 -
384
M r N c 9 N h N O Co O 1n N OD N 'p h O M aD
t0 O O M M h Co R p (d M
U) N Co C) lA Of e' h in
c0 O O f- O M Co r M Co N a at r
O) O cli h Co h M N r In Ol M 4n Of r
01 `7 <O F- F- Co at
CD Co N N CO r -t o)O n 0 N r
M r r r r 00)
p ul Ln LO
F- M Ln N N M 'if r
N r r
O N
O M h Of 10 N O O 1A O
O CO r N Izt h N O CO le N
In o Co r r M 3
Co O r O C') P Co 0 . r r t 1f)
CO O tC cO M O (O M
N M ^ M N M O 0) O M 0
0) O N O O C) N M N P
CD h N '? 0) Co (D r N r
r p> (O .r r CO r 'r M N 0 O)
_ Co - - M !
N) N N r
r4
(D C') N
0) C) N t0 'd' O r h W O O Of
0 h N of t[) 0) Co N O N lt
et Vf d 00 N t le N N
CO O O p 10
c) o cO Oi fz ( h r
M tf r N N M O r` Co i-
c3) p Co N (3) M t0 N 0M CO et
"., in (00 (00
M -t
C) A Co O Cy)
O 0 tl) r r 1[) r r O) 0 O)
(O N ^ r - r
N N N N eF r
'R M
O M
Ln N
r- O O r CO e '? N t0 f- O O 10
O M O ^ M N M O O Co co
M p C9 N N O'
l0 O N r M ^ CO CO O CO <O M P cd M in
0 M M F` t0 e M r N M CD O Co M CO N r
O Co N U) r CO CO er r r p Q' 0 C) (0 r
N r r
r CO M r r p M M C0 co CO
'7 U) Co r 0) N Cl
"% N N
% M M (p
Co M N
M O 10 N U) h '- U) U) (0 0 N C)
N (0 0 U) G) Co C) C) 00 0)
v N 0 P - M O r'
O c0 F- N N O r /z Of OD N r
CD
0) O M
n rZ ui ei N M M N to o
0
00 M e N h 0
9
to
0
tA O
o Co
t
Co r r to to " r
Co r N CO CO r r r 0 M 90 P.
0 h
N N N M
0 M N O r r r
L M N
y..
f, IA O N Ce O 1l Of
V
O U) r r f- Co O CO h O C sT Co 0 r to

Il
0 0
r r 0 <O Co N 0 'P M
Co 0 an N tC O to CO O 'cf a0 r
M in Co 0 N N M C) N P 0) N - 0)
O Co M N C-2 M O) Co M N Co Co
= A le er 0 M Co
CY n h r O r r O M UA at to
" IA r Co
ti) r r r Co r N N M "
0) N
M N
Co
(D W) 0 Co (0 0 1 19 r O 1l 9
P C) LO 1- h t0 0 1- t0 r r
Co 0 0 N Co 0 p
M M M O (D M 0 O UA Cl O Co Co
a0 O h O p Co Co N O <O N M
I- N -e M M r N f0 r 1A N 0)
O O CO M 0 CO f') M M U) (0
i.. r O r N O r r - r r r r" lt -cr r
r Co I[) CD 00 <O
Q M M """ CO r r p
M N
r r N N l`') r
y.. M r r r a
L O M N
LO
Co O 0 1' P 9) (o r '7 OD O Of 10
O Co N in U1 h N (O F- CO f- " N N
0 O O Ql fp U) N M . 1n
h O O 1p I[)
.a o - 0
Os C) N - m - Ln - to le N M MO'
In tn 0) r le C) CO CO P.
Co
CY
l' Co
O "
Oi
CM
c
r
0 (0 0 C) 0 f-
r N
N r' r N l r
r M r r N
Ri O M N
V Ci C> Co O 0) cg IA 0 le C Kn r O to O
O CO N M r P M M W
CO f ' (0 fC Oi EO
h O <O In p M M O CO 10 f OD " "
r M r O M N M ''! N `C 1- f- 1- O
Of O O - uj N st 'V" sf at " - N C) M N r` r
CD Co -ir CD
c CD r r r"
N
r" le
OY O a
16. h O C W N Co
yr -r N 'V ' r
r r ,e r r r ,e
CO)
M N
O

O p m
Co ^ O 2
O N Co
O n
73 c: M N r "'
C N
tl1 n N
. ' O y
L n n m
0 p
c) O i0' 47
O
a
0 a o
C. 0) m o r ,
, r n
O 0 C K
0 0 r r - . 1
0 T:

Y
c
c S 34 Co

x m t a "- g
cO i O ' 00 p w r 6 -
S ayi
n
T o
aci v1 S it
Y- r o r m W II
E II E m m>, 0 cm ` CY)
V 0 S 5 t m n O tu 0)
O) m T Of O c Z
c O rn in o ao , m E E a ' n :
O .. n ' o a C E
CL u c "" >
ò' E - n g c aEi
m 0i 90 V E nE c o O Co
m _O N t 0 90 V d m li 0
Y i d ui -= c.
-5
C E 00 1 ..
w to z a So tai g 'c 5 5
d a co .M M 5 5 5
CL n d > > > > > > >
Q d
c7
a

c7
d

c7
n
o
c7
a

C7
> > > > > > z
E
-
h t
F
cg o 9 9
m 0
0
385
N LO N uY O O M Co e O O N O 01 N
M O 1. CO c7 't N Co to to
O C ; 0 t0 O LA T
0) ei O; h ei O; P et et lA
M Co M N
O CA N. N. M at N. CO r- N r)
O N N tt O) r r `7 r t U) r
O O W r r r U) CO N
N
CM CO M ^ d' mot' Co)
r M v N
O (D IA M N r r
r- Co
at O Ln
N
O O) N
Co
O CO N O) CO O O M O i0 N
C) C1 O1 - Of 0
N 0 CO CO -e - N N N V
ID O
O M O ^ O CO d n N N d '11 O N f.:
CO 0 Co M N N 10 -e
2
O Co of N in r N te r r 1n I. v
r - e r r O O Co
O O O N. M '7 N
O KD U (0 N r
Co M '-T O N
N
to N. O 0) O IA O p) CO
0) N 'Q n C) O ^ 0 N Co O r l
r O h M Co O O O) CO CO O
O O eh r_ C) c0 c0 f0 In cm t0 r 9
Of c6 Co t0 o C) O Co Y M
o h M r Co U) to c4 sf O CO Cl
c33 N O O m Co Cl O)
e t P Of r Co 1() r r
c> Of r r r r r r r O
O N. T M to h r
+ 1, le le M M N N
r CO to st st
CO " 0 Cl (0
N M N
O
0 Co N. U) O 19 O ei O P. CO
t Cl M to CO Cl 4) CO N. 'd' 0 0
O O O to - h CM 0 P
O to C) O h O in r
C) ( IlY I O CO Co O le 0 Cl M
( O) o m N O
e2 CO M M CO (0 P CO
) lit le
M O
n O M r r le et Cl r C9
yam, r O O O p) CM r r N r r O N. O so N
CO M LO Co
KO U) N. U)
O Co t st V' U) N
n U) N in N ", ', N
Co
V r
O r !7 O M V
1 O h Co CO CO O in to CO Co r
h N. Cl r Co ( (0
O 0 O . O M N Co O
M O1 r 0 O 1 r f0 (0
CO O r i} In N M
E
01 O M rj
C
N
CO N
LC)
O
0 Kn e O O
Co Co v2 in
to r
N. r r to Co
r KO
M N r ,p ^ r 0) N r
" - O IA N in M Cl - r Co N
4f) - M N. N. Co N N
CI) r "sr er M
(N) N O M r CO
V Cl
i N. M N O CO IC) t[) CO O r OD M O U) O!
E oD
CO
O
O
N
"t
A f0 r
O Co
of CO I
O
M
N
r
O O N
r
P
Co O
M Of CO N (0 Co If) CO Co Of
M
M
O O to O O U1 N U) -f -! O Cl c Y
1.. O O M O Cl '- r CO r r 0) t gA r
- N 0) N r In O r O 01
r O r M
C) N r r 0 -ir o M f Ln N
o N O le cri LA N
f
M N
Q LO Ict
. O O O O r 't
O O O O V' CO O IA Co O to co d 00 t0 r V' O1

N. 00 m Ol M CO
13 CO O O O O
Co O
CO I. -e N 0) O O O O
1n
r
Co
r
Co) 0
r
Co O r N
C 0 In O CO .-r C < N
O r N CO Co r - -r - M 1n e to e IO r
O N -
m - O Co
N
p) r N r r r
Co Ilt C 4n -9r d
Co co Co o M N
-,
M
N N CD CM
C7 Co)
Co
yam.. O
N
a y
C o 8 cu
c Co
O y CO o
r y
(, i 0 d N C Q
x
0 (4 to
T]
h- C Co co a ' 8

