Professional Documents
Culture Documents
A DEA APPROACH
Burcu Anadol
National Library
Bibliotheque nationale
du Canada
Acquisitions and
Bibliographie Services
Acquisitions et
senrices bibliographiques
Abstract
-
ABSTRACT
Traditionaily Company valuation methods are based on discounted cash tlows and
liquidation values, but aii have a number of known shortcomuigs. The application of
Data Envelopment Analysis (DEA) for finding comparable fims and predictng the
market values of companies is an extension of the market-based approach and replaces
the traditionai method where DEA is used tor assessing relative eEciencies.
For the inefficient units, the DEA-Valuaton analysis correctly classified the
market ranges for 70% of the companies. An upper bourid was predicted for the
inefficient companies based on their efficient peers. Of these, 75% were below their
upper bound. For the efficient peers, DEA was able to tind a lower bound for their
market value based on peers being compared to them 50% of the t h e . Although some of
the ranges specified by the peen are quite wide, the method still does a relatively good
VALUINGPRIVATECOMPANIES:
A DEA APPROACH
II
Dedk ation
- -
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
III
Acknowledgements
ACKNOWLEDGEMENTS
1 wish to sincerely thank rny supervisor Dr. Joseph C . Paradi for his advice,
guidance, and insighti comments throughout this research. It is a very wondemil group
he has set up here at the University ofToronto and 1 have had a fntastic time working at
the CMTE. The deepest appreciation also gocs to the other members of my cornmittee,
Professor McCabe and Professor Vicente. for their comments and suggestions.
1 would aiso like to thank Fai. Sandra. Heather. James, Robin, Taraneh. Kerry.
Zijiang, Paul. John, Ozren. Audrey and al1 the CMTE members for their support and
Inendship. Also many thanks to great friends. Elcin. Cigdem, Aviva, Nihan. Wendy, Joe,
Eric, Jot, Semih, al1 the Indy 9T8's, and to Def and Fun.
Very special thanks to my farnily and iiends. My father, mother and brother
Onur were of great support during this work and 1would not be here without them. With
the exception of the bout with pneumonia this has been a wonderhl experience.
VALUINGPR~VATE
COMPANIES:
A DEA APPROACH
Table of Contents
TABLE OF CONTENTS
2.1. I
Intrnsic Vulue............~...~~~~.~~~~~~~.....~~..~~~~.~.~~.~~..~.
7
2.2 VALUEPURPOSES
.................................................................................................7
2-3 VALUATION
APPROACHES
..................................................................................
2.4 ASSETAPPROACH....................................
.
.
...............................................9
5
DISCOUNTED
CASHFLOW (DCF) .......................................................................... 10
75.1 DCF Pupers ..................................................................................................
2.1 2
11
12
2 - 5 2 1 CAPM ....................................................................................................... 12
...........................15
16
17
19
-7-7
DEA APPROACH
Table of Contents
2-832 issues with this study ................................................................................ 26
Ls.sue.s with the CF Appruuch ........................................................................ 16
2.8.4
2.9 PRIVATE
AND PUBLIC
COMPANIES
.......................................................................... 27
2.9.1
2.9.2
INITIALPUBLICOFFERINGS
(IPOs) .................................................................... 31
2.1 O
34
2.1 1
INTASGIBLE
ASSETVALUATION
........................................................................ 34
2.12
SCOPEAND L I ~ ~ I T A T IOF
O NTHIS
S STUDY...................................................... 35
...................................................... 37
3
3
.2 IDENTIFICATION
OF THE FACTORS
.......................................................................... 40
3
............................................................................
MATHEMATICAL
FOR~IULATION
41
. ~ . -2
1 CCR Output Onented .............................................................................. 43
-)
3.3.2
43
..................................
.
46
PRACTICAL
R E Q C I R E ~ IOF
~ TDEA
S .................................................................... 48
AND FIRMVALL'ATION
...........................................................49
DEA ANALYSIS
NOVELAPPLICXT
IONS OF DEA .................... .
.
.
.
............................................... 50
4
................................................
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
54
VI
Table of Contents
. . . . . . .
4.4 SELECTION
OF FACTORS
.......................................................................................
57
The Inputs......................................................................................................
58
44I
TheOutpntts...................................................................................................
59
63
..................................................................................................
67
5.
DATA ANALYSIS
5.4 CORRELATION
........................................................................................................73
5.5 PRIVATECOMPANIES
................... .-..................................................................... 74
......................................................................... 76
5.6 PL'BLICAND PRIVATECOMPANIES
................................................................................
78
OF COMPANIES
WlTH SIC CODE3600s ........................................ 79
6 1 BCC ANALYS~S
6.1.1
6.1.2
.
Ineffrcient Cornpuny dnulysis ......... .............................................................. 79
79
..
IO0
101
6.7 1
6.2.3
v-
VALUINGPRNATECOMPANIES:
A DEA APPROACH
VII
Table of Contents
6.7 4
I ~ u / ~ ~ u tof
i o!SC
n 3 700s...............................................................................
IO4
6.3 ANALYSE
OF THE COMPANIES
(S1C2000) ........................................................... 108
6.4.7
6.5 PRIVATECOMPANIES
........................................................................................... 111
..
63-I I>t@cient Privutr Cornpunies.....................................................................
112
6.3.2
...................................................
113
115
......................................................................................
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
133
--
List of Tables
LIST OF TABLES
Tubk 2-1 VuiuutioniIpprouch - Advuntuges und Disudvuntages
of
I8
69
6Y
73
-4
-3--
Tuhk 5- 7 List of'the MC codes und the Ci,rre.sponding Nume of the i'ndus~ry
,'
80
XI
82
83
83
86
94
95
Y9
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
List of Tables
104
1IO
113
VALUINGPRIVATE
COMPANIES:A DEA APPROACH
List of Figures
LIST OF FIGURES
Figure 3-1 DEA Frontier
39
42
46
51
61
70
7.5
Fsrre 6-1 Ineflcient Cornpunies t ~ n dtheir Predicted und Act ual Mwket Vulires - 88
Figure 6-2 InefJicirnt Cornpunies- Red MC versus Predicted Upper Bound
89
90
91
92
Y7
IOI
IO2
103
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
List of Figures
Fibwre 6-11 Range for Pre Jicted Murket Vutues und the Reul Vuiue (SIC3 7110) -,105
--- - -
107
--
XII
List of Appendices
LIST OF APPENDICES
Appendix A Public Company Data - 1998
AppendLx B DEA R e d t s & Analysis of heflcient Cornpunies SIC 3600
135
145
188
199
203
207
212
VALUCNG
PRIVATE
COMPANIES:
A DEA APPROACH
IX
II
Chapter 1 Introduction
1. INTRODUCTION
This chapter presents an overview of this thesis by providing background
information on the corporate valuation tield, and establishes the motivations for this
research. Detailed problem definition and research objectives are presented as well as
solution approaches and contributions. The chapter concludes with a summary of the
organization of the remaining document.
1.1
Background
Valuation arose out of the need for appnisals of land. its management. and use.
throughout history, but in the recent years, as businesses have become more cornpiex, a
(USD) was invested in mergers and acquisitions (M&A). This figure was double that of
the peak M&A boom of the 1980s [ETT097]. The reasons for business valuations difkr.
but most businesses are appraised to obtain an irnpartiaI opinion on the value of the
business. Irnproper valuations c m lead to losses adding up to hundreds. thousands. and
o%enmillions ofdollars. Moreover. the range in prices that different buyers will oRer is
large. Ofien, the difference between ofirs may be 200 percent or more [DENNOO].
Chapter 1 Introduction
Company vaiuation studies iaclude the use of the asset-based approach, the
discounted cash flow method and the comparable f i m s approaches- In comparable t-ms.
the subject Company is benchmarked relative to specific comparable companies fiom
which usetl public vaiuation data can be obtained. When using discounted cash flows.
public companies' risWreturn characteristics are analyzed. adjusted and applied to the
subject Company.
I -2
Motivation
The motivation for this research stems fiom studies in several areas of corporate
finance and from applications of Daia Envelopment Analysis (DEA). The importance of
business valuation has received considerable attention over the past years. however. a
comprehensive review of using "comparable" firms has not been considered in the
present literature. For this reason, this work concentrates on tinding a peer group of
comparable tims using DEA; a unique application for this powertiil tool.
DEA is ideal tbr tinding comparable tirms in many ways. The basic underlying
methodology is a linear progamming based cornparison of the units under analysis. be it
b d s . hospital, or in this case businesses. DEA identifies best practice (efficiency
fiontier) units, aflows multiple inputs and outputs. is non-parametric. and deterministic
(does not reqitire the specification of a tnctional tom). DEA provides an efti~ciency
score for each DMU (Decision Making Unit), identities inefiicient units, h d s a peer
croup of similar companies for the units under analysis. and it projects the inefficient
units to the eficient kontier (targets for improvement). It incorporates al1 relevant
factors in a single modrl and no single factor is given prerence (non-biased). It
masures relative efficiency fiom the best-observed perhrrnance rather than fiom
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Chapter 1 Introduction
averages. Also, it does not make judgments on the importance of the variables
(unrestricted weights).
This paper continues the approach started by S h a k [SIMAO], but diflrs in
several respects. First, the sample includes over 400 publicly traded Canadian tims and
over 40 private ones. The models for vaiuation differ with respect to the inpudoutput
variables- as these variables appear to be more appropnate to measure value. Simak does
not carry out the actuai analysis of the private cornpanies with DEA. so this work wiIl fi11
that gap- The data onsisted of known SIC (Standard Industrial Classitication) codes
with analysis carried out only for the companies that had the same 4-digit SIC code.
Further. one can propose to determine an upper bound on the market values of the
ineficient companies and a lower bound for the efficient companies as determined by the
Iarnbda values (resulting tiom the DEA model) found fkom the analysis.
1.3
Problem Definition
The p r i m q objective of this research is to value private companies ushg DEA.
A secondary objective is to tind the tctors that determine market capitalization for
companies and in particuiar. to tind the tirms that can be used for cornparison purposes in
a vaiuation study employing DEA. The first goal of the project is to develop a procedure
For vaiuing public companies using DEA by tinding comparable timis. hence. appropriate
market value ranges, and then to validate the ranges with the known market values of the
public companies. The nest goal is to appiy the procedures available in the public
domain tor valuing pnvate companies. Some of the most pressing issues in developing a
model are to detennine a set of relevant inputs and outputs that serves the objective and
ensuring that important variables are accounted for.
Chapter 1 Introduction
1.4
Thesis Contributions
The major contributions of this thesis are:
To gather a comprehensive data set consisting ofdata for 452 public companies and
49 private cornpanies-
Finding a market value range for the subject (inefficient) company based on its
comparative peers.
For the ineflicient companies, establish an upper bound on their market value based
on the eticient companies they are being compared with.
For the etticient companies, tnd a lower bound for their market value based on the
companies being compared to them.
To provide an additionai tool for business valuators and othea to utilize in order to
detennine the value of a company.
1.5
Thesis Oreanization
The remaining chapters of this thesis are organized as follows:
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Chapter 1 Introduction
Chapter 3 explains the DEA solution approach. The development of the methodology
and the diKerent DEA models are presented. The compatibility with valuation is
discussed.
Chapter cl demonstrates the valuation mode1 ilhstrating the important variables to
utilize. [t demonstrates how each variable was selected and why certain ones were
not included.
Chapter 5 summarizes the data acquisition process and describes the data treatment
that had to be performed for the research.
Chapter 6 presents the results of the analysis and provides a discussion of the results.
An analysis of al1 the public companies with a SIC code in the 3600s is provided
using the BCC ( B d e r , Chames and Cooper) model. A complete analysis of
cornpanies having a SIC code in the 3700s and 2000s is examined including
- -
VALUING PR~VATE
COMPANIES:
A DEA APPROACH
Chapter 2
Literature Rwiew
2. LITERATURE REVIEW
"ln pricing an acquisition, ive c m alwuys agree on the numhers. W%af vulue is
makes it an ideai application area for DEA. This section highlights the literature related
to Company valuation and discusses the issues and limitations of the approaches.
The vaiuation concept can be dated back to the time of Napoleon and even King
Henry the VIII, although. its main purpose at that time was for inheritance taxation
[OVENiI]. As the business world developed. starting with the Industrial Revolution.
new methodologies for valuation became necessary. Present day valuations of IPOs
(Initial Public Offerngs) and srlling and acquirng businesses have taken centre stage in
the valuation world. but there is still a significant need to value tirms for estate. taation.
and divorce settlements. As there are many diffrent applications of vduations. so are
there methods to analyze them.
2.1
Value Definition
Value is the highest price a knowledgeable buyer. at armaslength. is willing to
pay in an open and unrestricted market. Pnce and value are ofien used interchangeably
but value generally means a willingness to sel1 or to buy. while price is the value at w-hich
a deal is carried out. The US. interna1 Revenue Service defines fair market value as
"the pnce at which a property would change hmds betwern a williny buyer and a willing
VALU~NG
PRIVATE
COMPANIES:
A DEA APPROACH
Chapter 2
Literature Review
seller when the tonner is under no compulsion to sell, with both parties having
reasonable knowledge of the facts" FRAT961.
as distinguished from the current market price of the stock. Tt is a subjective value in the
sense that the analyst must apply his or her own individual background, expertise and
s k i h to detemine it, and estimates ot'intrinsic value will Vary corn one analyst to
anothrr l [PRAT961.
2-2
Value Purnoses
The need for business valuation may arise fiom a variety of circumstances and the
diffierent circumstances wilI require diftrent valuations. Valuations of estate and divorce
settlements are quite different than valuations for IPOs. Here is a listing of the varying
needs for a valuation:
IPOs - the company intends to sel1 shares to the public
Mergers and Acquisitions (M&A)
Selling a Business
Estate and Gik Tax Purposes
Re-cripitalizing a Business
' From the Handbovk t'or Financiai Decision Makers. John J. Hampton, Reston VA. Publishing Co., 1979
p543-1.
VALUINGPRIVATE
COMPANIES: A DEA APPROACH
Chapter 2
Literature Review
There are many diffrent methods that can be used to assess the value o t
business. but these methods can be grouped under three main approaches. namely. assetbased. income-based. and comparable 8 m s or market value based. The two main
approaches for valuation are the comparable firm approach and the income-based
approach. with discounted cash tlow being the most popular of the latter method. There
have been articles expouding the benetits of both methods. Whether one method is used
or several. each method has associateci benetits and disadvantages as illustrated in
Approach
Comparable
tims ( CF)
Advantage
Works best when a highIy
comparable group is a\.ailable
Firmer theoretical fooring
DCF
Asset-based
1
(
Disadvantage
There is a chance that an entire sector is
under or over valued
Difiicult to estirnate future cash tlows
and an appropriate discount rate
More relevant when a signi-ifcant portion
of the assets can be liquidated readily.
For small. high-tech IPOs. this method
has littie relevance since it does not
consider gowth opportunities.
[BOOTOO]
The purpose for the valuation may intluence the method appiied. Of the diffrent
value approachrs the method most popular with companies is the DC F method. Booth
examined a survey conducted by Mohan et al on what valuation methods large conipiinies
use and recorded the results shown in Table 2-2 (based on a scale fiom 1-7, where a score
of 7 is the most desirable): [BOOTOO]
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
I
1
Chapter 2
Literature Review
iMETHOD
SCORE 1-7
DCF
6-1
5.7
MarketValueofComparable
Projected Earnings
Revenue hl ultiples
Earnings !Multiples
Book Value
2.4
4-4
4.1
4-0
3 -6
Asset A ~ ~ r o a c h
This rnethod is rnainly employed on companies that are asset intensive. for
example manutacturing and retail businesses. Arnong asset valuations, book value.
replacement cost. and Liquidation can detennine value. One of the most basic approaches
is to calculate the book value. hence the shareholders' equity. One of the problems of this
method is that the assets are based on historical costs. Another method that c m be used
relies on established benchmarks, which are used to estimate the value of a business. For
example. a small investment advisory sells for 3-4 times the current year's revenue.
Factors analyzed in asset valuation are the market value for tixed assets. leasehold
improvements. owners- benefits, goodwill (intellectual property. patents, trademarks,
copyrights. etc.) and inventory. The sum of these factors will give an rstimate of the
market value for the business.
Asset valuation is the least fvoured method and considered by some to be the
most difticult. Sometimes earnings have k e n "stored up tor a number of years ...to the
point at which two valuations are required, one for the operational side of the business
including sufiicient assets tor the maintenance of earnings. and the other for what has
become the investment side of the business" [OVEN59].
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Chapter 2
2.5
Literature Review
discounted back to present value, at a risk-adjusted rate. It is based on the Gordon Price
Model and is mathematically represented as:
Where P is the price, D, is the dividend in year n. and k is the cost of equity or discount
rate. On the other hand. the Gordon Price Model assumes constant growth so the above
equation is adjusted accordingly. In this instance. g represents the growth, k and d are the
same as in the above equation:
Another way of representing DCF is to show that the DCF fonnuIation values
only the assets that give rise to the fiee cash flow:
Where T is the t h e penod, C represents the free cash flow, K is the discount rate. and V,
is the terminal value.
Price Earnings RatioIMultiple (PER)Formulation
Another variation of the Gordon Model called the Price Earnings Ratio/MultipIe
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
10
Chapter 2
Literature Review
2.5.1 DCF PapeRuback @LJBA96]prefers the DCF approach fther than the comparable fimis
method. He investigated the valuation of 5 1 highly leveraged transactions. He showed
that the DCF valuation method produces more reliable estimates of market value. He
Kaplan and Ruback studied highly leveraged buyouts and re-capitalization's using
the DCF approach m L 9 5 ] . They found that a CAPM based DCF approach has
ripprosimately the same vaiuation accuracy as the CF approach. using the capitalization
ot- EBITDA?.
For most companies, tture eamings potential is more imporiant than past
performance, that is why calculating DCF is such an important issue in valuation. The
DCF method is used predominately for companies that are not asset intensive. like the
service sector. In the DCF method, certain variables and unknowns must be caiculated
which include: the amount of tture cash flow. growth potential of the cash tlow,
durarion of the tture cash Bow, risks associated with them. and the discount rate to apply
to arrive at the net present value F[CILL90]. Other factors to analyze inctude risk within
the industry. quality and size of the customer base. length of time the Company has been
EBlTDA - Eamings before interest. taxes, depreciation and amortization. also known as operathg cash
tlow
-
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
11
percentage added to the current 90 day T-bill interest rate- For better-established
businesses the discount rate may be 5- 10% above the T-bill rate, and for riskier
companies about 2 0 3 % . The riskier the business the higher the discount rate. For
determining an appropriate discount rate several methods can be employed: they include
CAPM
regression of the security's return against that on the market portfolio" [BOOT98a]. Et
was Sharpe who introduced the Capital Asset Pricing Mode1 (CAPM). otherwise h o w n
as the market equilibrium model. which c m br used to derive the required rate of retum
for DCF calculations. It states that the required rate is equal to the risk free rate plus
VALUINGPRWATE
COMPANIES:
A DEA APPROACH
12
some sort of premium. It measures the nsk of a security by relating the expected return
on an asset to its beta:
K, = Rf+ (R, - Rf)P
Where. K, is equal to the cost of equity capital. Rr is the risk-tiee rate of retum R, is the
market required rate of return. and P is a measure of the systematic risk of an asset due to
the CO-movementwith the market, If a company has ri beta of one then the stock has the
same volatility as the market index it is being compared to. if beta is 0.5. than the
company is half as volatiie. and if it is two than the company is twice as volatile- In order
to estimate beta, one needs to regess the stock-s price against the level of some broad
market index such as the S&P 500 and the beta factor is just the slope of the regession
line.
2.5.2.2
WACC
Another method to calculate the capitalization rate is to use the weighted average
cost of capital ( WACC); it is represented mathematicaily as:
In the rquation. K, and & ( 1-T) are the costs of aquity and debt capital respectively,
where the cost of debt is on an after tax basis. The weights. E N and D N . are the equity
and debt tinancing proportions. WACC is the weightcd average of the costs of
borrowing and of selling shtlres. The weights are the fractions of totd capital that corne
tiom the diffrent sources. WACC is not reaily a separate method; it combines two or
more capitalization rates to gst one. It assumes that the purchase of a business c m be
partly financed with debt and partly with equity. It weighs the cost of capital to the cost
of indebtedness.
--
VALUING PRIVATE
COMPANIES: A DEA APPROACH
Chapter 2
Literature Review
--
2.5.2.3
For the selection of the capitalization rate. vduators have intbrmation about the
capitalization rates of several companies, and they try to appiy these to the business under
consideration. The parameters they consider in a DCF approach are the time horizon, the
discount rate. and the cash tlow forecasts. Most malysts choose a time horizon of less
than 10 years and 55% use less than 7 yem. For the discount rate. 55% of analysts used
the acquiring tirrn-s WACC. and 4 1% used management judgment, Also. most analysts
apply industry growth trends and -'expecte&' pertormance to cash tlow forecasts
[BOOTOO].
is in future growth opportunities, which are very uncertain, and their employees.
Problems manutcturing plants can have include:
Plant may become obsolete
Manufacturing and distribution outdated
Products sutliering due to cornpetition or lack of sales
Location might n o longer be suitable. closeness to market
Problems with employees
[OVEN59]
Chapter 2
Litenture Review
the most widely used methoci in Canada. He quoted Mr, Justice Douglas of the United
States Supreme Court. who urged caution when applying this method:
-'Since its application requires a prediction as to what will occur in the future,
an estimate, as distinguished from mathematical certitude, is al1 that can be
trm for longer time horizons. The base CAPM ignores issues such as taxation and it
does not consider multiple time periods. It also does not model the real world. by
ignoring transaction costs and assuming al1 risky assets are traded. Selecting the
appropriate discount rate is a key issue in valuation. These and many other reasons make
it difticult to apply the DCF approach in trying to value a Company.
2.6
Each Company tiling an annual report has to adhere to their countrirs' Generally
Accepted Accounting Principles ( G A M L In Canada many of the larger public
Chapter 2
Literature Review
--
Capital structure
Discounted rate differences (risk, time value of money)
For these reasons. when conducting a hl1 valuation it is important to adjust the
(CFO) to get the m e cash generating ability of the firm. The result is what is now known
as Free Cash Flow (FCF). [t is mathematically represented as:
FCF = net income + non-cash items - increase Net Working Capital (NWC) Capex
Non-cash items include depreciation. goodwill. and dekrred income taxes and
afier
ta^
interest. FCF "captures whether or not the tirm actually generated cash tiom its
operations. afcr adjusting the eamings for non-cash items, the need to reinvest in NWC
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
16
Chapter 2
-
Literature Review
2.7
premiums one needs data and the data are only available in public markets. Booth
[BOOT98a] examined the eEect of size and transaction costs on the discount rate. He
looked at the total retum data on size based US portfolios going back from 1926 to 1995.
This information is available from the University of Chicago's Centre for Research into
Security Prices (CRSP). The data showed that smaller tirms constantly outperformed
larger ones. They eamed close to double that of the larger firms 22.62% versus 1 1.4 1%
respectively. Smaller tirms performing better than CAPM predictions are referred to as
the "small tirm efkct." Private companies are usually valutxi (discounted) on lower
PERS than public tirms are since they are considered to be riskier endeavon. Rules of
thurnb approximatr this discount to be around 20% for private companies.
[BOOT98a] suggested two alternatives. the pure play approach (PPA)and the
VALUING PRNATE
COMPANIES:
A DEA APPROACH
17
Chapter 2
Literature Review
instrumental variables approach. The PPA "tinds public companies with sirnilar
operations and use(s) their beta.. .(or) use(s) the broad industry classifications of the TSE
or S&P market indexes" [BOOT98a]. From this he found that -'a hin the general
industrial products area has ti beta of 1.2 or a CAPM risk premium of 4.M.8%. whereas
a pnvate retailer would have a CAPM equity risk premiurn of 2.7-3.l%" [BOOT98a],
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
18
Chapter 2
2.8
Literature Review
2.8.1 P !Ratio
The P/E ratio is the relationship between the current pice of a stock to its
earnings. It shows the shares selling at so many times their actual or anticipated annual
earnings. The P/E is calcultitcd by dividing the current market pnce of comrnon stock by
the earnings per share. lnvestors use it to ascertain the market demand for stocks.
Companies with a high P/E have greater eaniings potentiat than low P/E stocks, but also
are more volatile and somewhat riskier investments.
PIE ratios enable one s hare to be compared with another. they also retlect the
views of thousands of investors on the quaiity of an issue. They indicate liquidity
earnings trend, profitability. and dividend payout dong with intangible hctors such as
quality of management. and prospects for the industry or segment [CSI89]. To rate the
P/E for a company's common shares with those of others. the companies must be
comparable which usuall y means they must be in the same industry. PIE ratios for
difirent industries are available in various media and industry sources. for example the
VALUING
PRIVATE
COMPANIES: A DEA APPROACH
19
C hapter 2
Literature Review
The P/E ratio is probably the most usefui and widely used of ail financiai ratios
because it is. in k t , ail the other ratios combined into one figure since it represents the
ultimate evaiuation of a company and its shares by the investing public [CS189].
2.8.2 PIE Ratio and Comparable Finns
iMany studies examine the use of the P/E ratio for comparable ttirms. Boatsman
and Baskin [BOAT8 11 compared the accuracy of two diffrent types of P/Ernodels. The
tirst mode1 tinds a random company in the sarne industry. The second mode1 uses a
company h m the same industry but with similar ten-year average growth rate for
earnings. They found the accuracy of the second mode1 to be better.
tirm size and earnings growth- He found that the 4-digit SIC code is the most effective in
choosing comparable t h s . It is also interesting to note that Alford did not find
improved accuracy when size and industry membership were used togethcr in the PE
valut ion [KIM99].
He did this by examining working papers and timess opinions on management buyouts.
He showed that undenvriters base their pricing decisions on an analysis of the market
price ratios. with adjustments for fimi-specific ditrerenes. and determine a range of
minimum and maximum ottring price, then, after gthering more information they set
the market price.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
20
Chapter 2
Literature Review
Ratios other than P/E can be used in conjunction with the CF approach. One can
have any sort of ratio. even those that are specific to an industry, for example. market
value divided by cable subscriber. Amir and Lev. cited fiom [KIM99], provided a
valuation study of the wireless communication industry. They showed that the primary
determinant ofcompany value is the population of the fianchise temtory. rather than
financiai variables.
The value relevance of P/Eand market to book (ME3) ratios has been
demonstrated by empiricai studies. Kun quotes two articles. one by Zerowin and the
other by Liu and Ziebert [iUM99], who examined cross-sectional variability in P/E and
found a significant relationship between P/E and growth, dividend payment and size.
According to Campbell "trading statistics with respect to shares of so-called
comparable public companies cm be. and moreover should be, a sipiticant factor to
consider" [CAMP91]. He stated that it is important to use CF where the privately held
business under valuation is similar in size. business structure and business operations to
one or more of the publicly traded companies.
In the US.. the Interna1 Revenue Service (IRS) considers the analysis of public
companies an important tool in determining the value of private companies. although this
is mainly tor income-tau purposes. Ruling 59-60 urges "using cornparisons fiom
quotations h u n d in the prices at whih similar companies are selling in tiec and open
markets. including not only listed securities but also actively traded over-the-counter and
unlisted securities" [OVEN59].
Moms [MORRV] reatirrned the tct that the best multiple may corne from a
public market comparable. If it exists it should be given full and cornplete considention.
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
21
Chapter 2
Literature Review
He points out "it is diffTcult to depart from price. or multiple, set in the public
marketplace. These public companies are one's best empuical data" F[ORR91].
WB. and price to sales, and may need further adjubtments. The best way to employ PIE
ratios is through the use oF forecasted eamings rather than trailing eamings. Another
ratio that can be used in valuations is enterprise value to sales and enterprise value to
operating cash tlow. Enterprise value is equal to the market value of equity plus the book
value ofdebt minus cash. "Ratios seem more usetl especially when adjusted to reflect
dit'trences between pro titability and growth rates of the [PO in question and the
comparable firms use&' [KIM99]. However. it is dificult to forecast the future cash flow
for very young companies, which is the category most IPOs faIl into. That is one of the
main reasons why the DCF method is imprecise, and the key advantage of the
comparable tirms method. The CF method is supported by academics and practitioners
and widely used in case studies in business schools [KIM99].
Kim and Ritter [KIM991 use sales. eamings. operating cash tlow and book value
to determine value. as well as comparable firm multiples. They tested the accuracy of the
comparable firms method by using two sets ~Ccomparablefirms. The tirst set consisted
of recent IPOs in the same industry and the second set was chosen by a research
--
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
22
Chapter 2
Literature Review
-
- -
boutique3. The results showed that the second set worked better than the fim, and
applying forecasted earnings worked better than using historical eamings. They also
tested the hypothesis that young t i m s are harder to value by splitting the data into 'age'
goups and applying the second set. They found that the -taluation errors ofthe
comparable firm multiples are noticeably smaller for the older t i m s than for the younger
rms. especiaily when using emings" wM99].
