Professional Documents
Culture Documents
enforced. SHRM research is seldom based on samples from emerging economies where
there may be lapses in the institutional and legal coverage of core labor standards and
therefore poorer working conditions. For example, minimum (2011) and average (2010)
monthly wages in US$ in emerging economies such as China ($183; $461), Mexico
($108; $503), and India ($55; $138) are significantly less than those in the US ($1,257;
$3,340) (International Labor Organization, 2012).
Second, the principal focus of SHRM research has been on differentiation through the
adoption of high-road HR strategies such as HPWSs (Becker & Gerhart, 1996). A HPWS
may not always be a realistic or viable HR strategy alternative for establishments in
emerging economies where resources available to provide extensive inducements such as
100% health care and on-site child care may be limited. Moreover, the broad employee
contributions (Tsui, Pearce, Porter, & Tripoli, 1997), in terms of participation in decisionmaking and exercise of discretion, that are expected in return for these extensive
inducements, may not be attainable in emerging economies due to lower quality of
human capital (Noe, Hollenbeck, Gerhart, & Wright, 2013).
Voluntary labor code adoption neither requires the extensive inducements nor depends on
the broad employee contributions that characterize HPWSs. By meeting core labor
standards in an emerging economy context where labor standards are below average and
HPWSs are less viable, it represents an HR strategy that could potentially differentiate an
adopting establishment from establishments operating under poor and exploitative
conditions.
Traditional HR Strategy-Performance Mediated Effect
Prior SHRM research has demonstrated that HR strategies such as HPWSs, may translate
into greater operational (e.g. quality) and financial performance (e.g. sales volume)
through improvements in employee outcomes such as reduced voluntary turnover and
absenteeism (Batt, 2002; Harrison & Martocchio, 1998). Similarly, as an establishment
voluntarily adopting a labor code commits resources towards improving labor standards
(Barrientos & Smith, 2006), norms of reciprocity may develop (March & Simon, 1958)
leading to positive job attitudes among workers and a greater sense of identification with
the workplace (Mowday, Steers, & Porter, 1979). Consequently, job withdrawal, either
temporarily through absenteeism, or permanently through voluntary turnover, may
decrease. Therefore, I hypothesize the following:
Hypothesis 1: Voluntary labor code adoption is related to improvements in employee
outcomes.
Withdrawal of essential workers can result in loss of vital firm-specific human and social
capital (Hausknecht & Trevor, 2011). In labor-intensive manufacturing environments,
high voluntary turnover and absenteeism may result in increased machine downtime as
new workers are recruited and trained. This disruption to critical operations may in turn
result in delays in meeting manufacturing targets, a reduction in product quality, and
increased wastage. Therefore, by reducing withdrawal among workers, code adoption
Sri Lanka currently produces the most apparel per capita in the world (World Trade
Organization, 2011), and the Garments without Guilt (GwG) labor code was introduced
in 2007 with the leadership of the largest trade association, the Joint Apparel Association
Forum (JAAF), and other industry stakeholders in Sri Lanka. A convenience sample of
122 apparel manufacturing plants was drawn from the population of 312 plants in Sri
Lanka. Data on GwG labor code adoption from 2007 to 2012 were obtained from the
JAAF and matched with the 122 plants for which data on outcomes from 2007 to 2012
were collected using a questionnaire.
In the first stage of data collection, site visits and semi-structured interviews with
industry stakeholders assisted in developing context-based measures and in identifying
the Sewing Machine Operator (MO) as the "core occupational category" (Osterman,
1994: 175). In the second stage, a questionnaire was administered among multiple
respondents at each plant to acquire longitudinal annual data on plant outcomes and
relevant control variables. As a basis for comparison of effect size, and as a control, I also
measured HPWS use as of 2012. Information on plant performance, employee outcomes,
and other plant attributes were reported by the plant's General Manager. Information on
HPWS use was reported by both the Human Resource Manager and the Operations
Manager (or an Industrial Engineer).