C 0 Co
C
Co 0 r
a r o
Cl. - Ei
a . V c K
O O r
O o r c a
C) c t r
a x m o h M
c c v
be
M.
a

0 v

r
T rn
o r r

r
0 0 ai oi F a +S m
r
11
c \ E $ w
c 0 a m II 0 .5 5
V ' rn
H n a rn W i rn o a s .2
cp 03 y .! CO - (c o co CO d d E E E
' a tu ' > o a a
K CL 0 E 0
` a,
v o a c d e c a
:p ..m a
1; N m v vd ' c ai o o 0 0, o ' b
C 15 ( m
N c v v c m . a
E v t a o r 5
O io v u i co a '
E
W 5 i ' 5 -
O' o _ > > ; > > > ? ,
a a a a n 0 0
> > > > > > > z
KL t(7 (9 0 c(7 c(7 F- uF7
386
Appendix 7
Construction output statistics for 87 countries
3871
M In to (D d' CO M CO C'') N N O M to CO r
y r O (0 In CO r" "
4 N (0 lt U() r r sf 0) GO CL) 0) 1L) CV 0 <O
h (0 0 r,: M 0
c C") to C7 M !O e O) - - O) V' r CO ' 40 Co
r, Co Co M
C N f- Co 10 to M Co CM C) (D e t0 r
10 to Co (D
0 r ^ o M M r LA st
4'
0 0 M 0 " N - to Co to .'
N n r
U
(0 Co S r h t CO CO CO .- 7
C Co e2 to N co a W W Of at r r-
1 h f0 1 f r' M M r' h
O c O IA Co In N Co M O M ^ 9 i

0 h 0) t0 4 SC) Co N 0 Il ) 0 to O U)
= O f- 't r to
.- M
7a E
o c
0 - 0) M O W (0 7 N If) - to 0) 0) M 0) P CO
(y O O r to r- Co CD O ve
tA 1- N M et Q O N h 00
f l M O O lt Co r Co " 4
y M h 4 N
N N K6 N N M (0 '- ' O to 4 C9
ii Z N .- 0) M
-
jr. - r N P r V N
N r.
y
y
y O 0y C 7
7. y o
C C y NC
r U a G. C m y fl
y y y O y
y ry. C 0 r
C - E C - = O = O
ld cd y j t
v E E E E S 0
5
U y
c Con-.
u..
a V o
a
O
G
O > >
- - -
_
0 _
E E E . E E E E E
E E E E E E E E E E E
N et M M et O> O f- r Of N f0 M r
- ;2 e c O O N w Co r LA h O M
M CO N CO N r O to C) O) Of 7 'e
CD h M r t
CL N f. 01 CO N O M fz N M Iz
h (O Co r to O> O vA O) h h of
In
M e N 40 CO CO 0) to IA O V' M CO r
to O M CD - r 0
N CO
V ld r r r r r r,
CU
N 0
7 U
O)
Q)
f- P h N LA N Co at h - N e
1 M Co O> Of CO M _ r
d N to _ M CO to to CO tf 0 N) CO h - 0 CO CO le r
tt Co h tp Co 0
d
c2. CO r 1z f0 (6 N -t 1p Co tf P 1z O
CO Of kA - 10 r f0
N a :3 1- M t0 Co e h M et " Ol O Co t 01 Co h 'e et M h Co IA O
d h P CD Co Co C') a C) N Ol .- h M CO O
(. U) N N P 0 0) P
F. to to e2 0) N 0 CO N N CO N CM IA
^ 0 CO CO d. N CO
- M M
y. + r M t
O
7 .
0 N
>
V
O t- N N 0 IA Co 0 N 0 0 M Of CO r
ti a N Co ^ F- Co CO . - to C') N O e
N p) CD "7 tp r to O> r C') e r) N 0 r
O to O P
oo N h M Ci t

r -
4 1Z
C`3 = tC M t0 O LA a0 N N cm Oi
r) O ( st
A- O D to r 1- M 'V' M C'9 N e le of
M Ir N N N N h n SO Co N .-
0 .- N Of u) "-
f U(1 `-
o N M
r r r
Cu
e M to I. N e0 0 14t O Co N
cu O LA Of N O O O 4A c
et r N Co CO M - '- N CO Of (0 0 -if CO Co to to
oD h et t r
"0 N N M
N M [i M M of ei c4 <O ui vf N O;
y t 1
c .S
0
-
c
CL
tG Co 0 ' le M M 0 In 0) CO N 0) t0 CO
N -. et CO f- OD 'C st U) !9 CO C') O!
CO F- r' C') N f. In of h e- CD C')
0 et M 0
O
Co N
[f 'd'
M
tp to O
N (G at OO of !) O cD 1A c 1A
Z) CO O t0 a0 U) F- 0 M C') IA .- IA .- i-
w M in 0 t0 h in CO C9 .- .- N V) N f- -f "
N N 40 Co
Q
c 1n U) o to O in 0 0 N 0
N 0 to
N Z-- C (O n to lcr N " C)
N N

'a e N C) C) 0
OD 0 00 F- F- 0 0 0
c_ C C O) 0 0 (0 (D 0 In to -.
0O N Of 0 0 to 0 O! N to Co
O 0 0 0 0 0 1A 0 er I r
"c "O
O
O CO Co 0 CO 1- i- 0 0 0 0 N O/ P '
- 0 0) N O In tD N 0 N C'') r' In Co
O in U) Co Co U) Co .- Co 0 er
M N f7 0 CO -.
O ti M N CO N Co in if
y=. c r 0 CM -t
N
u) u) In 002 0
N CO
cm y M
C)
V c r
3 O
V
a.
y.. Z
N
c
O
U
Co
C
Co
`
to g ti
Co c E A.
v
' _
CL
Q. 0 N 9 ' E

0
a U Q Q 2 m Cocu in m m 0 Co Co m U U (( 0 w w w w 1L
388'
Ih 0) Co M I() 0h li! N 1'- U) N
I- N .-N 0) CI) 99, UA
.-N0 Ul) 0) Co e Co N u) 4Y 0 Cl tD u) s'
N (+) 0 V' "0P u) CO 'j
O0h ei 0) O) MM Co N ob h e) eP 0)
CV (0 N (0 CD 1. Co (D
Co a Co ONP r if C9 MN-0 IA rn
1- ^ 't t q Co Co LC) tn
0 0-N .-0N t0 rM Co (0
Co Or" 4) CO
a O F- (n a0 C') N' U) ^NN rj
0 c0
NNN
N Co N to h It) i- et e Co to " Ol M
N Co Co Co Co 90 CA -e
C tf) Co P- 4NO! O C) f') r 01 9N C')
c4 h 09 N 19
0 O CO rOOh N
Ul) l7 lt t- -e -! to Co) O1
id f_ 0 C) C> Co An LA OO0h
a{ In NN at
.- le CO
u') CO .-
W
0
CL 1- CO 0 0) Co 0)
r. (p O of O0 C) 0 CO U[) tO N ir -ON
.-. 110 N 10 'et O e- CO 19 10 NNO (n
PN O) f, 10 ' a0 7Nh
0 (0 00 t)
c) se to h CO C) 40 f- CO 0N U) N CO 0
e .-ON N r- O Co N Co N
CO h LA
m
of Eyt
y td y C
r_
Ndd
0
U
C
Co (p tu 00dyG
m
U= C cr Co
_= EvEv
0M EE o rn E EEENE v "Nc
v
0
EEEEEE ra EE ra EEfEEE
E E
d
1, Co t0 M .- U) Co 0M 1-. 1- a' 0N CO If)
h (3 C) 00 Co M C) llf 1p C9 at Co
O G1 .-
C m hN 10 10 O) h C9 Ir-

