Economic Modelling 35 (2013) 565–575

Contents lists available at ScienceDirect

Economic Modelling
journal homepage: www.elsevier.com/locate/ecmod

Compliance and multidimensional firm performance: Evaluating the

efficacy of rule-based code of corporate governance
Yasir Bin Tariq a,⁎, Zaheer Abbas b
a
Department of Management Sciences, Mohammad Ali Jinnah University, Off Islamabad Expressway, Kahuta Road, Sihala, Islamabad, Pakistan
b
Head, Department of Accounting and Finance, International Islamic University, Sector H-10, Islamabad, Pakistan

a r t i c l e i n f o a b s t r a c t

Article history: This study seeks to evaluate the efficacy of the Pakistani Code of Corporate Governance by finding out its impact
Accepted 13 August 2013 on firm's performance and efficiency. This exploration is done in the context that Securities & Exchange Commis-
Available online xxxx sion of Pakistan's choice of corporate governance regulations is heavily influenced by Anglo-Saxon approach,
whereas de-facto realities of Pakistani corporate environment are quite in contrast.
Keywords:
Using a panel data of 119 firms for the period of 8 years i.e. 2003 to 2010, and using a multidimensional perfor-
Corporate governance
Codes
mance framework i.e. financial performance and technical efficiency, we find that the extent of compliance has
Panel data increased since the issuance of code in 2002. After controlling for firm size, growth, dividend payout, age and le-
Data envelopment analysis verage, we find significant positive impact of compliance on firm's performance (ROA, ROE and ROCE). We also
Financial performance find a weak positive relationship between compliance and technical efficiency. We suggest that compliance is not
Technical efficiency linearly related with financial performance and we find that high compliant firms are less profitable than average
Emerging economies or low compliant firms. This implies that one-size-fit all approach along with mandatory compliance is a ques-
tionable approach for Pakistani firms.
This study is first in Pakistan in providing empirical evidence on efficacy of the rule-based Code of Corporate Gov-
ernance and also adds to growing but underdeveloped literature on compliance and firm performance in emerg-
ing/developing economies. Further, this study offers insight to policy makers on the efficacy of current corporate
governance regulations and offers a research framework for assessing the extent of compliance, effectiveness and
economic impact of code of corporate governance.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction
The series of high profile corporate collapses (Claessens and
Yurtoglu, 2012; Davies and Schlitzer, 2008; Lavelle, 2002) due to fraud
and inadequate systems of check and balances brought the issue of gov-
ernance in corporate form of the business in the spotlight. This world-
wide attention made corporate governance reforms and regulations, a
priority agenda for governments and market regulators around the
globe. This emphasis is evident from the number of published reports
around the world, during the last two decades, e.g. Cadbury Report
(1992) and Higgs Report (2003) in the UK, Viénot Report (1995) in
France, King-I (1994) and King-II (2002) in South Africa, Peters Report
(1997) in the Netherlands, Cardon Report (1998) in Belgium and
Preda Report (1999) in Italy.
As a result, as these reports have formed basis, governments around
the globe have started addressing the problem of corporate governance
by introducing and implementing governance codes for corporate form
of business, using different implementation protocols i.e. rule based or
⁎ Corresponding author.
E-mail addresses: yasirbintariq@gmail.com (Y.B. Tariq), zaheerabbas@iiu.edu.pk
(Z. Abbas).
principle based. The rise and wide acceptability of codes of good gover-
nance can be seen from the fact that there were 72 codes in 24 countries
in 1999 (Aguilera and Cuervo-Cazurra, 2004) and in 2008 the website
for European Corporate Governance Institute reported 189 codes in 63
countries (Zattoni and Cuomo, 2008). As per information on ECGI's
website1 (as of February 2013), there are 92 countries and many inter-
national organizations (e.g. World Bank, OECD) that have issued one or
more codes of good governance.
Following the international demand for governance reforms in the
capital markets, Pakistan has introduced a number of reforms, which in-
clude introducing new legislation to strengthen the equity market liber-
alization process and development and implementation of corporate
governance codes. In March 2002, Securities and Exchange Commission
of Pakistan (SECP) issued the code of corporate governance (SECP,
2002) for publicly listed companies. The requirements of the code are
heavily influenced from the UK and US governance regulations. In con-
trast to widely applicable “comply or explain” approach, Pakistan
followed the US rule based approach and made the requirements of
the code as mandatory.
1
www.ecgi.org (as on 18th February 2013).