Measures
Dependent Variables. Employee outcomes included Absenteeism, which is the number of
whole days of missed work per MO per year, and Voluntary Turnover which is the log of
the percentage of the total MO cadre that quit working at the plant in a typical month in
each year. Operational performance was assessed by four indicators: Efficiency, which is
the log of the number of earned minutes per day (total number of units produced per MO
per day multiplied by Standard Minute Value1) divided by the total number of minutes
worked per day per MO, averaged over the year; Productivity, which is the log of the
number of units produced per MO in a year; Defect Rate, which is the logged percentage
of finished garments produced during a day that were defective, averaged over the year;
and On-time Delivery, which is the percentage of orders received throughout the year that
were delivered to retailers on time. Financial performance was assessed by four
indicators: the log of the volume of garments sold during the year (Sales Volume) and its
annual percentage change (Growth in Sales Volume) and the log of the average price
charged to retailers at point of shipping across all orders produced during the year
(Average Selling Price) and its annual percentage change (Growth in Avg Selling Price).
Independent Variable. GwG Adoption is a binary variable coded 1 if a plant had a
certification date within a particular year from 2007 to 2012. Following this first
adoption, all subsequent years are coded 1 to denote continued maintenance of the labor
code's stipulations.
Control Variables. I controlled for Plant Size, Plant Age, the number of Other Labor
Codes non-voluntarily adopted by the plant, the number of Other Non-Labor Codes
relating to environmental, quality, and energy standards adopted by the plant, and
whether the plant was located in an Export Processing Zone. I controlled for product
variation, using the log of the average Order Quantity across orders and the Average
Standard Minute Value across all products manufactured. The HPWS use measure was an
additive index of 29 standardized items based on previous studies of HPWSs (e.g. Batt,
2002) and of HR practices in the apparel industry (e.g. Appelbaum, Bailey, Berg, &
Kalleberg, 2000) with specific items tailored to the context. The scores on the HPWS use
index were transposed onto a 0-100 scale. The mean ICC (1,1), mean ICC (1,2), and
were .47, .62, and .51.
Analysis
I tested all hypotheses using Generalized Estimating Equations, an extension of the
generalized linear model that accommodates data with repeated measurements on
subjects (Zeger, Liang, & Albert, 1988). I assumed an unstructured working correlation
matrix that does not constrain the dependencies among measurements across time
periods. The unit of analysis is plant-year. Due to missing data on some variables, the
actual number of observations varies between 169 and 732. Tests of indirect effects in
mediated models were based on Kenny, Kashy, and Bolger's (1998) procedures and the
Sobel (1982) test of joint significance of indirect effect coefficients.
RESULTS
Hypothesis 1 was partially supported. GwG adopters had significantly lower absenteeism
than non-adopters ( = -2.98, p < .01) which represented a decline in annual absenteeism
of 3 days per MO. However, there was no significant difference in voluntary turnover
between adopters and non-adopters. Hypothesis 2 was also partially supported. Although
the total effect of GwG adoption on on-time delivery was not significant, its direct effect
was statistically significant ( = .08, p < .05) once absenteeism was included as a
mediator. This was in addition to a significant indirect effect of GwG adoption on on-time
delivery through absenteeism ( = -.004, p < .01). Also, adopters were 32% more
productive than non-adopters ( = .28, p < .01). There was no evidence of an indirect
effect of GwG adoption on productivity through absenteeism. GwG adoption was
unrelated to efficiency and defect rate.
Hypothesis 3 was partially supported. GwG adoption resulted in 33% higher sales
volumes ( = .28, p < .01). When on-time delivery was included as a mediator, there was
evidence of a statistically significant direct effect of GwG adoption ( = .28, p < .01) on
sales volume but no evidence of an indirect effect. The effect of GwG adoption on the
growth in sales volume was negative and statistically significant ( = -.096, p < .05)
implying that adopters experienced a 9.6% lower growth rate in sales volume compared
to non-adopters. This relationship was stronger (more negative) ( = -.102, p < .05) when
on-time delivery was added as a mediator, and the indirect effect through on-time
delivery ( = .43, p < .001) was also significant. There was a statistically significant and
negative direct effect of GwG adoption on the log of average selling price ( = -.25, p < .