The recent IPOs set consists of companies that went public not more than 12
months prior to the IPO under investigation's ot'fer date and that has the same 4-digit SIC
code. if there were more than 5 comparable tirms to compare with they suggested using
the 5 tirms with the closest sales and ignored the rest (firm = recent [POs). Earnings for
the tirst year after going public typically include substantial amounts of interest incorne
thus in the analysis EPS is used and sales numbers tiom prospectuses only, rather than
more recent financial statements that may be skewed by interest income.
The variables thry sekcted were offer date (cannot be more than 12 months
before the offer date of the new IPO), EPS, Sales (13 month revenues prior to IPO as
stated in prospectus). and book value per share (BPS) (both pre and post IPO). Another
condition placed on the variables is that they must be positive. They then computed the
following ratios: PIE. P/S, and W B . The median of these market multiples is then used
as the CF multiple for the Company. The ompany's ratios are computed based on its
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
23
Chapter 2
Literature Review
retum in the sample is 12%). one would cxpect the [POmultiples to be discounted by
about 12% wM993. The fimctional fixation hypothesis asserts that the market
mechanically capitalizes reported EPS numbers. without adjusting for the quality and or
persistence of the eamhgs. Friedlan et al IIUM991 suggested that the market does not
h l l y incorporate the inibnnation content ot'discretionary accruals in valuing new issues,
consistent with the fiinctioniil fixation hypothesis,
The tirst mode1 of the CF approach is surprisingly weak for the tollowing reasons.
First. because historical eamings of IPOs are transitory in nature they have little value
relevance. Second, using market multiples and past data have an intnnsic limitation in
that they do not retlect future growth. This leads to the third issue, which is allowing for
adjustment of diEerences in growth and protitability. Finally, there is an element where
the comparable firms may have been chosen improperly w M 9 9 J .
Kim and Ritter used --the geometri mean of the Renaissance Capital comparable
firm PIE multiples as the explanatory variable in regressions using the three P/E ratios of
the IPOs as the dependent variables" [KIM99]. The three P/Eratios. using the midpoint
of the POP (preliminary oEer price) for the IPO. are the last 13 months. current fiscal
year's forecast. and next year's forecast of the EPS. Comparing the regressions, the
Chapter 2
Literature Review
average prediction error fdls from 55% to 43-7% to 28.5% respectively, and the
percentage of rms that are valued within 15% of the actual multiple increases. Younger
firm valuation using forecasted eamings works beer than using historical earnings. as
there is a potential ~ o w t opportunity
h
that needs to be addressed,
They also tcsted the presumption that older t-urns will be easier to value since
capitalized earnings represent more of their value than expectations about ture growth
rates. Consistent with that presumption. the mean absolute prediction error t'or the twms
that are less than 10 years old at the time they go public is higher than tor older fimis.
3 1.9% and 23% respectively.
2.8.3.1
LM/BRatios
Renaissance Capital does not utilize iM/B ratios- They find 1WB ratios to be poor
valuation metrics. They daim that M/B ratios have a disadvantage beause the book
value is arbitra-; for IPOs since there is a large change fkom before the issue to aier
(intlux of new capital tiom new shareholden). Rather, they recommend using enterprise
value (EV) where they look at market value to sales. EV to sales and EV to operating
cash tlow (EBITDA) in their andysis-
To sumniarize Kim and Ritter [KIM99], the variation in these ntios is JO large,
both For public tirms and IPOs. that they have only modest predictive value. Many
idiosyncratic factors are not caphired by industry multiples unlrss various adjustments for
diffrences in grow~hand protitability are made. In practice. analysts place more weight
on a given multiple for some industries than others. Taking this into account would
probably show that CF multiples result in more accurate valuations. In practicr. one
starts with industry multiples and adds or subtracts adjustments of 10-10% to retlect
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
25
Chapter 2
Literature Review
ditierences in growth rates. protitability. and quaiity of eaniings. Investment bankers use
the midpoint of the oflr pnce range resulting in smaller prediction errors than using
comparabIe tirms. They canvas the market place tor demand before setting the final
issue price N M 9 9 1 .
2.8.3.2
One issue with this study is the use of a maximum of 5 comparable t i m s and the
deletion of the rest. The comparable tirms approach works best with a hi& number of
comparable Lims and exclusion of tirms to simpliQ a study may not give meaningfl
results. tt is better to have as many comparable companies as possible.
2.8.4 Issues wit the
CF Approach
T h e e main issues with this ripproach are the P/Eratios employed, the use of the
right multiplier for the industry. and the availability of a highly comparable group of
companies. PIE ratios are dynamic and Vary widely over time and among companies.
The problems with PIE ratios and the right multiplier c m be remedied by looking at
industry trends and the broader market to forecast trends and adjust historicai P/E ratios
accordingly. Companies can be public and not usell for these comparisons because:
--
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
26
--
Chapter 2
2.9
Literature Review
have to be made to account for the differences that exist between public and private
companies. One of the issues to keep in mind is that private companies are not as well
established. and most of their value lies in their growth opportunities. Exceptions to this
include Goldman Sachs and Eritons, which were both very large and well established
betore their respective IPOs, Private companies have kss than 50 shareholders and their
shares are hiyhly concentrated, leading to liquidity risk oncems. Private companies are
under no obligation to report their earnings or to produce annual reports and hence their
value is often times subjectiveOn the other hand, information on public companies is readily available. through
t-mancial statements? annual reports. and such. Public markets rire organized places of
exchange- their p k e s are inkluenced by coverage in thc press and by analysts, and there
is more intorrnation avaiiable about them. When compared to private firms. they are:
They usually have better developed products. management trams. and strategirs
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
27
Chapter 2
Literature Review
VALUING
PRIVATE
COMPANIES:
A DEA APPROACH
Chapter 2
Literature Review
Products/services
Corporate profile
Customer pro tile
Or~anizationaichart
Opportuni ties
Market analysis
Cornpetition
Supplier pro tile
Ernrrlovee base
15% of the price. must be discounted in arriving at the ir market value of the business.
Finally. income-ta.. discount considers the impact on share valuation of axes and takes
into account the recapture of tau depreciation, capital gains. and tax on proceeds of
intangibles.
Redundant assets are assets owned by the Company that are not required in its day to day operations,
examples include propertirs, investment in subsidiaries that rire not of the core business, and investment
porttolios. [HEND88]
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
29
Chapter 2
Literature Review
Weaver looked at discounts caused by lack of control and lack of liquidity and
marketability F A V 9 8 1 . Perception dictates that risk is less if the investor has the ri@
to control the business- An adjustment to the pnce of a business has to be made for lack
ofcontrol. He quoted many studies that try to determine an appropriate amount for this
rate. The important issue here is that a rate actually sxists but quoting values fiom other
studies is futile; each case is unique and must be exmined individually to arrive at the
discount rate. Studies included Lyons and Wilczynski KYON891 who found the range to
be 27-28%. This amount was tound fiom premiums paid in M&A tiom 1968- 1987.
Pratt IPRAT931 studied takeovers of companies and the premium paid by the company
taking over. The premium is converted to a percentage that acts as an estimate for lack of
control. and is directly abstracted tiom actual market transactions. The discount level
was found to be 26-33%. Weaver noted that it is "interesting to see this level of discount
(that) exists even for SEC regulated companies where minority shareholders have more
WAV981.
protection than with small partnership~'~
Lack of marketabiky and liquidity is concerned with the lack of a real market for
the shares and the fact that investors pretr liquidity. Again. quoting many studies. one
can conclude that this rate exists b u t it is specific for each situation and there is no easy
formula to apply. ChaTe [CHAF93] used Options Pricing Theory to tind this amount.
He applied a European option. which results in a Iower price than using Amcrican
options. and showed a minimum applicable discount. AppIying the Black-Scholes Put
Formula the amount is 28-4 1% when the marketability is constrained for 2 years. It is
32-49% using 4 years. Srnriller trms with weaker earnings trade at much higher
discounts than larger tinns- Management Planning Inc. a valuation company, prescnbed
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
30
--
Chapter 2
--
Literature Review
live factors that are predictive o f restricted stock discounts. The more stablehigher these
variables are the lower the market discount @4ERC97]:
Revenue
Recent earnings
Market pnce per share
Price stability
Eamings stability
2.10
Companies generally issue an IPO to raise money and to increase their f i m i a i base and
ability to make acquisitions. Most IPO literature recommends that to pnce an IPO one
should initiaMy look at and compare its operational and tinanciai performance with that of
several public companies in the same or a similar industry. In order to compare private
manutacturing companies one shouid look at the industry as a whole. and then at the
specitic 4-digit SIC Code. For companies going public. the factors that id--uencepnce
include the ownership retention in the business. and the risks associated with the POs.
Al1 this information should be applied back into and included in a complete valuation
procedure.
IPOs and their average first day retum and the s i g n a h g tctors associated with them.
Signaling factors are issues such as percentage retention of stock by the original
owner(s). A number of studies tind that IPOs espenence large increases in value dunng
tiieir tira day of trading. Ibbotson et al PBB0941 tound an average retum of 16%, while
others have tound 1.9 2% [LECr195]. Most found that IPOs underperhrm by the end of
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
31
value o f a h n ' s [PO is detemined by its ownership structure. If larger investors buy
stakes in the company the offering price and revenues raised are higher- Firth also tound
a positive ownership retention signal [FIRT97]. Jan [JAiN99]. on the other hand.
company as this can inf'2uence an increase in monitoring. Warrant holders are tied to the
long-term pertrmanee of the company, and will want to insure that the company is doing
well,
12% over the tirst three years of trading when compared with the market (exclusive o f
the initial underpricing retum).
VALUING
PRIVATE
COMPANIES:
A DEA APPROACH
32
Chapter 2
-
Literature Review
-- --
-- -
in an issue following an [POI both spread and commission revenues could be enhanced
by setting a lower offer price. It is at least conceivable that the increase in revenues due
to a lower ofir price would be sufficient in magnitude to outweigh the additional
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
33
Chapter 2
Liaerature Review
--
return on assets declines tiom the year previous to going public to the end of the first year
of public trading, but thereat'ter. performance does not decline in ten years. However.
they found that oprnting pertbrmance both within one year of the offerng and during the
valuations [RITTW].
2.1 1
assets such as brand names, trademarks, and patents [MULL90]. Intangible assets are
valued j ust like regular assets: find the net present value of the future cash tlows for the
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
34
Chapter 2
Litemture Review
-
- -
[OVEN59].
Although the market-based approach tinds comparable
it is not a
cornprehensive search. the DEA-Valuat ion method will fil1 this gap.
2.12
S c o ~ and
e Limitations of this Studv
Al1 these factors should be taken into account when conducting a private company
vaiuation. However. al1 this information is not readily available and only applies to a
complete valuation study. As well, the earnings outlook and financial conditions of the
pnvate company are not known. other than what can be interpreted from the financials.
This fr exceeds the scope and premises of this study. This study is intended to be a
guide for valuaton and should be considered as another tool they have at their disposal.
Use of the DEA-CF method is to determine the best and most comparable public
companies. Calculation of the market value is attempted by considering the lambda
values to be a percentage, and multiplying this by the known market values in order to
tind unknown market values and also by infemng a market value range from the
VALU~NG
PRIVATECOMPANIES:
A DEA APPROACH
35
Chapter 2
. - -
Literature Review
groupings. The methodology is discussed in detail in the anaiysis and discussion section.
Another application of this approach could be to first, fmd the similar companies then,
instead of trying to calculate the market value, the known multiples/earnings of the public
companies can br applied to the pnvate companies. Thus one c m project the earnings of
the private companies using the multiples from similar public companies- Again this is
information beyond the scope of this study but it is an issue to consider in a fture
research project. The valuzition approaches e x h e d and the andysis conducted herein
are intended mainly ror companies that are being acquired or sold and for P O s -
Valuation of prokssional practices, dentist oEices, travel agencies and such, is not
included in this analysis as more ofien than not, rules of thumb procedure is applied to
these businesses.
VALUMG PRIVATECOMPANIES:
A DEA APPROACH
Chapter 3
--
--
DEA -'compareso' the int'ormation rivailable from a group of DMUs with al1 others in the
group optimizing each individual DMU with the objective of calculating a discrete.
piece-wise linear, &%estpractice" frontier.
The origin of DEA is tiom a Ph-D. dissertation done in 1978 by Rhodes
[CHAR781 supervised by Cooper and Charnes. In this dissertation, Rhodes exarnined the
efficiency of educational institutions in the New York City district with regards to their
abiIity to improve the developrnent of disadvantaged children. The resulting model.
named CCR after the three authors, allows calculation of relative efficiencies on the units
under investigation based on a constant-retums-to-scale mode[. The CCR model of DEA
was an extension of Farrell's single outputlinput technicd efficiency rneasure [FARR57].
Following the CCR model was the introduction of the BCC model in 1984. by Banker.
C h m e s . and Cooper [BANK84]. The BCC model contains an added constraint of
convexity. which allows DMUs to be measured on a variable-retums-to-scale basis.
In this way, DEA's original design was to measure the relative eficieny within
not-for-protit organizations. Through the yeae. DEA has been expanded to sni- the
eficiency of over 50 industries inciuding banking, tst food retail. healthcare. insurance.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
37
Chapter 3
~reditunions. highway road maintenance. and capital budgeting projects. Over 1.300
papers have been written by academics on a wide variety of theoretical and pncticd
advances in DEA. DEA can be considered to be a combination of econornics, statistics.
engineering, operations research. and management techniques.
3.1
METHODOLOGY
DEA is a t o m of frontier analysis that recognizes that some DMUs pertiirrn
below optimum 1eveIs. In this method, each DMU is vptimized against al1 othrr DMUs.
That is. the linear pro_mam assigns weights to the variables so that each DMU looks the
best it can. DEA constnicts the eficiency tiontier and alculates the efficiency score for
each DMU based on its distance from the fiontier. The score indicates the --proportion by
which the unit c m increase its outputs without consuming any more inputs, or conversely
the proportion by which it c m decrease its inputs and maintain its current output lever
[S 1MA001The graphical representation, refer to Figure 3-1. of the efficiency scores
identities the efficient as well as the inefficient DMUs in the observed population. The
efticiency iiontier. or envelopment surace, consists of the best practice DMUs.
Connecting the efticient units forms the frontier and the ineficient units are enveloped
within. DEA links the estimation of the technical et'ticiency and the production Frontier.
Somewhere above the efticiency fiontier is the theoreticai tiontier. This tiontier is
beyond the reach of human pertormance. as there is always a chance for perhrmance
improvement when human labour is nvolved. The DMUs that are below the best
practice tiontier art: considered to be inefficient. For each ineflicient DMU, the DEA
anaIysis identifies the sources and the levels of inefflciency for both the inputs and
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
38
--
Theoretical Frontier
est-Pradice Frontier
INPUT
This allows the DEA mcthodology to maximize the relative etficiency score of each
DMU, given that a set of assigned weights for a particular DMU is also feasible for al1
the other DMUs included in the analysis.
VALUING PRIVATECOMPANIES:
A DEA APPROACH
Chapter 3
3.2
- -
- -
environmental tctors or exogenous factors that may intluence the formulation of the
model. For any of the non-quantifiable or qualitative factors, certain proxies may be
assigned and then be included in the analysis.
The generai desire by management when conducting a DEA analysis is to
ma~imizetheir outputs while minimizing inputs used in producing the outputs. This is a
particularly important issue dunng the formulation of the mode1 and the identification of
the factors.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
40
3.3
Mathematical Formulation
The Charnes et ai. [CHAR941 model for DEA was formulated to maximue the
efllciency of each DMU subject to the constraint that the efficiency of al1 DMUs be less
C Ur-,
2 1
C
subject to ,-
Y b 1,
pc.Vi
where
2E
The two most prominent models of the DEA methodology are the CCR and BCC
ratio models. The CCR model is also referred to as constant returns to scaLe (CRS), and
the BCC model is used interchangeably with the name variable reninis to scale (VRS).
Both of these models address managerial and economic issues and provide usetl results,
but the orientations are somewhat dieerent. Within both the CCR and BCC model a
decision can be made as to whether an input or output orientation is required for the
analysis. This iniormation is usually derived from the objectives given by the
management or the executives of a panicular organization. The brmulation of the model
can be oriented towards output maximization or input minimization. If a par<icular
output level is demanded. o r more specitically. i f a corporation has a production standard
for these DMLTs.an ineficient unit can be made efficient through the reduction of its
input levels while holding its output levels constant; this is referred to as an inputVALUINGPRIVATE
COUPANIES:
A DEA APPROACH
41
...
--
Chapter 3
oriented model. I f the corporation has established amounts of input resources available to
these DMUs. the inefficient units can be made eficient through increases in its outputs,
while holding input resources constant: this is the output-orented model.
Figure 3-2 represents how ineficient units are projected ont0 the frontier
depending on whether the analysis requires input rninimization or output maximization.
The peers for the inefficient company C are company A and company B for input
minimization and company B and company D for output maximization.
Figure 3-2 Projecting the inefficieat Unit to the Frontier
-=
3Y
=
6
C,
y c
thit
.----------I
I
I
I
.1
I
I
Efficient DMUs: A, B, D
Peer Grout, for DM&
Input minkization: (A.8)
Output rnouimization: (B.D)
E ~ c i e n c vof DblU,
Input minimization = X,,,JX,
Output maximization = Ywg,JYc
Inputs
The CCR ratio model focuses on the identitication of the overall efiiciency of the
DMUs studied. it also identifies the sources and estimates the potential improvements
available for the inetticient DMUs. It is important to note that the CCR model provides
the envelopment surfafe that represents CRS. This assumption illustrates that any
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
42
Chapter 3
3.3.1.1
- E- 1
-*
Min.
z, = 0 - E. 1 S-
Yh - s-
s-t,
S-
= Y,,
Mms.t-
BX,-IU-s-=o
w,, =
dy,,
vTx,= 1
$Y-vrxs0
3.3.1.2
The two approaches o f CCR lead to the same envelopment sufice. the only
s-t.
$Y,, - Y A + sXh+s-=Xo
.
s.t.
=O
-pTY,, = 1
opTy+ v-'x 2 O
s-. s- l O
present for further exploitation. This mode1 interprets the etftciency of the DMUs with
the underlying assumption that a VRS mechanism exists among the population of
observations. It incorporates the notion that an increase in the inputs does not necessary
translate to the same proportional increase in outputs throughout al1 scales of operation.
-
--
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Chapter 3
This is different tiom the CCR model. which assumes the input and output relationships
of al1 DMUs e'xhibit a proportionai or a linear relationship for al1 scales of DMUs. The
BCC model allows the DMUs with varying scale sizes to be compared in the same DEA
analysis. This is a result of the BCC mode1 being technically efficient while CCR is scale
and technically efficient. Technical efficiencies are efticiency measures that ignore the
impact of'scale-size by comparing a DMU only to other units of sirnilar scale. On the
other hand. scale eEciency depends on the size of operations and makes this optimal.
A n y moditkation to the size will cause the unit to be less efficient.
3.3.2.1
The input oriented BCC model focuses on maximizing the DMUs peribrmance by
reducing the inputs 05 in other words. produces the observed outputs with a minimum
resource level. The linex proprnrning mode1 for the BCC model is:
Input-Oriented BCC Prima1
s.t.
Y h - s - = Y,,
ex,-Xh-s-=O
s-t.
v T x ,= i
JY - vTx +
-iC
(Constraint of Convexity) 1 h 2 1
, sC. s- 2 O
15 O
-pr 5 -c- 1
4
- v'r s -&. 1
Where s-and s-are the input slacks and output slacks respectively; p and v are the
multipliers: p,, is the measure of scale et'ticirncy. If pl, is negative then it indicates
increasing retunis to scale (IRS). If it is positive then it is operating at decreasing retums
to scale (DRS),and if it is zero then it is operating CRS. Theta. 8, is the proportional
reduction which has to be applied to al1 inputs of the DMU beinp evaluated to improve its
efliciency. The E allows optimization to occur in two steps. First. the inputs are
-
--
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Chapter 3
minimized by the optimal 8.Then the optimization involves a movement to the tiontier
by the input and output slacks. Lambda, ktindicates the level o f contribution o f efficient
firms to the virtual DMU created by the projection of an ineficient DMU ont0 the
fiontirr. Those units that contribute to the construction of the imaginary composite unit
will have non-zero dual weights. AL. and make up the reference set for unit DMUo (the
DMU under consideration). The refrence set for DMUo is thus solely comprised o f
efiicicnt iinits and serves as a basis for computing the eficiency score of unit DMUO.
The magnitude of k indicates the "degree to which the characteristics of the efficient
DMUs art:used to constnict the virtual DMU on the tiontier to which the inefiicient unit
is projected" [SIMAOO].
3.3.2.2
q , = V'X,
- v',
-p-r Y', = L
s.t.
-.
v' 2 s 1
v,, tiee
VALUING
PRIVATE
COMPANIES:
A DEA APPROACH
dS
Chapter 3
3.4
fiontier does not have to start tiom the origin while the CCR does. In the CCR model.
there are Iess efficient DMUs. the values o f 6 are generally smaller, and the values of $
are generally larger. To convert the BCC input oriented mode1 to the CCR input orientcd
4
model the constraint of convexity, 1h 2 1. is removed from the prima1 model, dong with
+
the p4,fiom the dual model. in a similar manner, 1h = 1 and v,, are removed kom the
BCC output model to arrive at the CCR output model. Figure 3-3 depicts the diffrences
between the models.
Figure 3-3 Graph of VRS and CRS
Output
Input
3.5
Inter~retationof DEA
Using these formulations, DEA identifies the potential improvement for the
inefficient DMUs. The intormation derived ffom the analysis is intended as a guide for
policy making and managerial decision making in an organization. It provides a better
-
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Chapter 3
3.6
Are value fiee and do not require specification or knowledge of a priori weights or
Produce specific estimates for desired changes in inputs andor outputs for projecting
fiontiers
12. Satistj. strict equity criteria in the relative evaluation of each DMU
[CHAR94J
Disadvantaees
Major issues with DEA are that there are no possible improvements o r targets for
Chapter 3
3.7
DMUs. the environment, the weights, and the Merence between the data First, DEA
works best when there is at Ieast three times as many DMUs as the sum of the number
inputs and outputs. if this requirement is not met, the majority of DMUs will be on the
efficient tiontier and the assessrnent will not convey proper results. Second- it is essential
that the DMUs evaluated corne from a sirnilar environment. it is not appropriate to
compare manufacturing and service industries against each other since they both have
very diffrent objectives and management styles. Some frther insights can be gained
about DEA ikom the weights used. DEA assumes equally proportional improvements of
al1 inputs or al1 outputs. This assumption becomes invalid when a preference structure
over the improvement of inputs (outputs) is present in evaluating inefficient DMUs
[THAN92]. This leads to the third issue. the unrestricted weights may mean some of the
inputs or outputs, especially if the DMU is doing poorly in that dimension may be
assigned a weight of zero, wherc in reality al1 of the variables contribute in some way to
the overall efticiency. This stems fiom the fact that each unit is assigned weights so that
it looks the best it can. To address this problem, the weights can be restrcted using
analyst and management input. Bounds c m be placed on the importance of each of the
factors so that the weights assiped are more realistic- Finally, if some of the data points
are extremely large when compared to the rest of the variables, these DMUs have
potentials to be outliers and may be anomalies in the study.
VALUINGPRXVATE
COMPANIES:
A DEA APPROACH
48
Chapter 3
3.8
been discussed, but so fm a discussion of the usetlness of DEA and valuations tberetkom
has not been covered, The main motivation for this research cornes tiom a Ph-D.
dissertation by Sunak [SIMA001 on this subject. In this work. Simak examined DEA and
private company valuation and proposed a DEA model to identif'y similx public
companies to the one under scrutiny. The resulting peer groups fiom the DEA analysis
identifL the sirnilar companies. which when combined with their efticiency scores, the
"peer" cornpanies can be tound. The model uses data on 5 1 public manufacturing
companies, The peer group for each inefflcient Company is identitied and other
companies that have the same peer groups are collected. A difference indicator variable,
'
dij=
k
The lower the dij value. the more similar the companies are. He also examined the
closeness of the efficiency scores. By looking at the market capitalization (MC) of these
public companies and testing if the DEA analysis found peer companies with similar
MCs. the factors c m be combined to reved the best possible peer(s) for the subjectcompany. The results were encouraging provided adequately sirnilar public cornpanies
were availabie for reference.
DEA coupled with valuation will allow multiple inputs and outputs to be used in
the same model. The DEA comparable tirrns model is an extension of the market
approach in order to ind comparable cornpanies. DEA is ot'ten used to h d the most
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
49
Chapter 3
efficient unis but now it can be used to nd proper peer groups. This paper will extend
the approach started by Simak [SIMAOO], but d a e r s in several respects. The ttors
used for valuation are different, Also, two new methods are introduced. one to value
efficient companies and one to detemine upper bounds for the ineficient companies. No
estimation of the value of efficient companies has ever been attempted betore.
3.9
each member nation in order to determine relative size of their net-burden. This is a
novel application of D E A which examines a net-burden index for each member nation
(the ratio between contributions to an alliance and benefits received therefiorn). instead
of calculating relative cfficiencies. In this work, two different models are evaluated and
three ofthe nations appear on the frontier in both models. hese nations are givrn a rank
of one. Subsequently. each mode1 is examined individually and the remaininy nations are
ranked according to descending net burden starting at four. Next, a cluster analysis is
applied to both the models to determine the relative position of the DMCs on the t'rontier.
ALso. the et-ticiency score fiom both the models were =paphed against e x h other to get a
simulation of the fiontier. The results from the cluster analysis support the tiontier found
%om the plot of the et'ticiency scores.
Zhu [ZHU001 conducted one of the best studies on DEA and Company efticiency.
In this work Zhu examines tinancial petiormance of Fortune 500 companies using DEA.
The data is From the April 1996 issue of Fortune Magazine. The investigation found that
the top-ranked companies by revenue do not necessarily perform etriciently. and that
only 3% of these companies were on the best-practice fiontier.
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
50
Chapter 3
-
single DEA model for performance satisfactory Instead a diffrent combination of these
tctors was sought. Zhu tested three models in ordrr to obtain an overall perfbrmance
index. nie tint model. called protltability. had three inputs (employees. assets. and
stockholders' equity) and two outputs (revenues and profits). The second rnodel. named
marketability, had two inputs and three outputs. (namely the outputs fiom the tirst rnodel
were considered as the inputs into this one). and the new outputs were market value, EPS
and total return to investors. The third rnodel combined the tirst two models, it used the
inputs tiom the tirst model and the outputs tiom the second model. and thus it had three
inputs (employees. assets, and stockholders' equity) and three outputs (market value.
earnings per share and total retuni to investors). This third model was meant to represent
the overall performance/efficiency- The models are shown in Figure 3-4.
Figure 3-1 Zhu's DEA Mode1 for Company Performance
Stage 3
Em pioyees
Market Value
Profitabil ity
Shrireholders'
Equity
Eamings Per
Share
4
Stage 1
Stage 1
- -
VALUINGPRIVATE
COMPANIES:A DEA APPROACH
51
Chapter 3
Zhu showed, with this DEA analysis. that the larger companies exhibit decreasing
rehrrns to scale (DRS). Testing the data for errors using a sensitivity analysis proposed
by Seiford and Zhu reveals that the best-practice tiontier is reliable [ZHUOO]. An input-
onented CRS model was employed in which the DEA rnethodology was modified to
show performance measures that better retlect the way businesses nnk against each
other. Results showed that substantial input savings could be achieved as evidenced by
the distribution of the CRS scores. Most industries perbrmed better with model-1
(prot'itability) than
with mode1 2 (marketability), and the model 2 scores were better than
In addition to the CRS rnodel Zhu dso investigated the VRS mode1 in order to
examine returns to scale. The VRS version of model 1 revealed an eficiency of 0.46 as
opposed to 0.4 for CRS. In model 2 it was 0.29 with CRS and 0 5 4 with VRS and in
rnodel 3 it was 0.25 with CRS and 0.3 1 with VRS. The t-test is applied to the CRSineftcient companies, "the results show that the mean of the paired difierences betwern
CRS and VRS scores for the three-mode1 evaluations are si~mificantlygreater than zero
(the values are equal to 13.5.2 1.5 and 14.3 for model 1, model 2 and model 3,
respectively. at ievei 0.0 1 )" [ZHWO]. The t-test indicated that the ratios were
signifcantly less than 1. Therefore serious scale inefficiencies were present in the
Fortune 500 cornpanies. 4 summary of the above results is displayed in Table 3-1 Table 3-1 Summary results for Zhu analysis
-
VRS
CRS
t-test (0.01)
iModel 1
0.46
0.4
13.5
Mode12
0.54
0-29
21.5
Mode13
0.3 1
0.25
14.3
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
52
Chapter 3
Upon further investigation into scale eficiency and returns-to-scale. Zhu found
that the top-70 companies were not oniy scale inefficient, but they also operated in a DRS
region. while most relatively smail companies operated in an IRS region.
For the pertormance of inefcient companies, "one may use the average factorspeci tic scores within each industry to characterize the tactor-specific industry eficiency.
However. dit'trent companies with difierent sizes may exist in each industry. Therefore
arthrnetic averages may not necessarily be a good way to characterize the industry's
efticiency. usually. one expects bigger input and output levels, tiom relatively big
companies-' [ZHUOO]To summarize, DRS are found mainly among the top ranked companies by
revenue. Sensitivity analysis shows which production factor will not affect an efficient
company's financial pertormance given that ail other production tactors are kept at their
current levels. This information is usetiil when a Company considers its tiiture policy
against other companies.