0264-9993/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.econmod.2013.08.015
566 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575
The corporate governance structure in any country is determined by
several factors; the de facto realities of the corporate environment in
the country; the legal and regulatory framework that defines the rights
and responsibilities of stakeholders involved; and each firm's articles of
association. While developing the code of corporate governance
Pakistan was also influenced by the international advocacy of Anglo-
American model2. Pakistan presents an interesting case for studying
the impact and efficacy of the code of corporate governance the de
facto realities of the corporate environment are quite in contrast from
the system of corporate governance adopted. The Pakistani capital mar-
ket is characterized by concentrated ownership structure through cross
shareholdings and pyramid ownership structure (Claessens et al., 2000;
Thillainathan, 1999), family based business groups, debt (from banks)
as a preferable form of financing instead of equity, an underdeveloped
equity market and inactive market for corporate control, i.e. takeovers.
There exists a vast anecdotal and empirical evidence both in devel-
oping and developed world that good corporate governance can cause
or lead to improved financial performance and benefit shareholders
through access to more capital, reduction in cost of capital (Reddy
et al., 2010), free cash flow be distributed among shareholders rather
than being expropriated (Jensen, 1986; La Porta et al., 2002), reduction
of control rights of managers and the probability that managers make
investment decisions that can enhance shareholder value (Shleifer and
Vishny, 1997). It is argued that poor quality regulations can increase
compliance costs for business and other groups, cause unnecessary com-
plexity and associated uncertainty as to regulatory obligations and re-
duce the ability of government to achieve its objectives(OECD, 2008).
Therefore a corporate governance system which is compatible with cap-
ital markets and corporate culture will, theoretically, facilitate the busi-
ness and thus enables a firm to realize the previously noted benefits, in
sum, improved financial performance and efficiency.
Therefore in this context, the main purpose of this study is to find out
the effectiveness and efficacy of the current code of corporate gover-
nance by using finance approach to corporate governance i.e. by finding
out the relationship between compliance with firm's efficiency and fi-
nancial performance. This is accomplished by constructing a compliance
index to measure the extent of compliance with the code of corporate
governance and comparing it with multi-dimensional firm performance
measures. The firm performance is measured through financial perfor-
mance measures (using accounting ratios) and efficiency measures
(using DEA).
1.1. Significance and justification of the study
Neither Securities and Exchange Commission of Pakistan (SECP) nor
any other study in Pakistan has attempted to evaluate the effectiveness
of the code of corporate governance or addressed compliance perfor-
mance relationship. Therefore there is a dire need for evaluating the
economic impact and effectiveness of the code. The evaluation method-
ology in the field of law is quite different and as in general “Codes” are
less like laws and more like guidelines/best practices. Also they are
equally applicable to all listed firms and every firm is required to comply
with it within its unique firm setting, therefore it makes sense to evalu-
ate outcomes of firm's core activities while complying with the code, to
use as indicator of the impact and effectiveness of the code. Therefore in
line with economics and social sciences research we are using outcomes
(financial & operational performance) to measure the effectiveness of
the process i.e. the code of corporate governance. Also this is the first
study which offers evidence on the economic consequences of comply-
ing with the code of corporate governance for Pakistani listed firms.
Further, the extant research in Pakistan has suffered from three main
limitations. First, mostly the use of only large firms e.g. KSE-100, second
2
Soederberg (2003) noted that “despite the claim that the international standard of
corporate governance embodies ‘universal principles’, the definition advanced in the
ROSCs intentionally draws on the Anglo-American variant”.
shorter study periods, and third use of market or market based hybrid
measures e.g. Tobin's Q. The problem with using market based mea-
sures is that our stock markets are inefficient, very volatile and there
is clear evidence of insider trading and price manipulation by colluding
stock broker (Hameed and Ashraf, 2006; Iqbal, 2012; Khawja and Mian,
2005; Mirza and Afzal, 2009; Tariq and Butt, 2008). This problem is
common with stock markets of emerging economies (Mei et al., 2004;
Siddiqi, 2007). Results and findings based on such measures can be mis-
leading and questionable.
This study is first to attempt to overcome these shortcomings by an-
alyzing a large sample of 119 firms for Pakistan (large, medium and
small sized firms with wider presence) over a longer period of time,
2003 to 2010, i.e. 8 years and employing multi-dimensional perfor-
mance measures to assess the impact on both firm's financial and oper-
ational performance. Further this study contributes towards a smaller
branch of literature that combines productivity literature with corpo-
rate governance literature.
2. Literature review
Ariff and Ratnatunga (2008) assert that the concept of corporate
governance can be looked upon from four viewpoints i.e. economic,
legal, societal (social) and applied finance. This study is attempting to
assess the efficacy of SECP's code of corporate governance by employing
applied finance viewpoint towards corporate governance.
The theoretical foundations for corporate governance research can be
found in Agency theory (Alchian and Demsetz, 1972; Jensen and
Meckling, 1976), which pointed out the conflict of interest between man-
agement and owners due to separation of ownership and control. To min-
imize this divergence between interest of mangers and shareholder and
to reduce the agency cost, agency theory suggests the installation of inter-
nal and external control mechanism. This is what recently known as cor-
porate governance (Haniffa and Hudaib, 2006). Shleifer and Vishny
(1997) affirmed that the main purpose of the corporate governance is
to provide a sense of security to capital providers. The two other theories
that have been used separately or with agency theories are; information
asymmetry and managerial signaling theory (Kapopoulos and Lazaretou,
2007; Rhee and Lee, 2008) and convergence theory (Clarke, 2004a,b).
Mega corporate failures, globalization and increased awareness of
investors are the triggers of worldwide popularity and diffusion of
good governance codes (Iturriaga, 2009; Mallin, 2007). The extant of re-
search on codes can be divided in to three main categories.
First category includes those studies that have examined the extent of
compliance with codes of corporate governance within a single country or
across multiple countries (e.g. Akkermans et al., 2007; Alves and Mendes,
2004; Berglöf and Pajuste, 2005; Chizema, 2008; Hooghiemstra and van
Ees, 2011; Seidl et al., 2012; Talaulicar and Werder, 2008; Werder et al.,
2005). Some of these studies have attempted to identify reasons for
non-compliance. Second set of studies are those that have examined the
effect of compliance with codes of corporate governance on corporate be-
havior like CEO turnover, earnings management and board structure (e.g.
Chen et al., 2011; Dahya et al., 2002; Dedman, 2003). And finally the third
category includes those studies that have investigated the relationship of
compliance with the codes of corporate governance and financial perfor-
mance of firms. Here we will take a brief account of these third types of
studies.
There is evolving but still premature (Talaulicar and Werder, 2008;
Teh, 2009) literature aiming at exploring the relationship between com-
pliance and performance. Reddy et al. (2010) investigated the empirical
question regarding code compliance and financial performance for New
Zealand. They carefully concluded that there is evidence that the recom-
mendation of code has positive influence on firm performance. The pres-
ence of remuneration committee as recommended by the code is also
positively related to the firm performance. Goncharov et al. (2006) re-
ported for Germany that firms with higher degree of compliance are
priced at premium in contrast to the firms with low degree of compliance.
 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575
The author's findings also support the hypothesis that due to capital mar-
ket pressure, boards adopt codes' recommended changes. For European
firms the evidence is mixed for compliance–performance relationship.
Bauer et al. (2004) reported that though firm value is positively related
with governance ratings, firm performance as measured by ROE and
Net Profit Margin is negatively related with governance standards. Fur-
ther they reported substantial difference between UK markets and
Eurozone markets. In contrast to Bauer et al.(2004), Bauwhede (2009) re-
ported a positive relationship between operating performance and extent
of compliance with international best practices.
Cadbury committee recommendations in 1992 are one of the first in
good governance codes/recommendations landscape. There are a num-
ber of studies which have tried to evaluate the efficacy of these recom-
mendations by comparing & evaluating the financial performance of
listed companies that have adopted or followed those recommendations
(e.g. Arcot and Bruno, 2006; Dahya and McConnell, 2007; Dedman,
2003). The overall results are inconclusive; however, there is some evi-
dence that due to implementing these recommendations, overall board
oversight capacity has increased.
de Jong et al. (2005) examined the relation between firm value and
corporate governance structure before and after the implementations
of “The Peter Committee” recommendations in the Netherlands. They
concluded that The Peter Committee recommendations had no effect
on corporate governance characteristics or on firm value. Alves and
Mendes (2004) concluded that there is a positive relationship between
compliance with some of these recommendations of Portuguese Securi-
ties Market Commission and abnormal stock returns. Fernández-
Rodríguez et al. (2004) reported inconclusive findings for Spanish firms
in compliance with the Olivencia Code.
For the US there are numerous studies that have tried to measure
the investor reactions on mandatory adoption of Sarbanes–Oxley Act
by listed companies. As with other countries, the results are mixed.
Larcker et al. (2005) and Li et al. (2004) reported a positive relationship.
In contrast, Zhang (2007) and Litvak (2007) reported a negative reac-
tion of investors towards adoption of recommendations of Sarbanes–
Oxley Act. Kouwenberg and Phunnarungsi (2012) found no significant
difference in market reaction for Thai listed firms. Price et al. (2011)
stated no association between compliance with code and operating per-
formance in Mexican firms.
There is a growing stream of literature, which combines productivity
literature with financial and corporate governance literature. This alter-
native measure of firm performance, i.e. DEA efficiency scores has sever-
al advantages over traditional firm performance measures and has been
employed by various studies (e.g. Bozec et al., 2010; Lin et al., 2009; Su
and He, 2011; Wang et al., 2007, 2012).
Table 1
Industry-wise distribution of sample firms.
Industry No. of sample firms
Auto manufacturers and assembly 8
Auto parts manufacturers 5
Cement 11
Chemical 13
Communication & networks 1 and stockbrokers give more weight to the amount and quality of infor-
Energy 5 mation regarding patterns of shareholding than percentage of indepen-
Engineering 5
Food & beverages 6
Glass & ceramics 5
Manufacturing (misc) 11
Oil & gas 10
Paper & board 3
Pharmaceutical 6 3
Services (misc) 1
Sugar 7 4
Textile 22
Total firms 119
567
On the basis of our review of prior theoretical and empirical litera-
ture, we formulate the following two main hypotheses with two sub-
hypotheses.
Hypothesis 1a. There is a positive impact of compliance with the Code
of Corporate Governance on firms' financial performance.
Hypothesis 1b. Firms that showed higher level of compliance with the
Code of Corporate Governance have superior financial performance
than firms with average and low level compliance.
Hypothesis 2a. There is a positive impact of Compliance with the Code
of Corporate Governance on firms' efficiency.
Hypothesis 2b. Firms that showed higher level of Compliance with the
Code of Corporate Governance are more efficient than firms with aver-
age and low level of compliance.
3. Research design & methodology
3.1. Sample and data
To avoid large firm's size bias and to ensure a wider presence of the
firms under study, we sampled those non-financial firms that are com-
monly listed on at least any of two stock exchanges out of three (i.e. KSE,
LSE & ISE). Final sample consisted of 119 firms, as for only these firms,
complete annual reports were available for the study period i.e. 2003–
2010 (8 years).
Financial data were obtained from State Bank of Pakistan publication
i.e. Balance Sheet Analysis of Joint Stock Companies (SBP, 2007, 2011)
whereas corporate governance and compliance data were obtained
(hand-picked) from published annual reports. Table 1 reports the in-
dustry wise distribution of sample firms used in this study.
3.2. Construction of corporate governance compliance index
SECP released the code of corporate governance for publically listed
firms in March 2002 and revised it in 20073. All stock exchanges were
required to add the requirements of the corporate governance code as
listing requirement. The code addresses six main areas i.e. board of direc-
tors, corporate and financial reporting framework, corporate ownership
structure, audit committee and compliance with the code of corporate
governance (SECP, 2002).
A custom corporate governance compliance index (as no such index
is available) is developed to measure the extent of compliance with the
code of corporate governance by the sample firms. In order to develop a
compliance index, first we break down the requirements of code of cor-
porate governance (clauses) into measureable blocks or units. Since not
all requirements are equal in importance, we assigned weights to differ-
entiate among important and just for reporting requirements4. Then
this index is shared with people from academia, stock exchanges (bro-
kers and regulatory staff) and regulators (SECP) to get feedback on the
adequacy of scoring criteria and weights. Based on their feedback, the
scoring criteria and weights were adjusted. For example, regulators
dent directors on the board.
The latest revision came in 2012, which is not covered by this study. We used the re-
quirements of the code applicable until 2011.
It is argued that a binary index can avoid subjectivity, but much of the quality informa-
tion is lost in this manner, therefore despite risk of subjectivity, we prefer a weighted com-
pliance index.
568 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575

3.3. Measurement and scoring criteria
The score on each requirement (clause) ranges from zero to five. A
score of zero is assigned in case of non-compliance and a score of 1 to
5 is assigned depending on the quality of information reported. Given
all clauses are applicable to a firm in a given year; the maximum score
a firm can secure on this compliance index is 500. Minimum score a
firm can get is 101 as many requirements of the code are also statuary
requirements of the Companies Ordinance 1984, to which companies
always shown compliance.
3.4. Measuring firm's efficiency using non-parametric data envelopment
analysis
trading and inefficiency, it is very difficult to estimate true market
value of firms. The results and findings based on such market measures
could be misleading. Therefore, we measured firm's performance in an
alternate framework of standard production function. This type of anal-
ysis allows us to look at the root of the corporate governance problem
i.e. inefficient usage of resources that is usually cannot be detected by
the traditional financial ratio analysis in the short run. As Copeland
et al. (2000) argued that based on the theory of value creation, firms
with higher productivity are more likely to create more value than
those with lower productivity. Therefore following Zheka (2005), we
argued that firm's efficiency and value go together and firm's efficiency
is more preferable measure when stock markets are inefficient, rigid
and volatile.