01) when on-time delivery was added as a mediator. This implies that adopters priced
their garments around 28% lower than non-adopters. The statistically significant, positive
coefficient on on-time delivery ( = 2.22, p < .001) implied that there was also an indirect
effect of GwG adoption on average selling price in the opposite direction. The growth in
average selling price did not differ between adopters and non-adopters.
Hypothesis 4 was partially supported. GwG adoption was significantly and positively
related to sales volume ( = .20, p < .05) when on-time delivery and absenteeism were
included as mediators. Because neither of these mediators had a statistically significant
coefficient, there was no evidence of an indirect effect of GwG adoption on sales volume.
However, since the direct effect was both statistically significant and positive, Hypothesis
4 is supported in the case of sales volume. GwG adoption resulted in a 22% increase in
sales volume during a year by way of signaling alone. The direct effect of GwG adoption
on the growth in sales volume was statistically significant but negative ( = -.12, p < .01),
while it had an indirect effect through on-time delivery ( = .50, p < .001) that was
statistically significant and positive. There was limited support for a direct effect of GwG
adoption on average selling price ( = -.20, p < .10) and no support for one in the case of
growth in average selling price when on-time delivery and absenteeism were included as
mediators. There was evidence of a significant indirect effect of GwG adoption on
growth in average selling price through absenteeism ( = -.01, p < .05).
DISCUSSION
The findings of this study are indicative of improvements in establishment-level
employee outcomes, operational performance, and financial performance from voluntary
adoption of a labor code that parallel or exceed that of other HR strategies. Voluntary
labor code adopters experienced lower absenteeism, higher rates of on-time delivery, and
higher productivity compared to non-adopters. Code adopters also had sales volumes that
were higher than that of non-adopters, and this difference was explained mainly by a
direct signaling effect rather than by an indirect effect involving internal operational
improvements. Although the sales volume advantage diminished over time due to
adopters' comparatively lower annual growth in sales volume, adopters continued to
realize a sales volume advantage over non-adopters during the 6 year period of the study.
Annual growth in sales volume may be less for adopters due to a decline in the potency
of the signaling effect as code adoption becomes more common, or due to retailers'
redistribution of orders among manufacturers differentiated by other attributes such as
lower cost and speed as a result of declining growth in the demand from consumers
willing to pay a price differential for ethically manufactured goods.
The findings on average selling price and its annual growth differed from what was
expected. The benefit of voluntary labor code adoption was from reducing production
costs and allowing for lower prices to be set, rather than through charging premium prices
for ethically manufactured products. Prices were less among code adopters compared to
non-adopters, and this difference may be driven by operational improvements from
reduced absenteeism and increased productivity. These findings on average selling price
are consistent with cost based pricing strategies predominantly used in outsourced
manufacturing in emerging economies (Pathirana & Heenkenda, 2011) in order to remain
competitive over manufacturers in advanced economies. Based on the findings on the
indirect effect of code adoption on growth in average selling price, the cost savings
associated with lower absenteeism may have allowed code adopters to decrease prices
every year at a higher rate than non-adopters.
In comparison with voluntary labor code adoption, the impact of HPWS use was weaker
or non-significant for sales volume, average selling price, on-time delivery, productivity,
and absenteeism. This suggests that in contexts where resource limitations might prevent
the joint adoption of both HR strategies, a larger payoff may be expected from adopting a
labor code than from an equal investment in adopting a HPWS. Based on the evidence
presented in this study, both researchers and practitioners may benefit from understanding
the effectiveness of voluntary labor code adoption as an HR strategy in emerging
economies.
ENDNOTES
1. Number of minutes to produce one unit at standard (100%) performance (Kanawaty,
1992).
REFERENCES AVAILABLE FROM THE AUTHOR
~~~~~~~~
By MEVAN M. JAYASINGHE, School of Human Resources and Labor Relations,
Michigan State University, 417 South Kedzie Hall, 368 Farm Lane, East Lansing, MI
48824
Copyright of Academy of Management Annual Meeting Proceedings is the property of
Academy of Management and its content may not be copied or emailed to multiple sites
or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.
Back