VALUING PRIVATECOMPANIES:
A
DEA APPROACH
--
4.
selection of relevant inputs and outputs. Afier the appropriate variables are ascextained?
the DEA valuation mode1 is constnicted. The variables and the model are based on
available literature. expert opinion and known valuation criteria Possible new DEA
applications to business are presented as well.
4.1
Mode1 Criteria
Due to insuftiient arnount of knowledge about the geographic locations, social
contexts and structures and other qualitative inIomation, this study of the tiontier
analysis tocuses more narrowly on the available financial information. A study of just
the tinancials leads to a much more uniform study and compensates for the dit'ferent
management policies and procedures that may exkt among the cornpanies.
4.2
problem is the specification of the model's inputs and outputs. The three sources of
information that will be used for this purpose are available literature, expert opinion and
known vaiuation criteria,
Articles aimed at identitjring the most important indicaton of value include Zhu.
[KIM991 makes use of the: prelirninary offer price (POP). the ot'fer price (OP). first
VALUING PRIVATECOMPANIES:
A DEA APPROACH
--
market price. proceeds. EPS, preissue book value per share (BPSp,i,uc).
sales. P/E (OPEPS). M/B,, (OP/ BPSp,,).
postissue book
ME!,,
P(IM991. Most of these variables are specitic to IPOs and were not available to this
study. Many other studies quoted earlier in the literature search section examine similar
financial ratios such as P/E and W B . Investment bankers utilize accounting intortnation.
comparable trms and also market dem@
key accounting variables as detennined by them are sales. earnings. operating cash flow
and book value [KiM99].
Another source for idornation regarding value was tiom communications with
professors that teach DCF and M&A courses. Free cash tlow (FCF), 7- year revenue and
profits. competing companies were some issues identified as deterninants of value.
There were also some qualitative issues to consider such as the quality of manqemrnt.
strength of market position. quality of products and services. level of customer
satisfaction. strength of corporate culture. quality of investor communications. and
effectiveness of new product development PARB971. The collection and quantification
of these types of value indicators will need to be evaluated if they are to be included in
the DEA modefs.
4.3
c m evaluate the service efficiency of a company and as an added feature the conversion
ability of the company. or One can examine issues related to predicting market value or
rkk. Each of these perspectives require different variables. The choice of inputs and
----
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
55
outputs has to retlect the objectives of the business. In the case ofvaluation, one wants to
examine the variables and factors that optimize the ability of the company to turn its
assets into profits.
Borrowing fiom studies done on DEA and banking eficiency. one c m apply the
sarne models to cornpanies with some adjustments. If a bank is rneasured for service
efticiency. i-e. the Production Modet. one looks at variables such as labour. oftce space.
loans and deposits. Thus. the cost to produce services is considered as an input and the
and loans. Thus al1 types of costs should be treated as inputs and al1 types of revenues as
outputs.
Tulkens and vanden Eeckaut [BERG981 list five approaches to efficient bank
management. The approaches are:
1. Profit maxirnization
2. Service provision
3. Intermediation
1. Utility provision
5. Risk management
One c m adjust the above approaches to retlect the concerns of companies instead
of banks. When detemining the value of a Company, the kry variables are not objects
such as number of employees. number of officeslstores. square footage of ofice space. or
number ot'products produced, time to produce and amount of sales. These are more
appropriately associated with the service efficiency of a company. and thus will be
deemed the Company Model. This measure can identi* how good the company is in
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
56
--
- - -
- -
marketing and selling their products. For example, this model can be applied to
companies such as McDonald-s and other fast food restaurants to compare efficiencies.
The Value Model on the other han& consists of the variables that make a
company more profitable, valuable, and competitive. These variables can consist of net
income. net sales. total assets. total liabilities. shareholders' equity. cash tlow. capital
expenditure. retained eamings. working capital and other quantitative and qualitative
factors.
AIso, risk management is a crucial concern for companies. One can view risk as
an input. and the return tiom the risk as an output. Thus. if given a hi@ risk more
income is preferred to less, "minimize risk for a given level of g r o s revenues.'
[BERG981. For a company, the DEA Company Risk Model can utilize the net income
and net sales as outputs. and the credit rating, i-e. credit risk. will be the input to the
approach.
4.4
Selection of Factors
n i e following were deemed to be the important factors used to establish the woah
of a company: total assets. total liabilities. shareholders' equity, net income, net sales and
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
57
The Inputs
Total Assets & Total Liabilities
Total assets and total liabilities are included as inputs to the operation because
these are items one wants to minimize. There are some -'off-balance sheet" issues with
assets that need to be addressed. Total assets are composed of things such as tked assets
and some of these tixed assets may be perceived to distort the study, A separate appraisal
may be required for certain tixed assets, for example, real estate. This is due to the fact
that the valuation approaches used for the operating entity might not be appropriate for
the appraisal of the real estate. Most srnail businesses and prokssional practices do not
own the real estate they occupy. Real estate should be taken out of the balance sheet so
cornparisons can be made more readily. Thus. the related items fiom the incorne
statement should be adjusted and instead a rent value included in the expenses.
Total assets have the characteristics of both inputs and outputs. Total assets
consist of items such as cash, receivables. loans. inventories. raw matenals, work-inprogress. prepaid expenses, total current assets. tixed assets (plant and equipment.
leasehoId improvement) and intangible assets. It can be a raw material in providing the
funds for investment or collateral, while it c m also be seen as a major contributor to
liquidity. a highly desirable state for the company. Since the primary objective is to
establish company value, it is more important to analyze how one can convert these
assets to cash and hence profit. As a result. it is considered an input into the operation.
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
SS
--
--
-- -
of the mode1 works well with the pnmary objective of the DEA analysis. output
maximization and input minimization. Since the rnaximization of it on the output side of
the efiiciency model is liighly undesirable, this item is appropnately detined as input.
Cash flow. as detined in a business and finance sense, is equal to the net income
afier al1 cash expenses ignoring the effect of taxes. This means d e r interest costs but
before non-cash items- such as depreciation and amortization. This c m include
amortization, amortization of detrred financing costs included in interest expense. noncash assets paid for services rendered. and others (minority interests. gain on disposal of
capital assets, development costs written off. issuance of shares to settle outstanding
debt). Defined another way. it includes items such as amortization of goodwill.
amortization of detrred costs. amortization of capital assets and income producing
property. Cash How from operations does not include changes in non-cash workinp
capital items, such as accounts receivable. inventory. income taxes payable. accounts
payable. accrued liabilities. prepaid expenses. share proceeds receivable. and deferred
revenue. Cash tlow from operations is an item that any Company would desire to
maximize. Accordingly. with the central theory of DEA of output mavimization and
input minimization. it tits well on the output side of the model.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
59
--
The outputs of the financial operation include net income and net sales. Net
incorne includes revenues, interest and other income, depreciation and amortization,
investment income. and income before tax. Net sales are revenue minus cost ofgoods
sold- These two terms are the cornmon items in studying the earnings potentid of a
company. Generally, the income tiom these two sources determines whether a business
is profitable or not. And, of course. in the formulation of the DEA model, the
maximization O t' these two items indicates a better pertormance of the business.
Moreover, the inclusion of these items also allows the analyst to study the conditions of
operations in these corporations and to determine their viability.
1.4.2.3
Book Value
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
60
Finally, the model constructed aims to represent the opentions o f the companies.
Inputs and outputs selected for the model are represented in Figure 4- 1. This operational
representation is the basis for the formulation of the DEA model. The specitication of the
formulation of the DEA model to measure the relative et'ficiencies is discussed below.
Figure 4-1 DEA Valuation Modcl
1 Total Assets -l
4 Shareholders' Equity 1
1 -
1 Companies
Total Liabilities
l-
Net incorne
Net Sales
Cash Flow
Items that are important but w-ere not included in this analysis are market value.
price to earnings ratio (PIE), retained earnings, and working capital.
Market capitalization was not directly included in the model. It is another key
variable that should be examined in a valuation analysis. However. by de finition, market
value only esists for publicly traded companies. To be able to adapt the mode1 for both
public and private companies the variables selected have to exist in both domains.
Tlieretore. market value was not included in the initiai analysis. Total assets can be used
to account for the size differences among the companies. Market capitalization is added
to the analysis after the DEA-Valuation models are run. The market values are used to
VALUINGPRIVATECOMPANIES:
A DEA APPROACH
61
validate the results. by not including them in the analysis they can be used for tbis
purposeFinancial ratios are often utilized to evaiuate corporate performance. The firm's
ratios are compared with some benchmark value. Size is ignored and difterent sized
tims are compared. This leads to the issue of the choice of the value to compare with. of
the dozens of ratios in existence, which ratios to use. and the choice ofthe iinns to
compare with. Companies should preferably operate in the same industry and when
constnicting the ratio there is an assumption of proportionality among the numerator and
the denorninator. However, there may be instances where this does not hold [SIMAOO],
and this leads to problems when cornparin- difkrent sized companies.
Retained earning is another very important variable in determining value. It
represents the reinvested eamings over the yean that have not been paid out in dividends.
Many corporate raiders bought companies ihat had a lot of unused retained earnings. then
afier taking dividends resold the Company. In general one wants to use the least amount
ofcash in the company while maintaining profit levels. Aithouyh retained earnings is not
directly incorporated. the inclusion ofshareholders' equity in the rnodel accounts for it.
Next, finns need cash to pay for al! their day-to-day activities. They have to pay
for wages. bills. raw materials, and so on. The money available to them to do this is
known as the tirni's working capital. Working capital consists of current assets: these are
the short-tenn assets that the firm c m use to convert into cash. Current liabilities have to
be taken into account as well to determine how much working capital a firm has at its
disposal. Working capital is theretbre current assets kss current liabilities. An
interesting issue e s e s when exmining the assets of a company. Are the assets owned or
the model, so the smaller it is the better the DEA results. The companies that own their
own property will have l q e r total assets than companies that lease. It is important to
know if a Company owns land or not. since it inadvertently adds to the value of the
company. Thus it is justifiable to use total assets and totai liabilities rather than cwrent
assets and current liabilities.
4.5
rnodel that would best describe the behaviour of the industry. The available data on these
companies will ~ l s o
be evaluated to help detemine the most appropriate DEA mode1 for
valuation measurement.
A private company is a company that is not a "public" company as detined by the
various securities acts. it tends to be smaller in size. and typically has one or few owners.
There may be signiticant diftrences in the size of the total assets ofprivate companies
versus public companies. Actually. a private company may very well have larger assets
because it does not have the motivation to use al1 the xcounting tricks avaihble to show
a larger pcr share emings. ALthouh the market values of these companies are not
directly involved in the evaiuation oftheir cfiiciency scores. their behavior can be
inherently difierent due to diffrences in size. Applyiny some banking theory. it was
suggested by Piesse [PIES951 that the scale and technical eficiency rneasurements of
tinancial institutions are inversely proportional, that is. the larger institutions are
commonly less scale efficient than their smaller counterparts. This may be due to the
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
--
requirement for a significantly larger employee group in the larger unit~.One assumes
the sarne holds true tor companies. With such a significant range of magnitudes in asset
values of these companies, it is unwise to assume that they c m al1 operate with the same
scale efticiency. Thus. the focus of this thesis will be only on the technical efficiencies of
the companies. using VRS technology.
retums to scale where the proportional increase in the input does not translate to a
proportional increase in the outputs for the entire range of DMU sizes. The efficient
fiontier produced by the BCC ratio model is typicdly a piece-wise linear surface. The
slopes of the fiontier representing the ratios of the eficiency scores are different in the
various combinations of input and output. But al1 of the DMUs that are on the frontier
are considered to be technically efficient.
Afier the determination of the appropriate DEA model, its orientation should also
be decided. Companies could improve their performance by an increase in their output
VALUING
PRIVATE COMPANIES:
A DEA APPROACH
64
levels. or by a decrease in their input levels. It is observed that with the limited
intiormation on Company performance. industry trends can not be reliably predicted.
Financials are available only for a single year, therefore, information regarding the
gowth of individuai companies can not be determined. With this limitation on the data,
input-orientation becomes the most appropnate approach. The output-oriented mode1
wouid project an inefficient unit by increasing its total output. But, the lack of
In the analysis, the CCR model is not exarnined due to the existence of negative
variables in the data set and because it is scaie invariant. The BCC model can handle
negative data because it is both scale and translation invariant. Scale invariance deals
with scaling the inputs o r outputs of a DMU by a certain amount where the efficiency
scores are not affected- Translation invariance retrs to the efiiciency scores being
invariant to the transiation of inputs and outputs by a scalar. The BCC input oriented
model is translation invariant in the outputs and vice versa. Thus the efficiency ofa
DMU wiil not be affected if the outputs of al1 the DMUs are translated by a scaiar
amount.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
65
Summary
Fomulating the DEA model with the available financial intonnation is one of the
goals of this thesis. Different articles cite different methodologies and uidicators that can
be included in a valuation model. The most accurate DEA model was identified dong
with the corresponding variables. In the analysis of the model no weights will be
assigned to the variables since DEA does not require pnor knowledge of the importance
of each variable and will assign weights in accordance with the Company appearing the
best it cm.
--
--
VALUINGPRIVATE
COMPANIES:A DEA APPROACH
66
5. DATA ANALYSIS
The data consist of financial statements for 457 Canadian publicly traded
companies in the manutacturing sector al1 having %ancial records from the year endng
in 1998. Each Company has a SIC code between 2000-3999. which entails the entire
manuhcturing sector. The industries range fiom machine rnanufacturing to food and
kindred products (refer to Table 5-7 tor a complete list). Most companies have a
different tiscal year-end, however. for this study the year-end was assumed to be
standard. As well. data on both private and public companies were collected.
The manutcturing sector was chosen for the purposes of this analysis because of
the availability of the data. A substantial number of the companies listed on the TSE are
in the manutcturing sector. Service companies. specifically hi&
monitor since most of these companies have not been in business long enough and
reliable tinancials do not exist or go back tr enough.
The data from the year 1998 were chosen because it was the most recent,
available and cornpIete. Also, when thdint comparable h
the bigger the set of data. the better- According to Booth [BOOTOO], it is better to
exclude companies that deal in resources such as pulp and paper, since they are very
cyciical. and he suggested looking at the 7-year prothbility of the companies.
Availability of this inf'orxnation is an issue. Perhaps for future research someone can
examine only a specific group ofcompanies (50 public and one pnvate Company, al1 with
the same SIC Code) and get data on the past or fture 5 years for these companies.
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
67
Stock Prices
5.1
Public Com~anies
5.2
The data for the public companies were gathered fiom a varety of sources.
Mainly. ~rirnark's' --Compact D" for December 1999 and January ZOO0 for Canadian
publicly trded ompmes. The inf'ormation consisted of total assets, total liabilities, net
income, sharehotders' equity, net sales, date of fiscal year end, SIC codes, and so on.
Orignally. there were over 700 cornpanies' financials available. Unt'ortunately. due to
irregularities in the data. some filings from the Iate 80's & early 90's were still included.
These companies were omitted fiom the study and h a l l y 586 companies were lefi in the
data set. Missing and suspicious accounting numbers were checked and corrected if
necessary. After resolving the variable selection (inputs and outputs to use in the model),
the missing information was collected- Most of the rnissing information pertained to net
sales, shares outstanding. and net income. For the case of a Company having two SIC
codes. chietly the primary and secondary SIC codes, only the pnmary was considered.
Some of the cornpanies reported in diffrent currencies so everything had to be canverted
to Canadian Doilrs using idormation on historicd exchange rates. This insured that the
data were consistent,
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
The tollowing sources in Table 5-1 were used to find the missing information:
Table 5-1 Financial Information Sources
tnformation
Financial statementrr,
1 annual l financial renods
1 CDNX homepage
Source
[ Oanda www.oanda.com
Statistical Testing
Statistical testing of the data was c&ed out by calculating the minimum,
matimum, average. standard deviation, and percentiies (25. 50. and 75) of the data. This
information is recorded in Table 5-2. Total assets (T4), total liabilities (TL),
s hareholders' equity (SE). net income (Ni). net sales (NS). and cash flow (CFl) values are
iMin.
iMax.
Mean
Stm, Dev.
Pcrc, 25
Perc, 50
Perc. 75
TA 1
SIC
code '
0.006
2-01 1
3.999 395.298
2.368
3.153
572 30,934
5
2.759
33
3.3 12
23
1
3.663
TL
0.006
372.186
1.71 7
18.530
2
14
108
SE
NI
S(in millions)
-367
-841
64.1 75
9.7 18
66
65 1
3.993
599
-2
18
104
0.021
5
NS
-0.927
236.54 1
1.933
14.806
3
28
22 1-
CF1
-225
18.969'
180
1.42 1
-0.548
1
15
As c m be seen fiom the variation in the standard deviations. across ail factors. the
data set is large and accounts for a large number of smail and large companies, that
causes the variation in the data. The data contain negative values since some of the
companies had negative eamings, The range in total assets is wide and varied. from a
minimum of $6.000 (tbr Primero fndustries LTDJ to a maximum of$395 billion (for
GenenI Motors Corp.). Figure 5-1 represents the distribution of the totai assets (log scale
is used for the graphs to show a clearer picture):
-
--
Total Assets ( $ 0 0 0 ~ )
----
--
TotalAssets
-
. --
Figure 5-1 Distribution of the Total Assets for the Public Companies
As can be seen from the graph, most of the companies. 305 to be exact, have total
assets of Iess than S 100 million. Looking more closely at these 305 companies, one can
see that 68 companies have total assets of less than $2 million and 45 have assets bqeater
than $56 million and the rest lie in between, as demonstrateci by Figure 5-2.
Statistical analysis of the firms' market capitalization(M1) was also carried out,
as recorded in Table 5-3. Market capitalization and market value are used
CURRENT O / S
M.C
SHARES
(SOOOsI
P R K E TSE
Dec-98
mea an
Standard deviation
Percentiles 25
Percentiles 50
Percentiles 75
1
156.723 1
3 1,133.2131
9.58
The distribution of the market value varies widely tiom $190 thousand to $6 19
bilIion. The top tive companies by market value are BP Amoco, IBM, Mobil Corp.. GM,
and Sony, their market capitaiization are $619. $240. $105, $86 and $53 billion
respectively. Ignoring the top tive companies which are multinational conglomerates. the
maximum market value in the analysis becomes $50.5 billion, which represents Norte1
followed by Thomson Corp. at $22 billion. The tive conglomerates will d l be included
in the DEA analysis because most likely they wiil be self-evaluatoa shce they are so
large and they will thus not have any negative effects on the analysis. Figure 5-3 shows a
gaphical representation of the market values.
Figure 5-3 Graphical Representation of the Distribution of Market Value:
100
80
60
Ei Frequency
-- - -
40
20
O
Log of Market Value (in thousands)
There were 16 companies that have been de-listed by the TSE, CDNX or the other
Canadian exchmges. so no stock prices could be found for the year 1998. These data
were removed from the analysis. Upon M e r examination of the market value some
anomalies were found. For example, there were tive companies with market values of
less than $100.000. Subsequent analysis showed that at l e s t one of these companies.
Scott Paper. had been bought out by another Company in the preceding year ( 1997 in this
case) and still had 100.000 of its shares outstanding. The other four. mainly Consolidated
General Diarnond Corp. Difiacto Ltd., Lenox Polymers Ltd.. and Vendtek Systems.
upon M e r investigation were also excluded from the analysis.
Another important step is the examination of potential correlation arnong the data
items. In a typicai DEA analysis. correlation among the data means that some of the
variables c m be eliminated tiom the anaiysis in order to simplifj. the model. Their
inclusion in the analysis may place extra importance on these factors and skew the study.
Table 5-1 Correlation Between the Public Company Data
some of the populations of the data. Specitically, total assets to total liabilities have a
correlation coefficient of 0.98. High correlation exists because the data set includes only
the manufactwing industry. Manufcturing companies tend to mostly fnd their assets
through borrowing, thus total assets and liabilities will be highly correlated. Because
they are highly correlated, they would have similar correlations with equity. Neither
assets nor liabilities tend to be highly correlated with equity. as equity tends to be much
smaller. The amount of total assets gives an indication of the credit worthinesdequity of
a company. while total liabilities give an indication of the debt of the company. They are
both necessary in the valuation sheme. Net sales and cash tlow are also highly
correlated. This is partly due to the fact that companies cannot have cash tlow unless
they have revenues. More notably. cash flow is an indicator of tinancial well being,
ibqoring the impact of this variable on market value will render an ineffective analysis.
VALUINGPRIVATE COMPANIES:
A DEA APPROACH
73
together in an analysis of this nature. The same argument can be used to contend the use
of the other variables in this DEA-Valuation mode!.
5.5
Private Cornnanies
One of the big major Canadian Banks provided the private company information.
[t contained statistics on the companies tkom the year 1988-1003. including tinmcials.
SIC codes. statement date. and other financiais. From this data. al1 the information on the
companies for the year ending 1998 were extracted. For companies with SIC codes
bctween 1000-3999. No information about the real names of the companies was known,
due to the contidentiality of the intorrnation. Thus, the companies are listed by number.
The data were cleaned and any duplicates elirninated. There were a total of 84 private
companies alter al1 anomalies were sorted through. Statistical intocmation about the
private companies is located in Table 5-5.
Table 5-5 Private Company Statistical Information
SIC
Code
TA
1 TL 1
NI
SE
NS
CF1
(SOOOs)
I ~ t n Dev.
.
1537
I
Perc. 50
Perc. 75
3271
3577
1324.739
1 113,326 1225.9 1 4
13,599
26.803
15 1.298
1326.308
165.855
595
27.614
40.598
1933
12.866
7,099
12.138
5.567
14.534
2.471
with the ssception of a w large private companies. The existence of this tirly large
pnvate companies will enhance the DEA analysis since there will be ample public peers.
The large private tirms are namely company #156 and #198 with total assets of $2.8
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
74
--
--
billion and $1.2 billion respectively. and net sales of $2-6 and $1.5 billion. Their net
income is quite reasonable at $457 million and $103 million respectively- The next
highest values for total assets, total liabilities, shareholders' equity, net income, net sales.
and cash tlows are smaller at $145. $11 7. $72, $68. $416, and $68 million respectively.
Below in Figure 5-4 is a histogram of the total asset distribution for the private
companiesFigure 5-4 Distribution of the Assets of the Private Companies
A correlation analysis for the private companies was carried out in Table 5-6 and
it shows a high degree of correlation among the variables7especially net income to cash
tlows (0.99). The preceding section on correlation among public companies pointed out
the main reasons for this. Ayain, the purpose of this analysis is to determine the value of
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
5.6
companies. The first group included companies with SIC codes between 2000-2099:
there were 18 other groups. thus aitogether 19. In the tirst g o u p there were 44 public
companies and 4 pnvate ones. Each of the private companies were added one at a time
into the DEA model, so in this case there were 45 DMUs in the DEA analysis. The
companies were analyzed one at a time in the DEA analysis so as not to change the
tiontier too drastically. Aiso. one does not want a private company to have another
private company in its peer group since this would not yield any usefl Iformation. The
grouping chart is shown below in Table 5-7:
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
76
1
. -L...u"...
- --
. -7 , .,.
- .J.
;>.-+.:.
O
--;."6.;
. ..
. . ,'. -0 .-..; - .
-.
..-.:;$
:
..,;
;'
..:-:-.,;
3..
: .- ( -:"-? - - ..
..
-x
20
19
22
3
5
47
85
9
3
22
36
5
5
a+,
-*;
4.2-
'
- . ..-.
i'
. .
g
452
-
;7'
;"'.
II_
6
84
Since it is essential in a DEA study to have at least three times the number of
DMUs as the sum of the inputs and outputs (6 in this case), some of these private
companies can not be investigated for market value. Therefore, one needs at Least (6'3)
18 DMUs for the DEA models. For the ones with less than 18 DMUs, a DEA analysis
could not be carried out, as was the case for 36 of the private companies and 91 of the
public companies. The shaded areas in Table 5-7 represent this. There are some models
with ody 19-22 D W s . This is rather small but nevertheless a DEA analysis would be
able to be carried out. The data on the public companies c m be found in Appendix A.
VALUlNG
PRIVATE
COMPANIES:
A DEA APPROACH
77
--
--
public companies with SIC code 3600s only. and calculates market vaiues for the
companies. Section two is a comprehensive study of the companies that have a SIC code
in the 3700s. Section three reports on companies that have a SIC code in the 2000s.
Section Four examines an etlkient and an inefficient company compared with d l 452
companies. Finally, section tive examines the methods developed tor valuing public
companies and appIies them to private companies.
The BCC model is run to classi@ the companies as efficient or inefficient. Then
the two classifications are examined separately. For the inet'ticient companies, peers are
found based on the other companies ha\:ing the same e f k i e n t companies as their peers.
The distance equation is applied and a range for the market value for the subject company
is revealed. An upper bound is found for the inefticient companies as well. The upper
bound is based on the lambda relations of the inet'ticient Company to the eflicient
companies multiplied by the market values of the efficient companies. For the efficient
companies, a lower bound c m be found by tinding al1 the companies using it as a peer.
summing these values and nonnaliziny them then rnultiplying these by the market values
and summing them to get the lower bound.
Since there are negative variables in the data, the most negative variable is found
and one plus this value is added to al1 the outputs in the BCC model, making them
positive. This c m be done since BCC is both scale and translation invariant.
VALUINGPRIVATE
COMPANIES:A DEA APPROACH
78
6.1
DEA BCC model. The BCC model was nin for the 452 companies gouped by their SIC;
thus 9 models were run in total. one for each SIC grouping. The tirst malysis was
conducted on only cornpanies with SIC codes in the 3600s (3600-3699) and consisted of
85 DMC's. The data For these companies can be round in Appendix B- For each
cornpany one now has their associated eficiency. lambda, slack and multiplier values.
The market capitalization is incorporated back into the model to validate the t-mdings.
Once a value is estimated for the company. it can be compared back to the known market
capitalization of the company to see if the results are redistic. In this grouping, there are
also three private companies with the sarne SIC code. These are added one by one into
and a median of0.79. There were 23 eficient DMUs and the eficiency scores ranged
tkom as littlr as 0.0653. for company Noron, to 1. Noron had $57 1 thousand in total
assets and $4.4 million in total Iiabilities along with -% 1.3 millions in net income and
-$ Z .2 million in cash tlow. These account for why its efficiency was so poor.
Subset Hammond
A subset of the 85 DMUs was chosen for this anaiysis. The peer p u p for
cornpany. Hammond was found. The peer g o u p consisted of al1 the efficient companies
Hammond was being compared to as well as al1 the inefficient companies that were k i n g
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
79
- -
IDMU
6ILG TECH
7 C'4iUCO
32 TS TELECOM
1
!Score
1 0.58651
0.73021
1
O
1
741ELECTROHOME 1 0.67331
27,7071 6.03E-031
11 ~0O.i)OO
1
il
11.832
O
O
O
1 1.3987 7.52E-021
0-43571
0.499
Carnco. TS Telecom and Primetech were on the efficient frontier and thus had a
score of 1. h e inefticient companies had scores mging fiom as low as 0.59 to 0-8. The
lambda values going across the table for crich DMU should sum to unity in the BCC
model; this is not the case with the CCR niodel. In the BCC mode[, this is due to the
convexity constraint. the CCR model does not have this constraint so its lambda values
have to be adjusted: refer to section two for an explanation of the adjustment. This subset
was then d y z e d with the Simak [SIMAOO] distance equation. The distance for
'
(0.33154.2637) = O. 1315
VALU~NG
PRIVATE
COMPANIES:
A DEA APPROACH
80
Table 6-2.
Table 6-2 Distances to Hammond
1 Hammond
10
LG Technologies
Camo
TS Telecom
i Calian
1 Primetech
1 Electrohome
1 0.1315
/ 0.9152
1 0.4652
1
1
0.048
0.6956
0.0490
Al1 the companies that had a distance of less than 1-0 tkom Hammond were then
compared in terms of their eficiency scores and market capitalizations. Caiian was the
closest Company with a distance of0.048 followed by Electrohome, LG Tech., TS
Telecom, Primetech and Camco. Caiian and Electrohome were also closest in tenns of
efficiency score, 0.80 and 0-67 respectively. compared with Hammond's efticiency OF
0.76. EIectrohorne was second closest by market value. 9 1 . 4 million. followed by TS
Telecom at $1 1-8 million. Although Primetech and Carnco had a distance of less than
one in relation to Hammond. they were greatly diffierent from Hammond in ternis of their
efficiency scores and market values and thus can be ignored as peers. nie analysis
deterrnined a range for Hammond's market value basrd on its peen of $1 1.1 to $33
million. Hammond's reai market capitalization of$lll.4 million is in this range.