We used data envelopment analysis (DEA), a non-parametric linear
programming application to measure the firm's efficiency. Technical ef-
ficiency (TE) is measured through input oriented CCR model to deter-
mine the best practices frontier that is based on constant return to
scale (CRS) technology (Charnes et al., 1978). The ratio of the observed
to potential production is its efficiency level (Simar and Wilson, 2007).
The CCR score is called global technical efficiency.
We used total assets and costs (cost of goods sold and operating
cost) as inputs and revenue as output for DEA analysis. Our selection
of inputs and outputs are consistent with previous studies using DEA
to measure efficiency in multiple industry contexts.
3.5. Reasons for using firm's efficiency as a measure of performance
There is an increasing interest in determining the impact of corporate
governance by using a performance measure (Technical Efficiency) that
is directly linked to the production process (Nanka-Bruce, 2011). In the
perspective of agency theory, Shleifer and Vishny (1986) suggested that
controlling managers (or owner-mangers) sometimes undertake pro-
jects that do not add value or productivity to the firm. Technical efficiency
is helpful in this situation to gauge this productivity loss which otherwise
could have gone unnoticed in the short term, when financial ratios are
used to measure performance (Destefanis and Sena, 2007). The use of
frontier efficiency approach like DEA captures the evaluation of invento-
ries and depreciation in the short term than with financial ratios
(Destefanis and Sena, 2007; Pi and Timme, 1993; Sheu and Yang, 2005).
As we noted in chapter one, due to the very volatile nature of the
Pakistani stock market, the evidence of price manipulation, insider
3.6. Measuring firm's financial performance
In line with the corporate governance research, we used Return on
Assets (ROA), Return on Equity (ROE), and Return on Capital Employed
(ROCE) and Earnings per Share (EPS) to measure firm's financial
performance.
3.7. Control variables
Control variables help to neutralize the varying effects of other firm's
related characteristics, which are not directly addressed in the research
design, but they can exert either downward or upward pressure on the
dependent variable. Based on the review of literature, we used firm's
size, firm's age, firm's growth, leverage, dividend payout, industry and
year dummies as control variables to isolate the impact of compliance
on firm's efficiency and financial performance.
3.8. Summary of variables and measures
Table 2 reports the summary of all dependent and independent var-
iables used in this study.
3.9. Econometric models
The following econometric models are tested by employing econo-
metric techniques for pooled and time-series cross section (Panel
data) data.

Table 2
Summary of variables & measures.

Variable Symbol Proxy for Operationalization Expected sign

Independent & control variables

ðTotal Assetst Þ−ðTotal Assetst−1 Þ
Asset growth AG Growth ðTotal Assetst−1 Þ
+
Compliance score CGCI Compliance with code Score calculated from Corporate Governance Compliance Index for Pakistan & Malaysia +
D/E ratio DE Leverage Ratio taken directly from SBP's BSA for Pakistan and from DataStream for Malaysia –
DPS DPS Dividend policy Dividend per share (in Pak Rupees) +
Firm size FSz Firm size Natural log of total assets +
Firm's age Age Learning curve Number of AGM +

Firm performance variables: financial & efficiency

Return on assets
Return on equity
Return on capital employed
Earnings per share
DEA efficiency scores
ROA Financial performance EBIT divided by total assets
ROE Financial performance Net income divided by shareholder's equity
ROCE Financial performance EBIT divided by capital employed
EPS Financial performance Ratio taken directly from SBP's BSA for Pakistan and from DataStream for Malaysia
TE Firm's efficiency Inputs: total assets & costs (cost of goods sold and operating expenses)
outputs: revenue TE calculated through CCR model

Dummies
Industry Ind Industry ±
Year Year Time ±
Type Indtype Industry types Whether a firm is service or manufacturing firm ±
Compliance HCF Compliance Dummy for high compliant firms +
Compliance LCF Compliance Dummy for low compliant firms –
 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575 569

3.9.1. Financial performance model Table 4

Descriptive analysis- CG compliance index (2003–2010).

Mean Median St. dev Minimum Maximum

FPit ¼ β0 þ β1 CGCIit þ β2 AGit þ β3 Ageit þ β4 DEit þ β5 Fszit þ β6 DPS
CGCI score 364.21 365.76 48.64 203.76 459.80
þβ7 Year it þ β8 Indit þ β9 Indtypeit þ εit ð1Þ No. of ED 2.99 3.00 1.75 1 10
% of NEDs 0.62 0.66 0.22 0.10 0.93
where FP is the financial performance measured through Return on No. of BoD meetings 5.46 5.00 3.042 0 33
No. of days to AGM 110.13 117.00 20.26 55 210
Assets (ROA), Return on Equity (ROE), Return on Capital Employed
Audit committee size 3.35 3.00 0.73 3 7
(ROCE) and Earnings per Share (EPS). %NED on AC 0.77 0.75 0.26 0 1.00
AC meetings 1.03 0.00 1.98 0.00 9.00
3.9.2. DEA efficiency performance model

control. The board of directors of Pakistani sample firms, on average

TEit ¼ β0 þ β1 CGCIit þ β2 AGit þ β3 Ageit þ β4 DEit þ β5 Fszit þ β6 DPS
þβ7 Yearit þ β8 Indit þ β9 Indtypeit þ εit :
4. Data analysis, results and discussion
met 5 times a year with a standard deviation of three. On the other
ð2Þ hand, the average number of audit committee meetings is one. This dif-
ference indicates that audit committees, the average size of which is
approx. three are not very active. This also indicates the lack of spirit
in following the code of corporate governance.

This chapter presents the results and discussion of bi-variate and 4.1. Measuring efficiency through non-parametric data envelopment
multivariate data analysis. analysis (DEA)
Tables 3 & 4 report descriptive statistics of financial, control and
compliance index variables. We employed a two stage methodology for our efficiency analysis. In
Average assets' growth (AG) is 16%. The average ROA of sample the first stage we employed non-parametric linear programming tech-
firms is approx. 11%, whereas, the average ROE and ROCE is 14% and nique DEA to calculate the efficiency score of the sample firms. We
19.4% respectively. The average EPS is 8.793 rupees with a standard de- used Total Assets and Costs (Cost of goods sold and operating costs) as
viation of 13.12 rupees, which indicates the volatility of earnings among inputs and Total Revenue as outputs.
Pakistani sample firms. The mean DPS is PKR 4 but median DPS is PKR 1, For DEA analysis to be meaning full, a general rule of thumb is that the
this difference along with the value of standard deviation i.e. PKR 6 in- number of DMUs analyzed should be greater than three times the num-
dicates that most of the firms in our sample are either not paying divi- ber of inputs and outputs (Bozec and Dia, 2007; Cooper et al., 2001;
dend or paying a low amount. The average debt–equity ratio of 1.462 Wang et al., 2011). To fulfill this requirement we borrowed firms
indicates the higher levels of debt in the sample firms. The standard de- (which we called reference firms) from the same industry with approxi-
viation figures revealed the extent of volatility in the data. The average mate similar size and used them along with our sample firms in order to
age of firm in our sample is approx. 33 years with a standard deviation conduct a meaningful analysis.
of approx. 17 years. The median firm's age is 29.50 years. Table 6 reports the distribution of sample and reference firms used
CGCI scores are obtained from a custom-built compliance index that for data envelopment analysis to calculate efficiency scores.
measures the extent of compliance with the code of corporate gover- Except for Communication & Networks, Glass & Ceramics, Paper &
nance. Majority of the clauses of the Pakistani code of corporate gover- Board and Services (Misc), the number of firms analyzed in all other sec-
nance are statuary requirements of the Companies Ordinance 1984, so tors are at least nine. For these four sectors, the total population of firms
companies do show some level of compliance (at least on papers) is less than nine. In total, we used 166 Pakistani firms for data envelop-
with these requirements. This is why the minimum score (theoretically) ment analysis. Our correlation analysis indicates a high positive correla-
a firm can get is approx. 101. The average CGCI score is 364.21, which is tion among inputs & outputs (Table 5).
just below median. The minimum score obtained by any firm is 203.76 The DEA efficiency score ranges from zero to one when they are
and the maximum is 459.80. The average number of executive directors computed on the assumption of input orientation i.e. the inputs are
during 2003–2010 is approx. 3 and presence of non-executive directors minimized to achieve the current level of output. DEA measures relative
(percentage) is in the range of 10% to 93% with an average of 63% efficiency. An efficiency score of one means that the firm is relatively ef-
approx. There is not much increase in the percentage of non-executive ficient than its peers in the same industry. In other words, a firm with a
directors (independent directors included) since the implementation
of the code of corporate governance. A year wise analysis of board com- Table 5
position indicates that the average percentage of NEDs was 61% in the Sample and reference firms used for DEA.
year 2003, increased to a maximum of 65% in 2008 and then reduced Sectors (industry) Sample firms Reference firms Total firms
to 63% in 2010. This may indicate that family/group oriented firms
Auto manufacturer & assembly 8 3 11
(which makes majority at stock market) are not willing to let go of Auto parts manufacturers 5 6 11
Cement 11 0 11
Chemical 13 0 13
Table 3 Communications & networks 1 5 6
Descriptive analysis of financial variables (2003–2010). Energy 5 5 10
Engineering 5 4 9
Mean Median Standard deviation Minimum Maximum
Food & beverages 6 5 11
AG 0.16 0.13 0.23 −0.46 0.79 Glass & ceramics 5 2 7
DPS 4.05 1.00 6.60 0.00 25.72 Manufacturing (misc) 11 0 11
D/E 1.46 1.13 1.56 −2.25 5.03 Oil & gas 10 1 11
ROA 0.11 0.09 0.12 −0.15 0.36 Paper & board 3 4 7
ROE 0.14 0.14 0.21 −0.32 0.61 Pharmaceutical 6 3 9
ROCE 0.19 0.17 0.21 −0.32 0.70 Services misc 1 5 6
EPS 8.79 5.02 13.12 −21.95 37.41 Sugar 7 4 11
Fsz 8.10 7.99 1.59 3.93 12.33 Textile 22 0 22
Age 33.25 29.50 16.84 5.00 116.00 Grand total 119 47 166
570 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575

score of one is converting inputs into outputs more efficiently than its Table 7
peer firms. Criteria for defining compliance based groups.