r a Co N Co N t h at to V' N .-rN CO CO e'MPPN
K) 1-
(0 Co CD N .- CO .hh 1N0
NM
C I[) i- O
Co
(0
lit N (0 r- t0 r CO t- sf .-ON IL)
0)
0
V m
v m
N
C) m
C)
r to Cl) qr (0 Co 0) CO 0N to CO 0) co co 1f P) C) 0) CO off N0
h 01
O C9 OO ul OR NrYOO cl CO
0) 0 co prhO IA
M
m 4 Nh CO CD 't VV 00
N
U[) 1- t Ui !hN
Cl) a aD u'1 e- 1-
0) V, Nh 0) (7 h 0) 1() hVV 0) t0 of N ""
MP CO U) In r (0 N0 to P0
U) OD
d m O 1l) O0 (D co It F- 0 It r Go It .0r
MN 4) f- CO N
p 0) U) rN (0 CO NNt IL) 1- CO V (0
Cl) N CO IT to WN c')
v 'rr
Co
m
0
1- N V) P- -t IA h Of {fl N)
N C) to (0 CO (0 CO '7 C) it
It) r c0 VON (D 0
a (D O (0 O CO CO W) r CO N 0) CO !7 l4! N l4t O) O)
co 0 4 C') N in
6NONr Oi OO! in N tG t r66 in M
a- c7 00 f0
co 0rV C) N-0 (0 co -Nh i- N0N 00 N C') r in
rr O) Vr co to r CO U) VNN ! ') f0 7 to
0 r U) co CO
(0 N C') 0 to V
N U) qw Cr) rC
C)
0 N -.
U MN- t") N E0 IA W Ol V C)
o0 N LA d' 0) r It (0 co
Cu co co C9 N 1A OD O OD W T- ID 0r C) NO
NO CD IN V' Ul) Co C7 rNrr
6 4m to ^644^ 40 l7
N OD IA U) 1n 1n f0 N)
In 1n 1n ui
C
O
0
CL Cl
0) Co Co Co 0N Co Co Co O00N Co r g?
ON CO M Of CD Co cm Co Of NM Ob
0 (s h c4 O Il O9OeNO
h
Co Ol
MN
vA
Ql 1 ED wi
d Uf u;
CL NrO CO Co O Co Co Co Co Co O/
U1 NO
M Co et hON 1A N) 0h ( CO Co -t
O t0 N 1A N K) 40 (- V'
CO d' U[) O) NO aT (O h Co) MMrN -e r
h Cl O) r CO r
CO U) r C') N Of errr00N
N 1l) 0 Co NN
"' NrNN
7 N C9
U) M Of OOOOO U) OO t' Or
C_ OOOOrrrOOO (2 0000 N) 0 O
C9 0 01 0000 Co Co O0
"C OOOO N0 Co CD 0 (0 0 at 0 Uf Co r
(0 N0 0) Ul) 0 lit 'ct 0 `7 'V'
CI) .-O N` CD OMN 1 02
NMr Co M CO r CO hN Oi~r In IA
O rn N CO Nkr 0 1
m F. (O et () sh N^0 CO Co
1A 0N lam
rn N in N'r CO
V
i
N
C
O
ti
X
' NNm
rn CL
C
61 Y(d
d Z, co
m mmv v 0) VS
a C
7 vd
a O
U C7 ti C7 C7
20Z5. hp =- YX
a
389
CO CO (2 N to le M N O) r IA IA CO
M 0 t0 CO 0 - 10 CO et T
rA T M O O UA T M N N p n (O
N U) U) e N O Cl O) CO .
<O O M d' to eh O
N c
t0 01 N Co O Ict CO N Co
U) N
f
N - ^ r o N
C M c O d" (0 C) t0 CO N N fN M O CM 0)
CV C) le in C> V) to ^
0' e Q (0 n Cl I, C%t Co M r r M
CO N M Co M CO C9 kn "
U M of r CO
r
N T
7
T
U
N CO LO <C CO O M r P 1n N N
M N et 10 CO T tt M CO 01 T N U) r O O CO CO
d " N 19) CO N to U) T e C7 CO
Q (O N O U) U) r 4
r N
4
L6 N -c! N U O N fz O O N M
i0 '- OD O O ui IA M ui ni
E N IcF
Co T T r
Q
0
O C
(0 CO 0 N CO N M Of Co 0 Of Co) O1 ""'
0 Co N M M O) N N r - N Ol l! NN N O 1n Co
(q 0 Co C) K) CO fN CD CO to O O N r
to Ul) N N
O (n - N - .- N U) 9 0) O u; r Ul) N U) CO r
N t In IA O) r N
r N r j N r N
` r r r `
d% CO
r T
CI
i O
O
T y (q _ M
y O
to 0 a 0
-a ei 0
(P C cis
9 'a 2 l
2 c 22 - . =
Q w a a . 3 t= c
_
o _ _
o 0 E E E c "E 1; E E E E
JG E E E E E E
CO M CO t0 tN 0 Of (1 ""'
C N O U) et 01 0 CD 0 CO 1N T
N C! U) G) fr
- - Co of r T 0 CO CO Co M N N CO 01 C') M U) CO
O) r U) e
r r f, Cf OD O> 01 tG O N at N Co CO
+ M I M O h O> O 'd' "t r r N 0 N
00 N U) i r U) 0 N
= M U) N to "- M l N CO
aai r r r
r r r r
a T
:. i cd
cl.
N w
C)
N ' N O/ N U) CO M O
T N C) 0 00 0 Co of CO CO
N Co O) t0 Co et 1N 10 to T N CO N T
N '0 O CO M 1n M 40 fN U)
'ch 0 le -
1 r T r
of rn r: of ri ui m m 1i
- ui of O T vi
d ? a of of
p
vn ui
N Co CO O T M M O N CO N O 1A N M in
t0
M Co
N
"t M
Q: Co r 1N U[) C1 Co 0 N sP U[) 'R 01 O et of 'C
t N Of T in (0 e0
M O) U) CO N 0 .- U) 'C - r
"> *a 0 Lo N N Co a 'C N CO N r
N
,. o
4.0
C Co >,
y O
O.
V > N N N N N N
C n Co N Co ' N 0 C) ^
Co a `t M Co N 0) Omi CC) M u0.1
N h O n d O;

.4
0) 01 cD o ut t0 UA N if N c0 Lf)
C3 M N a0 t aD Cl M
O) aD r 0 Of M 'C' O U tO V" r N T r r
Co O - N 'C (0 N
O p) C) 'C 0 M N U) W T to
o M M n - M Co)
y r' r pp r lt N M N `Nt
y. N 1n r
O Pn
Cu T
V
N N "- M Of 00 CO <O N <O aC N "d 0
Cu CO CO r- N M M T T N O N CO I.
CO r M N 0 0) N O CO N W W 0 r
" CO O O h
N 0 CO M P N
` t M M IC) vi Of '4 Cd it c C0 ui r
C M 01 <O t r r In st T

CL 0
0 1N UO 1N CO O O N CO tO 00 d "'"
N O O N 'C M O UA 1N le N O O
0 N t0 h N r CO h r r O N OD CD N
A O T O h N CO LA r N M h O
M M to I f' N Co 0
v = N .
4 M t f0 M N eP fG O
N O C) C M 10 1n c> 0) T
Co) CO O Co (D t0 Of -
'C N M IA 01 M
p2 O) P M O 10 M h e) N O O
(0 O (D M N N Co -t to N
y. r C O e -e r U) N N - T 40 M T 0
: i M N M - r' N
0
Z (9
10 0
U) fO O 0 M 0 1A r 0 O
C 0 0 0 0 0 0 r O
C C 0 r O r CO 0 0 CO r N 0 0 at O M 0 O O
O M Co 0 0 0 0 0 0 O N
" 0 IA O 0 0 M 00 M 0 0 T 0 1A 1O M
'0 O 0 O) (0 ei
' O) - 0 N Co O O> O U) t0 U) N O) %t O O CM CO M r N T
N
M c Co -e -. N 0 CO CO l ") O
0 2 - N O N CO Co è N N M M
",
N N M M IA r C) r T M
C) j; -e CO r
r r
c T
0
0 U
N Z
C
0
0 C
U
ti
N aC "
C d Al
(f3
0 CU (U Co
C. E ed d -
mss. Qi i A
tu Co '" 0
Q- la O cm
0 O 0 z in v) 'm 0
z z
390
N to u) 0 0 n N n N C9 .- e! U) n e CO
y O) O) Cl Cl O1 eh Co Cl '
C N U) M a:
Cl (0 - M O CO
Co h Co C) O> st et Ln
C = 0 O> Cl C) O Cl U) -
C)
to r cl) 9 N Cl CO O O) r- ^ N Cl
CL u) CIJ CI) -
O to (0
U M C
C r N r '
y
U
N U) Co IO M N M N 'f n
C N 0) O (O r
C N O> n <O CO O - N tD N
O O 9 O 01
O n cj CO O
Co Cl to r n CO n n to M
E U) U[) U) N N Cl r
M r
a
o C
IL N r t0 O O OO N tO n
(+) at C1 O "f M O
N N O to U) O O C7 M
O C7 N M r
O N
i - U) e 0 U) 0 - 1A Of in at
Z O at U) r . N CO h
(3) r N 0 cl)
GC "," -t N
O n N
CL
lto y
d
!q
>. y c y O y
C tu
v C
d `' y 'y0 y "yp
0) C) c: C7 to c) cu
Co
o c -t i 2
CL Co V Se
j :3 t 12 12 to a c V
^ 0
E E E E E E E E E E E E E
.
d
Z lf) t0 CO Co n
(0 CO e O 00 CO -
".. - M CO N V) h Of h h
= h t O h a0 to
h f st r
*+ U) I- O) e) O> -e 01 Cl Co v) r t 0)
to IA r f 10 Co to 0 O1 r
C 0 Co CO Cl O r 0 10
N O CO - r
O a C9 O O - 7 -
0 0 N N
o
v
N
C)
h (3 N C) h r tO Cl
Lo U) N. r r -e c3)
d O N to h t n O 19 CO r 19
1 ' M - M O) tF
a CO " r- of ri ei Lci t:
to ri of ui
O)
n
Cl C)
r,
r r-
-- - n to IA 0 0 C Co
n 0 0 n c3) O N C)
M IA CO n O CO C)
M '"t to
<O
L V M O) CO r GO M N to M CO (0 r
m M Cl It) C) M Cl
. 4+ al M (O U)
tu >
O,
Z CL
V > E
0 to 1n
u) O M h n CO O CO n
N r at (- N 1- M
(O 0 Co N 1 O Co O
Co Co r