6.1.2.2
In this part. a new method ofdetemining value basrd on the reference units tor
the inefficient companies is introduced and it is argued that this is an upper bound for the
inefficient Company. To calculate an upper bound for Hammond one uses the market
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
values of the efficient companies that it was k i n g compared to. namely, Carnco, TS
Telecom. and Primetech, multiplied by the lambda values of Hammond in relation to
these efficient peers. The lambda values can be found in Table 6-1. This formula is
exemplified by:
= (0.22 l7* 100,000,000)
The $75 million value thus round is considered as an upper bound on the market
capitdization of Hammond, since it is based on the market values of the efficient
companies. Similar analysis is carried out on the rest of the ineft'icient companies in this
subset. The results are summarized in Table 6-3:
Real iMarket
Cao
Company
i~lectro
home
Minimum
Maximum
1 1-39%-O001 1 1,832,0001
Predicted Upper
Bound
33.191.000(
84.103.0001
The tint two companies' actual market values falls within the predicted range. so
the mode1 appears to work well. Calian on the other hand is not within the predicted
range with a diflerence of SZ.5 million fiom its predicted upper range value. so it could
be overpriced, but it is less tliiin its predicted upper bound. Electrohome seems to be
undervalued, but again it is less than the predicted upper bound. It is only $434,000 less
than its predicted minimum range value.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
82
The next subset in the 3600s group consists of Inverpower and its peen. which
are Avante. MM, TS telecom, Advantedge, Fifty-Plus, Kasten Chase, Wi-Lan. Plaintree,
Ne.xtron. WE XL. in total there are 11 companies in this subset- RIM. TS telecom and
Fiflty-Plus are the efficient companies. Table 6-4 represents this set,
Table 6-1 Company Inverpower and Distance to other p e r s
Score !Market Efficient P e r s (A):
Cap.
S(00s) JO-MM 132-TSTel.
ZU
Name
1 . I6E-02
INVERPOWER
0.545
7,179
AV ANTE
2-302 2-76E-03
0.583
RESEARC14 IN MOTION 1
1 585,308
If
TS ELECOM
1
1
11.832
Ol
ADVANTEDGE
0.180
19.928 1. I 7E-O2
OI
FIFTY-PLCS
2.404 1
1
1
1
M S T E N CHASE
1 O-9951 23.5691 O. 1 129991
I
42-Fife
0,920842 6.76E-02
2.1 O
O
1
8.9 1 E-02
O1
0-76 17091
0.976265
6.18E-16
2.78E-17
0.8992 19
1
O . 12529 1
I~istance
(from
D~MUrnvct.)
O
1 -635
1,829
0-0109
I -383
1.717
0.0389
1.122
0.308
1 -548
I-679
TabIc 6-5 shows the predicted market values for these companies.
Table 6-5 Subset Inverpower and the Predicted Market Values
Inverpower tlls short of the predicted range for its market vatue, this could mean
it is undervalued. It is also very inefficient and an improvement in etliciency score may
VALU~NG
PRIVATE
COMPANIES:
A DEA APPROACH
83
enhance its market value. The predicted upper bound is 2.5 times as g e a t as
Inverpower's real market value. Avante's market value falls below the predicted range.
however. DEA found two very similar peers to it. namely, Nextron and Fifty-Plus.
Especially Nextron. which had a similar score and market capitalization. Avante does
seem slightly undervalued but again it is highly inefficient an increase in efticiency
might help Avante improve its market value. The predicted upper bound is 1.8 times
greater than the real market value of Avante. Ignoring Wi-Lm. which was greatly
diffrent fkom Avante in terms of the efficiency score and its market value. one c m find a
more appropriate range for Avante of $2.3 to $20 million. The same is true for
Advantedge; if one ignores the etTect of Wi-Lan. which was not even close in t e m s of
etticiency or market value. Its range becomes $2.3 to $17 million. Also. there are closer
companies like Avante. Nextron. and WE XL in terms of distance. Advantedge had the
Iowest eficiency score in the g o u p at 0.18, so this seems to be a very poor Company in
t e m s of tuming its assets into profits. That also explains why the upper bound is so 10w.
By al1 accounts, Advantedge is overvalued. It is just 0.5 times less than its upper boundKasten Chase. too is not very sirnilar to Wi-Lan. in terms of distance, eficiency score,
and market value. The closest peers are TS Telecom and Plaintree. Thus. ignoring WiLan the upper range becomes $14.9 million. Kasten Chase then becomes ovrrvalued, but
it is still operating less than its upper bound and it is highly efficient at 0.994. The
predicted upper bound is 3.2 times greater than Kasten's real market capitalization.
In all. 62 inefficient companies were exarnined and a range for market values
predicted as displayed in Table 6-6. Subsequently. it was found that eleven of the
companies did not yield any information due to too few peers. Of the remaininp 5 L
VALUINGPR~VATE
COMPAN~ES:
A DEA APPROACH
84
- -
--
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
Score
Market Cap
1998
Predictd Range
Maximum
Minimum
, UPP-
Bound
0.757
0.545
0-435
1 HAMMOND
3 INVERPOWER
4 UNITEC
6 LG TECH
8 NORON
9 AVANTE
I O SEMI-TECH
l IIDANBEL
0.587
0,065
0.583
021 5
0.755
0347
0.656
0,SU
I2'AVVA
I 3 TIR
14 DIGITAL PROC
18 ABL
19 UCS TECH
20 BELL CAN
21 ClRCA
22 DBA
23 EIGER
34 KING
25 MEMOTEC
26 MICROCELL
28 NHC
3 1 - R A D E WIND
33 ADVANTEDGE
34 CALIAN
36 CELL-LOC
37 COM DEV
39 CYGNAL
40 DIMENSIONS
41 DYNAMIC
43 GEMSTAR
45 KASTEN CHASE
46 MARATHON
38 SENSE
49 SIMMONDS
50 SPECTRUM
51 SR TEL
52 STRATEGIC
53 SYSCAN
54 TEKLOGIX
55 TRIPLE CROWN
O .Z 99
1
1
0-677
0,254
0-859
0,640
0-81 5
0-472
0.649
0.066
0.67 1
0.676
O- 180
0.800
0-980
0.863
0-589
0.1 I O
0.939
0.232
0.9Ujl
0.604(
0.3 7-1
0.766
0.787
0.447
0.983
0.657
0.8 1 O
0.526
74,988.000
33-191.O 18
18,426,080
I 1297.939
17,848,000
7,179,058
1 1,832,000
33-569.233
5,057.000
2,404.000 one peer
5,570,707
47,055,000
33.19 1 .O 18
27,706.969
l 1.397.939
2-383.000
1,104,000 one peer
7,039.3 1 O
4.2 1 1.000
2,404,000
329.339.3 22
3.3 02.125
I00,000.000 one peer
720,885.000
10.025.976
44-09 1,000
30, 1 56.000
289,000
5,797.5 76
1 3 -3 82,000
2,669,000
1 -710.000
3,429,279
I
5
-3
83,000
3-53
7,000'
1,709.64 1
1.956,OOO
1 54.005.000
28.123.144
1 1.832-000 one peer
7.609.000
1 1,832,000
1 -7 10.000
13.381.932
1 5.097-570
3-376.000
622.999
1 -026,687
3.60827 1 .O00 2,448,639.000
1-371-1 19,948 1-031.559.000
46-09 1 .O00
22,797.000
6,600,085
3.855.000
1 3-3 82.000
3,782,000
1-7 10.000
1,956.000
5-512,000
3,878.000
2,404.000
8289.000
1,854,080
1,072,000
307.000
7-738-941
76.67829 I
48,488,000
4,952.53 1
1 1,832,000
2 1-615,000
3.878.432 one peer
307.086.660
44.4 1 1 .O00
2,404,000 one peer
7.3 1 7.082
1,485.000
15,697.570
307.000
9.699.54 1
I0,064.000
329.339322
19,927,644
2,302,425
62.705.000
27-706.969
33-191.018
1 1.397.439
23,996.000
2,404.000
1 -969.999
28,669,000
85
1.037.000
1
.
3
9
.
0
3533 15.304
1 12,896.7 I O
50,650,000
32.734.620
1 1.832,OOO
8-637.174
13.34 1-000
28.669.000
2.404.000
1-969.999
89.29 1-000
80.6 15.000(one peer
1 10.353.695
3 -327.000
6 1220.528
1,729.263
4,887,873
75.454,000
329.339.322
7.1 79.058
23369.233
2.447-000
15,697570
7 1 7.0 1 0
622.999
4.887373
1 5.085.000
1,739263
6 1220.538
7-73
8-94
I
3 -334.000
3.878.000
2.854080
76.67829 1
1 00.858,OOO
38,637.268
4.952.52 1
1 .O3 1.559.000
532269.000
44.404,l 12
1 12896.710
l3,222.000
6,597.000
2.404.000
3,943. 1 54
5.65 4,000
3,855.000
2.404.000
739.505
1 79,974.000
143.365.000
243.50 1 .J58
44.404.1 12
4.8 18.000
I 3-382.000
1.7 10.000
4978.95 I
56 TSI TELSYS
57 WESCAM
1
61 SIERRA
63 STRATEGIC
64ISY MPLEX
651WI-LAN
WORLD
661C'IRC'UIT
,
69 COREC'O
71 SHELLCASE
71 T U S D M
73 c'-MAC
71 ELECTROHOME
75 LAMBUS
77,PL.4INTREE
78 TRI-VISION
80 BLCESTAR
8 1 COMPRESSION
83 GSI LLMONICS
83 NEXTRON
84 ROYAL LASER
I
0 21 1
0,628
0.7571
0,996
0,534
0,562
0,873
0.753
0,983
0,085
0,983,'
0-817
0,673
0-7i 0
0.866
0,743
0,555
0.3 I O
0.9 t 7
0.534
0.684
-- --
622,999
12,192,753'
15,697,5701
2.3 04,000
76.678.29 1
4-952-521
I43.365.0001
169,2 12.000
9-421.6651
2,404,000 1one peer
1
14, 108.00a
3,855.000
32,734,620
1 1.832.000
1 07.539-000
622,999
1,719,718
15,697.570
2,004,000
5,879226
1-71 0.000
1 3-382.000
2-657,000
339,339,323
23,569.233
24.838-000
2.3 02,425
188.73 7
46-001,000
24.592.000
5,798,000
.dl-922-938
2,404.000
5 1-361.300
1 53.377.000
622.999
15,697.570
13.192.753
2-14 1.000
120,705.32436,523,000
3.855.3001one peer
29 1,126.880
100,000,000~ 1 .O3 1359.000
969.673,OOO
1 1,397,939
339108
84,103,000
1 1,832.000
2,404,000
1,729,263
6 1220,528
7.566,000
14,946,155
329-339.322
58,607,000
2,302.425
2,404.000 one peer
39,930,177
53,470,OiiO
44,404,112
112,896.7lO
717.010
622.999
1 5,697,570
3-391-000
129.625.600 1-031,559.000 one peer
1-037,907.000
4,707,000
2,426, I 40
2302,425
329.339.322
5 1.56 1.300
1 5 1,693.O00
2,404.000
U.922.93 8
1
- -
VALUINGPRIVATECOMPAN~ES:
A DEA APPROACH
The real market values were also graphed against the predicted upper bound- Of
the total 62 companies, 41 companies had r d market caps that were less than the
predicted upper market capitalization. thus DEA is correct 66% of the time in predicting
an upper market cap that is truly higher than the market cap of the Company. This is the
result one would hope for since the upper market value is based on the etticient
companies. and intuitively, the market caps of the inefficient companies should be below
these values. Another 13% of companies were within 35% of their predicted upper
bound.
Figure 6-2 represents the predicted upper bound versus the a c ~market
l
capitalizations.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
In this part. a new way to analyze eficient companies is introduced. The anaiysis
for the efficient companies was carried out in a diflferent mamer than for the inefticient
cornpanies. Since these efficient cornpanies are rarely compared to other companies. one
has to calculate the market value fiom the cornpanies being compared to them, whether
they are inefiicient or efficient. One c m hypothesize that this way a lower bound on the
market values can be tound. Since it is an efficient company, it should presurnably be
doing better than the market value cdculated from the inefficient companies in its peer
group. The lower bound was calculated based on the lambda values of the ineficint
companies being compared to the efficient company. Adjustrnents must be made for the
h c t that the lambda values on the column of the matrix do not s u m to unity.
6.1.3.1
Camco was analyzed tirst: it had 17 cornpanies being compared to it. The lambda
values of the ineficient companies in relation to Camco were surnmed then this sum was
used to divide each lambda value to get a new normalized Iarnbda (A / Ch). For example.
Harnmond has a lambda relation of 0.222 to Camco. the sum of ail the lambda relations to
Carnco equaled 1 -961. Hammond's new lambda score is equal to (0.22l7/l .%OS) 0.1 13 in
relation to Camco. This same procedure was repeated for al1 the other companies' lambda
values, thus the new sum of the lambda values equals one. The new, normalized tambda's
of the firms were then multiplied by their market values and summed. The resdts are
s h o w in Table 6-7:
VALUINGPRIVATE COMPANIES:
A DEA APPROACH
93
--
DMU
Score
#
1 HAMMOND
6 LG TECH
10 SEMI-TECH
15 SONY CORP
17 AASTRA TECH
23 EIGER TECH
26 MICROCELL
24 CALIAN TECH
37 COM DEV
5 1 SRTELECOM
0.7568
0.5866
0.2 145
I
i
0.8151
0.0664
0.800 1
0,8633
0.447
52 STRATEGIC
0.9832
54 TEKLOGIX
0.809V 1
1
CELESTICA
INC
68
73 C-MAC
0.8171
71 ELECTROHOME
0.6733
80 BLUESTAR
0.5547
X2 GSI LCMONICS
0.9 167
SUM:
18426
27707
10026
5.3&+07
4609 1
8289
307087
33 191
152315
1 12897
3943
24350 1
2854867
29 1 127
1 1398
44404
129626
0.2217
0.006(
0-7746
3.1E-15
6.3 E- 1 8
0.0107
0- 1845
0.0599
0.08231
0.13631
0.00 111
0-0755
2.8E- 16
0.2202
0.0752
0,0926
0.0 196
1.96053
0-1 13 1
0,003 1
0,395 1
1-6E-1 5
3.2E-18
0.0055
0,094 1
0.0305
0,042
0,0695
0,0006
0-0385
1-4E-16
0.1 124
0.0384
0.0472
0-01
1
M.C. J(0s)
Calculation
2083-92
85.22
396 1.35
8.343E-OS
1 -483E- 1 3
45-2-1
38902-05
1 O 14.08
10604.66
7835.93
2.45
9377.23
4.092E- 1 O
32707.43
437.19
2097.3
1295.9
S100,460
For Camco, the predicted market capitalization is $100.459.000; its real market
value is $100.000,000, DEA did an exceptional job at predicting the value of the public
company. Similar analyses were carried out on the remaining 22 eficient companies and
are recorded in Table 6-8. A complete malysis can be found in Appendix C.
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
94
Company
(9
Turbo Genset
Global Therm
Camco
Sony
C inrarn
Aastra Tech
Lui itel
Nortel
Res. In Motion
TS Telecom
Cdn Marconi
Comm. Sys
Fi@-Plus
In t. Datacast.
Newbridge
Jemtec
Pacitk E-link
Snif Security
Primetech
Celest ica
Gennum
Leitcii
Bal lard
80-615,000
1 3,121,000
100,000,000
53.93 5-385,000
1.075-69 1,000
46-091 .O00
1.257-3 78,000
50,495-406,000
585-3 08.000
1 1 ,x33.000
426,508,000
3.855.000
2,404.000
3,893,000
8.125-48 1,000
330,000
3,878,000
3 07,000
1 1 3-365.000
2,854,867,000
576,578,000
1 .O3 1 ,559.000
3.60827 1.O00
--
Comment
# of
Peers
Above lower
Above lower
Above lower
no info
no inb
Xbove lower
no info
no into
Above lower
Undervalued
N o info
Cndervalued
t'ndervalued
L ndervalued
L'ndervalued
Cndervalued
Undervalued
L'ndervalued
A bove lower
A bove lower
Ahove lower
Ahove lower
A bove lower
6
4
17
O
O
5
O
I8
29
O
9
45
4
2
1
18
11
15
6
5
21
Excluding the tive self-evaluators, DEA does a tirly good job of tinding a
reasonable or appropriate lower bound for the companies. as evidenced by the majonty of
companies having real market values hiyher than the predicted lower market value. DEA
finds a lower bound that is less than the tme market value 56% of the tirne. In some cases,
there are too few peers and the calculation maybe exaggerated or the peers may be vasrly
ditferent fiom the Company under analysis. The seven companies that had market caps
below the predicted lower bound cm be classitied as undervalued, since accordinp to the
--
analysis their value can be greater. Efficient Company Snif Securty was the most
undervalued. It had a predicted lower market cap of $7 million when in reality it only had a
market cap of $307.000. It had a 2234% difkierence between its predicted and real market
value. Figure 6-6 displays the red market values compared to the predicted lower bouod?
and Figure 6-7 shows the percentage dimerence between the real market value and the
lower bound.
--
VALUINGPRJVATE
COMPANIES:
A DEA APPROACW
Company Name
--
m ~ e aM.C.
l
m~redictedLower Bound MC
VALUING PRIVATE
COMPANIES: A DEA APPROACH
--
-.
--
- .
-- -
6.2
A similar anaiysis as tor SIC 3600 was caned out on the companies with a SIC
code in the 3700s- This group consisted of 22 DMUs. This group was tested to see if the
number of DMUs in the analysis really has an effect on the number of peers that DEA
tinds. AH 22 DMUs were analyzed. The average efficiency of the BCC mode1 was 0.8 1
with 8 efficient DMUs. The efficient DMUs are listed in Table 6-9.
Table 6-9 Efficient DMUs SIC3700s
#
-7
5
6
7
Il
I I
21
-73
IDMU
GM
Budd Canada
Linamar
# of Peers
O
12
6
Magna
Spectra
Wescast
Templar
Bornbardier
6
8
1I
O
DMUs Budd and Templar are the most employed peers with 12 and 11 peers
respectively, however Templar has a higher degree of intluence over its peers. DMUs
GM, Mabqa and Bombardier do not act a reference unit for any of the inefticient units.
This could mean they are either efficient in a very narrow sector or that they possess a very
uncornmon input/output mix or they are just very different. The cornpiete eficiency scores
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
DMU
1 ~fficientP e r s (As):
1 3
Il
14
21
market capitalization. It is evident fiom this figure that the larger compaties, the ones
with the white points, are al1 efficient or nearly so. The correlation coeficient is 0.26.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
100
22
The correlation coefficient between eficiency and total assets of the DMUs is O. f 5 ,
illustrated in Figure 6-9. Thus there is great difference between the total assets of the
various companies.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
<O
IS
20
T o t a l A s s e t s (LN S c a l e )
indicates the importance attached to a given output. In the case of the CCR model. the sum
of the virtual outputs is equl to the unit's efficiency score. In the BCC model. the sum of
the virtual outputs plus the mesure of scale efficiency, %. is equal to the unit's eficiency
score. For each DMU, a stacked column that has the unit's virtual outputs as its senes is
plotted in Figure 6-10. It displays the relative amount that each output factor contributes.
It is interesting to note that Company 21 has no conmbution t'rom its outputs. it is solely
made up of the contribution tiom its input. namely total liability. Figure 6-10 also shows
that the factor shareholders' equity is the most influential output.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
102
Table 6- 1 1 surnmaxizes the predicted values for the companies with SICs in the
3 700s and Figure 6- 1 1 shows the predicted ranges. There were 1O peer groups in total
with some companies only being compared to efficient companies. For most DMUs
there were at least duee peers. with the exception of Trailmobile and Devtek. which oniy
1 Real Market
Cap
A-E Ventures
Avcorp
Merch
Rose
Stackpole
Glendaie
627.000
45.730.000
f,869.000
25.150.000
1 3 1,497.000
-17.948.000
4,576,000
627.000
4576.000
2,869,000
1 16,772,000
4576,000
4.5 76.000
4.576.000
525.280,OOO
----------4, ~76,000
1 16.772.000
574300,000
i 16,772,000
55.3 27,000
16 1.222.000
One peer
52 1-65 1.O00
3 1 4.7 1 0,000
1 16.772.0001
96 1.8 1 7.000
-----------
Northside
Heroux
Serena
Mage1lan
Devtek
14.056.000
63,162,000
572.000
496-389.000
99.709,OOO
Tesma
570.121(
45.730.000
525,280,000
25.150,OOO
525.280.000
574.3 00,000
1 16,772,000
-----
1.792.5 1 7.0001
1 2.58 1 ,000
23 5.632,OOO
I 4.730.000
3 1 6.2 13.000
43 7.63 8.000
60,753.000
Real M.C.
Preicted Market Cap. Lower Bound Comment
I 16.772.000
Budd Canada
2441 1 6.000
Undervalued
Linamar
S pectra
Wescast
Templar
GM
Magna
Bombardier
1 -792.5 1 7.000
502,534.000
64.3 57.000
383-838.000
16.3 59,000
Self Evaluators
Self Evaluators
525280.000
5 74.300.000
4376.000
86.392.582.000
7.3 74.1 03.000
7,502,049.000
Self Evaluators
Above lower
A bove lower
Above lower
Undervalueci
No info
No infb
No info
- -
Figure 6-1 1 Range for JDredictedMarket Values and the Real Value (SlC3700)
Ali Ven
Avcurp
Mrch
Rose
Stuck
VALU~NG
PRIVATE
COMPANIES:
A DEA APPROACH
Glend
r u
1 Icroux
12
10
Mugcl
Dccoiiia
Tesmil
1OS
Figure 6-12 Inefiicient Company-Real & Predicted Upper Market Value (SIC3700)
A.E
Ventures
Avcorp
Merch
Rose
Slackpole
Glendale
VALUINGPRIVATE
COMPANIES: A DEA APPROACH
m ~ e aMarket
l
Cap
Trallmoblle
Norlhslde
Heroux
P r e d i c t e d u 3per bound
Magellan
Oevtek
Decoma
Tesma
m ~ e a M.C.
l
Budd Canada
Llnamar
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
Spectra
Wescast
Templar
Il
in the case o f the inefticient companies, in Figure 6-12. DEA accurately classifies
64% ofuie companies, but without exception the range from the minimum to the
ma~imurnvalue is very large. In the instances when DEA camot c l a s s e the companies,
it is due to there being not enough similar peers. DEA correctly t'inds an appropriate
lower bound 60% of the time for the efficient companies. F i p r e 6-13, with a SIC of
3700. Appendix D contains the information on the companies with a SIC in the 3700s.
6.3
~ ~ o represented
u p
food and kindred products. There were 44 DMUs in this subset, with
15 on the efficiency fiontier and an average efficiency of 0.8 1. The most popular
efficient peers are A. L. Van Houtte Ltd. with 21 peers tollowed by AMT Fine Foods
Ltd. with 19. Analysis of the lower bounds of the efficient companies shows that 67% of
the companies are tound to be undervalued, they have actual market value less than the
predicted lower bound. Canada Bread Company Ltd.. S e a g m and Petroreal Ol
Corporation are Lound to be self-evaluators. For the 29 ineftcient companies. 10 of them
do not reveal any information due to the tct they have one or no peers. The remaining
19 companies have a correct predicted market range 68% of the tirne. Also. 86% of the
tirne the inescient companies have an actual market value less than the predicted upper
bound. TabIe 6-12 shows the range of market values for the inefticient companies and
Table 6- 13 demonstrates the market values for the eEcient tirms.
--
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
Company
3 SEPP'S
6 T M L FOODS
7 SAPUTO
8 DELICIOUS
10 LASSONDE
I l GLOBAL COMM
14 AGRO PAClFIC
15 RI DLEY
18 BETA BRANDS
19 S M A L L FRY
121 BANFF
Maximum
Predicted
Upper Bound
4-5, f 2
27,532,000
1 1.173.0 1 5 one peer
4-33 1
1,782,000 l
1 1.1 73 .O 15 1one ~ e e r
2-24-431 2.1 07.8 19.000 632.44 1262lone peer
302.0001
3 8,002.000
4 1 -43 1
1 7.246.000
1 7.825.0001
15.889.0001 1
632.44 1.O001
I
11.4 10.000
1 -9 1 7.000
3 -93 8.000 1 6.4 1 -43
8,570.000
2-45
L 0.874.000
1 1 - 1 73,000'
2 - 4 4
i h0.05U.000 no peers
no peers
4 1
47,999.000
1 1.1 73,000 one peer
1 1.1 73 -000
15,889.000
4
36,740.000
1
1
1
I
1
1
18.037-000
10,684,000
586.998-000
19.742,OOO
328-5 12.0001
5367.000
19.65 4,000
48 1.98 1,000
19.308,OOO
29.793 .O00
w
10.33.4 Ilt
4 14
4-12-43
44.43
1 6-41 -431
1.3 78,00011
19.445.000l
1 5-42 1 .O00
126.836.000 1
1.9 1 7.0001
2.879.000I
3 -938.OOOl1
1 1,173 -0001
13,878.000[
1 1.173.000 one peer
1I,173.000
632,441.000 1
302.0001
14-41 0,0001
2.88 1 .000I
7 1 -738.0001
52-8 1 1.O00
205,764,000
2.368.0001
4,924.43
4 - 2 4
9 1-1
4.1 6.4 1-43
4- 16.43
79.826.000
1 3 -310.000
197.1 14.0001
2,373.000
962.000
1 1.173.000
633-44 1.000
126-836.000
1 1 - 1 73.O00
13.889.0001
632-44 1.0001
3.93 8,000 one peer
1 1.1 73.000 one peer
139259.000
49,850.000
288.746.0001
6.5 17.000
13,002.000
36 COTT
37 LEADING BRAND
139 URBAN JUICE
1
40 VlVANT
42 LEF MCLEAN
1-14ARCTIC GROUP 1
I
I.LX 330.647.000
8,570,000
2.4.5
4.4 1-431
12.878.000~
33
876.000)
3,381.000)
-1.5.9.1 2
4 1.43 1
3 8.002.000)
I
139.628.000
1 0,874,000
1.3 74.236.000[
1 1-1 73.000f
1 1,173,000~
19-445.0001
2.879.000 Similar to Aqua [
11,173.000onepeer
1
302.000(
17,246,0001
I
VALUING PRIVATE
COMPANIES:A DEA APPROACH
552.839.000
10,947,000
14.83 4.0001
2,879.000
14,939,000
66.865.0001
109
Table 6-13
Real M.C.
1
1 1 MAPLE LEAF
2 SCHNEIDER
4 AVT FINE
5 FLETCHER'S
9 SLN-RYPE
12 DOVER
I
1 1 7 CAN BREAD
138 PETROREAL
JI EVEREST
3 3 A. L. VAN
L
Appendix E contains the information on the companies with a SIC in the 2000s.
6.1
Some may argue that using different SIC codes in the same analysis is incorrect.
However. there are studies that do in fact use companies with difirent 4-digit SIC codes
in a valuation analysis, such as Renaissance Capitai [ZHUOO]. To test the theory that a
relevant peer group would be found when al1 452 companies were included in the
valuation analysis. an efficient and an inefficient DMU were analyzed with al1 452
companies.
2 1 peea of which 4 were efficient. The range for its market value based on these
p
.
..
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
--
companies was between $3.4 to $142.8 million with a predicted upper bound of $57.5
milIion. Hammond's real market value is $18.4 million. The most similar companies
were Patheon. Sico, and Lassonde with SIC codes 0~2834,2851and 2034 respectively
and market values off 142.8, $71. and 5 17.8 million. respectively. The only other
company with the same SIC code in the 2 1 peers thus tbund was Aastra. Even though this
analysis resulted in a bigger peer group for Hammond. the results did not conclusively
tind a better. tighter range for its market capitalization.
6.4.2 Efiicent Company Wescast Comparecf to a11 452 Companies
An efficient company. Wescast, was also tested with al1 452 companies. This time
around Wescast had 97 companies utilizing it as a peer. The lower bound calculated
based on these peers was $408 million, with Wescast having a real market value of $574
million. This m o u n t was better than the $383 million predicted as a lower bound for
Wescast when it was compared only to cornpanies with an SIC code in the 3700s just like
itself. In this cornparison, the peer group consisted of companies with SIC codes ranging
f?om 2000 to 3900, The most comparative group were companies with SIC in the
3500as,numbering 15 in total. There were also 8 companies with SIC in the 3700s
compared to Wescast. including 6 that had the same identical SIC code. Appendix F
contains the information on this analysis.
6.5
PrivateCompanies
The private companies were analyzed by addng one at a time and re-running the
DEA BCC mode1 with the 85 DMUs. Once similar reference peer units are found, the
private companies can be assumed to have the market values of these retrence DMUs;
thus one has found a market value for the private companies based on the similar public
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
111
- --
companies. The private companies were analyzed in the same manner as the effrient
and inefficient public companies. As mentioned, there are three pnvate companies
having a SIC code in the 3600s,namely. company #55?85, and 220.
6.5. f Inefficient Pnvate Companies
Mer the initial DEA analysis, private company #55 was found to have an
efiiciency score of 0.77 in BCC. It was then analyzed using the procedure applied for
public cornpanies above. Untbrtunately. the company had a very unique dekition of
peer groups. consisting oniy of 4 other companies. ail efficient- The peer group included
Astra Technologies. Ts Telecom, Communication Systems Internat. and Fitly-Plus.Net
International. Afier the analysis, it was found to be most similar in tenns of distance and
efticiency score to TS Telecom. The a c d values found are listed in Table 6- 14.