The average technical efficiency score for sample firms is 0.90. Phar- Group Criteria Group title
maceutical sector has the highest average technical efficiency score
Group 1 Firms with compliance Low compliance
whereas the energy sector is the least efficient in terms of technical ef- score ≤ 40th percentile
ficiency with an average score of 0.73. Textile (0.95), Engineering (0.94), Group 2 Firms with compliance Average compliance
Auto Parts (0.94), Sugar (0.93) and Paper & Board (0.91) are among the score N 40th percentile ≤ 75th percentile
Group 3 Firms with compliance High compliance
sectors with an average technical efficiency score of more than 0.90.
score N 75th percentile
Auto Manufacturers, Services (Misc), Communications and Networks,
and Energy are sectors with the least technical efficiency scores.

4.2. Pair wise Pearson correlation
Table 6 reports the correlation among all variables for Pakistan. The
Corporate Governance Compliance Index (CGCI) score is positively and
significantly correlated with ROA, ROE, ROCE and EPS. The highest is
with ROA, where it is 36%. The low magnitude of the coefficients can
be attributed to high dispersion in the data set. CGCI is also positively
correlated with age, firm size, DPS and TE. ROA, ROE and ROCE are pos-
itively correlated with firm's age, but the magnitude of the coefficients is
low. Debt to equity ratio is negatively correlated with financial perfor-
mance measures, whereas it is positively correlated with firm size. TE
showed significant positive correlation with age.
4.3. Multivariate analysis
For multi-variate analysis of our panel data, we employed common-
effect and random-effect specifications of regression. To ensure
robustness of our results we also employed a more robust regression
specification i.e. PCSE Prais–Winsten regression with AR1 and panel
specific AR1 autocorrelation. PCSE–PW regression is robust to both au-
tocorrelation and heteroscedasticity problems (potential) in the data.
For DEA efficiency models, we employed pooled Tobit regression,
random-effect Tobit and a more robust bootstrapped regression.
Tables 8 to 11 reports the results of financial performance model for
three regression specifications i.e. common-effect (pooled OLS) regres-
sion with robust standard errors (regression No. 1, 2, 7, 8, 13, 14, 19 &
20), random-effect GLS regression (regression No. 3, 4, 9, 10, 15, 16,
21 & 22) and PCSE Prais–Winsten regression with correction for panel
specific AR1 autocorrelation (regression No. 5, 6, 11, 12, 17, 18, 23 & 24).
Here we will discuss results of our two robust specifications i.e.
random-effect GLS and PCSE Prais–Winsten regression only. CGCI is
found to positively and significantly relate with all four measures of fi-
nancial performance and under both specifications except for EPS
under PCSE–PW regression, where it is positive but insignificant. Fur-
ther except in EPS model, the magnitude of CGCI's coefficients is very
low. This may be an indication that the relationship between extent of
compliance and performance is not very much linear. To investigate
further, we divided the sample firms in three categories i.e. high compli-
ant firms (HCF), average compliant firms (ACF) and low compliant firms
(LCF). Table 7 outlines the criteria for defining these three compliance
groups.
We ran the regression again with high and low compliance dummies
included and reported the results side by side of the original analysis. The
findings are quite interesting. The dummy for low compliant firms, as
expected, the coefficient is negative in all cases except in regression 10
where it is positive but insignificant. The interesting observation is in re-
gard to the dummy of high compliant firms, where against our expecta-
tions they are negative. In case of ROA under GLS specification the high
compliant dummy is negative and significant. It means that highly com-
pliant firms are less profitable than average and low compliant firms.
There are two explanations for this unexpected result. First because of
the mandatory nature of compliance, the increasing level of compliance
may have put more pressure on firms in terms of cost and because of
conflict of indigenous governance structure with impose structure. Sec-
ond the relationship between compliance and firm performance may
not be linear. As a result beyond a certain point the increase in compli-
ance does not translate into improved performance.
Except for dummies based models, our findings are in line with
Bauwhede (2009), Dahya and McConnell (2007) and Alves and
Mendes (2004) reported positive association between compliance and
firm performance. Further we note that Arcot and Bruno (2006) reported
that firms that have deviated from code's requirement have performed
well, thus confirming the notion that optimal governance structure for
each firm is different.
Asset's growth has shown consistently positive and significant asso-
ciation with measures of financial performance and under all regression
specifications except with ROA where it is positive but insignificant.
Similar results have been cited by previous empirical studies e.g.
Klapper and Love (2004).
Firm's age is negatively and significantly associated with ROA, ROE
and ROCE. In contrast firm's age is positively associated with EPS but in-
significant. Andres (2008) and Tuschke and Sanders (2003) found neg-
ative correlation between firm's age and firm's performance of German
firms measured by return on sales, return on assets.

Table 6
Pair wise correlation among dependent, independent and control variables.

AG Age CGCI D/E DPS EPS Fsz ROA ROCE ROE TE

AG 1
Age .014 1
CGCI .060 .236a 1
D/E .074b −.039 .033 1
DPS .038 .360a .292a −.105a 1
EPS .216a .269a .255a −.091a .664a 1
Fsz .130a .118a .379a .226a .193a .242a 1
ROA .102a .169a .359a −.177a .524a .637a .162a 1
ROCE .130a .113a .311a −.144a .529a .573a .119a .785a 1
ROE .182a .094a .258a −.369a .439a .546a .147a .614a .725a 1
TE .047 .151a .127a −.020 .229a .361a .047 .522a .343a .217a 1
a
Correlation is significant at 0.01 level.
b
Correlation is significant at 0.05 level.
 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575 571

Table 8
Compliance & financial performance (ROA).

1a 2a 3b 4b 5c 6c

Return on Asset (ROA)

Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

(standard error) (standard error) (standard error) (standard error) (standard error) (standard error)

CGCI 0.00039⁎⁎⁎ 0.00030⁎⁎ 0.00029⁎⁎⁎ 0.00027⁎ 0.00020⁎⁎⁎ 0.00012

(0.00008) (0.00014) (0.00011) (0.00015) (0.00008) (0.00013)
AG 0.02553⁎ 0.02437⁎ 0.01209 0.01096 0.00220 0.00107
(0.01444) (0.01442) (0.01076) (0.01072) (0.00935) (0.00940)
Age −0.00058⁎⁎ −0.00057⁎⁎ −0.00001 0.00001 −0.00073⁎⁎ −0.00076⁎⁎
(0.00023) (0.00023) (0.00049) (0.00049) (0.00035) (0.00036)
DE −0.01165⁎⁎⁎ −0.01191⁎⁎⁎ −0.00642⁎⁎⁎ −0.00663⁎⁎⁎ −0.01062⁎⁎⁎ −0.01089⁎⁎⁎
(0.00239) (0.00237) (0.00207) (0.00206) (0.00267) (0.00264)
Fsz 0.01243⁎⁎⁎ 0.01300⁎⁎⁎ −0.00814⁎ −0.00746 0.00212 0.00295
(0.00316) (0.00317) (0.00490) (0.00493) (0.00404) (0.00402)
DPS 0.00791⁎⁎⁎ 0.00793⁎⁎⁎ 0.00860⁎⁎⁎ 0.00857⁎⁎⁎ 0.00800⁎⁎⁎ 0.00797⁎⁎⁎
(0.00068) (0.00069) (0.00069) (0.00069) (0.00065) (0.00065)
HCF −0.00667 −0.01837⁎ −0.00616
(0.01078) (0.01043) (0.00766)
LCF −0.01496 −0.01432 −0.01433⁎
(0.01048) (0.00938) (0.00759)
Constant −0.143⁎⁎ −0.14506⁎⁎⁎ −0.00162 0.00826 −0.00510 0.02409
(0.065) (0.05615) (0.05252) (0.06708) (0.03919) (0.05189)
Obs 950 950 950 950 950 950
R2 0.42 0.42 0.38 0.38 0.55 0.55
Adj R2 0.41 0.41
F 24.60⁎⁎⁎ 22.96⁎⁎⁎
Wald χ2 308.2⁎⁎⁎ 330.3⁎⁎⁎ 187039⁎⁎⁎ 504172⁎⁎⁎

Coefficients for year, industry type and industry dummies are not reported. Standard errors are in parentheses.
a
Common effect (pooled OLS) regression with robust standard errors.
b
Random-effect GLS regression.
c
PCSE Prais–Winsten regression with PSAR1 autocorrelation structure.
⁎ p b 0.1.
⁎⁎ p b 0.05.
⁎⁎⁎ p b 0.01.