C) to e) Cl r Co Cl d' to eh
C7 U) r e
O 0 n Of 0 O N to ft 0
to Cl r (3) U)
N n to le Co O C0 M CO 1- r-
v = O l0 ^ -
t0 O -
- r Cl
IA
c Cl N
0 O) Cl to Q!
*W t0 'V aD IA r r M
(p
O

C)
et
OI
M
C)
O
r
O h v <C N
O 0 O M Co Co
Y 0
O) ! t0 N at ( to "t N
y C Iti t0 td oD M t0

C. U
M u) n 0 Co Co 1B 1A N N
y M e{ p Co Cl O)
Q fA Cl Cl O N to M CO 40 r Cl
A 0 U) r U) to
Ip
2-. cn 1n M O M CO) r r M h N M (D h
cco, .a Ip n n 0 N M
C) -e CO N
U D . Cl ' LO tn CO
*+ 10 q
to M 0 U) h r N Ln c9 - r n C) M if
C et Cl Co Cl Cl Cl
r r u) Cl "a r N
M M r
O O O O O O O r O O
O O O O
C C O O O O O 00 O O O O O
" O O M O O O O O
U) O O O O O O N M O O
O CO O O O
n to n 0 Ol 0 "" M n M r'
U O Co M 0
O Co 2 h O F- CM h CO N to
C) to n f
C')
N
M
O
n
it
Cl v 141
y;, r C) V; A N Cl
r r Co
V y C -
V 0
O
N Z
o
n
x d in p mod
.
1- ..
CL Co(nm 0 0 P F-~ > > >
391
Appendix 8
Site productivity data for eight construction activities in

Tanzania
392
Notes to Appendix 8:
1. Code: These represent the East African Standard Method of

Measurement for Building Works (1977)
classifications for building work and the study
" number. For example, Code A1.12 represents
activity sampling study number 12 for blocklaying
item (230 mm).
2. Site: The first letter in the code representsa contractor.

The number attached to the letter representsthe site
number for the contractor. For example, Site N2
representssite number 2 for contractor N.
Contractorswere coded for confidentiality.
3. Productive time: This refers to the percentage of time the operatives

were observed active during the observation period.
Productive time included both direct and indirect
work.
4. Unproductive time: This is the percentage of time spent on unproductive

work.
5. Productivity: This was computed by dividing the total output during

observation time by the total number of labour hours
(Ihrs) expended.
6. Productivityrate: This was computed by dividing the total output during

observationtime by the total productive labour hour
expended.
7. Statistics: The averages, standard deviation, skewness and

kurtosis are presented.
393
A: BLOCKLAYING (in sm)
Code Site Productive Unproductive Productivity Productivity

rate
time (%) time in (sm/Ihr) (sm/Ihr)
.%
A1.1 K1 54.10 45.80 0.43 1.33
A1.2 Fl 67.60 32.40 0.34 0.51
A1.3 Dl 70.00 30.00 0.96 1.39
A1.4 Fl 65.00 35.00 0.29 0.44
A1.5 A3 61.50 38.48 0.18 0.29
A1.6. L1 87.00 13.00 0.70 0.87
A1.7 Fl 79.00 21.00 0.81 1.03
A1.8 Fl 68.00 32.00 0.38 0.54
A1.9 N2 72.00 28.00 1.40 1.94
A1.10 N2 39.00 61.00 0.68 1.73
A1.11 N2 42.40 57.60 0.64 1.48
A1.12 N2 54.60 45.40 0.91 1.67
A1.13 N2 65.20 35.30 1.23 1.91
A1.14 Ni 82.00 18.00 1.29 1.58
A1.15 Ni 88.90 16.10 1.36 1.63
A1.16 Ni 89.70 10.30 1.12 1.25
A1.17 Ni 91.20 8.80 1.19 1.30
A1.18 N2 68.90 31.10 1.66 2.05
A1.19 Ni 78.80 21.20 2.82 3.58
A1.20 Ni 77.00 23.00 1.22 1.58
A1.21 Ni 60.00 40.00 1.05 1.75
A1.22 Ni 77.70 22.30 1.20 1.54
A1.23 Ni 71.50 28.50 1,02 1.43
A1.24 Ni 76.00 24.00 1.22 1.61
A1.25 Ni 73.80 26.20 1.26 1.71
A1.26 Ni 77.00 23.00 1.26 1.64
A1.27 Ni 74.00 26.00 1.33 1.80
A1.28 Ni 73.00 27.00 1.22 1.67
A1.29 N2 61.30 38.60 0.71 0.43
A1.30 N2 58.90 41.10 0.39 0.62
A1.31 N2 57.40 42.60 0.31 0.53
A1.32 N2 58.90 41.10 0.27 0.46
A1.33 N2 62.30 37.70 0.40 0.67
A1.34 N2 70.10 29.90 0.81 11.16
A1.35 N2 41.20 58.80 0.42 1.02
A1.36 N2 76.30 23.70 1.14 1.49
A1.37 N2 82.30 17.70 1.03 1.29
A1.38 N2 80.00 20.00 1.13 1.41
A1.39 M1 36.00 64.00 0.43 0.76
A1.40 M1 50.20 49.80 0.39 0.77
A1.41 M1 98.90 1.10 1.08 1.09
A1.42 M1 97.90 2.10 0.92 0.94
A1.43 A3 81.60 18.40 1.16 1.42
A1.44 A3 73.30 26.70 0.92 1.26
A1.46 A3 69.00 31.00 0.95 1.37
A1.47 A3 95.80 4.20 1.01 1.05
394
A: BLOCKLAYING (continued)
Code Site Productive Unproductive Productivity Productivity rate
time (%) time in % (sm/Ihr) (sm/Ihr)
A1.48 G3 53.60 46.40 1.23 2.30
A1.49 N2 67.70 32.30 0.49 0.72
A1.50 Ti 60.20 39.80 0.64 1.06
A1.51 Ti 70.00 30.00 1.85 2.65
A1.52 Ti 65.80 34.30 1.33 2.02
A1.53 Ti 58.30 41.70 0.86 1.48
A1.54 Ti 69.90 30.10 0.48 0.68
A1.55 Ti 73.10 26.90 0.82 1.12
A1.56 V1 43.50 56.50 0.22 0.49
A1.57 V1 45.60 54.40 0.35 0.78
A1.58 V1 65.80 32.20 0.66 1.21
A1.59 V1 50.50 49.50 0.49 0.96
A1.60 V1 59.50 40.50 0.52 0.88
A1.61 V1 59.40 40.60 0.26 0.44
A1.62 V1 13.00 32.00 0.31 0.46
A1.63 V1 56.40 43.60 0.24 0.43
A1.64 Al 62.80 37.20 0.28 0.68
A1.65 Al 56.20 43.80 0.20 0.36
A1.66 Al 50.90 49.10 0.25 0.48
A1.67 X 43.50 56.50 0.31 0.71
A1.68 X 49.00 55.00 0.28 0.61
A1.69 X 51.36 48.64 0.24 0.46
A1.70 X 45.53 54.47 0.18 0.40
A1.71 X 63.10 36.90 0.27 0.43
A1.72 H3 34.30 67.50 0.19 0.55
A1.73 H3 68.00 32.00 0.38 0.56
A1.74 H3 19.80 80.20 0.15 0.75
A1.75 H3 30.30 69.70 0.20 0.67
A1.76 H3 59.85 40.15 0.32 0.53
A1.77 H3 66.25 33.75 0.33 0.50
A1.78 H3 60.00 40.00 0.33 0.55
A1.79 X 45.50 54.50 0.10 0.22
A1.80 Y1 62.80 37.20 0.41 0.65
A1.81 Y1 70.00 30.00 0.29 0.42
A1.82 S2 76.04 23.96 0.35 0.46
A1.83 Z1 64.30 35.70 0.12 0.18
A1.84 Zi 50.46 49.54 0.11 0.21
A1.85 N5 65.90 34.10 0.25 0.39
A1.86 L3 50.82 49.18 0.31 0.61
A1.87 L3 36.89 63.11 0.98 2.66
A1.88 B2 59.38 40.62 0.43 0.73
A1.89 B2 59.97 40.03 0.40 0.66
A1.90 B3 64.50 35.50 0.44 0.68
A1.91 B2 59.33 40.67 0.44 0.75
Average 63.06 36.43 0.69 1.15