Private Company #220. coincidentally, had the same peer grouphg as subset
Hmmond above. In its peer group were LG Technologies, Camco Inc, TS Telecom,
Calian, Primetech, Electrohome and Harnmond. it was found to be most similar to TS
Telecom. which had a market cap of $1 1-8 million.
Table 6-1 4 Inefficient-Range for Market Values (Private Cornpanies)
VALUINGPRNATE
COMPANIES:
A DEA APPROACH
Analysis Summary
DEA is traditionally used to measure relative efficiencies and this novel application
expands its tnction. DEA does a firly good job in determining appropriate peers for the
companies when utilizing the BCC model. The CCR model was not employed due to the
existence of negative variables in the model. The DEA-Valuation approach works better
with more DMUs. There must be more than 22 DMUs for DEA to find a correct range
for the company under analysis. Private companies are valued by tinding suitable public
peers and intmng their market values ont0 the private companies.
Although the range specitied by the peers is sometimes quite wide. the method
still does an above average job of classitj6ng market value. For companies that have only
one peer or none (the self-evaluators), the resdts are not binding. A larger data set, if
possible. should be collected. This will increase the likelihood of correct classifications.
--
Cr
value is a si@icant part of any business. be it for selling a business, divorce settlements
or for mergers. Value estimates c m vary widely and to insure a f ' r value for the
business a number of issues need to be addressedValuation is O ftentimes a subjective task performed by valutors. business
analysts. accountants. lawyers and investment bankers. The existing methods for
detennining this value consist of forecasting g o w t h and fture cash tlows, or optimizing
the value of owned assets by liquidation or finding similar companies and adjusting their
forecasts and growth expectations to arrive at a value. The goal of this thesis is to
develop a methodology that evaluates the market value of corporations utilizing Data
Envelopment Analysis to find comparable t - m s and to apply this methodology to assess
the worth of private companies.
A DEA andysis of rnanufacturing companies was studied. Valuation models for
predicting market ranges were developed. DiRerent approaches were constnicted based
on the performance ofthe Company in the DEA rnodel. Ineficient companies had their
most similar peers identified which lead IO a predicting of the market value range. An
upper bound for their market value was determined as well. For the efficient companies.
.- -
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
114
the DEA-Valution method, and quite a number of companies were scarcely outside of
their individual predicted range. Overall, the calculated upper bounds were greater than
the real market value for the inefficient companies 75% of the tirne. For the efficient
companies the Iower bound was less than the real market value 50% of the t h e , however
there were instances where it was not and this could mean that the companies are
undervalued.
7.1
CONTRIBUTIONS
The contributions of this thesis include novel ways to predict lower bounds for
market value for eficient cornpanies and upper bounds for inefficient ones. This work
expands on the use of DEA for finding peer groups instead of its traditional role of
h d i n g eficirncy measures. Valuation models were developed for both pnvate and
public companies and eflficiency scores, difference indicators. market value ranges. and
comparative peers were provided where appropriate. DEA was able to handle the
multidimensional nature of companies because it ailows multiple inputs and outputs. The
work consists of a group of data. more than 500 companies al1 toether, which can be
used to determine the peer groups.
7.2
number offirms in the peer group. The number of DMUs in the analysis varied tiom 22-
--
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
115
85. The number of similar peers should at Least be three in order to conduct an
appropriate analysis. lncreasing the number of firms in the andysis may improve and
tighten the ranges for predicted market values. IdealIy one wants a set with many DMUs
but with very sirnilar DMUs. Having more DMUs would not p u a n t e e more peers
however a bigger group would increase the probability of having more similar
companies.
7.3
AREASOFFUTUREWORK
The methods discussed in this thesis to determine market values represent an
excellent start and show encouraging results. The above methodology cm be enhanced
and made stronger by applying it to different areas and comparing its results with other
methods. In the following list, areas of future research are addressed and exarnined to
advance the DEA-Valution methodology.
Obtain data on ve-year growth forecasts for cornpanies and nin a DEA analysis
Analyze diflierent sectors rather than just manutcturing to see if the results hold for
the other industries.
Attempt to analyze Internet Company valution, need to somehow rank key-men and
technology contributions, and it is necessary to look at other qualitative and
quantitative factors. Although information of this nature was not available for this
analysis, if it were, the possible models would consist of the important quantitative
tactors that Internet start-ups rely on. These include their burn-rate of cash and the
amount of cash or investments on hand. While one wants to irease the amount of
cash investments in the business, and thus maximize it, it has to be in balance with
VALUINGPRIVATECOMPANIES:
A DEA APPROACH
i 16
--
For example, Molson's Canada is a brewer. i-e. Manufacturing, but at the same time
it owns the Montreal Canadians Hockey CIb. Lc. Non-manutcturing. This might
lead to an interesthg study if a lot of companies with varying SIC codes can be
included in the same studyA complete valuation using diffrent methodologies such as DCF. CF. asset
valuation. with al1 the appropriate discounts applied would be the best study. As
well as being the best way to compare al1 the methods against each other
Find and set optimal cutsff values for the distance and the efficiency similanty
when analyzing the minimum and maximum of the inefikient companies. Right
now the distance is set to 1, and only companies that have a distance of less than 1
VALUING PRWATE
COUPANIES:
A DEA APPROACH
117
that is appropriate for peer companes, For example, if the subject Company has an
efficiency score of 0.75 only examine other cornpanies that are within 10% o f this
score,
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
118
References
-
REFERENCES:
[BANK841
Banker. R-D.. Chames, A., Cooper. W. W., "Some Models for Estimating
Technical and Scale Efficiency in Data Envelopment Analysis." Management
Science. ( 1984). 30(9): 1078-1092.
[BERG981
[BOATS1 1
[BOOTOO]
[BOOT981
(BOOT98al
Booth, L., "The Capital Asset Pricing Model: Equity Risk Premiums And The
Privately Held Business," Journal of Business Valuation- Proceedings of the
3rdJoint Business Valuation Contrence of the Canadian Institute of Chartered
Business Valuators (CICBV) and the Arnerican Society of Appraisers (ASA).
Montreai, (1998), 1-36-
[BYRN94]
Byrne. A., Rees, W., "The Time Series Behaviour of Initial Returns on U.K.
IPOs." Journal Of international Financial Markets, Institutions & Money,"
(1994), 4(3-4): 81-100.
[CAMP921
[CBCOOl
[CDNXOO]
[ C H U 9 3j
(CHAR78)
[CHAR941
Chames, A., Cooper, W. W., Lewin, A-Y., Seiford, L.M., Data Envelo~ment
Analvsis: ~heorv:Methodoloev and Atmlications, Boston: Kluwer ~ f a d e m i c
Publishers (1 994)
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
119
References
[CS1891
[DENNOO]
[ETT097]
[~-571
[F[RT971
Firth. M.. Liau-Tan. C.K. --Signaling Models and the Valuation of New
Issues: an Examination of IPOs in Singapore," Pacitic-Basin Finance Journal.
( 1997). 5(5): 5 11-526.
[FISOOO]
[FITCSS]
[-971
mEND88l
[HUNT751
[IBB094]
Ibbotson. R., Sindelar, J.L.. Ritter, J.R., "The Market's Problems With The
Pricing of IPOs..' Journal of Applied Corporate Finance, (1994), 756-74.
(JAiN991
Jain, B.A., Kini, O., "On Investment Banker Monitoring in The New Issues
Market," Journal of Banking & Finance. ( t 999) 23:49-84.
[JENS76]
[KAPL951
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
120
References
[KIM951
Kim, J-B.. Krinsky, 1.. Lee, J., T h e Role Of Financial Variables In The
Pricing Of Korean Initiai Public Offerings." Pacific-Basin Finance Journal.
( 1993, 3(4):449-464.
[KIM981
[KIAM991
Kim, M-. Ritter J.R.. "Vaiuing IPOs." Journai of Financial Economics. ( 1999).
53:4O9-43 7.
(IUE1961
[LELA77
[LEVI951
Levis, M.. Thomas, D.C., "Investment Trust IPOs: lssuing Bebviour And
Price Pertormance. Evidence From The London Stock Exchange," Journal of
Banking & Finance, ( 1995). 19(8): 1437-1458.
[LITT591
(LIUN971
[LYON891
[ME001
[MERC971
(MORR921
[MULL901
Mullen, M., "What is a Brand Name or Trademark Really Worth - How Can
that Vaiue be Measured?" Journal of Business Valuation- Proceedings of the
2nd Joint Business Valuation Cont'erence of the CICBV and the ASA,
Vancouver, (1WO), 203-2 12.
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
References
[NEED861
Needhant. J., Young, A., "The Leading Edge In Mergers And Acquisitions."
Mergers & Acquisitions. November I December 1986.
[OANDOOI
[OVEN591
(OVEN721
Ovens. G.. "lntroduc tion to the Business Valuation Field." Solvin- the
Valuation Problem, Toronto: h e CICA (1972). 7-9-
(PIES951
[PRAT931
Pratt. S.. valu in^ Small Businesses And Protssional Practices, znded.,
Homewood: Dow Jones-Irwin (1993).
[PRAT961
Pratt, S.. Valuing A Business: The Analvsis and Amraisal of Closelv Held
Companies, jrded.. Homewood: Dow Jones-Invin ( 1996).
[RITT841
[RUBA961
Ruback. S.N.. Richard, S., --The Market Pncing Of Cash Flow Forecasts:
Discounted Cash Flow vs. The Method Of Comparables." The Bank of
Arnerica Journal of Applied Corporate Finance. ( 1996) 8(4):45-57.
[SEDAOOI
[SEGU971
[SIMA001
Simak, P.C.. Paradi. J-C., "How To Estimate The Market Value Of A Private
Company - A DEA Approach." University of Toronto working paper (1999).
[STOCOO)
[STOU981
[THAN921
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
i 22
References
VALU~NG
PWATE COMPANIES:
A DEA APPROACH
123
Bibliography
BIBLIOGRAPHY:
AL,AL. Lerme, CS., Seiforci, L.M., 'romponents of Eficiency Evaluation in
DEA." European Journal of Operationai Research. (1995), 80:462-473
Berger. A.N.. Hurnphrey. D.B.."Efficiency of Financial Institutions:
International Survey and Directions For Future Research.' European Journal of
Operational Research. ( 1997). 98: 175-2 12
Brocken. P.L.. Charnes, A.. Cooper. W.W.. Huang. Z.M., Sun. D.B.. "Data
Transfomation in DEA Cone Ratio Envelopment Approaches for Monitoring
Bank Pertormances," European Journal of Operational Research. ( 1997). 98:
350-268.
Campbell. I.R.. Low. RB., Murrant, N.V.. The Valuation & Pncinrr of
Privatelv-Held Business interests, Toronto: Canadian Institute of Chartered
Accountants ( 1990).
Chi1ingerian. J.A., "Evaiuating Physician E fficiency in Hospitals: A
Multivarate Analysis of Best Practice." European Journal of Operational
Research. ( 1995),80:548-574.
Cook. W., Roll, Y., Kazakov, A.. "A DEA mode1 for Measuring the Relative
Eficiency of Highway Maintenauce Patrols." Journal of Information. (1990),
?8(2):113- 124.
Dauderis, HI.Financial Accounting: An intro to decision rnakinq, Toronto:
Holt, Rinehart and Winston of Canada, Ltd. (1987).
Droms. W.G.. Finance & Accounting for Nonfinancial Manarrers. al1 the
basics vou need to know, 4'h ed., Reading: Addison Wesley Longman Inc.
( 1998).
El-Mahgary. S., "Data Envelopment Analysis: Visualizing the Results,"
European Journal of Operational Research. ( 1999, 85 :700-7 10.
El-Mahgary. S., -'Data Envelopment Analysis - A Basic Glossary." Journal of
Operational Research Society, (19 9 3 , 8(4): 15-22.
Mikkelson, W. H., Partch, M.M., Shah, K., -'Ownership And Operating
Pertbrmance Of Companies That Go Public," Journal of Financial Economics,
(1 997) W:281-307.
Morck, R., Shleifer, A., Vishny, R.,"Management Ownership And Market
Valuation: An Empirical Analysis,?' Journal of Financial Economics, (1988)
20293-3 16.
-
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
124
Bibliography
[PAPA971
Papagapoou, A., Mingers, J., Thanassoulis, E.,"Would You Buy A Used Car
With DEA?" Journal of Operational Research Society, (1997), 1O( 1 ): 13- 1 9.
[PAST97]
[TATJ971
(TRAC991
GLOSSARY
Another name for the efficiency measure as obtained tiom the CCR model.
also known as eff?ciency
The act or process of detemuning the value of something
BCC Model
Beta
Binomial
Model
Binomial Tree
A tree that represents how an asset price can evolve under the binomial
model
Black-Sc ho les
Model
BPS
Business
Enterprise
Business
VaIuation
The act of determinhg the value of, or the estimated value of, a business
enterprise. or an interest therein
Cal1 Option
A hancial instrument that gives the owner the option to buy an asset at a
certain price by a certain date
Capitalization
Rate
The required rate of retum that is used to convert income/cash tlow into
value. The inverse of the cap. Rate is referred to as the multiple or
multiplier.
CAPM
The capital asset pricing model is a model relating the expected retum on
an asset to its beta.
. .
126
Giossary
CCR Mode1
The CCR ratio mode1 calculates an overall efticiency for the unit in which
both its pure technicd efficiency and scale efficiency are agzegated into a
single value- The obtained efficiency is never absolute as it is always
measured relative to the tieid
C losely-Held
Corporation
A term used to describe a corporation whose shares are neither listed for
trading on a recopized stock exchange nor traded in an over-the-counter
market- Technically it has less than 5 1 shareholders (Canada).
Comparable
Firrns
Constant
Returns to
Scale (CRS)
DEA
Depreciation
Discount Rate
Discounted
Cash Flow
Dividend
DMU
Decision Making Unit. This term emphasizes the fact that the focus was not
on profit generatuig enterprises. but rather on decision making entities
DRS
EBIT
EBITDA
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
127
Glossary
Economies of
Scale
Efficiency
Efficient
Frontier
Enterprise
Value
The market value of equity plus its book value of cash minus cash.
Envelopment
Surtce
EPS
Exogeneously
Fixed Factor
Fair market
value
Financial Ratio
Financial Risk
The portion of overall business risk associated with the financial structure
of a business
Forced
Liquidation
The sale of the assets where an imrnediate cessation of the business and
disposition of assets is assumed on an 'as is/where is' basis. The forced
liquidation of assets at auction.
Going Concern
Going Concern
Value
The present value of al1 future benefits expected to accrue from ownership,
where a business operation is expected to continue to operate either for a
finite period, or indefinitely into the tture
VALUINGPRIVATE COMPANIES:
A DEA APPROACH
128
Glossary
Goodwill
The diEerence between going concern value and the sum of the net angble
assets and identifiable intangible assets
Degree of similarity between the units in the field
Indicates that an inefficient unit is made efficient through the proportional
reduction of its inputs while its output proportions are held constant. The
CCR mode1 yields the same efficiency regardless of whether it is input or
output-oriented, this is not the case with the BCC mode1
Intrinsic Value
IPO
IRS
Lambda Values Results Liom a DEA analysis that indicate the level of contribution of
efficient firms to the virtual DMU created by the projection of an efficient
of an inefficient DMU ont0 the fiontier
Liquidation
Liquidation
Value
Liquidity
Market
Capitalization
Market to
Book Value
WB)
Marketability
Discount
Minority
Discount
The reduction tiom the pro rata portion of the en bloc value of the assets or
ownership interests of a business as a whole to reflect the disadvantages of
owning a non-controlling interest
VALUINGPRNATECOMPANIES:
A DEA APPROACH
129
Glossary
Multipliers
Net Book
Value
Nominal Rate
of Return
Notional
Market
Open n market
OutputAn inefficient unit is made eficient through the proportional increase of its
oriented Mode1 outputs, while the inputs proportions remain unchanged
Peer Group
The set of efficient DMUs against which inefficient units are compared.
The efficient targets are a linear combination of these peers.
Piecewise
L inearity
Price
Price to
Eamings Ratio
(Pm
The relationship between the current price of a stock to the eamings over
any earnings period. The PIE is calculated by dividing the current stock
price by the earnings per share or dividing the market value by earnings per
share for the most recent 12-month period. It is used by investors to gauge
the market demand for stocks. Companies with a high P/E have greater
earnings potential than low PE stocks, but also are more volatile and
somewhat riskier
Privately-Held
Business
Interest
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
130
Privately-Held
Company
Public
Company
Put Option
Rate of Return
Real Rate of
Return
Research
Boutique
RTS
Returns to scale
Rules of
Thumb
Scale
E fficiency
Shareholders'
Equity
Slack Variable
Technicai
Efficiency
DMU only to other units of similar scale, computed using the BCC mode1
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
131
Glossary
Threshold Rate
of Return
Treasury Bill
Valuation Day
(V-Day)
The date proclaimed in the Income Tax Act (Canada) on which capital
gains taxation in Canada came into et'fect. With respect to privately held
business interests. this date is December 3 1. 1971.
Value
Variable
Variable
Retums to
Scale (VRS)
Warrant
Weihts
Results fiom a DEA analysis and indicates the level of contribution each
variable has in the determination of the DMU7sefficiency score.
Working
Capital
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
132
List of Abbreviations
-
- -
LIST OF ABBREVIATIONS
BPS
Capex
Capital Expenditure
CAPM
CDNX
CF
Comparable Firm
CF1
Cash Flow
CF0
CICA
CICBV
CRS
CS1
DCF
DEA
DMU
DM
EBITDA
EPS
ESOP
EV
Enterprise Value
FCF
GAAP
IPO
Initial Public O f f e ~ g
---
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
133
List of Abbreviations
IRS
M&A
m
l
Market to Book
MC
Market Capitalization
NI
Net Income
NS
Net Sales
W C
OP
Offer Price
Pm
Pnce to Earnings
PER
POP
PPA
RA
Redundant Assets
RFR
SE
Shareholders' Equity
SEC
SIC
TA
Totai Assets
TL
Totai Liabilities
TSE
VRS
WACC
APPENDIX A
Public Companies DATA 1998
SIC 2000-3999
-- - C
TOTAL ISHARE~~NET
NET SALES CASH
LIAB-EQUITY
-ME
/REVENUE FLOWS
(in thousands)
206,480
40,532
165,948
13,952
207,067
1,995
1,779
1,181
598
-283 (
TOTAL
ASSETS
1 A. L. VANHOUTTE
L 2095
A.E. VENTURES LTD 3714
1 3661
11
If
PW
1 55.6 17
44,467
9,435,500
9266
108,383
563,100
3,019
6,577
7,100
340
76,024
11,173
79,826
140,743 1
2,879 1
87.1 12
434.862
15205.956
84,550
1-521,000
1.77 1,800
1,238
1,956
15,160
2.146
75-913
1,592
126,085
27,508
2 10
10.887
357.966
6.729.876
63220
1.1 85,000
8 14.000
886
613
2.5 1 3
1.945
33.760
696
72.58 1
767
97
76.225
76.8%
8,476,080
2 1330
336,000
957.800
352
1.343
12,647
20 1
42,153
896
53,504
26.741
113
4,625
-9,479
612.784
-15,700
-59,000
25.900
-308
-1 10
-13.1 16
-667
596
164
6,732
-6.73 1
-50 1
5.6 13
17.425
-3,522
18.732
-2.395
-1.016
78,478
7,62 1
349.056
24.354
1 1,962,346 l297,7S 1
196,136
-3,467
l372,OOO
74,000
1,085.100
138,200
1,927
-227
107
3,123
-12,736
50
O
-6 18
3,609
49.655
25 1
6,188
1 1,078
1 18,668
2277
-6.325
387
-485
- 1,869
AQUARISJS COATIN~3479
ARCTIC GROUP INC 2097
7
2.1 15
38.002
2.09 1
36.157
--
- -
.-
2,497
1.171
187.708
268,2061
523
35,053
240
70,004
-1.818
627,692
605
82
538.325
9,399
7337
4,196
17.73 1
478.62 1
4,209
78,6 19
2,s 17
--
-803 1
-1,881
10,768
27.3621
-625
2-810
-374
1,540
1,885
765
-43 2
-1.21 1
-65.6 1 1
5,432
-697
-3,4601
289 1
1 14,774
-3,014
69,755 1
-5,143 [
1
4.3 74
9.574
-3 67
1
769 1
469 [
229.976 f
27.984
777
66.903
1,875
43.404
2,498
25.078
1 79
2,176
77 1.506
69.155
6,280
63,870
25.185
33532 1
1,356
1 73294
2,955
-563
5,032
-3 1 1
2,240
-100
- 18,826
-3 72J
-9 13
-7 I,29 1
6,808
157
-1 16
2,3 84
129,087
-1,723
77.367
-1,3 19
-84
- 1.65 1
51 1BLUESTAR B
A-!
3692
44,4041
133,332
83,377
49,955
-8.816
135,366
-3,62#
521 BOLIDEN LIMITED 3339
40 1.3 121 2-831370 1,828,276 1,002,994 -1 16,239
1,618-176
4,195
53 ( BOMBARDIER [NC 3743 7..502..049 ,1 L 0-575200 7,685,900
2
.
8
8
9
.
3
0
0
420200
8
.
5
0
8
.
9
0
0
9
19-70f
= - -.
-54[BORDERCHEMICAL 28 19
3,7251
IO,174
1,515
8,659
352
14,387
79(
55 i BP AMOCO P.L.C
29 1 1 619,712,0001 129,775.100 65..6O0..161 64,174.939 5.006,708 128.595.606 14,732,,17- 5
56 BRAMPTON BRICK 4325 1
33,4661
64,523
20,389
44,134
7,042
34,9491
10,335
57 BRICK BREWING CO 2082
1 5,421 '
19,835
7,775
12,060
694
19,9231
1,57C
58 BUDD CANADA W C 3714
116,772
121.603
40.977
80,626
7,359
349,036) 13.72I
59 BUHLERIMlUSTRlE 3523
66.665
61,139
16.349
44.790
7,1 13
89,1941 11.194
60 CAE INC
3812
917,070
928,179
651,726
276,453
70.236
922.369
87,541
G I 1 CALIAN TECHNOLO 3663
33-191
45,533
22.936
22.597
1,408
79,928
3-00;
621 CAMCC iNC
363 1
100,000
192,375
117,134
75.2411
8,979
565.305
24.1 1 S
63 iCANADA BREAD CO 205 1
449,753
338272
107.736
230.5361
2,896
508,445
28,883
46,8981
97,932
30,659
ij52l
100.589 1
64i CANADJAN BANK ~d 2796
8.50
67273f
370
10,155
58 1
65 iCANADIAN MANOIR 3433
3.5 1 1 '
11.815
9,474
2.34 1 1
426.808
1 13,940
38 1.988 1
28,003
495.928
29 1.544
39-501
66j CANADIAN MARCO? 3663
177.740
645,753
445,882
199.871
26,066
1 ,000,700
52.0 11
67 1 CANAM MANAC GR^ 3441
68f CANFIBRE GROP LI 263 1
9 1.809
362265
33 1.5 15
30.- 749
-6.334
284
-3,564
27415 1
1.864295 1,438,247 426.0481 -203.731
1.632.273
69 !CANFOR CORPORAT 2421
'7 1,921
38.300
10,159
70i CANGENE CORPORA 2836
33 1,802
30,1831
1 1,000
45.784
15.60 1
7 1 i CANHORN CHEMICA 2869
1,7921
6.3 82
4,858
1.524(
-2.701
4,155
-1,827
-46
8.490
85
4,825
72i CANTOL LIMITED 2842
2,108
6.508
1,683
73 ;CARBITE GOLF [NC 3949
24.054
1.424
9240
12.271
11237
367
1,034
78
-2,401
212
-2.5 1 5
2.065
279
67
74 i CARDIOCOMM SOLI.384 t
75 1 CASCADES INC
26 1 1
595,430 2485.000 1,867,000 6 18,0001 45.000 7.527.000r l S5.00C
1.469.1951 1 32.4 1 8
571.417
76 CCL KNDUSTRIES IN(2844
44.394
1 .128,531
857.1 14
698,733
471
296
109
1861
-206
289)
-1 la
77 : CD ROM NETWORK 1 273 I
78 ;CEAPROINC
1 2844
10.546
- 1.654
300f
-1.088t -14,260
479
1,5671
79 CELESTICA MC
3674 2,854,867 2,513222
1,193,561( 1.3 19.66 1
-74.442 4.990.12 1 1 130326
3081
-1,403
80t CELL-LOC INC
3663
1421
2,271
-1,441
28,669
2.4 13
3.835
28 1
4.2
1
2
-191
2,8981
2,765
7-110
8 1 CERAMIC PROTECTI 3259
O
-518
-133
-79
3.053
3,1863
82!CFE MDUSTRIES
3999
7,887
5.368
150,2411
10,106
42,542
22,4651
65,007
83 :CHATEAUSTORES 233 11
45,927
15.001 1
-150
1.106
-150
8.792 1
54 j CHEMBOND LI
2.9 17
9,898
- - MIT^
- 289 1 1
85 :CIE-NERGY PLY-FOI3086
918
542
528
14
-873
2.066
436
86 !CIMATEC ENVIRON 3564
4,040
1.9 10
283
1,627
148
1.493
232
111.912
42.303
642.714
87 ' CMRAM LNTERNATI 3652 L,0%,69
433,063
423,7I1
856,774
1,472
1 ,O31
2 1.464
88 1 CIRCA ENTERPRISE 366 1
6,601
1 1.660
5,113 7..6
1.942
22.00 1
5,330
1,561
89 1 CIRCUIT WORLD COI 3672
8,962
289
14392
3.841
476
6 1.320
50,508
65,775
15,2671
90 i CLEARLY CANADIA:
2086
5,135
9 1 :C-MACINDUSTRIES 3679
29 1.127
607.83 1
347,s141 260.3 1 7
25.4 18
642,983
47,363
92ICME TELEMETRIX 384 1
27.749
5.024
1,295f
3.729
4,139
1.250
-3.826
93 !CML INDUSTRIES LI 2621
7,l09
19,702
12,2241
7,478
21
33.152
1,159
-1,511
94 I COM DEV INTERNAI
3663
252-315
105-496
56,6731
148,823
- 1 8.439
175,298
95 !COMMUNICATION 3663
3,855
4.635
1-1 101
3,525
424
8.350
717
1 96 1 COMPRESSION & EN13695I
7171
763I
753 1
101
-1.2881
791 1
-278
97! CONCERTINDUSTRI 262 1
48,917
57.951
35.977
2 1.974 1
1.46 1
12.672
3.27 1
36
-460
-71
1
-53
8
708
636
936
98 j CONSOLIDATED EC( 2844
24,7801
1,974
-2,984
8.124 1
22.3 17
14.193
991 CONSOLIDATED EN7
2875
5.178 I
1,368,150
97,730
184,255 47.985
204,373
1.63 1.982 1,447,727
1 O0 i CONSUMERS PACW 3221
14,349
4,379
2,778
3,830
24,788
28.6 18
1 O 1 1 CORECO LNC
1 3674
44.923
378.139
-2,781
43.789 I -46,646
170,934
1 14.724
362.8 16I
1 03i COREL, CORPORATI~3575
z-
- - 7 - - - - - -
-.
T.
.-.
SI
- -- - -
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
137
104 1 CO-STEEL MC
3312
385,413
330,647
60.85 1
3,765
7.6 16
12,657
8,637
5,798
35.889
1.956
560,656
1,705,264
1,829,043
3 1.709
9.222
1,732
3.608
27,797
12,460
45.5 1 1
2,369
681.894
1,032,327
1333.12 1
2 1.973
5.636
674
1.094 I
14,5631
9.024 1
18,897
1.468
427.1 16
672.937
495.922
9.736
3,586
1.O60
3514
13,234
3.436
26,6 14
901
254.778
13.124
28.133
1.970
13,304
39.6 15
7,806
5,159
10.309.
7,Ol 1 '
8.145(
19,306i
-7951
12 1 ID~VERSITYCORPOR~2899 !