Table 9
Compliance & financial performance (ROE).

7a 8a 9b 10b 11c 12c

Return on Equity (ROE)

Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

(standard error) (standard error) (standard error) (standard error) (standard error) (standard error)

CGCI 0.00026⁎ 0.00035 0.00028 0.00050⁎ 0.00033⁎⁎ 0.00029

(0.00014) (0.00024) (0.00019) (0.00028) (0.00015) (0.00024)
AG 0.13138⁎⁎⁎ 0.13269⁎⁎⁎ 0.11933⁎⁎⁎ 0.11868⁎⁎⁎ 0.09390⁎⁎⁎ 0.09447⁎⁎⁎
(0.02421) (0.02421) (0.02145) (0.02150) (0.02078) (0.02055)
Age −0.00167⁎⁎⁎ −0.00168⁎⁎⁎ −0.00160⁎⁎ −0.00160⁎⁎ −0.00247⁎⁎⁎ −0.00239⁎⁎⁎
(0.00044) (0.00045) (0.00070) (0.00069) (0.00044) (0.00046)
DE −0.05321⁎⁎⁎ −0.05292⁎⁎⁎ −0.06142⁎⁎⁎ −0.06135⁎⁎⁎ −0.06146⁎⁎⁎ −0.06193⁎⁎⁎
(0.00513) (0.00509) (0.00396) (0.00396) (0.00729) (0.00722)
Fsz 0.03321⁎⁎⁎ 0.03256⁎⁎⁎ 0.02485⁎⁎⁎ 0.02594⁎⁎⁎ 0.02161⁎⁎⁎ 0.02167⁎⁎⁎
(0.00529) (0.00529) (0.00793) (0.00788) (0.00677) (0.00685)
DPS 0.00900⁎⁎⁎ 0.00898⁎⁎⁎ 0.01025⁎⁎⁎ 0.01025⁎⁎⁎ 0.01017⁎⁎⁎ 0.01020⁎⁎⁎
(0.00113) (0.00112) (0.00123) (0.00123) (0.00127) (0.00131)
HCF 0.00799 −0.02781 −0.01412
(0.01819) (0.02017) (0.01438)
LCF 0.01639 0.00778 −0.01466
(0.01865) (0.01856) (0.01629)
Constant −0.10845⁎ −0.14372 −0.04853 −0.13580 0.00220 0.02119
(0.05764) (0.09535) (0.08244) (0.11504) (0.05739) (0.07939)
Obs 950 950 950 950 950 950
R2 0.44 0.44 0.59 0.58 0.45 0.46
Adj R2 0.42 0.42
F 23.50⁎⁎⁎ 22.17⁎⁎⁎
Wald χ2 495.3⁎⁎⁎ 500.9⁎⁎⁎ 8.6e + 08⁎⁎⁎ 5.23e + 06⁎⁎⁎

Coefficients for year, industry type and industry dummies are not reported. Standard errors are in parentheses.
a
Common effect (pooled OLS) regression with robust standard errors.
b
Random-effect GLS regression.
c
PCSE Prais–Winsten regression with PSAR1 autocorrelation structure.
⁎ p b 0.1.
⁎⁎ p b 0.05.
⁎⁎⁎ p b 0.01.
572 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575

Table 10
Compliance & financial performance (ROCE).

13a 14a 15b 16b 17c 18c

Return on Capital
Employed (ROCE)

Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

(standard error) (standard error) (standard error) (standard error) (standard error) (standard error)

CGCI 0.00060⁎⁎⁎ 0.00048⁎ 0.00045⁎⁎ 0.00049⁎ 0.00035⁎⁎⁎ 0.00035⁎

(0.00015) (0.00025) (0.00019) (0.00029) (0.00011) (0.00021)
AG 0.06508⁎⁎ 0.06693⁎⁎ 0.03536⁎ 0.03492 0.04765⁎⁎⁎ 0.04725⁎⁎⁎
(0.02615) (0.02620) (0.02127) (0.02133) (0.01791) (0.01814)
Age −0.00166⁎⁎⁎ −0.00167⁎⁎⁎ −0.00137⁎ −0.00137⁎ −0.00199⁎⁎⁎ −0.00193⁎⁎⁎
(0.00044) (0.00044) (0.00076) (0.00075) (0.00064) (0.00062)
DE −0.01414⁎⁎⁎ −0.01385⁎⁎⁎ −0.00875⁎⁎ −0.00896⁎⁎ −0.01107⁎⁎ −0.01113⁎⁎
(0.00528) (0.00525) (0.00398) (0.00399) (0.00538) (0.00533)
Fsz 0.01583⁎⁎⁎ 0.01496⁎⁎ 0.00997 0.01076 0.00115 0.00126
(0.00585) (0.00583) (0.00840) (0.00836) (0.00886) (0.00926)
DPS 0.01377⁎⁎⁎ 0.01369⁎⁎⁎ 0.01386⁎⁎⁎ 0.01386⁎⁎⁎ 0.01284⁎⁎⁎ 0.01282⁎⁎⁎
(0.00123) (0.00123) (0.00127) (0.00126) (0.00154) (0.00155)
HCF 0.02450 −0.01370 −0.00305
(0.01888) (0.02027) (0.01409)
LCF 0.00158 −0.00408 −0.00341
(0.01910) (0.01851) (0.01634)
Constant −0.08481 −0.03794 0.00346 −0.01565 0.13674⁎ 0.13498
(0.06725) (0.09765) (0.08796) (0.11918) (0.07407) (0.10229)
Obs 950 950 950 950 950 950
R2 0.40 0.40 0.40 0.40 0.45 0.44
Adj R2 0.39 0.39
F 21.37⁎⁎⁎ 20.44⁎⁎⁎
Wald χ2 295.9⁎⁎⁎ 300.1⁎⁎⁎ 2.8e + 07⁎⁎⁎ 9.4e + 07⁎⁎⁎

Coefficients for year, industry type and industry dummies are not reported. Standard errors are in parentheses.
a
Common effect (pooled OLS) regression with robust standard errors.
b
Random-effect GLS regression.
c
PCSE Prais–Winsten regression with PSAR1 autocorrelation structure.
⁎ p b 0.1.
⁎⁎ p b 0.05.
⁎⁎⁎ p b 0.01.

Table 11
Compliance & financial performance (EPS).

19a 20a 21b 22b 23c 24c

Earnings per Share (EPS)

Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

(standard error) (standard error) (standard error) (standard error) (standard error) (standard error)

CGCI −0.00234 −0.01466 0.02909⁎⁎⁎ 0.03484⁎⁎ 0.01107 0.01299

(0.00925) (0.01603) (0.01088) (0.01588) (0.00913) (0.01354)
AG 9.03536⁎⁎⁎ 8.95453⁎⁎⁎ 6.83929⁎⁎⁎ 6.76621⁎⁎⁎ 5.37251⁎⁎⁎ 5.28203⁎⁎⁎
(1.49087) (1.47389) (1.14360) (1.14423) (1.06989) (1.07982)
Age −0.00866 −0.00751 0.00319 0.00425 0.01050 0.01264
(0.02072) (0.02068) (0.04573) (0.04566) (0.02189) (0.02205)
DE −0.83324⁎⁎⁎ −0.85428⁎⁎⁎ −0.88925⁎⁎⁎ −0.90317⁎⁎⁎ −1.24344⁎⁎⁎ −1.23615⁎⁎⁎
(0.26253) (0.26041) (0.21712) (0.21728) (0.22128) (0.22005)
Fsz 1.84618⁎⁎⁎ 1.88743⁎⁎⁎ 1.06218⁎⁎ 1.13143⁎⁎ 1.52893⁎⁎⁎ 1.60294⁎⁎⁎
(0.31720) (0.32357) (0.48435) (0.48555) (0.46918) (0.47974)
DPS 1.13607⁎⁎⁎ 1.13619⁎⁎⁎ 1.08343⁎⁎⁎ 1.08285⁎⁎⁎ 1.00741⁎⁎⁎ 1.00597⁎⁎⁎
(0.07816) (0.07890) (0.07072) (0.07069) (0.09176) (0.09275)
HCF −0.14144 −1.68649 −0.95674
(1.15114) (1.10106) (0.89390)
LCF −1.56588 −0.41501 −0.43269
(1.12080) (0.99756) (0.81063)
Constant −8.35301⁎⁎ −3.66447 −12.83215⁎⁎ −15.11328⁎⁎ −9.27322⁎ −9.97712
(3.66166) (6.06169) (5.11920) (6.74521) (5.13401) (6.33886)
Obs 950 950 950 950 950 950
R2 0.53 0.53 0.52 0.52 0.61 0.60
Adj R2 0.51 0.51
F 42.89⁎⁎⁎ 42.90⁎⁎⁎
Wald χ2 454.1⁎⁎⁎ 457.9⁎⁎⁎ 1.2e + 07⁎⁎⁎ 7.8e + 06⁎⁎⁎