Std. deviation 16.05 15.15 0.48 1.24
Skewnesss -0.37 0.18 1.29 6.13
Kurtosis 0.74 0.32 2.85 48.23
395
396
A2: BLOKMAKING (in No. )

time (%) time (%) (No/Ihr) rate (No/Ihr)
A2.1 Fl 55.70 44.20 13.00 23.36
A2.2 Fl 55.00 45.00 2.00 4.70

A2.3 Fl 71.00 29.00 7.00 12.00
A2.4 K1 67.00 33.00 9.00 13.00

KI 52.90 47.00 6.20 21.40
A2.5
G1 56.60 43.00 3.50 6.30
A2.6
A2.7 Ni 75.20 24.00 16.00 22.00
A2.8 Ni 58.50 41.50 15.00. 25.00
A2.9 N2 57.00 43.00 13.00 22.00
A2.10 Ni 76.10 23.90 18.00 24.00
A2,11 N3 80.00 20.00 20.00 24.00

A2.12 N2 59.00 41.00 12.00 21.00
Ni 65.20 34.80 32.00 49.00
A2.13
Ni 53.70 46.30 10.00 18.00
A2.14
Ni 66.30 33.70 16.00 24.00
A2.15
A2.16 Ni 60.10 39.90 15.00 25.00
A2.17 Ni 61.80 38.20 14.00 22.00
A2.18 Ni 43.10 56.90 5.00 12.00
A2.19 N2 76.70 23.30 11.00 14.00
A2.20 N2 77.10 22.90 16.00 21.00

A2.21 A3 86.20 13.88 15.00 17.00
A2.22 A3 59.50 40.50 18.00 22.00
A2.23 A3 47.50 52.50 16.00 34.00
A2.24 A3 64.50 35.50 37.00 44.00
Q 77.40 22.60 21.00 26.00
A2.25
A2.26 Q 76.30 28.70 17.00 22.00
A2.27 Q 75.80 24.20 15.00 19.00
A2.28 0 83.30 16.70 15.00 18.00
A2.29 Q 73.70 26.30 16.00 23.00
A2.30 N 44.70 55.10 4.00 9.00
A2.31 N 46.50 53.50 8.00 18.00
A2.32 N 52.50 47.50 9.00 16.00
A2.33 A7 33.10 66.90 1.29 3.90
A2.34 A7 52.00 48.00 4.04 7.80
A2.35 A7 61.50 38.50 3.70 6.00
A2.36 A7 33.40 66.60 2.20 6.80
A2.37 Wi 47.50 52.50 15.00 30.00
A2.38 Wi 48.00 52.00 15.00 31.00
A2.39 W2 46.70 53.30 6.00 13.00
A2.40. W2 74.00 26.00 10.00 13.00
A2.41 W2' 64.75 35.75 11.00 17.00
A2.42 W2 80.00 20.00 12.00 15.00
A2.43 V 65.60 34.30 6.90 10.50
A2.44 V 64.30 35.70 9.00 14.00
397
A2: BLOKMAKING (continued)

time (%) time (%) (No/Ihr) rate (No/Ihr)
A2.45 X 51.00 49.00 7.00 13.00

A2.46 V 53.23 46.77 14.00 27.00
A2.47 S2 71.43 28.57 2.90 4.10
A2.48 X 58.70 41.30 1.00 2.00
A2.49 H3 56.50 43.50 9.75 17.25
A2.50 H3 66.25 33.75 8.00 12.00
A2.51 H3 61.30 38.70 13.30 21.72
A2.52 H3 52.30 47.30 6.00 11.00
A2.53 H3 53.29 46.71 9.18 17.21
A2.54 Z1 77.86 22.14 7.49 9.62
A2.55 Z1 65.70 34.30 7.46 11.35
A2.56 B2 64.46 35.54 5.00 8.00
Average 61.76 38.30 11.28 17.75
Std. 12.27 12.18 6.85 9.27

deviation
Skewness -0.12 0.12 1.29 0.94
Kurtosis -0.44 -0.39 3.34 1.86
398
A3: PLASTERING (in sm)

time (%) time in % in sm/Ihr rate In sm/Ihr
A3.1 K1 53.00 47.00 0.86 1.72

A3.2 K1 44.00 56.00 0.55 1.67
A3.3 K1 58.00 42.00 0.92 1.57
A3.4 K1 60.00 40.00 0.85 1.45
A3.5 D1 74.00 26.00 0.95 1.28
A3.6 Dl 91.00 9.00 2.24 2.46
A3.7 L1 60.00 40.00 0.54 0.91
A3.8 L1 80.00 20.00 0.66 0.87
A3.9 A3 80.00 20.00 1.50 1.80
A3.10 Ni 78.40 21.60 3.39 4.32
A3.11 Ni 71.10 28.90 1.92 2.69
A3.12 Ni 73.80 26.20 2.16 2.93
A3.13 Ni 80.30 19.70 1.94 2.42
A3.14 N1 80.50 19.50 2.13 2.65
A3.15 Ni 81.70 18.30 0.81 1.03
A3.16 Ni 76.10 23.90 2.79 3.64
A3.17 Ni 68.40 31.60 2.18 2.90

A3.18 Ni 73.00 27.00 2.40 3.30
A3.19 Ni 79.00 21.00 1.96 2.47

A3.20 Ni 55.60 44.40 1.76 3.17
A3.21 T2 53.00 47.00 1.77 3.35
A3.22 T2 61.00 39.00 1.83 2.98

A3.23 T2 50.80 49.20 2.07 4.08
A3.24 T2 44.30 55.70 1.40 3.16

A3.25 T2 45.70 54.30 1.57 3.44
A3.26 T2 49.50 50.50 1.34 2.72
A3.27 T2 49.80 50.20 1.23 2.47
A3.28 T2 48.30 51.70 1.60 3.32
A3.29 T2 50.30 49.70 1.48 2.95
A3.30 T2 54.10 45.90 2.15 3.97

A3.31 T2 55.40 44.60 1.24 2.24
A3.32 T2 60.50 39.50 1.14 1.88
A3.33 T2 52.30 47.70 2.27 4.34
A3.34 T2 50.90 49.10 1.69 3.32
A3.35 T2 61.40 38.60 4.95 3.17
A3.36 T2 53.60 46.40 2.13 3.98
A3.37 T2 50.50 48.50 1.79 3.55
A3.38 T2 49.00 50.90 1.23 2.51
A3.39 Ti 72.00 28.00 2.70 3.74
A3.40 N2 63.10 36.90 0.83 1.31
A3.41 N2 58.20 41.70 0.77 1.34
A3.42 N2 57.00 43.00 1.00 1.75
A3.43 N2 52.90 47.10 1.37 2.60
A3.44 N2 64.70 35.40 1.20 1.86
A3.45 N2 65.80 34.20 4.40 2.13
399
A3: PLASTERING (continued)

time (%) time in % In sm/Ihr rate In sm/Ihr
A3.46 N2 59.10 41.00 3.92 6.64

A3.47 N3 65.50 34.60 1.93 2.95
A3.48 N3 65.50 34.50 1.70 2.50
A3.49 N3 51.30 48.70 1.65 3.20
A3.50 N3 68.00 32.00 1.80 2.70
A3.51 N3 48.80 51.20 0.98 2.02
A3.52 N3 45.40 54.60 1.35 2.98
A3.53 N3 40.20 59.80 1.32 3.29
A3.54 N3 55.10 44.90 1.37 2.94
A3.55 N3 54.00 46.00 1.35 2.68
A3.56 V1 45.00 55.00 1.04 2.31
A3.57 N3 54.40 45.60 0.54 1.00
A3.58 N3 55.00 45.00 0.78 1.47
A3.59 N3 41.70 58.30 1.09 2.61
A3.60 N3 43.00 57.00 0.97 2.40
A3.61 01 68.90 31.10 1.48 2.15
A3.62 01 63.50 36.50 0.90 1.42
A3.63 01 66.80 33.20 2.01 3.01
A3.64 01 61.10 38.90 1.81 2.97
A3.65 01 74.80 25.20 1.80 2.41
A3.66 al 66.90 33.10 1.68 2.51
A3.67 01 69.40 30.60 2.03 2.93
A3.68 Qi 56.50 43.50 1.56 2.76
A3.69 al 60.00 40.00 1.27 2.12
A3.70 Ti 63.70 36.30 1.47 2.32
A3.71 Ti 68.74 31.26 0.38 0.55
A3.72 Ti 64.00 36.10 4.08 6.40
A3.73 Ti 60.05 39.95 0.95 1.59
A3.74 V 45.00 55.00 0.15 0.34
A3.75 X 96.50 53.50 0.64 1.33
A3.76 V 45.25 54.75 0.31 0.69
A3.77 N2 60.00 40.00 1.20 2.00
A3.78 U2 38.80 61.10 0.85 2.20
A3.79 U2 44.90 55.10 0.54 1.20
A3.80 U2 48.90 51.10 1.16 2.40
A3.81 U2 51.00 49.00 0.89 1.76
A3.82 U2 46.50 53.50 0.58 1.25
A3.83 U1 57.86 42.14 0.69 1.20
A3.84 U1 52.72 47.38 1.12 2.13
A3.85 U1 63.15 36.85 0.55 0.88
A3.86 U1 64.95 35.05 0.58 0.90
A3.87 S2 75.00 25.00 0.87 1.16
A3.88 N6 58.20 41.80 0.51 0.88
A3.89 72 45.50 54.50 0.78 1.72
A3.90 X 64.80 35.20 1.07 1.67
400
A3: PLASTERING (continued)