132 i DOFASCO iNC
1 33 121
1.O96
1,603,890
1,143
3,377,100
CON
CON
INTERNA
INDUSTRIE^
--
673
470 1
1.548.600 1 1,828,500 1
-80.745
-1 1.1 98
2,726
1,952
-5 13
-849
-824
1.O48
3,049 1
1
-722 1
29.307 1
107j10
67,969
4,160
865
-3 19
6641
-79
. - 1-
717
1322
4.685
-2.461
73.649
-
-6.052 1
1 A87 )
1.637 1
42.464 1
-7 1
-5.460-3 82
-129
-3,167)
175.700 1
983 )
2.982300 1
-484
4 16, IO0
329,lO 1
ll3,908/ 215.193
4,030,000 2,342,000l 1,688,000
3,456.865 ) 2.068,543 ) 1288.322
460.790
137,2081 223,582
54,546
19.766 i
74.3 12,
62 1,9771 1209.668
t ,831,6451
9,4331
7821
8,65 1 )
43,41 1
80,6551
37344 1
3611
996 1
635
3.064 1
9.22 1 1
6.157
68.9 16
39,775 1
39.14 1
16,8671
9,638 1
7.229
100.056
159.506
59,450 1
7.870 1
1.181
9.05 1 1
14.0 1 1)
50290
5,085 1
587
4,498
1.570 1
163;
1.406
125,4051
47,763 1
77.642 1
i.61~1
3,292 1
4,907 [
883
i .O05 1
1,888)
38,1931
45,013'
83.206
4,4071
4.674
9,080
-17
52 (
35
355.000 i 3.175.700
2.530.700
64,204
83.161 1
147,365
1 -050(
2,608
3,658
20.732
3 1.2671
5 1.999
1.5 1 1,432
2.1 52766
46.50 1
13.5071
2 2 141
2.585 I
24.553 '
24.099
82.783 1
4.3361
656.1 75 i
14.725
74,000
329249
26.985
4270
298.772
-2.062
4,364
189
84
5,116
1.817
22.568
93 /
6.5 12
-1.096
-1 -559
1 1.1 15
99
51
10.318
147
18
298.600
3.656
-3.457
-4.597
3 50.798
2,348.000
2295.805
766,607
131.987
2.025.072
3681
136.618
3,766
42.104'
2.0 1 2
168.685
1.336-7561
33,477
641.73 1
34.1 70'
58.000
450,728
47.477
8.0 I 1
282,480
-1,750
7.508
236
360
22,734
-4.342
97,873
3.208
1 6.447
3 14.496
16.042
3.3301
161
53.1 74 (
8.627
928 1
-337
OI
-527
2 1.238
1 1 8.045 /
3,027 1
311
330
958 1
13.185
177.585 (
11.7891
4 12
151;
17
284.1001 -105,000
394.0151
11.377
4.5221
-1,762
44.948 i
-4.3 71
559,473 :
19,557
2.425.988
1,153,587'
29,250
702,572
728.64 1 '
6,044
424.8 15
424.946i
232061
277.757 1
117.33r
3,163
223,762
VALUINGPRWATE
COMPANIES: A DEA APPROACH
138
IPL MC
IPSCO iNC
IRWIN TOY LIMITEI:
IVACO iNC
M CHECKM~TCO
JANNOCK LIMITED
JEMTEC MC
KAFUS INDUSTRIES
KASTEN CHASE APP
KELMAN TECHNOL(
KING PRODUCTS IN4 366 1
KRYSTAL BOND INd 289 1
LABOPHARM INC / 2834
~
LAFARGE C O R P O 3241
LAMBUS ENTERPRI~3679
LANDOVER EN ERG^ 2869
7.739
1,398
44,269
4,330,770
1,729
2846
'
3,337
24,493
307,970
496,389
7374.103
20,940
13.3 10
56,024 1
1374236
623
23,227
1,570
.310
-5.44 1
-5.108
-6.027.
36 1.68 1 1
473/
-1521
11,018
3,070
1-301
3,759,953 1
88
7
4.96 1
-2.556
-4039
6 14.461
-179
-47
6,39 1
3,376
3,996
7,041 1
8.1561
171.986 1
15.888
423.290
3041
-4, I O8 1
1,4171
-7.72 1
604
33,393
-9,463
84.3 851
40,142
4,742
22,106
35,667 1
3 1.232
147,743
45,6 16
998.3 12
95 1
-3.64a
2.822
- 1339
1,570
29,672
916
150.854
1210(
1,092
118
5,596 1
2.63 7
2,959
168.735 1
84,927
83,808
23 1.548
1 76,4 12 1
407.960
3,675,300 ( 4,945,400 1
8.620.700
18.599 (
5.532 1
24.13 1
1 7,764 1
8.766 1
26.530
96.0 1O 1
187.4 13
9 f -403
399.256
1,406.595 1-007.339
1.862,
434
3,296
-338 1
-2,548
1 2724 f
32,336 1
506200 1
4.289 1
1,432 1
3.80 I 1
-23 -257/
5,868
1,372
174.84 1
426,942
9,190,800
6,640 1
16.537 1
1 1 1.829 1
328 1,464
6,350
-2,096
20,084
4 1.9 18
803.498
-2,898
2-41 5
13,498
44280
-277
7,309
12.80 1
2,463
8,099
2.974
20,896 1
4.46 1.187 2387.763
243
634
6
3 16
24.0 18
4.799
14.625
37.899 1
17.037
193,542
38,659
669.64 1
17.627'
1.423
10,629
20,858 1
8.88 1
2 1 -556
2 - 7 71
216.3 5 1
- ---
-5,492
5,636
17,932
2,173,413
39 1
-464
4 1,445
1,797
9,428
923
2,093 1
32.0 17
874
-9
5 1,589
1,166
-232
-1.1 17
134
- ---
MCGRAW-HILL RYE~373 1 1
39.533 1
62.3481
11.1931
41.1551
5.2071
68.3 56 1
8.094
MDC CORPORATION 2759
242,559
740.0 t 3 1
54 1.869
198.144
14.1 O8 1
476.727 1
45,072
MEMOTEC COMMUb 366 1
4,953
30.28 1
15,913
14.368
-2,360 1
37,926
864
174.896
15,009
9.012j
153,0191
180.265
138.46 L
333384
MERCER MTERNAT 26 1 1
498
442
-754 1
597
1.3941
2,869
1.99 1
MERCH P E R F O3714
~
1.107.126
75,783
1
05,122
f
L
.
O
6
1
-9601
1.70
1,037
1,370,334
2.762.996
METHANEX CORPO~2869
143,412 -224.5 IO
307,087
720,969 I.087.6571 -366,688 1 408,920 I
MICROCELL ELECC 366 1
MILLTRONTCS LTD ,3823,
222.793 ,
105.188 1
15.320 1
89.868 1
10,006 /
125,053,
14,499
34,343
4,6 19
35.958
6,456
29,502 1
2.63 8
258 MITEC TELECOM IN(357 1
56,40 1
888.500
125,600
802300
435,500
9 1.900
1,237,700
359 MlTEL CORPORATIq 366 1
1 ,257,378
260 MOBiL COWORATI~
29 1 1 105,489,100 65.66 1,593 37,448,947 28.2 12,646 2.6 17.003 80.076.612 8,145,883
29,118
3,922
17.796
13.362
2,865
33,407
3 1,158
26 1 MOBILE CLMATE 3585
4,173,847
133,682
1,713,937
937,061
-841,413
2,650,998
262 MOORE CORPORA'TI{2761 1;428,580
1
'
- - -
CI
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
140
3663
2426
8,125,481
929,923
1,966,825
2,775,000
733305 1 133.620
1,493,O1 1,282,000
- 183 18
48,000_
1,620,620
2,426,000
323,961
260,000
-44 1
1,247
1,439
-3 66
274 NORDEX EXPLOS~~E( 2892
530
738
2,177
-1.317
159
-3.879
-1,201
275 NORON TECH LTD 363 1
7,029
4.450
57 1
276 INORTEL NETWORKQ 3661 50,495,406 30$04,406 12,542.879 17-761.327 -824,725 36,991,685 1,397,578
17.704
333
14.059
85 1
11.001
Z%o6or
7,036
277 NORTH AMERICAN f 3829
-59
1
2.306
-23
1
6
2,698
2,467
7.100
378 NORTH AMERiCAN 1i 3841
54,256
4.58 1
2.1 15
7.884
21295
29,179
279 NORTHSIDE GROW 1 3714
14,056
-3.1 92
-3 96
2.195
-133
693
560
380 NORTHSiDE MINE^ 3533
4,066
82.76 1
7,388
1.3 87
25.841
3 0,959 1
69,109
43,268
28 1 NORWALL GROUP 262 1
1 7O.OOO
2,075,000
18.000
1,838,4471 3,580,000 2.O68,OOO 1,5 12.000
182 NOVA CHEMICALS q282 1
612891
16.033
67.7321
9.807
34,3501
283 NQL DRILLMG TOOI 3533
102,082
7 1,671 1
34,030 1
5266
284 NTS COMPUTER S Y 3571
~
48,09 1
38.093
18.55 1 1
19,5421
3,593
1
8.033
9
1,653
4.763
C
O
R
P
O
R
~
3879
~
4
1,449
6
1,530
36,407'
25.1
33
W-GRO
285
69.887
16.039
1 2,242
286.013
82,129
2.3241
801 1
1,701 1
997
3,007
683
287 OMNI-LITE WUSTIg 3297
23,659
-2,745
158 1
-3,003
3411
7.828
8,169
2.93 1
288 1ORALEE GROI.JP, IN1 3834
2.247
20,186
-3,286
4,544
5,234
9,778
40.576
289 (OTATCOlNC
1 3533
784
47,950
-8 82
4.767
15,142
1 0,375
3,878
290 1 PACIFIC E-LINK COq 3669
10,611
213
1.490,
-59
6,048
4,558
771
29 1 / PACIFIC RANGER Pd 3443
-3 80
9,3 12
-522
9,1175,772
3345
292 1PACIFIC SAFETY PR( 3 842
5,040
-4-4
12
96.9
-39.299
1
5
4,373
38,736
34.364
5,466
293 [PALLETPALLET INd 2448
VALUINGPRIVATE
COUPANIES: A DEA APPROACH
141
PRIMETECH E L E 3672
~
PRIMEX FOREST PR^ 242 1
PRoCESS CAPlTAL 4 291 1
143.365
56,925
5,667
2,980
10,3 74
78-451
205,602
974.676
92329
3.82 1.070
6366
56,143
99,053
3,455
3,929
2,876
36,374
154.315
103,223
7,372
5,900,722
24369
--
14201
33,341
1.8 IO
1,737
402
16,463
26,004
8,145
33 1.7328
3,497,957
15.507
4 1 -942
65,7 12
1,645
2.1 92
2,474
19.9 I 1
128.3 t 1
95.078
7,039
3,402,765
8,863
1 1,074
6,447
-877
129
936
-7.497
12,886
-24,071
-1,900
245,052
851
IREX
1
1
89,O 13
168.079
1 50
6,766
3,144
10,586
175,656,
2.0041
3,244
5,848,564
36,090 1
1 1.2011
12.45a
571
19a
966
4,807
16,766
-22.854
-1,724
653,148
1.474
1 1,908!
-3.222
9.654
67,927 !
1.732
33,159;
1 1.3621
-2.538
398,004 1
27.007
14.783
450,942
2.788
78.712
3 1392
670,669
5,799.
1,944
196.8 19
1.050.1 O3 (
4.6 16
I9,36 1 i
3,333
70.1841
II
-3 8
2.3941
-11.239
735,4751
46.800
8 1 7355 (
63,972
905,908 1
1 8,180
104,8031
14,047
12,5951
-573
9.7 14.000
90.000
1.562.5381 -2 12,843
4391
-2.42 1
O1
-51 1
2.702 1
165
43.175 (
1.509
177
3.357
-289
O
95.05 1
769.0 17
-2.587
799
13-252
109.438
1 1.067
206.086
3,193
26.400
251
-4.30 1
40.79 1 1
-8.175
2 1-678
142.3 84
-1.1 12
1 1.466
10.407
199.433
10-927
509.6 10
11,283
76,023
VALUINGPRIVATE
COMPANIES: A DEA APPROACH
142
CO-4
-.
5 1,2341
5.1561
1-981
33,477
1,368
36,678 (
2,452
24069
2369
2.473
6,678
3,764i
32.529
29,44 1
4,340
45.35 1
8.95 1
6637 1
1 5.43 8
9,450
10.082~
2,496.900
583.665
4,576
7i
1,999
2.376
1 9.938
5-531
2 1.146
27.465
1.887
493
36.799
-2.3961
14.149 1
453
14.003
1.964
25.4 131
3.419
45,125
4
-33571
13,806)
1,535,200
580
96 1.700
-573
'
3.82 1
382
- 1,778
- 14309
-1221
1.8141
- 13231
-13,487
159
2,880
82 1
4.422
-8,9586 1,100
-94
82.0 141
7.3 181
3.408
2.626
5.064
83.8 1 1
4.352
633
5.778
78.064 /
16.5 171
3.6241
9.577
1,422,900
5.371
1.171
- 1.373
- 13.394
-1.302
5.556
425
-13.016
425
4.7 1 1
1.555
9.388
'>
-5,789
207.000
-94
6,006
20,142
-8.949
2.242
12,193
8.9 12
8,662
249
14.889
3,156
I 1-733
-9,3 131
473 i
~d2452
80,615
3,323,827
33,828
136,498
199
441
5371
1258,176
11,221
12,878
793,613
29,840
f 92.1 14
~ ~ ~ ! V I T R ECAPITAL
OUS
329 1
4381VIVANT GROUP INC 2086
J39!WAJAX LIMITED
] 3569
440f WARWICK COMMUB 273 1
44 I i WE X-L. HOLDINGS C 3699
442 1 WESCAM MC
3663
443 j WESCAST INDUSTRI'3714
t14) WEST FRASER TIMB.26 1 1
445 IN'FI DVDUSTRES L T ~
3567
446 ( WI-LAN INC
3669
447!WRJPAK LTD
2656
3481WOOD COMPOSITE 3089
449 1 WORLD SPORTS ME: 3949
4501EROX CANADA iNC 3579
45 1 XILLLX TECHNOLOG 3 84 1
452 ZEMEX COR~ORATI~
3299
6,325
8,162,777
29,284
409,914
3,O 16
402
23 13
3,245,637
12,415
4,055
256,957
47,0821
203.096 1
t
914
876
120,082
240
17,083
76,678
574,300
880.03
9,697
329,339
3 18.500
9,334
4,932
2.56 1,066
22,149
86,904
156
6,1691
-1,9L9
6,03720 2,125,547 -1 19.946
2,113
20.065
9,219
-22244
128296
281,618
-4,022
-979
7,03 8
-278
680
-3,096
17
580
1,933
153.10 1
1,867,593 1,378.044
8,743
171
3,672
2327
-2,342
1.728
89,449
167,508
34,156
4.129 /
100.753 1
1,689i
254.7371
122,150'
2733
4,504 [
538 1
1.075 1
2,575,6 18 2-061.952 1
16.596
1,4001
228,628
102,849
VALUING PRIVATE
COUPANIES:
A DEA APPROACH
284.483
-1,743
374.735
4,047
-8.208
-256
-108
166
220,582
1.774
1,904
42.37a
13,753
7.0 1 1
12,953
102,343
833
10,825
44.425
204.8 10
1,316
18,956
2,520
-476
20 1.827
796
1,338
42.4 19
23 7.773
897.63 1
11.338
6.333
132.587
1.77 1
-536
5 13.666 1
15,1961
125,7791
230
-1,611
9,469
36
-1.647
2,865 1
47,856 (
5,6251
-1,0171
-6,292
16.138
-732
-1.838
69,826
-2,984
8,240
S,Ol 7j
2841
992,248 1
46a
-1,517
203 13
79
-1,624
8,494
67.277
161.198
-1 19
-5,ZM
37,018
-688
-1,814
158.732
5.385
2,865 1
956
1.431!
637,973 4363461
505 1
130ll
1.523 1
1851
873 121
44,893 1
42,594 1
280.3671
2,108,930 ( 1211.799 3
16,541 1
37.88 1 1
8.022
35(
17,100.5 191
21,8141
292,357)
4.905 1
0[
2,474
3,102,890
I.I90!
593
79,9041
284,861
1.863399
53.595
5.414/
336.3971
851
1231
1,550,4 t 2
5,602
159,560/
-2,5 1 1
7,128
APPENDIX B
DEA Mode1 Results and Market
Value Analysis of the Inefficient
No.
0,9208 FIFTY-PL
0-0472 FIFTY-PL
0.0676
0.9285 PRIMETEC
0-01O1
0-0210 FlF7Y-PL
0.2254
0,621 0 FIFTY-PL
0,9525 PACIFIC
0-8776 PAClFlC
0,8591 LEITCH 1
0-3099 INTERNAT
0.0230
0.0027
O. 1375
0-4001 PAClFlC
0-0417 SNlF SEC
0-4090 BALLARI:
0.0737 COMMUC
0,8450 PAClFlC
0,1798 FI-PL
0-7857
0-1655 LEITCH 1
0-8155
0.0567
0.0428
0.4507
0.6009 PRIMETEC
0.0104
0.5594 PAClFlC E-
0.0279 LEITCH TE
0.2612 LEITCH TE
0.8763
0.761 7 FIFIY-PL
0.1285 SNlF SEC
0.8379
O. 1523
0.0760 LEITCH 1
0.41 92 LEITCH 1
0,0012 FIFIY-PL
0.4233 FIFlY-PL
0.8796 LEITCH 1
0.71 53 PAClFlC
0.1203 SNlF SEC
0.7437 LEITCH 1
0.2240 FIFiY-PL
0.0775
0.4445
0.4871 JEMTEC Iii
0.5220 PRlMEtEC
0.0449
0.0339
0.1 320
0.0584
0.7599 LElTCH TE
0.2631 COMMUt
0.6358 LEITCH TE
0-2359
0.0802
0.8042
0.1531 INTERNAT
0.01 19 LEITCH 1
0.01 13 SNlF SE<
0.7976 PRIMETE
0.1 599 CELESTi
0.4257 PRIMETE
0.9340
0.1 332
0.0710 LEITCH TE
0.0588 LEITCH TE
0.4990
0.4958 FIFTY-PL
0.141 5 COMMUF
0.8695
0.0131
0.01 35
0.0452
0.1021
0.0060
LEITCH 1
SNlF SEC
LEITCH 1
FIFTY-PL
LEITCH 1
FIFTY-PL
VALUCNG
PRIVATE
COMPANIES:
A DEA APPROACH
147
TRIPLE (
DBATEL
TS TELE
TIR S Y S
SYMPLE
A W A TE
FIFlY-Pl
0.526
0.64
1
0.656
0.562
0.847
1
4.979
1.956
11.832
O
5,879
2,429
2.404
0.1854
0.3590
0.3720
0.4101
0.5356
0.5753
0.6581
0.526
0.562
0.64
0.656
0.847
1
1
4,979
0.1854
5,879
1.956
1.710
2,429
11,832
2.404
0.5356
0.3590
0.4101
0.5753
0.3720
0.6581
Range for ABL CANADA's market cap. based on peer group is $1, T l 0,000 and $11,832.000
ABL is a little bit overvalued
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
148
DMU
Score
33 ADVANTEDC
9 AVANTE TEC
3 INVERPOWE
30 RESEARCH I
32 TS TELECOb
42 FIFTY-PLUS.
45 KASTEN CHI
65 WI-LAN INC
77 PLAINTREE :
83 NEXTRON C
85 WE X.L. HOL
Distance
42 Formula
0.1 80379
0.582956
0.544648
1
1
1
0-994883
0.87321
0-866344
0-534451
0.94369
Ranking, according to the distances that are less than 1 (for ADVANTEDGE TECHNOLOGIES I
Nextron
0.00474
Avante
0.01 06s
WE X-L. H0Id
0.01432
Wt-Lan
0.01463
Fifty-plus
0-01823
Plaintree
0.40491
Similarity of the efkiency scores (difference between ADVANTEDGE score (0.18):
Nextron
-0.354072
Avante
-0.402577
WE X.L. Hotd -0.76331 1
Wi-Lan
-0.692832
Fifty-plus
4-849621
Plaintree
4-685966
ADVANTEDGE's efficiency score is not close to any of the others, the closest is Nextron at (
Similarity of the Market Cap. (diffemnce between ADVANTEDGE's market cap. ($19,92
Nextron
17,502 (19,928-2
2.426
Avante
17,625
2,302
WE X.L. HoId
2,844
17.083
Wi-Lan
-309.412
329.339
Fifty-plus
17,524
2.404
Plaintree
4,981
14,946
Range for ADVANTEDGE's market cap. based on peer group is $2,302,000 $329,339,000
ADVANTEDGE's real market cap of $19,928,000falls between this range.
149
Score
9 AVANTE T
3 INVERPO\
30 RESEARC
32 TS TELEC
33 ADVANTE
42 FIW-PLL
45 KASTEN C
65 WI-LAN IN
77 PLAINTRE
83 NEXTRON
85 WE X.L. HI
42
0,5829556
OS446483
1
1
0-1803789
i
0.994883
0.8732104
0-8663444
0.5344511
0-9436897
Ranking, according to the distances that an, less than 1 (for AVANTE TECHNOLOGIES INC):
WE X.L_ HI
0.00034
Fifty-plus
0.001
Nextron
0.0012
Advantedg
0.01O 6
Wi-Lan
0.05
Plaintree
0.5468
Similarity of the efficiency scores (diffemnce between AVANTE's score (0.583):
WE X.L. H
t -0.360734
Fifty-plus -0.41 7044
Nextron
0.0485045
Advantedg 0-4025768
Wi-Lan
-0.290255
Plaintree
-0.283389
AVANTE's efficiency score is very close to Nextron's than Plaintree
Similarity of the Market Cap. (difference between AVANTES market cap. ($2,302,000):
WE X.L HI
-14,781 (2,302-17,083)
Fi@-plus
-101 (2,302-2,404)
Nextron
-124
Advantedg
-17,625
Wi-Lan
-327,037
Plaintree
-12,644
Range for AVANTE's market cap. based on peer group is between $2,404,000 and $329,339.000
AVANTEs real market cap of $2,302,000 falls below this range.
VALUINGPRIVATE
COMPAN~ES:
A DEA APPROACH
150
A W A (SIC 3600)
DMU
#
12 A W A TECH
32 TS TELECO
18 ABL CANAD
55 TRIPLE CRC
22 DBA TELEC
13 TIR SYSTEh
64 SYMPLEX C
42 FlFTY-PLUE
60 PAClFlC E-L
Score
32
42
60 Distance
0.84726
1
0.39914
0.52596
0-63987
0.65569
0.56217
1
1
Ranking, according to the distances that are Iess than 1 (for AWA):
A W A TECH 0.84726
2,429
o.oo00
FIFW-PLUS
1
2.404
0.0034
SYMPLfEX C 0.56217
5,879
0.0056
1.71O
0.01 51
TIR SYSTEh 0.65569
DBA TELEC 0.63987
1,956
0.0261
TRIPLE CRC 0.52596
4,979
o. 1076
ABL CANAD 0.39914
13,382
0.5753
Similarity of the efficiency scores (dhmnee between AWA's score ):
ABL CANAD 0.39914
13.382
0.5753
TRIPLE CRC 0.52596
4,979
o. 1076
SYMPLEX C 0.56217
5,879
0.0056
DBA TELEC 0-63987
1.956
0.0261
TIR SYSTEh 0.65569
1,ilO
0.0151
AWATECH 0.84726
2,429
o.oo00
FIFTY-PLUS
1
2.404
0.0034
Similarity of the Market Cap. (difhrenee between AWA's market cap.):
1.ilO
0.01 51
TIR SYSTEB 0.65569
DBA TELEC 0.63987
1.956
0.0261
1
2,404
0.0034
FIFTY-PLUS
o.oo00
A W A TECH 0.84726
2.429
TRIPLE CRC 0.52596
4,979
0.1 076
SYMPLEX C 0.56217
5,879
0.0056
ABL CANA 0.39914
13.382
0.5753
Range for AWA's market cap. based on pm group i.$1,710,000 and $13,382,000
A W A ' s mal market cap of $2,429,000 b right in this ange
DMU
Score MsrlcetCap
Efficient Peers (lambdas)
#
$(OOOs) b
42
60
62 distance
19 BCS TECk 0.676
1 . ~ 7 o.oo00 0-9156 0-0417 0.0428 0.0000
0.3210 0.4601 0-3397 0.2002 0.3210
9,700
31 TRADE WI 0.6762
0.0213 0.8016 0-1285 0.0699 0-0213
46 MARATHC 0.6045
623
56 TSI TELSI 0.21 12
12,193
0.0423 0.7477 0-1203 0.1320 0-0423
63 STRATEG 0.5342
1.720
0.0843 0.6999 0.0642 0.2359 0.0843
0.0127 0.8555 0.01 13 0.1332 0-0127
71 SHELLCA: 0.0847
is.696
81 COMPREE 0.3105
717
0.0047 0.9716 0.0131 0-0153 0-0047
42 FI+-PL1
1,0000
2,404
0.0107 1.0000 0.0000 0.0000 0-0107
60 PAClFlC E 1.0000
3,878
1.7584 0.0000 1.0000 0.0000 1.7584
62 SNlF SEC1 1-0000
307
1.7562 0.0000 0.0000 1-0000 1-7562
Ranking, according to the distances that are les8 than 1 (for BCS TECH):
BCS TECk 0.6766
1,027
0.0000
COMPREL 0-3105
717
0.0047
FIFTY-PLL 1-0000
2,404
0.0107
SHELLCA: 0.0847
is,sss
0.0127
MARATHC 0.6045
23
0.0213
TSI TELSI 0.21 12
12,193
0.0423
STRATEG 0.5342
1,720
0.0843
TRADE WI 0.6762
9,700
0.3210
Similarity of the eniciency scoms (diffemnce between BCS TECH'S score ):
SHELLCA: 0-0847
i5,696
0.0127
TSI TELS) 0.21 12
12,193
0.0423
COMPREE 0.31 OS
717
0.0047
STRATEG 0.5342
1,720
0.0843
MARATHC 0.6045
623
0.0213
TRADE WI 0.6762
9 . 7 ~ 0.3210
BCS TECk 0.6766
1,027
0.00
FIFP/-PLL 1.O000
2,404
0.0107
Similarity of the Market Cap. (difference between BCS TECH'S market cap.):
MAi3ATf-K 0.6045
623
0.0213
COMPRE 0.3105
717
0.0047
BCS TECk 0.6766
1.027
0.0000
STfWTEG 0.5342
1,720
0.0843
FIW - P U . 1.O000
2.404
0.0107
9,700
0,3210
TRADE Wi 0.6762
12.193
0.0423
TSI TELS) 0.21 12
I
5,698
0.01
27
SHELLCA: 0.0847
Range for BCS TECH'S market cap. b r s d on peer g i w p is $623.000 to $15,898,000
152
Score
20
68
76
79
79 distance
2,854,867
1,031.5~9
s,soa,271
0.0000
1.1097
0.5721
0.4887
Score
BLUESTAR BAT
SR TELECOM IN
COM DEV INTEF
TEKLOGIX INTE
7 CAMCO INC
67 PRIMETECH EU
76 LEITCH TECHNC
80
51
37
54
0.55479
0.44695
0.86326
0.80991
1
1
1
Distance
BLUESTAR BAT 0.55479
TEKLOGIX INTE 0.80991
PRIMETECH EL1
1
SR TELECOM IN 0.44695
76 distance
O
0.00044
143.365 0.02705
i12,897 0.36994
44,404
243.501
VALUINGPRIVATECOMPANIES:
A DEA APPROACH
i 53
Score
Distance
67 (formula)
0-00000
0-02846
18,426 0.04795
100,000 1.35618
11,832 0.28897
143,365 0.78275
11.398 0.05162
33.191
27,707
Calian's real market cap. of $33,191,000 falls just out of this range
Score
Efficient Peers
2
30
42 Distance
CIRCA
DMU
#
Score
Market Cap
$(OOOs)
6,601
46.091
1 1,832
3.855
143,365
Upper bound
$22,797 thousand
$22,797,000
VALUING
PRIVATE COMPANIES: A DEA APPROACH
155
Scom
42
0.4006
0.6210
0,0000
1.0000
0-0000
0.0000
0.1531
0.3099
0.0000
0.0000
1.0000
0.0000
66 CIRCUIT W( 0-75253
11 DANBEL IN1 0.75534
5 GLOBALTH
1
17 AASTRA TE
1
42 FlW-PLUS
1
44 INTERNATIC
1
0.4312
5.798 0.0556
13,222
1.0000
4s.0~1 0,0000
2,404
0.0000
3.893
0.0000
289
44 distance
0.0151
0.0135
0.0000
0.0000
0.0000
1.0000
0.0000
0.2142
0.5077
03689
1.0638
1-3399
Score
73
7
27
68
76
Market Cap
$(OOOS)
291.127
f00.000
1,257,378
2.854.867
l,osl.sss
27
68
0.2203
1.0000
0.0000
0.0000
0.0000
0.1 599
0.0000
1.0000
0.0000
0.0000
0.0588
0.0000
0.0000
1.0000
0.0000
76 distance
0.5610
0.0000
0.0000
0.0000
1.0000
0.0000
0.9518
1.0725
1.2747
0.2702
Upper bound:
$969,673 thousand
1 56
DMU
#
37
51
80
54
7
67
76
COM DEV lt
SR TELECO
BLUESTAR
TEKLOGIX 1
CAMCO INC
PRIMETECk
LEITCH TEC
0-86326
0,44695
0.55479
0.80991
1
1
1
0.0824
112,897 0.1363
44.404 0.0926
243.~01 0.0755
ioo,o00
1.0000
143,365 0.0000
1,031,559 0.0000
252.31 5
76 distance
O. 1168
0.4192
0.8695
0.8796
0.0000
1.O000
0.0000
Score
2.404
3.893
0.0556
1.0000
0.4312
0.0000
0.0000
0.0000
0.6210
0.0000
0-4006
1.0000
0.0000
0.0000
42
0-3099
0.0000
0.1531
0.0000
1.0000
0.0000
44 distance
0-0135
0.000
0.0151
0.0000
0.0000
1.0000
0.0000
1-3737
0.2142
0.2429
0.8652
1.4579
157
Score
81 COMPRESE 0.31046
71 SHELLCASE8.47502
63 STRATEGIC 0.53425
56 TSI TELSYS 0.21119
46 MARATHOh 0.6045
31 TRADE WIN 0.67621
19 BCS TECHh 0.67662
42 FlFlY-PLUS
1
60 PACIFIC E-L
1
62 SNIF SECUl
1
$(OOOs)
o.oaoo
717
15,698
1,720
12,193
0.0274
0,1251
0.0752
623
0.0452
9,700
1,027
0.4024
0-0047
2,404
3.878
307
0.0012
1.9182
1-9138
42
0.9716
0-8555
0.6999
0.7477
0.8016
0,4601
0.9156
1.0000
0.0000
0.0000
60
62 diaance
0.0131 0.0153 0.0000
0-0113 0.1332 0.0274
0.0642 0.2359 0.1251
0.1203 0.1320 0.0752
0.1285 0.0699 0.0452
0.3397 0.2002 0.4024
0.0417 0.0428 0.0047
0.0000 0.0000 0.0012
1 .O000 . 0.0000 1.9182
0,0000 1 -0000 1 -9138
Ranking, according to the distances that are Iess than 1 (for COMPRESSION):
COMPRESE 0.31046
717
o.ooo0
FIFW-PLUE
1
2,404
0.0012
BCS TECHh 0.67662
1.027
0.0047
SHELLCASE8.47E-02
15,698
0.274
MARATHOh 0.6045
623
0.0452
TSI TELSYS 0.21 119
12.1 93
0.0752
STRATEGIC 0.53425
1,720
0.1251
TRADE M N 0.67621
9.700
0.4024
Similarity of th8 efliciency scores (dmereiice between COMPRESSION'S score ):
SHELLCASE8.47E-02
15,698
0.0274
TSl TELSYS 0.211 19
12,193
0-0752
COMPRESE 0.31046
717
o.oo00
STRATEGIC 0.53425
1,720
0.1 251
MARATHOh 0.6045
623
0,0452
TRADE WlN 0-67621
9,700
0.4024
BCS TECHh 0.6762
1,027
0.0~7
FIFlY-PLUE
1
2.404
O.OOI2
Similarity of the Market Cap. (difference between COMPRESSION'S market cap.):
MARATHOh 0.6045
623
0.6452
COMPRESI 0.31046
717
o.ooo0
BCS TECHh 0,67662
1,027
0.0047
STRATEGIC 0.53425
1,720
0.1251
FIFTY-PLUE
1
2,404
0.0012
TRADE WIN 0.67621
9,700
0.4024
TSl TELSYS 0.211 19
12,193
0.0752
SHELLCASE8.47E-02
15,698
0.0274
Range for COMPRESSION'Smarket cap. based on peer group is $623,000 to $15,698,000
Score
42
70
$(OOOs)
0.01 19
CORECO IN 0.98338
44,923 0.8480
ROYAL LAS 0.68375
s i , m 0.8098 0.1021
FIFTY-PLU3
1
2.404
1.0000 0.0000
1
5 7 6 . ~ 8 0.0000
1-0000
GENNUM C
l
0.0000
76 LEITCH TEC
1
1,031,559 0.0000
Coreco's p e r s are Royal Laser and F'My plus
MV range $2.4 to $51 million
Upper bound estimate ( h e d on efficient pers):
$153,477 thousand
69
84
42
70
76
0.1402
0.0881
0.0000
0.0000
1.O000
distance
0.0000
0.0123
0.0429
1.7 150
1.4585
Qgnal
76 distance
0.0265 0.0000
0.0959 0.3260
0.0000 1.4945
0.0000 O. 1668
0.0000 1.O205
1.O000 1.4899
Siern
distance
0.32603
O
1-47229
UPPER BOUND:
0.9564
Market Capitalization of SIERRA based on EFFICIENT F 0.21112
$107,529 thousand
1.29087
-
---
VALU~NG
PRIVATE
COMPANIES:
A DEA APPROACH
i 59
Scom
22 DBA TELEC
18 ABL CANAD
13 TIR SYSTEh
12 A W A TECH
55 TRIPLE CR<
64 SYMPLEX C
32 TS TELECO
42 Flm-PLUS
60 PACIFIC E-L
0,63987
0.39914
0.65569
0.84726
0.52596
0.56217
1
1
1
13.382
1,710
2,429
4.979
5,879
11.832
2.404
3,878
Efficient Peers:
32
42
0 . m 0-14457 0.84501
0.3590 0,54321 0.40014
0 . ~ 1 70-11977 0.87756
0.0261 0.02450 0.95252
0.0286 0.25074 0-71533
0.0255 0.01430 0.90542
1 . ~ 5 91.00000 0.00000
0.04% 0.00000 1.OOOOO
1-7142 0.00000 0.00000
60
0,01040
0.05670
0.00267
0.02300
0.03390
0.08020
0.00000
0.00000
1.O0000
Distance
0.00000
0.35897
0.00173
0.0261 3
0.02864
0.02549
1.44591
0.04503
1-71425
Ranking, according to the distances that are les$ than 1 (for DBA):
DBA TELEC 0-63987
1.956
o.oo00
TIR SYSTEh 0.65569
1.ilO
0.0017
SYMPLEX C .OS6217
5.879
0.0255
A W A TECH 0.84726
2,429
0.0261
TRIPLE CR< 0,52596
4,979
0.0206
FIW-PLUS
1
2,404
0.0450
ABL CANAD 0.39914
13,382
0.3590
Similarity of the efficiency scores (difference between DBA's score ):
ABL CANAD 0.39914
13.382
0.3590
TRIPLE CRC 0.52596
4,979
0.0286
SYMPLEX C 0,56217
5.879
0.0255
DBA TELEC 0.63987
1,956
o.oo00
TIR SYSTEA 0.65569
1.71O
0.0017
AWATECH 0.84726
2,429
0.0261
FIFTY-PLUS
1
2,404
0.0450
Similarity of the Market Cap. (diflerence between DBA's market cap.):
TIR SYSTEA 0.65569
1,710
0.001 7
DBA TELEC 0.63987
1.956
o.oooo
FIFTY-PLUS
1
2.404
0.0450
A W A TECH 0.84726
2.429
0.0261
TRIPLE CRC 0.52596
4,979
0.0206
SYMPLEX C 0.5621 7
5,879
0.0255
ABL CANAD 0.39914
13.382
0.3540
Range for DBA's maricet cap. based on peer group is $1,710,000 and 513,382,000
~-
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
160
Score
Market Cap
$(OOOs)
30
32
76
14 DIGITAL PR 0.88428
28.123 3.45E-03 0.85907 0-13748
1
585.308
1
O
O
1
11,832
O
1
O
76 LEITCH TEC
1
i,o~l.sss
O
O
1
Digital is only similar to TS telecorn, minimum of $1 1-8 million
30 RESEARCH
32 TS TELECO
0.00000
1.75002
0.03877
1.48196
UPPER BOUND:
Market Capitalization of DIGITAL based on EFFICIENT PEERS
$154,005 thousand
Score
Distance
DIMENSION
CELL-LOC Il
FIFW-PLUS
Eniciency:
DIMENSION
CELL-LOC I!