Coefficients for year, industry type and industry dummies are not reported. Standard errors are in parentheses.
a
Common effect (pooled OLS) regression with robust standard errors.
b
Random-effect GLS regression.
c
PCSE Prais–Winsten regression with PSAR1 autocorrelation structure.
⁎ p b 0.1.
⁎⁎ p b 0.05.
⁎⁎⁎ p b 0.01.
 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575 573

Debt to equity ratio is significant and is negatively associated with all
measures of financial performance and in all regression specifications,
consistent with previous findings. These findings are consistent with
Mir and Nishat (2004) and Chen (2001).
Firm size is positively related with ROE, ROCE and EPS, whereas the
relationship is significant for only ROE and ROCE models. On the other
hand, firm size is negatively correlated with ROA under GLS specifica-
tion. Similar positive associations are reported by Shah (2009) and
Andres (2008). Dividend per share (DPS), a proxy for dividend policy
is positively and significantly associated with all four measures of finan-
cial performance. Similar findings are also reported by Reddy et al.
(2010).
4.3.1. Bootstrapped Tobit regression model: CGCI & firm's efficiency
In addition to comparing compliance and financial performance, this
study employed two-stage DEA methodology to explore the effects of
complying with the code of corporate governance on firm's efficiency
(technical efficiency). In the first stage, we computed efficiency scores
using DEA methodology. In the second stage we regressed DEA scores
with corporate governance compliance score and control variables.
Since the dependent variables in this case are efficiency scores
bounded between 1 and 0 (a case of limited dependent variable), there-
fore the use of OLS will result in biased and inconsistent estimation re-
sults. Therefore, we employed bootstrapped censored Tobit regression
(random-effect) for testing the empirical relationship between compli-
ance and efficiency. There is debate on the conditions under which
random-effect or fixed effect Tobit regression is appropriate. Greene
(2004) noted that in case of fixed-effects, the Tobit panel data estima-
tors are considered problematic. Further, the “incidental parameters
problem” is an issue of concern, as MLE in non-linear panel data models
(e.g. Tobit & Logit) with fixed effects can be biased and inconsistent, when
the length of time-series (T) is small and fixed and the number of obser-
vations (n) in the cross-section is large (Baltagi, 2005; Greene, 2004).
For comparison purpose, in addition to bootstrapped regression, we
also have reported results from pooled Tobit regression (with robust
standard error) and random-effect Tobit regression (see Table 12).
CGCI has a positive and significant impact on TE under both
random-effects Tobit (significant at 1%) and bootstrapped Tobit (sig-
nificant at 10%). Similar to financial performance models, we further
added dummies for high and low compliant firms in the model. Both
high and low compliant dummies are negative, though insignificant.
The positive and significant CGCI coefficient and in contrast negative
coefficient of high compliant firms' dummies may indicate that the
relationship is not plain linear.
Among control variables, the only consistent outcome is the positive
relationship between DPS and technical efficiency. This can be explained
by the fact that firms with substantial foreign shareholdings are the ones
that pay higher and regular dividends and it is argued that firms with for-
eign investors are more efficient. This efficiency can be attributed to the
advance technology introduced by foreign investors (Barbosa and Louri,
2005; Dimelis and Louri, 2002).
5. Summary and conclusion
The main purpose of this study is to find out the effectiveness and ef-
ficacy of the current code of corporate governance by using finance ap-
proach to corporate governance i.e. by evaluating the relationship
between code's compliance firms' performance (both financial and op-
erational). This objective was explored in the background that the de
facto realities of the capital markets in Pakistan are quite different
than what is promulgated by current Anglo-American driven code.
The two main questions explored are (1) does compliance with the
SECP code of corporate governance result in improved performance
and (2) do firms that showed high level of compliance with the code
of corporate governance are superior performers than average and
low compliance firms?

Table 12
Compliance & technical efficiency.

25a,d 26a,d 27b,e 27b,e 28c,f 29c,f

Technical Efficiency (TE)

Variables Coefficient Coefficient Coefficient Coefficient Coefficient Coefficient

(robust standard error) (robust standard error) (standard error) (standard error) (Bootstrapped standard error) (Bootstrapped standard error)

CGCI 0.00032⁎⁎⁎ 0.00020 0.00033⁎⁎⁎ 0.00032⁎ 0.00024⁎ 0.00028

(0.00009) (0.00018) (0.00013) (0.00018) (0.00014) (0.00024)
AG 0.00436 0.00292 −0.00063 −0.00099 0.00697 0.00681
(0.01639) (0.01629) (0.01297) (0.01298) (0.01328) (0.01215)
Age −0.00037 −0.00035 0.00015 0.00015 0.00050 0.00052
(0.00030) (0.00030) (0.00057) (0.00057) (0.00049) (0.00051)
DE −0.00444 −0.00477 0.00105 0.00097 0.00277 0.00273
(0.00302) (0.00296) (0.00249) (0.00250) (0.00431) (0.00413)
Fsz 0.01517⁎⁎⁎ 0.01589⁎⁎⁎ 0.00052 0.00086 −0.00932 −0.00917
(0.00392) (0.00394) (0.00601) (0.00602) (0.00632) (0.00601)
DPS 0.00519⁎⁎⁎ 0.00521⁎⁎⁎ 0.00419⁎⁎⁎ 0.00418⁎⁎⁎ 0.00339⁎⁎⁎ 0.00339⁎⁎⁎
(0.00057) (0.00057) (0.00083) (0.00083) (0.00081) (0.00082)
HCF −0.00800 −0.00498 −0.00732
(0.01118) (0.01266) (0.01109)
LCF −0.01962 −0.00458 −0.00038
(0.01502) (0.01138) (0.01425)
Constant 0.57457⁎⁎⁎ 0.62004⁎⁎⁎ 0.67348⁎⁎⁎ 0.67727⁎⁎⁎ 0.85806⁎⁎⁎ 0.84433⁎⁎⁎
(0.04544) (0.06435) (0.06303) (0.08034) (0.06713) (0.08521)
Obs. 948 948 948 948 948 948
Pseudo R2 −0.229 −0.231
F 14.68⁎⁎⁎ 13.20⁎⁎⁎
Wald χ2 83.26⁎⁎⁎ 83.81⁎⁎⁎ 40.56⁎⁎⁎ 40.70⁎⁎⁎