time (%) time in % In sm/Ihr rate In sm/ihr
A3.91 A8 64.65 35.35 0.69 1.07

A3.92 A8 41.25 58.75 0.56 1.36
A3.93 A8 58.75 41.25 0.57 0.97
A3.94 A8 61.75 38.25 1.11 1.80
A3.95 A8 71.50 28.50 1.02 1.43
A3.96 A8 64.20 35.80 0.28 0.44
A3.97 A8 62.10 37.90 0.30 0.48
A3.98 A8 60.00 40.00 0.62 1.03
A3.99 A8 67.80 32.20 0.49 0.72
A3.100 A8 51.50 48.50 0.48 0.93
A3.101 Y1 80.00 20.00 0.46 0.58
A3.102 Y1 67.53 32.47 1.67 2.47
A3.103 B2 51.52 48.48 0.40 0.78
A3.104 A8 64.93 35.07 0.63 0.97
A3.105 A8 59.14 40.86 0.25 0.42
A3.106 A8 56.63 43.37 0.36 0.64
A3.107 L3 56.00 44.00 1.83 3.27
A3.108 L3 42.00 58.00 1.00 2.38
A3.109 L3 33.00 67.00 0.08 0.24
A3.110 L3 56.80 43.20 1.51 2.66
A3.111 B2 76.89 23.01 1.82 2.37
A3.112 B2 60.35 39.14 1.42 2.35
A3.113 B2 67.94 32.06 0.09 0.13
A3.114 B2 62.50 37.50 0.66 1.06
A3.115 B2 62.84 37.16 1.35 2.15
A3.116 B2 74.09 25.91 2.06 2.78
A3.117 B2 68.80 31.19 1.02 1.48
A3.118 B2 60.19 39.81 0.42 0.70
A3.119 B2 50.91 49.09 0.39 0.77
Average 60.03 40.38 1.32 2.14
Std. deviation 11.57 11.12 0.86 1.15
Skewness 0.38 -0.22 1.59 0.87
Kurtosis 0.14 -0.31 3.93 1.98
401
A4: CONCRETING (cm)

time (%) time (%) in cm/Ihr rate In cm/Ihr
A4.1 KI 43.00 57.00 0.09 0.21

A4.2 Hi 28.00 72.00 0.08 0.29
A4.3 Ni 89.14 10.86 0.01 0.01
A4.4 NI 76.20 23.80 0.08 0.10
A4.5 N2 90.90 9.10 0.16 0.18
A4.6 N2 29.40 70.60 0.04 0.14
A4.7 N2 58.30 41.70 0.07 0.12
A4.8 N2 72.10 27.90 0.11 0.15
A4.9 N2 79.00 21.00 0.15 0.19
A4.10 N2 53.60 46.40 0.06 0.11
A4.11 N2 49.90 50.10 0.05 0.10
A4.12 N2 47.00 53.00 0.04 0.09
A4.13 N2 65.30 34.70 0.04 0.06
A4.14 L2 60.20 39.80 0.30 0.50
A4.15 L2 32.60 67.40 0.10 0.31
A4.16 L2 42.00 58.00 0.47 1.12
A4.17 L2 62.30 37.70 0.50 0.80
A4.18 L2 24.20 75.80 0.53 2.19
A4.19 L2 31.60 68.40 0.53 1.68
A4.20 A3 45.90 54.10 0.32 0.70
A4.21 A3 98.00 2.00 0.23 0.23
A4.22 A3 41.20 58.80 0.46 1.12
A4.23 A3 69.20 30.80 0.13 0.19
A4.24 A3 59.70 40.30 0.14 0.23
A4.25 A3 57.80 42.20 0.19 0.33
A4.26 P1 42.30 57.70 0.05 0.12
A4.27 M1 18.90 81.10 0.04 0.21
A4.28 M1 82.30 17.70 0.10 0.12
A4.30 Qi 84.60 13.70 0.16 0.24
A4.31 01 89.20 15.90 0.21 0.23
A4.32 01 34.90 26.00 0.19 0.28
A4.33 Qi 89.20 49.05 0.21 1.14
A4.34 01 35.75 42.50 0.58 0.09
A4.35 GI 69.90 43.20 0.05 1.18
A4.36 R1 44.40 45.20 0.67 0.29
A4.37 N2 66.90 59.30 0.16 0.07
A4.38 S 28.60 34.00 0.03 0.06
A4.39 V1 78.00 42.80 0.04 0.72
A4.40 V1 51.10 37.40 0.41 0.59
A4.41 H1 65.40 53.10 0.37 1.36
A4.42 S2 78.70 43.69 0.64 3.73
A4.43 Zi 63.00 37.00 0.03 0.05
A4.44 Z1 65.30 34.70 0.03 0.05
A4.45 Z1 67.00 33.00 0.03 0.04
402
A4: CONCRETING (cm)
Cade Site Productive Unproductive Productivity Productivity

time (%) time (%) in cm/Ihr rate In cm/Ihr
A4.46 Z1 66.50 33.50 0.03 0.05

A4.47 Z1 50.50 49.50 0.04 0.08
A4.48 H1 55.80 44.20 0.29 0.52
A4.49 H3 37.10 62.90 0.08 0.22
A4.50 H3 32.60 67.40 0.08 0.25
A4.51 H3 24.24 75.76 0.13 0.54
A4.52 A8 58.36 41.64 0.03 0.05
A4.53 B2 49.03 50.97 0.02 0.04
A4.54 B2 66.95 33.05 0.05 0.07
A4.55 B2 73.91 26.09 0.04' 0.05
A4.56 B2 70.11 29.89 0.04 0.06
A4.57 B2 64.96 35.04 0.03 0.05
Average 57.36 43.04 0.17 0.42
Std. 19.77 18.03 0.18 0.64

deviation
Skewness 0.00 1.37 3.23

-0.02
Kurtosis -0.32 0.76 13.06

-0.82
403
Cl: EXCAVATION (cm)
Oode Site Productive Unproductive Productivity Productivity

time in % time in % in cm/Ihr rate in cm/Ihr
C1.1 A4 35.00 65.00 0.36 1.03

C1.2 B1 70.00 30.00 0.71 1.01
C1.3 B1 60.00 40.00 0.55 0.92
C1.4 Gi 70.00 30.00 0.29 0.41
C1.5 G1 71.00 29.00 0.32 0.45
C1.6 G1 59.00 41.00 0.26 0.44
C1.7 L1 50.00 50.00 0.61 1.22
C1.8 A3 74.00 26.00 1.00 1.35
C1.9 A3 78.00 22.00 0.58 0.74
C1.10 ml 25.40 74.60 0.33 1.30
C1.11 M1 40.00 60.00 0.31 0.78
C1.12 M1 52.00 48.00 0.37 0.71
C1.13 M1 67.50 32.50 0.61 0.90
C1.14 M1 55.60 44.40 0.47 0.85
C1.15 M1 31.20 68.80 0.32 1.03
C1.16 Ni 60.90 39.10 0.22 0.36
C1.17 Ni 63.40 36.60 0.17 0.27
C1.18 N2 65.10 34.90 0.86 1.32
C1.19 01 66.70 33.30 0.50 0.75
C1.20 01 52.20 47.80 1.02 1.95
C1.21 X 41.50 58.50 0.16 0.39
C1.22 X 54.20 45.80 0.22 0.41
C1.23 X 30.80 69.20 0.32 1.04
C1.24 X 57.10 42.90 0.35 0.61
C1.25 X 41.80 58.20 0.19 0.45
C1.26 X 40.00 60.00 0.17 0.43
C1.27 X 46.60 53.40 0.31 0.67
C1.28 X 58.30 41.70 0.21 0.36
C1.29 X 34.10 65.90 0.14 0.41
C1.30 X 26.90 73.10 0.11 0.41
C1.31 X 51.00 49.00 0.93 1.82
C1.32 X 45.00 55.00 0.26 0.58
C1.33 X 48.40 51.60 0.51 1.05
C1.34 X 50.50 49.50 0.13 0.26
C1.35 Z2 46.80 53.20 0.67 1.43
C1.36 N2 62.40 37.60 0.07 0.11
C1.37 H3 50.50 49.50 0.23 0.46
C1.38 H3 45.30 54.70 0.42 0.93
C1.39 N4 62.00 38.00 0.24 0.39
C1.40 V 68.50 31.50 0.11 0.16
404
Cl: EXCAVATION (continued)
(ode Site Productive Unproductive Productivity Productivity