FIFTY-PLU3
0.1 1029
0.9803
1
28.669
80,615
585.308
2.404
1,970
28,669
2.404
0.11029
1,970
0.9803
1
28,669
2.404
Market Value
DIMENSION 0.1 1029
FIFTY-PLUS
1
CELL-LOC Il 0.9803
42 Distance
1,970
2.404
28.669
VALUINGPRIVATE
COUPANIES:
A DEA APPROACH
161
41 DYNAMI4 0.939
110,354
0.0000
0.0006
1.9377
1
~0.61s
2 TURBO (
sss,3as
1
30 RESEAR
Distance
DYNAMI4 0.939
110,354
o.oo00
TURBO (
1
80.615
0.0006
Minimum market value is $80,6lS,OOO
Upper Range
$89,291 thousand
0.98275
1.O0000
0.00000
30 Distance
0.01720
0.00000
'i.00000
0.00000
0.00059
1,93169
DMU
60 Distance
0.25006 0.00000
0100000 1-38654
0-00000 1-04832
0.00000 0-28908
1-00000 0.90782
42
0.75986
0.00000
0.00000
1.O0000
0.00000
76
0.00443
0.00000
0.00000
0.00000
1.O0000
Distance
0.00000
1-60433
1-17967
O. 10802
1-61887
UPPER BOUND:
Market Capitaliution of EX1 bsed on EFFICIENT PEERS
$14,108 thousand
VALUINGPRNATECOMPANIES:
A DEA APPROACH
162
Score
74 ELECTROHC 0.67333
6 LG TECHNO 0.586585
1 HAMMOND l 0.756773
7 CAMCO INC
1
32 TS TELECOI
1
34 CALIAN TEC 0.800142
67 PRIMETECH
1
27,707
78.426
ioo,ooo
t 1,832
33,191
143.365
0.00000
0.15288
0.04998
1.28554
0.58451
0.05162
0.43784
0.0752
0.0060
0.2217
1.0000
0.0000
0.0599
0.0000
Distance
67 (formula)
0.4257
0.7302
0.4467
0.0000
1.O000
0.5935
0.0000
And Efectrohomes real market cap. of $1l,398,OOO falls below this range
UPPER BOUND:
Market Capitalization of ELECTROHOME based on EFFICIENT PEERS
Looks like:[(0.075'100,000) + (0.43'1 1,832)+(0-5'143,365)J
VALUING
PRIVATE
COMPANIES:
A DEA APPROACH
163
DMU
~istance
0-0000
0.0452
0-0059
1-9531
0~0003
$3,327thousand
#
82
7
68
76
2.8s.867
LEITCH TECk
1,031,559
i29.626
ioo,ooo
0.0196
1,0000
0.0000
0.ooOo
68
76 distance
Score
1 HAMMONC
6LGTECHN
7 CAMCO IN
32 TS TELECI
34 CALIAN TE
67 PRIMETEC
74 ELECTROI
0.7568
0.5866
1.O000
1.O000
0,8001
1.0000
0.6733
Market 4 Distance
Efficient Pmm
S(O00s) (displayed better)
7
32
18.4%
27,707
1oo.m
ii,832
33,191
143,365
11,398
0,0000
0.1315
0.91 52
0.4652
0.0480
0.6956
0.0500
0.221 7
O.O60
1.O000
0,0000
0.0599
0.0000
0.0752
0-4467
0.7302
0-0000
1-0000
0.5935
0.0000
0.4257
Distance
67 (formula)
0.331 5
0.2637
0.0000
0.0000
0.3466
1.O000
0.4990
0.0000
0.1315
0.91 52
0.4652
0.0480
0.6956
0.0500
$74,988 thousand
VALUING PRIVATE COMPANIES: A DEA APPROACH
165
DMU
Distance
42 Formula
Ranking, according to the distances that are Iess than 1 (for INVERPOWER):
TS Telecoi 0.011
Kasten
0.039
Plaintree
0.31
Similarity of the efficiency scores (diffemnce between INVERPOWER's score (0.545):
TS Telecoi -0.455
Kasten
-0.45
Plain4.322
INVERPOWER CONTROLS LTD's efficiency score is not close to the efficiency score of it's p e r s
Similarity of the Market Cap. (diffemnce between INVERPOWER's market cap.):
TS Telecoi 4,653 (7,179-1 1,832)
Kasten
-16,390 (7,179-23.569)
PIaintree -7,767 (7,179-14.946)
Range for INVERPOWER's market cap. based on peer group is $1 1,832,000 to $23,569,000
INVERPOWER CONTROLS LTD real market cap of $7,l79,OOO falls below this range
VALUING
PRWATECOMPANIES: A DEA APPROACH
166
Score
Distance
42 Formula
45 KASTEN CHP
33 ADVANTEDG
9 AVANTE TEC
3 INVERPOWEl
30 RESEARCH II
32 TS TELECOM
42 FIFW-PLUSI
65 W-LAN INC
77 PLAlNTREE I
83 NUCTRON CC
85 WE XL, HOLi
Ranking, according to the distances that are Iess than 1 (for KASTEN):
lnverpower
0.0389
Ts Telecom
0.0852
Plaintree
0.1557
Wi-Lan
0.8240
Similarity of the eniciency scores (d'iflerence between KASTEN score (0.18):
lnverpower
0.4502
Ts Telecom
-0.0051
Plaintree
O. 1285
Wi-Lan
0.1217
KASTEN's efficiency score is very close to that of TS telecorn
Sirnilarity of the Matlcet Cap. (dfference between KASTENvsmarket cap. ($19,928,000):
lnverpower
16390.1 8
Ts Telecom
11736-78
Plaintree
8623.08
Wi-Lan
-305770.09
Range for KASTEN8smarket cap. based on peer group is between $7,179,000 and $329,339,000
KASTEN8sreal market cap of 523,569,000 falls between this range, lgnoring the biggest market
cap. company in the peer group (W-Lan), the range is between $7,179 and $14,946 million
KASTEN seems either slightly overvalued when one examines the market cap., however it is perforn
near the efficient frontier, with an efficiency score of 0.99, thus it may well be undervalued
VALUING
PRIVATE COMPANIES: A DEA APPROACH
167
Score
)(ln9
Sirnrnond
0.214263 0.785737
O 0.8025
f 7-ISE-16 1.2348 0.0464
2.854 0.847691 0.152309 0.8025
O
307 1-62E-18
1 0.0918 1.4372
7,739
3,878
UPPER BOUND:
Market Capitalization of KING based on EFFICIENT PEERS
$1,072 thousand
King is overvalued according to its peers
Score
Market Cap
S(OO0s)
60 distance
A DEA APPROACH
DMU
Score
2
o.oo00 0.0660
0.0185 0.1621
o.wss 0.01 18
1.7448 1.0000
0.0087 0.0000
42 Distance
75 LAMBUS E 0.709557
1,729
0-9340 0.0000
SENSE TE 0.374065
61,221
0.8379 0.0185
GEMSTAF 0.232398
4,888
0.9882 0.0059
TURBO GI
1
80.61s
0.0000 1.7448
FIFTY-PL1
1
2,404
1.0000 0.0087
Distance
LAMBUS E 0.709557
1.729
o.oo00
GEMSTAF 0.232398
4.888
O-0059
FIFIY-P LC
I
2.404
0.0087
SENSE TE 0.374065
61.22t
0.01 85
Efficiency :
Market Value
GEMSTAG 0.232398
4,888
0.~59
LAMBU 0.7096
SENSE TE 0.374065
61,221
0.0185
FIFTY-F
1
LAMBUS E 0.709557
1,729
0.~00
GEMST 0.2324
Flm-PL1
1
2,404
0.0087
SENSE 0.3741
Range for market value $2,44,000 to $61,221,000
Upper Range
$7,566 thousand
48
43
2
42
1,729 0-0000
2.404 0.0087
4.888 0.0059
6 1 ~ 1 0.0185
DMU
Score
i 69
DMU
Score
VALUINGPRIVATE COMPANIES:
A DEA APPROACH
170
DMU
#
DMU
#
Upper bound
$169,212 thousand
-
171
DMU
#
Distance
42 Formula
172
Score
28
30
42
44
76
44
76 D h n c e
UPPER BOUND:
Market Capitalization of NHC based on EFFICIENT PEERS
W,4ll thousand
Score
8 NORON TE 6.53E-02
42 FlFlY-PLU!
1
62 SNlF SECU
1
2.404
307
62
distancDisbnce
0.98995 1.Oq E-02 0.0000
O
1
O 0.0002 0.000203
O
1 1.9599 1.959903
Noron only has F i i plus close to it in terms of distance, therefore its market
value should be near $2.4 million
Upper Bound
$2,383 thousand
According to the upper bound Noron is ovenralued
Score
Market Cap
$(OOOs)
68 Distance
10 SEMI-TECF 0.214508
10,026 0.?74623 0.22538
O
7 CAMCO IN<
1 100,006
1
O 0.1016
68 CELESTICE
1 2,854,867 2.81 E-16
1 1.2001
In terrns of distance only Camco is close, therefon, lower bound of $100 million
Upper bound:
$720,885 thousand
VALUING
PRIVATE
COMPANIES: A DEA APPROACH
1 73
DMU
#
3 INVERPO\ 0,5446
7.179
0.30786 0.0116 0.9208 0.0676 0.3079
9 AVANTE T 0.583
2 . 3 ~ 0.54681 0-0028 0.0210 0.9763 0.5468
30 RESEARC
1 585,308
1.24976 1.O000 0.0000 0.0000 1.2498
32 TS TELEC
1
11.832
0.43498 0.0000 1-0000 0.0000 0.4350
42 FIFTY-PLL
1
2,404
0.59484 0.0000 0.0000 1-0000 0.5948
33 ADVANTE 0.1 804
19,928
0 . 4 ~ 9 10.01 17 0.0891 0-8992 0.4049
45 KASTEN C 0.9949
23.569
0.15575 0.1 130 0.7617 0.1253 0.1557
83 NEXTRON 0.5345
2,426
0.49728 0.0032 0.0452 0.9516 0.4973
85 W E XL. Hi 0.9437
17,083
0.57049 0.0088 0.0060 0.9853 0.5705
Ranking, according to the distances that are Iess than 1 (for PLAINTREE):
PLAINTRE 0.8663
1 4 , s ~ O-owoo
KASTEN C 0.9949
23.569
0.15~75
WI-LAN IN 0.8732
329.339
0.26639
INVERPO\ 0.5446
7.179
0.30786
ADVANTE 0.1804
19.928
0.4~91
TS TELEC
1
11,832
0.43498
NEXTRON 0.5345
2,426
0.49728
AVANT T 0.583
2.302
0.54681
WE XL, Hl 0.9437
17,083
0.57049
FIm-PLL
1
2.404
0.59484
Similarity of the Market Cap.:
Similarity of the efficiency scores:
AVANTl 0.583
ADVANTE 0-1804
FIFPI-f
1
NEXTRON 0-5345
NEXTR! 0.5345
INVERPO\ 0.5446
l NVERF 0.5446
AVANTE T 0.583
TS TEU
1
PLAINTRE 0-8663
PLAINT 0.8663
WI-LAN IN 0.8732
WE X.L 0.9437
W E XL. Hl 0.9437
ADVAN
0-1804
KASTEN C 0.9949
KASTEl 0.9949
TS TELEC
1
WI-LAN 0.8732
FIFPI-PLL
1
Range for PLAlNTREE's market cap. based on peer gmup is $2,302,000 and $329,339,000
VALUI
NG PRIVATE
COMPANIES:
A DEA APPROACH
174
DMU
48
43
75
2
42
Upper Range
$15,085 thousand
42
0.4871
0.0000
0.0000
1.0000
0.0000
59 Distance
0.1 133 0.0000
0.0000 0.6122
0.0000 1.4064
0-0000 0.4345
1.0000 1.1822
Upper bound
$6,597 thousand
VALUINGPRIVATE
COMPANIES:A DEA APPROACH
175
Scom
62 distance
71 SHELLCA: 8.47E-02
is.698
0 . m 0.8555 0.01 13 0.1332 0-0000
63 STRATEG 0.53425
1,720
0.0376 0.6999 0.0642 0.2359 0-0376
56 TSI TELSI 0.21 1188
12,193
0.0235 0.7477 0-1203 0.1 320 0-0235
46 MARATHC 0.604497
623
0.0207 0.8016 0-1285 0.0699 0-0207
31 TRADE Wi 0.676206
9.700
0.2686 0.4601 0.3397 0.2002 0.2686
1.027
0-0127 0.9156 0.0417 0.0428 0-0127
19 BCS TECk 0.67661 9
81 COMPREI 0.310462
717
0.0274 0.9716 0.0131 0,0153 0-0274
42 FIFlY-PL!.
1
2.404
0.0388 1.0000 0.0000 0.0000 0.0388
1
3,878
1.7271 0.0000 1,0000 0.0000 1.7271
60 PAClFlC E
62 SNlF SEC1
1
307
1.4832 0.0000 0.0000 1.O000 1.4832
Ranking, according to the distances that are lem than 1 (for SHELLCASE):
O-oooo
SHELLCA: 8.47E-02
I5.698
1,027
0.0127
BCS TECk 0.676619
623
0.0207
MARATHC 0.604497
12,193
0.0235
TSI TELS) 0.21 1188
COMPREZ 0.310462
717
0.0274
STRATEG 0.53425
1,720
0.0376
FIm-PLl
1
2,404
0.0388
TRADE Wi 0.676206
9,700
0.2686
Similarity of the eficiency scores (diHerence between SHELLCASE's score ):
SHELLCA: 8.47E-02
is.698
o.oooo
TSI TELSt 0.21 1188
12.193
0.0235
COMPREZ 0.310462
717
0.0274
STRATEG 0.53425
1.720
0.0376
MARATHC 0.604497
623
0.0207
TRADE WI 0.676206
9.700
0.2686
BCS TECF 0.67661 9
1,027
0.0127
FIFW-PL1
1
2,404
0.0388
Similatity of the Market Cap. (difference between SHELLCASE's market cap.):
MARATHC 0.604497
623
0.0207
COMPREZ 0.310462
717
0.0274
8CS TECk 0.67661 9
1,027
0.0127
STRATEG 0.53425
1.720
0.0376
FIFTY-PL1
1
2,404
0.0388
9,700
0.2686
TRADE WI 0.676206
12,193
0.0235
TSI TELS) 0.21 1188
15,698
O-OO
SHELLCA! 8-47E-02
Range for SHELLCASE's market cap. based on peer group is $623,000to $12.1 93,000
51 SR TELEC 0.447
i12,897
44,404
80 BLUESTAI 0.5548
37 COM DEV 0.8633
252.31 5
54 TEKLOGI) 0.8099
243,501
7 CAMCO It+
1
iw,m
67 PRIMETE(
1
143,365
76 LEITCH TE
1 1.031,!~9
0.1363
0.0926
0.0824
0-0755
1.0000
0.0000
0.0000
Upper bound
$532,289 thousand
0.3699
O
0.5534
0-3753
0.2213
0.5029
0.4192
0-8695
0.1 168
0,8796
0.0000
1.0000
0.0000
76 distance
0.4445
0.0379
0.8008
0.0449
0.0000
0.0000
1.O000
0-0000
0-3699
0.2213
0-3753
1-1194
0,5534
OS029
VALUING
PRIVATE
COMPANIES:
A DEA AFPROACH
178
DMU
#
64 SYMPLE
55 TRIPLE (
22 DBA TEL
18 ABL CAh
13 TIR SYS
12 A W A TE
32 TS TELE
42 FIFTY-PI
60 PAClFlC
0.562
0.526
0.64
0.399
0.656
0.847
1
1
1
60 Distance
Ranking, according to the distancm that are less than 1 (for SYMPLEX):
A W A TE 0.847
2.429 0.0056
FIFTY-PI
1
2,404 0.0156
TIRSYS0.656
1,710
0.0179
DEA TEL 0.64
1.9s
0.02~5
TRIPLE(0.526
4.979
0.0942
ABL CAh 0.399 13,382 0 . 5 s
Similarty of the efticiency scores (difference between SYMPLEX's score ):
ABL CAh 0.399 13,382 0.5356
4.979
0.0942
TRIPLE(0.526
3
r
-<.--&fi'
,--
13.382
0.5356
Range for SYMPLEX's market cap. based on peer group is $1,Tl 0,000 and $13,382,000
179
53 SYSCAN Il 0.6574
1
17 AASTRA T
38 COMMUN1
1
42 FIFTY-PLL
1
67 PRIMETEC
1
17
38
42
46.091
67 Distance
2-52E-03
O
O
O
1
0.0000
1.3500
0.6078
0.4104
1.4493
67
76 distance
54 TEKLOGI) 0.8099
243,501 7.55E-02 0-879623 4.49E-02
112.897 0.13625 0.419239 O.444511
51 SR TELEC 0.447
80 BLUESTAi 0.5548
44.404 9.26E-02 0.869458 3.79E-02
252.315 8.24E-02 0.1 16802 0.800779
37 COM DEV 0.8633
1
O
O
loo,ooo
1
7 CAMCO Ib
O
1
O
67 PRIMETEC
1
143.365
O
O
1
76 LEITCH TE
1 1,031,559
0.00000
0.37533
0.00044
1.15330
1.63045
0.02221
1.69165
Range for TIR'. market cap. based on p m group k S1,956,000 and S13,382,000
TIR'S mal market cap of $1,7t 0,000 falls just short of this amount
181
Scom
62 distance
31 TRADEW
19 BCS TECk
46 MARATHC
56 TSI TELS)
63 STRATEG
71 SHELLCA.'
81 COMPREZ
42 FIFTY-PLL
60 PAClFlC E
62 SNlF SECI
Ranking, according to the distances mat are le- than 1 (for TRADE WIND):
TRADEW
0:67@tL
6700? -=P- - 1.720
0-1346
0.53L)25
STRATEG
TSITELSiI
0.21119
12.193
0-1355
623
0.1781
0.60450
MARATHC
is.698
0.2686
0.08470
SHELLCA.'
1,027
0.3210
0.67662
BCS TECi717
0.4024
0.31 046
COMPREE
FIFW-PLL
2.W
0.4469
1.OOOOO
PACIFIC E
1-00000
3,878
0.6878
SNlF SECI
1.00000
307
0.9669
Similarity of the Market Cap.:
Similarity of the efficiency scores:
SHELLCA:
0.08470
15.698
0.2686
SNIF Sf
1
307 0.9669
TSI TELS)
0.21 119
12.193
0.1355
MARAT 0.6045
623 0.1781
COMPREZ
0.31 046
717
0,4024
COMPF 0.31046
717 0.4024
SlTWTEG
0.53425
1,720
0.1346
BCS TE 0.67662
1,027 0.3210
STRATt 0.53425
1,720 0-1346
MARATHC
0.60450
623
0.1781
TRADE
s . o o ~ j ~ ~ ~ UFIF~ l Y - F
1
2,404 0.4469
0.676Zf
PACIFI(:
1
3,878- 0.6878
-..1,027
0.3210
BCS TECF - 0.67662
2.0.4469
T R A D ~0 - 6 7 6 2 f 5 9,h1&.-- O ~..-I O -O
1-00000
FIFTY-PL1
PAClFlC E
1-00000
3.878
0.6878
TSITEL 0.21119 1z.193 0.1355
SNIF SEC1 1.00000
307
0.9669
SHELLC 8.47E-02 1S,698 0.2-
7
-
# T < F -
"
<
Range for TRADE WND's market cap. b a s d on peer group is $307,000 to $15,698,000
60
0.0339
0.0104
0.0567
0.0027
0.0230
0.0802
Distance
0.0000
0-0286
0.1854
0-0444
0-1076
0.0942
0-0000 1-0742
O.OOOO 0.1451
1.0000 1.5079
Ranking, according to the distances that am lemm than 1 (for TRIPLE CROWN):
TRIPLE CF 0.526
4,979
o.oo00
DEA TELE 0.6399
1.956
0.0286
TIR SYSTt 0.6557
1,770
0-.
SYMPLEX 0.5622
5.879
0.942
AWA TEC 0.8473
2,429
0.1076
FIFW-PL1
I
2,404
0.1451
ABL CANP 0.3991
13.382
0.1854
'
- $13,382,000
183
38
0.1094
0.0000
0.0000
f -0000
O.OOOO
42
0.6-1
0.0000
0.0000
0.0000
1.OOOO
distance
0.0000
1.3101
1.1973
1.2614
0.1521
UPPER BOUND:
Market Capitalization of TRI based on EFFICIENT PEERS
$53,470 thousand
DMU
#
67
0.0710
0.0000
0.0000
1.0000
0-0000
76 distance
0,031 1 0.0000
0,0000 1.4515
0.0000 0.0571
0.0000 1.5102
1.O000 1.5900
UPPER B9UND:
Market Capitalization of TUNDRA based on EFfIClENT PEERS
$46,523 thousand
42
0-9285
0.0000
0,0000
I,OOOO
0.0000
67 distance
0.0025 0.0000
0.0000 1.8215
0.0000 1.7705
O.OOOO 0.0078
1.O000 1.8598
Fifty-Plus is only simibr to Unitec in terms of distance, therefore it should have a market cap
near $2.4 miIlione
62 distance
0.1 320 0.0000
0.0699 0.0068
0.2002 0-1355
0.0428 0-0423
0.2359 0.0162
0.1332 0.0235
0.0153 0.0752
0.0000 0.0956
0.0000 1-3503
1-0000 1-3269
Ranking, according to the distances that are less than 1 (for TSI TEL):
TSI TELS)
a211191 ~ 1--=_-~-~?ZO:mo
MARATHC
0.60450
623
0.0068
STRATEG
0.53425
1,720
0.0162
SHELLCA!
0.08470
15,698
0.0235
BCS TECk
0.67662
1,027
0,0423
COMPREZ
0.31046
717
0.0752
FIFTY-PLL
1.O0000
2,404
0-0956
TRADE W
0.67621
9,700
0.1355
-%O
Similarity of the Market Cap. (difference between TSI TEL'S market cap.):
MARATHC
0.60450
623
0-0068
COMPREZ
0.31 046
717
0-0752
BCS f Eck
0.67662
1.027
0-0423
STRATEG
0.53425
1,720
0.0162
FIFTY-PL1
1-OOOOO
2.404
0.0956
TRADE WI
0.67621
9,700
0.1355
TSl TELSY
O 2 1119 -- 12,19%z:
---2! O.0
SH ELLCA!