*** p b 0.01, ** p b 0.05, * p b 0.1.

a
Pooled Tobit regression (with robust standard errors) with year industry and industry type dummies.
b
Random-effect Tobit regression with year, industry and industry type.
c
Bootstrapped Tobit regression.
d
All industry dummies are significant except Communication and Oil & Gas.
e
All industry dummies are significant except Chemical, Communication & Networks, Oil & Gas, and Paper & Board.
f
To avoid computational complexities, industry and year dummies are dropped in bootstrapped standard error calculations.
574 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575
The extent of compliance is measured through custom built compli-
ance index. This study used a multi-theoretical approach and multidi-
mensional firm performance approach i.e. financial performance and
technical efficiency. The investigation of extent of compliance revels
that the extent of compliance has increased since the issuance of the
code but it appears that spirit in complying with the code of corporate
governance is missing. We find significant positive impact of compli-
ance on ROA, ROE and ROCE. The very low magnitude of CGCI coefficient
may indicate that the relationship is not fully explained by linear regres-
sion models. Further we find that contrary to our expectations, the
dummy for high compliant firms is negative and significant at 10% in re-
gression 4. This is an interesting finding, and we argue that due to the
requirement of the mandatory compliance with the code, beyond a cer-
tain point the increase in compliance does not translate into improved
financial performance.
In the similar manner, we find weak evidence that CGCI positively
impacts technical efficiency. DPS is found to be significantly positively re-
lated technical efficiency. This significant positive relationship can be
explained by the fact that firms with substantial foreign shareholdings
are the ones that pay higher and regular dividends and it is argued that
firms with foreign investors are more efficient. This efficiency can be at-
tributed to the advance technology introduced by foreign investors
(Barbosa and Louri, 2005; Dimelis and Louri, 2002).
This study offers several contributions towards literature and policy
horizons. First, due to mandatory compliance requirement, it is expected
that firms are not complying with the code of corporate governance in
spirit. Therefore a high compliance score does not indicate that firms
are actually following the requirements of the code in high spirit. Thus
a positive relationship between mandatory compliance and performance
should be evaluated with caution. Second beyond a certain point, the
impact of compliance with the code of corporate governance on firm's
performance is negative. This suggests that the relationship between
mandatory compliance and performance may not be linear. Third it is
an alarming indication for regulators and policy makers that high com-
pliant firms appear less profitable. This indicates that the requirements
of the code of corporate governance are in conflict with unique optimal
governance structure of each firm which in return increases the cost of
compliance to such a level where it affects firm's profitability in a nega-
tive manner. This is a call for national policy makers to review their
approach towards corporate governance and ask questions about the ef-
ficacy of this “one size fit all” approach.
References
Aguilera, R.V., Cuervo-Cazurra, A., 2004. Codes of good governance worldwide: what is
the trigger? Organ. Stud. 25 (3), 415–443.
Akkermans, D., Van Ees, H., Hermes, N., Hooghiemstra, R., Van Der Laan, G., Postma, T.,
Van Witteloostuijn, A., 2007. Corporate governance in the Netherlands: an overview
of the application of the Tabaksblat Code in 2004. Corp. Gov. 15 (6), 1106–1118.
http://dx.doi.org/10.1111/j.1467-8683.2007.00634.x.
Alchian, A., Demsetz, H., 1972. Production, information, costs, and economic organization.
Am. Econ. Rev. 62, 777–795.
Alves, Carlos F., Mendes, V., 2004. Corporate governance policy and company perfor-
mance: the Portuguese case. Corp. Gov. 12 (3), 290–301.
Andres, C., 2008. Large shareholders and firm performance—an empirical examination of
founding-family ownership. J. Corp. Finan. 14 (4), 431–445. http://dx.doi.org/10.1016/
j.jcorpfin.2008.05.003.
Arcot, Sridhar R., Bruno, V.G., 2006. One size does not fit all, after all: evidence from cor-
porate governance. Working Paper. (Available at http://ssrn.com/abstract=887947).
Ariff, Mohamed, Ratnatunga, J., 2008. Do accounting and finance tools serve governance?
Int. J. Econ. Manage. 2 (1), 1–27.
Baltagi, Badi H., 2005. Econometrics Analysis of Panel Data, 3rd ed. John Wiley & Sons,
Ltd., West Sussex.
Barbosa, Natália, Louri, Helen, 2005. Corporate performance: does ownership matter? A
comparison of foreign- and domestic-owned firms in Greece and Portugal. Rev. Ind.
Organ. 27 (1), 73–102. http://dx.doi.org/10.1007/s11151-005-4920-y.
Bauer, Rob, Guenster, Nadja, Otten, Rog, 2004. Empirical evidence on corporate gover-
nance in Europe: the effect on stock returns, firm value and performance. J. Asset
Manage. 5 (2), 91–104. http://dx.doi.org/10.1057/palgrave.jam.2240131.
Bauwhede, Heidi Vander, 2009. On the relation between corporate governance compliance
and operating performance. Account. Bus. Res. 39 (5), 497–513. http://dx.doi.org/
10.1080/00014788.2009.9663380.
Berglöf, E., Pajuste, A., 2005. What do firms disclose and why? Enforcing corporate gover-
nance and transparency in central and Eastern Europe. Oxford Rev. Econ. Policy 21
(2), 178–197. http://dx.doi.org/10.1093/oxrep/gri011.
Bozec, Richard, Dia, Mohamed, 2007. Board structure and firm technical efficiency: evidence
from Canadian state-owned enterprises. Eur. J. Oper. Res. 177 (3), 1734–1750. http://
dx.doi.org/10.1016/j.ejor.2005.10.001.
Bozec, Richard, Dia, Mohamed, Bozec, Yves, 2010. Governance–performance relationship:
a re-examination using technical efficiency measures. Br. J. Manag. 21 (3), 684–700.
http://dx.doi.org/10.1111/j.1467-8551.2008.00624.x.
Charnes, A., Cooper, W.W., Rhodes, E., 1978. Measuring the efficiency of decision
making units. Eur. J. Oper. Res. 2 (6), 429–444. http://dx.doi.org/10.1016/
0377-2217(78)90138-8.
Chen, J., 2001. Ownership structure as corporate governance mechanism: evidence from
Chinese listed companies. Econ. Chang. Restruct. 34 (1–2), 53–72.
Chen, K.Y., Elder, R.J., Hsieh, Y.M., 2011. Corporate governance, growth opportunities, and
earnings restatements: effects of a corporate governance code. Asia Pac. J. Account.
Econ. 18 (2), 169–200.
Chizema, A., 2008. Institutions and voluntary compliance: the disclosure of individual ex-
ecutive pay in Germany. Corp. Gov. 16 (4), 359–374. http://dx.doi.org/10.1111/
j.1467-8683.2008.00689.x.
Claessens, S., Yurtoglu, B.B., 2012. Corporate governance in emerging markets: a survey.
Emerg. Mark. Rev. http://dx.doi.org/10.1016/j.ememar.2012.03.002.
Claessens, S., Djankov, S., Xu, L.C., 2000. Corporate performance in the East Asian financial
crisis. World Bank Res. Obs. 15, 23–46.
Clarke, Thomas, 2004a. International Corporate Governance: A Comparative Approach.
Routledge, London.
Clarke, Thomas (Ed.), 2004b. Theories of Corporate Governance: The Philosophical Foun-
dations of Corporate Governance. Routledge Taylor & Francis Group, New York.
Cooper, W.W., Li, S., Seiford, L.M., Thrall, R.M., Zhu, J., 2001. Sensitivity and stability anal-
ysis in DEA: some recent developments. J. Product. Anal. 15, 217–246.
Copeland, T., Koller, T., Murrin, J., 2000. Valuation: Measuring and Managing the Value of
Companies, 3rd ed. McKinsey & Company, Inc., New York.
Dahya, J., McConnell, John J., 2007. Board composition, corporate performance, and the
Cadbury committee recommendation. J. Financ. Quant. Anal. 42 (3), 535–564.
Dahya, J., McConnell, J., Travolas, N., 2002. The Cadbury committee, corporate perfor-
mance, top management turnover. J. Finance 57, 461–483.
Davies, M., Schlitzer, Bernadette, 2008. The impracticality of an international “one size fits
all” corporate governance code of best practice. Manag. Audit. J. 23 (6), 532–544.
http://dx.doi.org/10.1108/02686900810882093.
de Jong, A., DeJong, D.V., Mertens, G., Wasley, C.E., 2005. The role of self-regulation in cor-
porate governance: evidence and implications from the Netherlands. J. Corp. Finan.
11 (3), 473–503. http://dx.doi.org/10.1016/j.jcorpfin.2004.01.002.
Dedman, E., 2003. The Cadbury committee recommendations on corporate governance—
a review of compliance and performance impacts. Int. J. Manag. Rev. 4 (4), 335–352.
Destefanis, S., Sena, V., 2007. Patterns of corporate governance and technical efficiency in
Italian manufacturing. Manag. Decis. Econ. 28 (1), 27–40. http://dx.doi.org/10.1002/
mde.1310.
Dimelis, S., Louri, H., 2002. Foreign ownership and production efficiency: a quantile re-
gression approach. Oxf. Econ. Pap. 54 (3), 449–469.
Fernández-Rodríguez, E., Gómez-Ansón, S., Cuervo-García, Á., 2004. The stock market re-
action to the introduction of best practices codes by Spanish firms. Corp. Gov. 12 (1),
29–46.
Goncharov, I., Werner, J.R., Zimmermann, J., 2006. Does compliance with the German cor-
porate governance code have an impact on stock valuation? An empirical analysis.
Corp. Gov. 14 (5), 432–445. http://dx.doi.org/10.1111/j.1467-8683.2006.00516.x.