time in % time in % in cm/Ihr rate in cm/Ihr
C1.41 V 68.25 31.75 0.08 0.12

C1.42 V 57.75 42.25 0.12 0.21
C1.43 V 47.25 52.75 0.09 0.19
C1.44 Y1 71.75 28.25 0.46 0.64
C1.45 Y1 43.00 57 0.66 1.53
C1.46 V 57.25 42.75 0.07 0.12
Average 53.35 46.65 0.37 0.72
Std. 13.27 13.27 0.25 0.46

deviation
Skewness 0.25 1.03 0.83

-0.25
Kurtosis -0.66 0.43 0.13

-0.66
405
D2: FORMWORK FIXING (sm)

time (%) time (%) in sm/Ihr rate in sm/Ihr
D2.1 L1 47.00 53.00 0.33 0.70

D2.2 L1 38.00 62.00 0.44 1.16
D2.3 L1 66.00 34.00 0.54 0.82
D2.4 L1 46.00 54.00 0.99 2.15
D2.5 H1 52.00 48.00 0.79 1.52
D2.6 L1 42.00 58.00 0.60 1.43
D2.7 F1 68.80 31.20 0.24 0.35
D2.8 A3 78.30 21.70 0.92 1.17
D2.9 A3 85.30 14.70 1.01 1.18
D2.10 A3 86.40 13.60 0.34 0.39
D2.11 L2 63.20 36.80 0.25 0.40
D2.12 L2 63.10 36.90 0.97 1.54
D2.13 L2 53.40 -46.60 0.47 0.88
D2.14 L2 75.80 24.20 0.85 1.12
02.15 L2 56.20 43.80 0.46 0.82
D2.16 L2 56.20 43.80 0,19 0.34
D2.17 L2 63.50 36.50 0.14 0.22
02.18 L2 74.60 25.40 1.54 2.06
D2.19 L2 42.00 58.00 0.07 0.17
D2.20 L2 82.40 17.60 0.79 0.96
D2.21 L2 80.50 19.50 0.45 0.56
D2.22 L2 84.90 15.10 0.53 0.62
D2.23 L2 70.90 29.10 0.78 1.10
D2.24 L2 54.00 46.00 0.55 1.02
D2.25 L2 68.10 31.90 0.98 1.44
02.26 L2 69.90 30.10 0.92 1.32
D2.27 L2 57.00 43.00 0.24 0.42
D2.28 M1 81.10 18.90 0.72 0.89
D2.29 M1 84.80 15.20 0.40 0.47
D2.30 M1 93.10 6.90 0.22 0.24
D2.31 Ni 90.00 10.00 0.60 0.67
D2.32 N2 70.50 29.50 0.31 0.44
D2.33 N2 57.30 42.70 0.56 0.98
D2.34 N2 54.30 45.70 0.42 0.77
D2.35 N2 59.90 40.10 0.66 1.10
D2.36 N2 75.50 24.50 0.49 0.65
D2.37 N2 72.10 27.90 0.81 1.12
D2.38 N2 63.20 36.80 0.55 0.87
D2.39 N2 50.00 50.00 0.35 0.70
D2.40 N2 65.70 34.30 0.55 0.84
D2.41 N2 48.90 51.10 0.30 0.61
D2.42 01 72.90 27.10 0.91 1.25
D2.43 01 72.00 28.00 1.74 2.42
D2.44 01 64.90 35.10 1.90 2.93
406
D2: FORMWORK FIXING (continued)

time (%) time (%) in sm/Ihr rate In sm/Ihr
D2.45 01 75.40 24.60 1.19 1.58

D2.46 01 73.50 26.50 1.31 1.78
D2.47 01 63.90 36.10 0.85 1.33
D2.48 01 73.30 26.70 2.03 2.77
D2.49 01 66.50 33.50 0.04 0.06
D2.50 G1 46.90 53.10 0.44 0.94
D2.51 Gi 53.60 46.40 0.81 1.51
D2.52 G1 36.84 63.16 1.06 2.88
D2.53 G1 59.70 40.30 2.00 3.35
D2.54 G1 37.70 62.30 0.59 1.56
D2.55 Gi 63.40 36.60 0.87 1.37
D2.56 G1 45.90 54.10 1.40 3.05
D2.57 01 77.80 22.20 1.64 2.11
D2.58 al 51.00 49.00 0.47 0.92
D2.59 01 78.00 22.00 0.93 1.19
D2.60 01 45.90 54.10 0.11 0.24
D2.61 01 37.80 62.20 0.11 0.29
D2.62 01 40.50 59.50 0.48 1.19
D2.63 A7 42.70 57.30 0.76 1.78
D2.64 A7 49.00 51.00 0.71 1.45
D2.65 N2 52.60 47.40 0.46 0.87
D2.66 N2 49.80 50.20 0.46 0.92
D2.67 N2 53.40 46.60 0.42 0.79
D2.68 H1 78.60 21.40 0.62 0.79
D2.70 X 36.25 63.75 0.07 0.19
D2.71 H1 71.14 28.86 1.94 2.73
D2.72 H1 72.04 27.96 1.81 2.51
D2.73 X 55.50 44.50 0.04 0.07
D2.74 X 57.92 42.08 0.11 0.19
D2.75 H1 74.80 25.20 0.82 1.10
D2.76 H1 51.80 48.20 0.58 1.12
D2.77 H3 59.60 40.40 0.75 1.26
D2.78 H3 42.70 57.30 0.16 0.37
D2.79 H3 49.75 50.25 0.77 1.55
D2.80 X 52.00 48.00 0.81 1.56
D2.81 G1 72.89 27.11 0.92 1.26
D2.82 G1 78.36 21.64 1.99 2.54
D2.83 Z1 70.40 29.60 0.45 0.64
D2.84 Z1 50.50 49.50 0.42 0.83
D2.85 Z1 62.60 37.40 0.49 0.78
D2.86 Z1 72.10 27.90 0.47 0.65
D2.87 Zi 72.40 27.60 0.56 0.77
D2.88 Z1 75.12 24.88 0.55 0.73
D2.89 Z2 49.10 50.90 0.73 1.49
D2.91 S2 72.84 27.16 2.07 2.84
407
D2: FORMWORK FIXING (continued)

D2.92 S2 80.25 19.75 2.23 2.78

D2.93 G1 72.88 27.12 0.71 0.97
D2.94 G1 75.37 24.63 2.13 2.83
D2.95 N5 62.70 37.30 0.51 0.81
D2.96 S2 62.21 37.79 0.88 1.41
D2.97 Hi 78.60 21.40 0.62 0.79
D2.99 X 36.25 63.75 0.07 0.19
D2.100 H1 71.14 28.86 1.94 2.73
D2.101 H1 72.04 27.96 1.81 2.51
D2.102 X 55.50 44.50 0.04 0.07
D2.103 X1 57.92 42.08 0.11 0.19
D2.104 L3 29.37 70.63 0.96 3.27
D2.105 L3 37.00 63.00 0.43 1.16
Average 62.22 37.78 0.75 1.20
Std. deviation 14.41 14.41 0.55 0.82
Skewness -0.16 0.16 1.15 0.93
Kurtosis -0.82 -0.82 0.64 0.14
408
D: CEILING BOARDS FIXING (sm)
code Site Productive Unproductive Productivity Productivity