0.08470
15,698
0.0235
Range for TSI TEL'S market cap. based on peer group b $623,000 to $15,698,000
185
Score
#
85 WE X-L. Hl
83 NEXTRON
77 PLAl NTRE
65 WI-LAN IN
3 INVERPO\
9 AVANTE T
30 RESEARC
32 TS TELEC
42 FlFlY-PLL
33 ADVANTE
45 KASTEN C
Distance
42 Formula
0.98526 0.00000
0,951 58 0.00271
0.41 586 0.57049
0.80420 0.05722
0.06760 1-67913
0.97626 0.00034
0.00000 1.95326
0,00000 1.95894
3 .O0000 0.00033
0.89922 0.01432
0.12529 1.32157
Ranking, according to the distances that are less than 1 (for WE XL):
WE X-L. Hl 0.94369
17,083
0.OOO
FIFW-PLL 1.OOOOO
2.404
0.00033
AVANTE T 0.58296
2,302
0.000%
NEXTRON 0.53445
2,426
0-00271
ADVANTE 0.1 8038
19,928
0.01432
WI-LAN IN 0.87321
329.339
0-05722
PLAINTRE 0.86634
f 4.946
0.51049
Similarity of the efkiency scores (difference betweeri WE X.L's score (0.94):
ADVANTE 0.1 8038
19,928
0.01432
N W R O N 0.53445
2,426
0.00271
AVANTE T 0.58296
2,302
0.00034
PLAINTRE 0.86634
14,946
0.51049
WI-LAN IN 0.87321
329,339
0.05722
WE X. L. Hl 0.94369
17,083
0.00000
FIFTY-PLL 1.O0000
2,404
0.00033
Similarity of the Market Cap. (dwerence between WE X.L's market cap. ($17,083,000):
AVANTE T 0.58296
2.302
0.00034
FlFTY-PLl. 1.00000
2,404
0.00033
2,426
0-00271
NEXTRON 0.53445
14.946
0.57049
0.86634
PLAINTRE
17,083
0.00000
WE X.L. Hf 0.94369
19,928
0.01432
ADVANTE 0.1 8038
329,339
0.05722
WI-LAN IN 0.87321
Range for WE X.L's macap. based on peer group is $2,302,000 and $329,339,000
WE X.L's real market cap of $11,083,000 b rigtit in this ange
Especially after ignoring the effect of Wi-lan, the range is betwee $2.3 and $20 million
186
VALUING PRIVATE
COMPANIES:
A DEA APPROACH
187
APPENDIX C
Efficient Companies-SIC 3 600
Market Value Analysis
Based on BCC-1
For Company Aastm (17)
Score
DMU
#
17 AASlRA
4
53
21
66
11
TE
Market Cap
$(OOOs)
1
46,091
UNITECINT 0.43488
SYSCAN IN' 0.657407
ClRCA ENTl 0.859468
CIRCUIT W 0.75253
DANBEL IN[ 0.755337
17
1 nonnalized
4.72E-02
0-052205
0-278
0.400605
0-621007
1.399018
5.571
7.260
6.601
289
5.798
0.033738
0.037316
O. 198711
0.286347
0-443888
1
187.9442594
270.8941374
1311-689405
82.67916377
2573.473118
$4,427
46,091
Based on BCC-I
For Company Research in Motion (30)
Score
DMU
#
30 RESEARCH
Market Cap
$(OOOs)
1
585,308
79 BALLARD P.
1
40 DIMENSION 0-110285
3,608.271
9 AVANTE TE 0.582956
83 NEXTRON C 0.534451
2.302
14 DIGITAL PR 0.884281
28.123
SIERRA W l F
CELL-LOC Il
WE X.L. HO1
INVERPOW
33 ADVANTEDI
58 XI TECHNt
28 NHC COMM
41 DYNAMIC D
39 CYGNAL TE
65 Wl-LAN INC
78 TRI-VISION
77 PLAINTREE
45 KASTEN Ch
32.735
61
36
85
3
0.996348
0.980304
0.94369
0.544648
0.180379
0.756899
0.670855
0-939374
0.589195
0.87321
0.742628
0.866344
0.994883
1.970
2.426
28.669
17.083
7.179
19.928
9.422
7.317
110.3~4
8.637
329.339
29.930
14.946
23.569
30
1 normalued
3.12E-16
3.26E-04
2.76E-03
3.22E-03
7.07E-16
7.39E-04
6-25E-03
7.30E-03
3.45E-03
5. i9E-03
5.58E-03
8.79E-03
1.16E-02
1.17E-02
1.17E-02
1.32E-O2
1-72E-02
0.024202
3.59E-02
0.085046
8.84E-02
0.1 12999
4.41E-01
7.82E-03
1-18E-02
1-26E-02
1-99E-02
2.63E-02
2.65E-02
2.65E-02
2-99E-02
3.90E-02
5.48E-02
8.14E-02
7 -93E-07
2.00E-01
2.56E-01
1.OOE+OO
i 89
Market Cap
$(OOOs)
Score
DMU
#
68 CELESTIC
2.850.867
10 SEMI-TEC
20 BELL CAN
73 C-MAC IN[
82 GSI LUMO
47 NEWBRIDl
0.214508
0.254057
0.817145
0.916721
1
10,026
1,371,120
291.T 27
129.626
8,125.481
TOTAL-
1 Nom
0.225377
O. 140231
5.88E-02
1-35E-02
4.70E-18
0.437908
0.514668
0.320229
0.134275
0-030828
1-07E-17
1
MCCALC
5160.04695
439072.5952
39090.97203
3996.146357
8.72095E-11
$487,320 (000s)
W a r k a Cap
$(OOOs)
Score
38 COMMUN1
CIRCAEN'
CYGNAL 7
SYSCAN 11
EX1 TECHI
61 SIERRAW
62 SNlF SECI
72 TUNDRA 5
78 TRI-VISIOI
0.859468
0.589195
0.657407
0.756899
0.996348
1
0.98174
0-742628
21
39
53
58
38
1 NORM
3,855
MCCALC
DML:
Market Cap
$(OOOs)
79 BALLARD
3,608,271
1,371.120
8,125,481
0.450728
1
1.88E-17 4.17E-17
0.450728
1
1371119.948
3.38916E-1 0
$1,371,120
VALUING PRIVATE
COMPANIES:
A DEA APFROACH
190
Score
DMU
#
20 BELL CAN
25 MEMOTEC
28 NHC COM
35 CANADIAF
37 COM D W
39 CYGNAL T
50 SPECTRU
51 SR TELEC
54 TEKLOGIIC
57 WESCAM
58 XI TECHI
61 SIERRA*
68 CELESTIC
69 CORECO 1
70 GENNUM 1
72 TUNDRA I
73 C-MAC IN[
80 BLUESTAF
82 GSI LUMO
84 ROYAL LA
0.254057
0.649191
0.670855
1
0.863264
0.589195
0.78669
0.446954
0-809908
0-628467
0-756899
0.996348
1
0-983382
1
0.98174
0.817145
0-55479
0.916721
0.683747
TOTAL
76
0.409041
1.46E-02
3.33E-02
4.53E-17
0.800779
2.65E-02
7.75E-02
0.444511
4.49E-02
5.84E-02
4.43503
9.59E-02
3.97E-16
0.140155
1.21018
3.11E-02
0.56102
3.79E-02
0.966892
8.81 4 2
3-972509
1-41+O5
1.O3E-Ol
1.82E+01
3.68E-03
6.13E+Ol
8.38E-03
1.14E-17
4.87E-12
5-09E+04
2.02E-01
5.76+01
6.67E-03
7.54+02
1-95E-02
1.26+04
1-12E-01
2-75E+03
1-13E-02
1-47E-02
l.l3E+O3
1.05E+01
1.12E-03
7.90+02
2.41E-02
2.85E-10
9.99E-17
3.53E-02
1.58+03
3.05E-19
1.76E-13
9.45+02
7.836-03
4.1 1+O4
1.41E-01
4.24+02
9.54E-03
3.16+04
2.43E-01
l.l4E+O3
2.22E-02
$287,982 Lower bound
1
$t,O3l,SS9
(000s)
REAL
Market Cap
YOOfW
426808.3836
VALUINGPRIVATECOMPANIES:
A DEA APPROACH
191
Score
DMU
#
62 SNlF SECURi'
307
5.68E-03
2.41 4 2
4.42E-01
1-13 E 6 l
3.93E-02
8-57E-02
7.43E-O2
4.02E-16
1-33E-07
7.49E-02
8-61E-03
1.00E+00
-.-p.
MC Calc.
3.99E+01
2.47E+Ol
3.42E+03
1.OgE+O3
2.4SE+01
2.45E+02
9.06E+02
1S6E-12
2.28E+02
1.18E+03
6.17+00
$7,164 thouiand
Lower Sound
Undervalued
VALUINGPRIVATECOMPANIES:
A DEA APPROACH
192
DMU
#
67 PRIMETECH ELECTRO
143,365
53 SYSCAN INTERNATIOh
4 UNITEC INTERNATIONJ
21 ClRCA ENTERPRISES I
72 TUNDRA SEMICONDUC
50 SPECTRUM SIGNAL PF
37 COM DEV INTERNAT10
25 MEMOTEC COMMUNIC
6 LG TECHNOLOGIES GF
1 HAMMONO MANUFACT
34 CALIAN TECHNOLOGY
51 SR TELECOM INC
74 ELECTROHOME LIMITE
57 WESCAM INC
80 BLUESTAR BATTERY S
54 TEKLOG lX INTERNATIC
REAL CAP
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
193
Score
DMU
#
60 PAClFlC E
12 A W A TEC
13 TIR SYSTE
18 ABL CANA
19 BCS TECF
0.847256
0.655687
0.399144
0.676619
0.639866
0.815085
0.471945
0.0664
0.676206
1
0.604497
0.766307
0.525957
0-211188
1
0.53425
0.562167
0.0847
0.310462
22
23
24
26
31
32
46
49
55
56
62
63
64
71
81
DBA TELE
EIGER TE4
KING PRO
MlCROCEl
TRADE WI
TS TELEC
MARATHC
SIMMOND
TRIPLE CF
TSI TELSY
SNlF SEC[
STRATEGI
SYMPLEX
SHELLCA:
COMPREE
S(0OOs)
Lambda
60
TOTAL:
3.053191
$86,216 (000s)
3,878 (000s)
1
REAL
VALUINGPRIVATE
COMPAN~ES:
A DEA APPROACH
194
DMU
#
15 SONY CORI
1 HAMMOND 0.756773
2 TURBO GEt
1
3 INVERPOW 0.544648
4 UNITECINT 0.43488
18,426
80,615
7,179
5.571
1,075,691
18.426
80.61 5
7.179
1
O
O
O
1,257,378
15 SONY CORI
1
73 C-MAC lNDl 0.817145
52.935.385
1.38E-15
291.127 0.159901
0.1 59901
SELF evaluator
MlTEL Corp is a
Based on 8CC-l
For Company Nortel(29)
Score
DMU
#
29 NORTEL NE
Market Cap
$(OOOs)
1 50,495,406
80 BLUESTAR
0.55479
81 COMPRESI 0.310462
82 GSI LUMON 0.916721
44.404
717
129.626
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
195
MarketCap
$(OOOs)
1
1
52,935,385
50,495,406
8.125.481
47
1
4.25E-16 5.58E-01
2.96E+07
3.36E-16 4.42E-01
2.23E+07
7.61E-16 1.OOE+00 $51,858,075
51,858,075
Score
Market Cap
$(OOOs)
1
44 INTERNATIOb
3.893
44
1 norm
rnc calc
5 GLOBAL THE1
1 13221-8604 9.54E-18 1 -696-16 2.2342E-12
11 OANBEL INDL 0.755337
5797.576
0-0135 0.239124 1386.34114
28 NHC COMMUI 0.670855 7317.08222 0.027856 0.493411 3610.32741
66 CIRCUIT WOF 0.75253 288.73734
0.0151 0.267465 77.2271112
$5,074 (000s)
5.65E-02 1.00E+00
Market Cap
$(OOOs)
Score
70
576.578
16 CINRAM INTE
1
27 MITEL CORPC
1
47 NEWBRIDGE
1
69 CORECO INC 0.983382
84 ROYAL LASEf 0.683747
1,075,691
220E-17
5.55E-17
1.20E-16
1.19E-02
0.102096
1.f 4E-01
1,257,378
8,125.481
44,923
51,Si
1.SE-16
4.87E-16
1.OSE-15
1.ME-01
8.96E-01
t.00+00
VALUINGPRIVATE COMPANIES:
A DEA APPROACH
2.08E-1 O
6.12E-10
8.5SE-09
4.69+03
4.62+04
$50,868 fower bound
196
Score
(lambda)
FIFTY-PLUS.NET Il
lr
0.8732 1
Real
-
%2,404 (000s)
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
197
Score
85
-ricet
W E XL. H
c 0.94369
VALUINGPR~VATE
COMPANIES:
A DEA APPROACH
85.2 18
45.239
1295.904
10 14.082
437.190
9377.226
10604.658
1097.297
7845.925
28902.052
32707.433
2083.925
396 1.346
$100,460 REAL MC
$100,000
198
APPENDIX D
Analysis of Efficient and Inefficient
Companies SIC 3700
DMU
#
~arket
PRInUf Capiblbiblon
sic (ooor)1998
116,772
Efficient Peer
3
1.00000
Nom
1 GENERALM(
1
2 AE.VENTUR 0.8
4 DECOMA IN1 0.58
5 DEVTEK COF 0-67
6 LINAMARC0
1
7 MAGNAINTE
1
16 GLENDALEIt
17 AVCORP IN0
18 HEROUX INC
19 MAGELLAN P
20 SERENA REE
21 TEMPLAR RE
22 BOMBARDlEl
0.52
0.5
0.59
0.63
0.73
1
1
Sum:
3.75
MC Calc
0-0000
0.6006
68469.6354
23467,0717
0.0000
0.0000
0.0000
339.8526
0.0000
0.0000
8571.3271
61425.8952
0.0000
1i75.8158
5549.4522
1604.6554
4541.1077
68970.6190
0.0000
0.0000
0.0000
$24,116 Caltulated
1
$116,772 Real
This is a lower bound for the efficient Company, BUDD Canada, based on the
inefficient companies k i n g compand to it
- --
VALUING
PRIVATECOMPANIES:
A DEA APPROACH
200
DMU
HWIW
EfkientPeer
PWYI Cpiblizaa
6
SIC
6 LINAMAR CC
3 BUDD CANAI:
1 GENERAL Mt
2 AE-VENTUR
4 DECOMA lNT
5 OEVTEK COF
7 MAGNA INTE
8 MERCH PERl
9 NORTHSIDE
10 ROSE CORPt
11 SPECTRA PF
12 STACKPOLE
13 TESMA INTEf
14 WESCAST IN
15TRAkMOBlLf
16 GLENDALE lt
17 AVCORP IND
18 HEROUX INC
19 MAGELLANP
20 SERENA RES
21 TEMPLAR RE
22 BOMBARDIEI
1
1
1
0.79684
0.57823
0.66674
1
0.65439
0.89812
0.70626
1
0.63128
0.99948
1
0.97208
0.51504
0.50231
0-5871
0.62997
0.72956
1
1
3714
(ooos) 11,?9%5it
N o m MC Calc.
O
0.0000
3714
iis.nz 0.0000
0.0000
O
0.0000
3711 86,392,582
0.0000
3714
627
0.0000
O
3714
560.656
0.4793 0.386 216331-3983
3714
99.709
O. 1181 0.095
9481-6534
0.0000
O
0.0000
3714 7,374,103
0,0000
3714
2.869
0.0000
O
256.8636
37:4
14.056
0.0227 0.018
3714
25,150
0.0000
O
0-0000
0.0000
3714
525,280
0.0000
O
0.0000
O
0-0000
3714
131,497
3714
570.124
0.4697 0.378 215595-31O3
0.0000
3714
574,300 0.0000 7E-17
0-0000
3715
14.342
0.0000
0
0.0000
O
0.0000
3716
47,948
0-0000
3n8
#,no 0.0000
O
0.0000
O
0.0000
3728
62,462
60868.6475
3728
496,389
0.1 523 0.123
0.0000
3731
m 0.0000
O
0.0000
3732
4,576
0.0000 ".;
O
3743 7 . ~ 2 . 0 4 9
O
0.0000
Sum:
1.242f 4
1
$502,534 Calculated
$1,792,517 Real
This is a lower bound for the efficient Company, LlNAMAR CORP;, brsed on the
inefficient companies k i n g compared ta it
VALUING
PRIVATECOMPANIES:
A DEA APPROACH
20 i
Scon,
hrkut
PRIM ~apitai:
Dbtince
Efficient Peem
3
11
21 Distance
sic (ooos)1998
# AVCORP 0.502312 3728 45.730 0.0000
627 0.4565
2 AE. VEN' 0.796841 3714
3 BUDD CA
1 3714 116,772 1.1322
1 3714 525.280 0.5643
# SPECTR/
# TEMPLAF
1 3732 4,576 0.4889
0.131
0.004
1.000
0.000
0.000
0.415
0.015
0,000
1.000
0.000
0-453
0-982
0.000
0-000
1-000
0.000
0-456
1.132
0.564
0.489
is $235,632 million
--
VALUING
PRIVATE
COMPANIES:
A DEA APPROACH
202
APPENDIX E
Analysis of Efficient and Inefficient
Companies SIC 2000
Score
IY-
PRIMARICapiblitlti
SIC
(Os)1998
43.0000 Nom.
MC Calc.
7 SAPUTq 0.787
15 RlDLEL 0.9894
10 LASSO 0.9632
31 VINCOf 0.7498
25 SLEEM 0.8323
35 CLEAR 0.9535
28 ANDRE 0.7103
27 UNIBRC 0.9878
44 ARCTIC 0.6402
22 BIG RO 0.6985
23 BRICK 0.7797
29 MAGNC 0.6915
8 DELICI4 0.5398
39 URBAN 0.8089
11 GLOBA 0.5749
20 AFRICF 0.8332
34 ARROV 0.8705
32 HlGHW 0.71O3
26 TREE E 0.8365
1
24 FOSTE
1
17 CANAD
LOWER BOUND
VAUJINGPRIYATE
COMPANIES:
A DEA APPROACH
204
4 AMT FINE FI
Efficient
capiti~h;Peer
(ooor)is! AMT
nom.
mc calc.
ARROWHEP 0.8705
TML FOODS 0.2571
MAGNOTTA 0.6915
BETA BRAN 0.3039
BRICK BREC 0.7797
LEADING BF 0.4515
UNIBROUE 1 0-9878
39 UR8AN JUlC 0.8089
14 AGRO PACll 0.7703
42 LEF MCLEAI 0.7971
3 SEPP'S GO1 0.5941
25 SLEEMAN B 0,8323
22 BIG ROCK B 0-6985
19 SMALL FRY 0.7801
28 ANDRES WI 0.7103
32 HIGHWOOD 0.7103
24 FOSTERS E
1
12 DOVER INDI
1
1
9 SUN-RYPE F
34
6
29
18
23
37
27
Score
0.082
0.078
0,078
0.078
0.077
0.077
0.073
0.071
0.071
0-067
0.065
0-057
0-056
0.029
0.024
0.015
0.000
0.000
0.000
1-000
579
$139
$1,037
$3,732
$1,192
$659
$2.478
3917
S774
$228
91,778
$7,230
$1,098
$786
$1,917
$36
$0
$0
$0
$24,079 thouiand
lower bound
205
Efficient Peer
SCHNEICAMT FIN FLETCHI Diaance
10,874 6.91E-02 0.85666 7.43E-02
0.0000 0.0094
8.~10
1-23E-03 0,92561 7.32E-02 0.0094 0.0000
1.6060 1-8597
139.628
t
MC
S(000s)
11,173
5,936
1
1
0.0308
1.5956
0.0109
1,7158
Upper Bound
$ 10,947 thousand
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
206
APPENDIX F
Analysis Utilizing Al1
452 Companies
Score
Distance
Enicient Pwrs
73
322
0.0000
18,426,080
0.667 0.187
3612
175 HAMMOt 0.51431
0.0268
0.764 0.064
45,927,284
2331
83 CHATEA 0.82904
0.751 0-088
66,664,560
3523 0.0248
59 BUHLER 0.87389
0.0366
0.565 0.341
2834
296 PATHEO 0.52526 142,814,041
OS43 0.156
0.0463
17,824,705
2034
226 LASSON 0.63848
0.440 0.547
2092 0.1897
3,380,635
231 LEFMCL 0.58608
0.383 0.305
141 EXCO TE 0-73515 138,405,738
3544
0.1483
0.4791
0.249 0.732
5,177,589
2875
99 CONSOL 0.44399
0.245 0-675
2821
0.4206
302 PFB COF 0.94048 24,080,255
0.226 0.707
3571
0,4685
284 NTS COB 0.60877 48,091,341
0.5248
0.164 0.708
3716
163 GLENDA 0.31597 47,948,068
71,058,787
0.4547
0.158 0.625
2851
360 SIC0 INC 0.5653
0.6923
0.147 0.831
13,310,001
2084
242 MAGNO1 0.39093
80,755,275
3081 0.5509
0.1 10 0.658
210 IPL INC 0.69724
0,075 0.899
3661
0.8635
3 AASTRA 0.69365 46,090,505
0.7992
0.054 0.838
3086
308 POLYAIF 0.41828 27,260,000
0.029 0.944
0.9865
7,109,353
2621
93 CML IND 0.50334
0.013 0.937
0.9949
3585
261 MOBILE90,52422 33,407,325
1.000 0.000
9,240,498
3949
0.1564
73 CARBITE
1
0.000 1.000
3841
1.1160
322 PROMA1
1 10,374,300
0.000 0.000
1-3514
2522
1 87,643,098
372 SMED IN
0.000 0.000
1-3404
3714
443 VVESCAS
1 574,299,775
$57,473,975 Upper bound h e d on Enicient Pwrs
Initially, Hammond was found to have 21 peers
VALUING
PRIVATE
COMPANIES:
A DEA APPROACH
208
MAGNO1
POLYAlF
AASTRA
CML IND
MOBILE
0.39093
0.41828
0.69365
0.50334
0.52422
13.310,001
27,260,000
46,090,505
7.109,353
33,407,325
RANKED BY EFFlClENCY
GLENDA 0.31597 47,948,068
MAGNO1 0.39093 13.310,001
POLYAlF 0.41828 27,260,000
CONSOL 0.44399
5,177,589
CML IND 0.50334
7,109,353
HAMMOt 0.5 1431 18,426,080
MOBILE 0.52422 33,407,325
PATHEO 0.52526 142,814,041
SIC0 INC 0.5653
71,058,787
LEF MCL 0.58608
3,380,635
NTS COIi 0.60877 48,091,341
LASSON 0.63848
17,824,705
AASTRA 0.69365
46,090,505
IPL INC 0.69724 80,755,275
EXCO TE 0.73515 138,405,738
CHATEA 0.82904 45,927,284
CHATEA 0.87389 66,664,560
CHATEA 0.94048 24,080,255
CHATEA
1
9,240,498
RANKED BY MARKET CAPITALKATION
LEF MCL 0.58608
CONSOL 0.44399
CML IND 0.50334
CARBITE
1
MAGNO1 0.39093
LASSON 0.63848
HAMMOt 0.51431
PFB COF 0.94048
POLYAlF 0.41828
MOBILE 0.52422
CHATEA 0.82904
AASTRA 0.69365
GLENDA 0.31597
NTS COB 0.60877
BUHLER 0.87389
SIC0 INC 0.5653
IPL INC 0.69724
=CO TE 0.73515
PATHEO 0.52526
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
209
MC
443
1NORM
MCCALC
62 CAMCO INC
0.6324
0.6055
0.432
0.4815
0.9171
0.5143
0.5253
0.9405
0.7518
0.5659
0.5094
0.5592
0.4183
0.6088
0.967
0.5702
0.316
0.5242
0.4344
o.ni9
0.9085
0.4766
0.3909
0.5671
0-8532
0.5592
0.5033
0.444
0.6215
0.8818
0.6936
0.7287
0.8089
0.5777
0.6784
0.6288
0.4861
0.5216
0.6417
0.5427
0.55
0.5861
0.6802
0.2496
0.9812
0.7845
0.7962
1
1
VALUING PRIVATE
COMPANIES: A DEA APPROACH
$408,330 thousand
574,300 thousand
211
APPENDIX G
SIC 2800
DMU na1Score MC
min
max
INEFFICIENT COMPANIES
50 ACETUi 0.486
$71,059 one peer
15 ALTARE 0.858
$2,931
$57,159
1 AMR TE 0.697
$24,080 $153,632
5 AT P M 0.516
$24,080 $974,676
17 AXCAN 0.954
$24,080 $974,676
39 BESTAF 0.754
$3,725
$44,269
18 BIOTEC 0.554
$530
$32,776
51 CANHO10.385
$6,577 one peer
44 CANTO10.902
$3,725
$6,577
45 CCL INC 0.877
$767,740 one peer
46 CEAPRC 0.991
$936
$2,996
61 CHEMBt 0.752
$2,846
$3,207
47 CONS010.515
$2,846
$10,546
57 CONS010.629
$6,577 one peer
34 CV TEC10.778
$530
$32,776
35 DIAGNC 0.630
$2,846 $140,743
66 DIVERS:0.748
$2,846 one peer
12 FYTOKE 0.883
$936
$10,546
36 GENOM 0.823
$2,846 one peer
20 HAEMA( 0.474
$19,046
$77,292
58 HEDLnt 0.753
$530
$32,776
21 HILLES10.81 1
$2,846 one peer
37 IBEX TE 0.871
$2,931
$57,159
38 IMI INTE 0.960
$530
$32,776
22 INEX Pk 0.937
$2,931 $140,743
62 KRYSTP 0.745
$530
$32.776
23 LABOPF 0.886
$1,488 $140,743
4 LITHOS 0.489
$530
$32,776
7 MAGNIF 0.821
$3,725
$78,451
13 MARINE 0.621
$2.931
$77,292
- -.
59 NAPIER 0.880
$32,776
$530 $140,743
63 NORDE 0.727
$530
$936
$12,657
8 NOVA C 0.679 $1,838,447 $4,O81,7iO one peer
60 NU-GRC 0.866
$41,449
$24,080
$71,059
25 ORALIF10.985
$2,931
$6,895
$57,159
26 PATHEC 0.684
S142,814
$24,080 one peer
27 PHARMI 0.177
$4,651
$2,846 one peer
14 PROME 0.591
$78,451
$20,940 $140,743
29 SALlX P 0.861
$14,803
$530
$32,776
41 SIGNAL10.766
$17,484
$2,931
$57,159
42 STRES5 0.825
$44,932
$140,743 one peer
43 SYNSOF 0.489
$77,292
$2,931 $140,743
$107,756
30 TECHNI 0.871
$24,080 $153,632
31 THERAI 0.903
$57,159
$2,931 $140,743
32 UPLANC 0.868
$1 1,221
$3,725
$13,328
$80,431
$51,195
$15,545
$140,881
$72,892
$259,036
$87,578
$9,949
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
(mal-max) (mal-upper)
ireal
/mal
--
73.31 %
-79.40%
-2.52%
96.42%
-121-42% -245.85%
83.24Oh
-226.92%-213.23% #VALUE!
-213.54%
96-21%
-82.09%
42.57%
-140.39%
77.65%
-146.23%
-53.22%
94.87%
66.80%
-18.78%
28134 within range
33/45 upper
than mal
213
Real
2 BORDE1
3 DUPON'
9 PFB CO
10 POLY-P,
1IACTA NI
16 ANGlOT
19 BlOVAlL
24 NATRAC
28 QLT PH(
33 MOCHE
40 CANGE1
48 POLYME
49 SIC0 IN1
52 LANDO\
53 METHAI
54 AGRIUN
55 POTASF
56 TERRA I
64 AQUA-P
65 DIVERS
67 SHAW If
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Lowe?Boun real-lowerlreal
$3,725
$17,264 -363-51%
$4,081,770 $1,541,128
62.24%
$24,080
$102,332 -324M%
$1,251 self evaluatoi WALUE!
$6,577
$2,622
60.14%
$140,743
$36,862
73-81%
$1,404,647 self evaluatoi #VALUE!
$8,108 self evaluatoi #VALUE!
$974,676
$139,816
8566%
$4,708,260
$146,983
9688%
$221,802
$66,915
69.83%
$3,207
$3,600
-12.27%
$71,059
$73,206
-3.02%
$2,846
$9,042 -21 7-67%
51,370,334 self evaluatoi #VALUE!
$1,541,000 self evaluatoi #VALUE!
$5,281,848 self evaluatoi #VALUE!
$716,918 self evaluatoi #VALUE!
$6,895
$25,847 -274,87%
$13,328 self evaluatoi #VALUE!
$767,740
$539,758
29.70%
1/43 lower than mal
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
VALUINGPRIVATE
COMPANIES:
A DEA APPROACH
214