Greene, W., 2004. Fixed effects and the incidental parameters problem in the Tobit model.
Econ. Rev. 23 (2), 125–148.
Hameed, Abid, Ashraf, Hammad, 2006. Stock market volatility and weak-form effi-
ciency: evidence from an emerging market. Paper Presented at the 22nd Annual
General Meeting and Conference of Pakistan Society of Development Economists,
Lahore.
Haniffa, R., Hudaib, M., 2006. Corporate governance structure and performance of
Malaysian listed companies. J. Bus. Finance Acc. 33 (7), 1034–1062.
Hooghiemstra, Reggy, van Ees, Hans, 2011. Uniformity as response to soft law: evidence from
compliance and non-compliance with the Dutch corporate governance code. Regul. Gov-
ernance 5 (4), 480–498. http://dx.doi.org/10.1111/j.1748-5991.2011.01118.x.
Iqbal, Javed, 2012. Stock market in Pakistan: an overview. J. Emerging Market Finance 11
(1), 61–91.
Iturriaga, Felix J. Lopez (Ed.), 2009. Codes of Good Governance Around the World. Nova
Science Publishers, Inc., New York.
Jensen, M.C., 1986. Agency costs of free cash flow, corporate finance, and takeovers. Am.
Econ. Rev. 76, 323–329.
Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: managerial behavior, agency costs
and ownership structure. J. Financ. Econ. 3 (4), 305–360.
Kapopoulos, P., Lazaretou, S., 2007. Corporate ownership structure and firm performance:
evidence from Greek firms. Corp. Gov. Int. Rev. 15 (2), 144–158.
Khawja, A.I., Mian, A., 2005. Unchecked intermediaries: price manipulation in an emerg-
ing stock market. J. Finan. Econ. 78 (1), 203–241.
Klapper, L.F., Love, I., 2004. Corporate governance, investor protection, and performance in
emerging markets. J. Corp. Finan. 10 (5), 703–728. http://dx.doi.org/10.1016/s0929-
1199(03)00046-4.
Kouwenberg, Roy, Phunnarungsi, Visit, 2012. Corporate governance, violations and mar-
ket reactions. Pacific-Basin Finance J. 21 (1), 881–898. http://dx.doi.org/10.1016/
j.pacfin.2012.06.006.
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 2002. Investor protection and
corporate valuation. J. Finance 57, 1147–1170.
 Y.B. Tariq, Z. Abbas / Economic Modelling 35 (2013) 565–575
Larcker, D.F., Richardson, S.A., Tuna, I., 2005. Ratings add fire to the governance debate. Fi-
nancial Times Mastering Corporate Governance Part Two.
Lavelle, L., 2002. Enron: how governance rules failed. Bus. Week 21 (3766), 28–29.
Li, Haidan, Pincus, Morton, Rego, Sonja O., 2004. Market reaction to events surrounding
the Sarbanes–Oxley Act of 2002. SSRN Working Paper. (Available at http://ssrn.
com/abstract=475163).
Lin, Chen, Ma, Yue, Su, Dongwei, 2009. Corporate governance and firm efficiency: evi-
dence from China's publicly listed firms. Manag. Decis. Econ. 30 (3), 193–209.
http://dx.doi.org/10.1002/mde.1447.
Litvak, K., 2007. The effect of the Sarbanes–Oxley act on non-US companies cross-listed in the
US. J. Corp. Finan. 13 (2–3), 195–228. http://dx.doi.org/10.1016/j.jcorpfin.2007.03.002.
Mallin, C.A., 2007. Corporate Governance, 2nd ed. Oxford University Press, Oxford.
Mei, J., Wu, G., Zhou, C., 2004. Behavior based manipulation: theory and prosecu-
tion evidence. NYU-STERN Working Paper. (Available at http://dx.doi.org/10.
2139/ssrn.457880).
Mir, Shahid Raza, Nishat, Mohammed, 2004. Corporate governance structures and firm
performance in Pakistan—an empirical study. Paper Presented at the 2nd Annual
Conference on Corporate Governance in Pakistan, June 3–4 2005. LUMS, Lahore.
Mirza, Nawazish, Afzal, A., 2009. Speculative bubbles in karachi stock exchange. CREB
Working Paper.Centre for Research in Economics & Business, Lahore School of Eco-
nomics, Lahore (Available at http://dx.doi.org/10.2139/ssrn.1340414).
Nanka-Bruce, Douglas, 2011. Corporate governance mechanisms and firm efficiency. Int.
J. Bus. Econ. 6 (5), 28–40. http://dx.doi.org/10.5539/ijbm.v6n5p28.
OECD, 2008. Introductory Handbook for Undertaking Regulatory Impact Analysis (RIA):
Organisation for Economic Co-operation and Development. http://www.oecd.org/
gov/regulatory-policy/44789472.pdf (Retrieved from).
Pi, L., Timme, S.G., 1993. Corporate control and bank efficiency. J. Banking Finance 17
(2–3), 515–530.
Price, R., Román, F.J., Rountree, B., 2011. The impact of governance reform on per-
formance and transparency. J. Finan. Econ. 99 (1), 76–96. http://dx.doi.org/
10.1016/j.jfineco.2010.08.005.
Reddy, Krishna, Locke, Stuart, Scrimgeour, Frank, 2010. The efficacy of principle-based corpo-
rate governance practices and firm financial performance: an empirical investigation. Int.
J.Manag. Financ. 6 (3), 190–219. http://dx.doi.org/10.1108/17439131011056224.
Rhee, M., Lee, J.-H., 2008. The signals outside directors send to foreign investors: evidence
from Korea. Corp. Gov.Int. Perspect. 16 (1), 41–51.
SBP, 2007. Balance Sheet Analysis of Joint Stock Companies Listed on the Karachi Stock Ex-
change (2001–2006). Statistics and DWH Department, State Bank of Pakistan, Karachi.
SBP, 2011. Financial Statement Analysis of Companies (Non-Financial) Listed at Karachi Stock
Exchange (2005–2010). Statistics and DWH Department, State Bank of Pakistan, Karachi.
SECP, 2002. Code of Corporate Governance. Securities and Exchange Commission of
Pakistan, Islamabad.
Seidl, David, Sanderson, Paul, Roberts, John, 2012. Applying the ‘comply-or-explain’ prin-
ciple: discursive legitimacy tactics with regard to codes of corporate governance.
J. Manag. Gov. http://dx.doi.org/10.1007/s10997-011-9209-y.
Shah, S.Z.A., 2009. Corporate governance and financial performance A comparative study
of developing and developed Markets (Unpublished doctoral dissertation).
Mohammad Ali Jinnah University, Islamabad, Pakistan.
Sheu, Her-Jiun, Yang, Chi-Yih, 2005. Insider ownership and firm performance in Taiwan's
electronics industry: a technical efficiency perspective. Manag. Decis. Econ. 26 (5),
307–318. http://dx.doi.org/10.1002/mde.1228.
575
Shleifer, A., Vishny, R., 1986. Large shareholders and corporate control. J. Polit. Econ. 94 (3),
737–783.
Shleifer, A., Vishny, R.W., 1997. A survey of corporate governance. J. Finance 52 (2),
737–783.
Siddiqi, Hammad, 2007. Stock price manipulation: the role of intermediaries. CMER Work-
ing Paper No. 07-58. Lahore University of Management Sciences, Lahore, Pakistan.
Simar, L., Wilson, P.W., 2007. Estimation and inference in two-stage, semi-parametric
models of production processes. J. Econ. 136 (1), 31–64. http://dx.doi.org/10.1016/
j.jeconom.2005.07.009.
Soederberg, S., 2003. The promotion of Anglo-American corporate in the governance
South: who benefits from the new international standard? Third World Q. 24 (1),
7–27.
Su, D., He, X., 2011. Ownership structure, corporate governance and productive efficiency
in China. J. Product. Anal. 1–16. http://dx.doi.org/10.1007/s11123-011-0257-8.
Talaulicar, T., Werder, A.V., 2008. Patterns of compliance with the German corpo-
rate governance code. Corp. Gov. 16 (4), 255–273. http://dx.doi.org/10.1111/
j.1467-8683.2008.00696.x.
Tariq, Yasir Bin, Butt, Safdar A., 2008. Impact of corporate governance practices on
financial performance: empirical evidence from Pakistan. Paper Presented at
the 8th Annual Hawaii International Conference on Business, Honolulu,
Hawaii.
Teh, Chor Tik, 2009. Compliance and Impact of Corporate Governance Best Practice Code
on the Financial Performance of New Zealand Listed Companies. (Unpublished doc-
toral dissertation) Massey University, Auckland, New Zealand.
Thillainathan, R., 1999. Role of Shareholder, Directors and Managers in Corporate Gover-
nance. Paper Presented at The World Bank Conference, Washington DC.
Tuschke, A., Sanders, G., 2003. Antecedents and consequences of corporate governance
reform: the case of Germany. Strateg. Manag. J. 24 (7), 631–649. http://dx.doi.org/
10.1002/smj.324.
Wang, Jennifer L., Jeng, Vivian, Peng, Jin Lung, 2007. The impact of corporate gov-
ernance structure on the efficiency performance of insurance companies in
Taiwan. Geneva Pap. Risk Insur. Issues Prac. 32 (2), 264–282. http://
dx.doi.org/10.1057/palgrave.gpp.2510125.
Wang, Wei-Kang, Lu, W.M., Tsai, C.J., 2011. The relationship between airline performance
and corporate governance amongst US Listed companies. J. Air Transp. Manag. 17 (2),
147–151. http://dx.doi.org/10.1016/j.jairtraman.2010.06.005.
Wang, Wei-Kang, Lu, Wen-Min, Lin, Yi-Ling, 2012. Does corporate governance play an im-
portant role in BHC performance? Evidence from the US. Econ. Model. 29 (3),
751–760. http://dx.doi.org/10.1016/j.econmod.2012.01.021.
Werder, Axel V., Talaulicar, Till, Kolat, Georg L., 2005. Compliance with the German corpo-
rate governance code: an empirical analysis of the compliance statements by German
listed companies. Corp. Gov. Int. Rev. 13 (2), 178–187. http://dx.doi.org/10.1111/
j.1467-8683.2005.00416.x.
Zattoni, A., Cuomo, F., 2008. Why adopt codes of good governance? A comparison of insti-
tutional and efficiency perspectives. Corp. Gov. 16 (1), 1–15. http://dx.doi.org/
10.1111/j.1467-8683.2008.00661.x.
Zhang, I.X., 2007. Economic consequences of the Sarbanes–Oxley Act of 2002. J. Account.
Econ. 44 (1–2), 74–115. http://dx.doi.org/10.1016/j.jacceco.2007.02.002.
Zheka, V., 2005. Corporate governance, ownership structure and corporate efficiency: the
case of Ukraine. Manag. Decis. Econ. 26 (7), 451–460. http://dx.doi.org/10.1002/
mde.1258.