D4.1 Ti 73.20 26.80 1.46 1.99
D4.2 Ti 73.90 26.10 1.58 2.14
D4.3 Ti 61.20 38.80 1.58 2.58
D4.4 Ti 60.60 39.40 1.82 3.00
D4.5 Ti 66.80 33.20 2.08 3.11
D4.6 T1 68.40 31.60 1.66 2.43
D4.7 Ti 64.20 35.80 1.49 2.32
D4.8 N2 67.20 32.80 1.48 2.20
D4.9 N2 56.50 43.50 0.83 1.47
D4.10 N2 69.50 30.50 1.40 2.01
D4.11 N2 70.10 29.90 1.29 1.84
D4.12 N2 69.10 30.90 1.23 1.78
D4.13 N2 51.50 48.50 1.20 2.33
D4.14 N2 60.20 39.80 0.95 1.58
D4.15 N2 65.20 34.80 1.32 2.02
D4.16 N2 37.50 62.50 0.50 1.33
D4.17 N2 40.70 59.30 0.50 1.23
D4.18 N2 45.70 54.30 0.40 0.88
D4.19 N2 39.80 60.20 0.51 1.28
D4.20 N2 49.80 50.20 0.59 1.18
D4.21 N2 62.60 37.40 0.68 1.09
D4.22 N3 69.90 30.10 0.70 1.00
D4.23 N3 49.90 50.10 0.40 0.80
D4.24 N3 57.20 42.80 1.01 1.77
D4.25 N3 59.00 41.00 1.12 1.90
D4.26 N3 50.30 49.70 1.26 2.50
D4.27 N3 53.00 47.00 1.91 3.60
D4.28 N3 43.20 56.80 0.56 1.30
D4.29 N3 75.60 24.40 0.33 0.44
D4.30 N3 85.90 14.10 1.14 1.33
D4.31 N3 66.90 33.10 0.86 1.29
D4.32 N3 67.90 32.10 1.16 1.71
D4.33 V 65.90 34.10 0.82 1.24
D4.34 N4 54.00 46.00 0.38 0.70
D4.35 A7 56.86 43.14 0.78 1.37
D4.36 A7 55.65 44.35 0.61 1.10
D4.37 A7 71.10 28.90 0.81 1.14
D4.38 A7 71.40 28.60 0.83 1.16
Average 60.72 39.28 1.03 1.69
Std 11.07 11.07 0.47 0.70

deviation
Skewness 0.28 0.35 0.72

-0.28
Kurtosis 0.30
-0.28 -0.28 -0.78
409
E: PAINTING (sm)

time (%) time (%) In sm/Ihr rate in sm/Ihr
E1.1 B1 33.00 67.00 4.33 13.12

E1.2 B1 39.00 61.00 13.00 33.33
E1.3 B1 51.00 49.00 4.90 9.61
E1.4 A3 71.00 29.00 3.10 4.37
E1.5 il 50.00 50.00 2.50 5.00
E1.6 A4 53.00 47.00 2.04 3.85
E1.7 D1 59.90 40.10 5.32 8.88
E1.8 Dl 61.50 38.50 5.15 8.37
E1.9 D1 38.10 61.90 2.29 6.01
E1.10 N2 76.00 24.00 6.69 8.80
E1.11 W2 62.00 38.00 3.75 6.05
E1.12 W2 72.80 27.20 2.29 3.15
E1.13 W2 38.50 61.50 1.83 4.75
E1.14 W2 79.70 20.30 6.55 8.22
E1.15 W2 76.20 23.80 3.03 3.98
E1.16 W2 53.50 46.50 1.77 3.31
E1.17 02 66.30 33.70 10.00 15.08
E1.18 Ti 72.10 27.90 14.00 19.42
E1.19 Q2 6.60 93.40 3.58 54.24
E1.20 02 52.50 47.50 6.75 12.86
E1.21 N2 64.50 35.50 5.69 8.82
E1.22 N2 64.60 35.40 5.18 8.02
E1.23 N2 74.27 25.73 5.47 7.37
E1.24 N2 53.75 46.25 0.73 1.36
E1.25 V 68.29 31.71 7.08 10.37
E1.26 X 46.77 53.23 1.70 3.63
E1.27 X 46.75 53.25 4.34 9.28
E1.28 Ti 39.50 60.50 4.89 12.38
E1.29 V1 43.00 57.00 6.71 15.60
E1.30 N2 54.00 46.00 5.60 10.37
E1.31 V 56.70 43.30 5.50 9.70
E1.32 N2 71.20 28.80 6.50 9.13
E1.33 N2 78.30 21.70 6.80 8.68
E1.34 N2 67.20 32.80 8.20 12.20
E1.35 N2 72.40 27.60 8.20 11.33
E1.36 V 77.42 22.58 4.80 6.20
E1.37 X 60.00 40.00 2.64 4.40
E1.38 X 52.70 47.30 0.34 0.65
E1.39 A7 86.20 13.80 8.80 10.21
E1.40 A7 80.85 19.15 9.38 11.60
E1.41 A7 74.50 25.50 3.20 4.30
E1.42 A7 82.59 17.41 9.84 11.91
E1.43 A7 79.85 20.15 3.12 3.91
E1.44 A7 75.90 24.10 2.80 3.69
E1.45 A7 83.60 16.40 9.19 10.99
410
E: PAINTING (continued)

time (%) time (%) in sm/Ihr rate In sm/Ihr
E1.46 A7 71.81 28.19 x. 92 4.07

E1.47 A7 65.00 35.00 0.71 1.09
E1.48 A7 73.20 26.80 2.74 3.74
E1.49 A7 76.69 23.31 2.21 2.88
E1.50 A7 43.64 56.36 2.09 4.79
E1.51 A7 67.52 32.48 2.31 3.42
E1.52 A7 64.34 35.66 2.68 4.17
E1.53 A7 75.90 24.10 2.62 3.45
E1.54 A7 80.91 19.09 2.14 2.64
E1.55 A7 54.42 45.58 2.36 4.34
E1.56 A7 85.19 14.81 1.97 2.31
E1.57 A7 66.36 33.64 2.08 3.13
E1.58 A7 60.49 39.51 1.57 2.60
E1.59 A7 72.29 27.71 2.08 2.88
E1.60 A7 64.50 35.50 2.10 3.26
E1.61 A7 78.50 21.50 2.50 3.18
E1.62 A7 72.70 27.30 2.80 3.85
Average 63.56 36.44 4.48 7.97
Std 15.39 15.39 2.96 7.97

deviation
Skewness 1.07 1.19 3.85

-1 .07
Kurtosis 1.68 1.68 1.22 19.46
411

PR

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PR

Uploaded by

Copyright:

Available Formats

Loughborough University

This item was submitted to Loughborough University's Institutional Repository

A Doctoral Thesis. Submitted in partial fulllment of the requirements

Metadata Record: https://dspace.lboro.ac.uk/2134/7444

Publisher: c Ninatubu Mbora Lema

Please cite the published version.

For the full text of this licence, please go to:

Ninatubu Mbora Lema, BSc. (Hons), MSc.

A Doctoral Thesis submitted in partial fulfilment of the

Ninatubu Mbora Lema, 1996

List of tables XIII

This research aimed at investigating strategies for construction performance

Labour productivity was selected as the key construction performance

Factors that influence operatives productivity were identified through an

I wish to express my most sincere gratitude to Dr. Andrew Price, my

I wish to acknowledge the support from Association of Commonwealth

This research was made possible by a number of people and organisations.

Many friends and colleagues at Loughborough contributed to the research

I have received invaluable support from my family during this research. My

2.0 Construction industry characteristics and performance

5.0 Research methodology and benchmarking framework

5.1 Introduction 142

6.1 Introduction 174

7.0 Tanzanian construction industry performance

7.1 Introduction 196

8.0 Analysis of labour productivity at site level

10. Conclusions and recommendations for further research and

Appendix 1: Construction operatives' and contractors' questionnaires

Appendix 2: Cost breakdown of selected building projects in Dar es

Appendix 3: Contractors' questionnaires - Phase II

Appendix 4: Operatives' site questionnaires

Appendix 5: Construction executives' interview guide

Appendix 6: Construction industries performance statistics - Tanzania,

Appendix 7: Construction output statistics for 87 countries

Appendix 8: Site productivity data for eight construction activities in

Figure 1.1 Thesis logical framework 12

Table 2.1 Number of registered building contractors - December 1993 33

Construction productivity has been an issue of concern in most countries for

Historically, the Tanzanian construction industry suffered major setbacks soon

In 1977, a major study was performed aimed at identifying factors contributing

As a result of this study, the National Construction Council (NCC), was

In a recent study focusing on cost trends of both construction inputs and

More recently, the government unveiled a policy document aimed at guiding

" to promote the economic and efficient utilisation of local capacities;

1.2 Justification and significance

Two issues had to be addressed in justifying this research:

This research focused mainly on the benchmarking of construction labour

" Identify, test and establish a practical methodology for construction

Establish productivity benchmarks, best practices and productivity gaps

Other benefits may include initiation of a research culture in the construction

Although serious construction productivity research in developed countries

The concept of benchmarking, which has its roots in the Japanese

1.3 Significance of the building sector

1.4 Objectives of the research

" develop a framework in line with modern performance improvement

" justify the use of labour productivity as a measure of construction site

" develop a simple methodology for labour productivity data collection on

" explore the effects of selected factors on productivity., Concreting will be

" pour size and type;

" create a basis for labour output database development in Tanzania.

The achievement of these objectives would result in identifying some actions

The initial stage of this research involved a literature review to confirm

Macro-construction industry performance indicators were identified and

Extensive labour productivity data had been gathered under a separate

A Doctoral Thesis. Submitted in partial fulllment of the requirements