A multi-methodology on building a corporate governance index from the perspectives of academics and practitioners for firms in Greece

Author:Michail Nerantzidis
Publication Date:04 Apr 2016
A multi-methodology on building a
corporate governance index from the
perspectives of academics and
practitioners for firms in Greece
Michail Nerantzidis
Michail Nerantzidis is Dr
in Corporate Governance
based at the Department
of Public Administration,
Panteion University of
Political and Social
Sciences, Athens,
Purpose The purpose of this paper is to look inside the “black box” in corporate governance (CG)
measurement, and shed some light on how to construct a transparent, reliable and valid index,
considering equally both the academics and practitioners’ perspectives.
Design/methodology/approach A synthesized literature review is presented and a CG index is
developed combining the strengths of three different methodologies: the Delphi method, the classical
test theory (CTT) and the analytic hierarchy process (AHP). This approach helps authors to break the
process into separate steps and to select the appropriate techniques to support their decision
regarding the norms, the criteria, the variables and the weights that someone should use to construct a
CG index.
Findings The authors’ analysis indicates that a well-designed CG index requires a combination of
research methods to identify the best options to solve several methodological issues in index
construction. For the application of this multi-methodology in Greece, the authors used two equal and
independent samples to explore the different perspectives regarding the importance of the index
criteria and sub-criteria. This process provides evidence that the opinion of academics and
practitioners in Greece tend to converge. Moreover, it is found that this multi-methodology produces the
highest variation in CG scores and ranking orders, as opposed to a traditional approach, in measuring
CG disclosure, an important issue with econometric implications.
Research limitations/implications The limitations of this study are associated with the methods
Practical implications This paper provides practical implications for investors and commercial
vendors. For the former, it highlights the need to be more cautious and/or suspicious when they use CG
ratings, meaning that they should comprehend the base of the ratings models, and for the latter, it
demonstrates the importance of enhancing the transparency in CG indices construction.
Originality/value The value of the paper lies in improved understanding of the methodological issues
in constructing CG indices. This is quite interesting because this approach could serve as a roadmap
for other researchers.
Keywords Experts, Analytic hierarchy process, Delphi method, Classical test theory,
Corporate governance index
Paper type Research paper
1. Introduction
In recent years, the concept of good corporate governance (CG) is considered as the
hallmark of a well-run company (Wilkes, 2004). Indeed, institutional investors, shareholders
and stakeholders in general are pushing firms to adopt best practices of CG (Koehn and
Ueng, 2005;Wei-An, 2008)[1], an approach that has been promoted mainly because of the
rapid increase in theoretical (Jensen and Meckling, 1976;Shleifer and Vishny, 1997;Denis
Received 11 August 2015
Revised 27 November 2015
Accepted 15 December 2015
The author would like to thank
the Heraclitus II: Investing in
knowledge society through the
European Social Fund, as this
paper draws on data collected
from the authors PhD thesis.
The author is also very grateful
to Dr John Filos for his
comments and supervision, as
well as to all academics and
practitioners who participated
in his research. Furthermore,
the author would like to thank
Dr Ioannis Tsalavoutas for his
recommendations and the
participants in the 10th
European Academic
Conference on “Internal Audit
and Corporate Governance
and in the International
Conference – International
Business ICIB, for their
comments. Last but not least,
the author would like to thank
the anonymous reviewers and
the editor from the Corporate
Governance: The International
Journal of Business in Society
for their invaluable
suggestions for improvement.
DOI 10.1108/CG-08-2015-0107 VOL. 16 NO. 2 2016, pp. 295-329, © Emerald Group Publishing Limited, ISSN 1472-0701 CORPORATE GOVERNANCE PAGE 295
and McConnell, 2003) and empirical research in CG (Gompers et al., 2003;Larcker et al.,
2005;Bebchuk et al., 2009).
In response to these expanding results, stakeholders in general have started to demand
governance ratings (Sherman, 2004;Wei-An, 2008, p. 12) to avoid undesirable outcomes
(Daines et al., 2010), a new trend that has introduced the issue of measuring the CG quality
(Aguilera and Desender, 2012) both by academics (La Porta et al., 1998;Tsipouri and
Xanthakis, 2004;Cheung et al., 2007) and commercial agencies (GMI Ratings, 2013;Risk
Metrics/Institutional Shareholder Services, 2014;Credit Lyonnais Securities Asia, 2010;
Standard and Poor’s, 2007).
In particular, there is an extensive empirical literature examining the relationship between
CG and a company’s performance either as dependent or independent variables,
respectively (Bauer et al., 2004;Black et al., 2006;Bhagat and Bolton, 2007;Bebchuk et al.,
2009;Cheung et al., 2010;Cheung et al., 2011), indicating that there is no clear evidence
regarding the governance ratings and their performance. To address these mixed results,
an increasing number of studies criticized the way of constructing CG indices and
highlighted the practical problems that needed to be understood and solved (Bhagat et al.,
2008;Bozec and Bozec, 2011;Aguilera and Desender, 2012).
Undoubtedly, the majority of the studies which construct CG indices (for example, Drobetz
et al., 2004;Black et al., 2006;Renders et al., 2010;Lazarides and Drimpetas, 2011)donot
take the criteria of validity and reliability into consideration, an approach that casts doubt
on previous efforts regarding measuring CG.
For this reason, this study shed some light on how to construct a transparent, reliable and
valid index, considering equally the academics and practitioners’ perspectives. In
particular, this is the first study – as far as we know – that provides advance knowledge
regarding the design and implementation of CG index. This is quite interesting because it
contributes to the discussion on the forward research agenda of CG index construction.
The remaining of this paper is organized as follows. The next section reviews the literature.
Section 3 describes the theoretical background in CG index development, while the
application of the proposed model is presented in Section 4. Section 5 discusses the
contribution of this methodology, as opposed to another approach in measuring the CG
disclosure. Finally, the conclusions and research perspectives are exposed in Section 6.
2. Literature review
2.1 Literature review on measuring corporate governance
In the aftermath of Enron’s collapse, varieties of academic and commercial indices are
being constructed to measure the CG (Bozec and Bozec, 2011). This widespread CG
evaluation reflects the stakeholders’ desire to improve transparency, accountability,
responsibility and, of course, economic performance (Aguilera and Cuervo-Cazurra, 2009;
Ntim et al., 2012). Therefore, CG rating systems could be used as a powerful indicator of
the extent to which a company is currently adding, or has the potential to add in the future
shareholder value (Mallin, 2001, p. 257). Thus, the premise of CG indices could be
summarized as follows:
[. . .] Companies that focus on CG and transparency will, over time, generate superior returns
and economic performance and lower their cost of capital (Sherman, 2004,p.6).
This means that the basic underlying assumption of CG indices is that they adequately
capture the quality of CG. However, what is the meaning of capturing the quality of CG?
Does it mean that the construction captures the firm’s level governance quality or the
disclosure quality (Bhagat et al., 2008;Aguilera and Desender, 2012)? In literature review,
there is a variety of studies in emerging markets (Klapper and Love, 2004), as well as in
Europe (Drobetz et al., 2004;Tsipouri and Xanthakis, 2004;Florou and Galarniotis, 2007),
the USA (Brown and Caylor, 2006;Bhagat and Bolton, 2007;Bebchuk et al., 2009) and Asia
(Black et al., 2006;Cheung et al., 2007), which use different CG indices to capture the
“latent variable” of CG quality. This illustrates that these studies benchmark the voluntary
CG disclosure, with the expectation to capture the quality of the firms’ governance.
However, despite the great effort of the empirical studies to measure the CG disclosure,
there is an extensive literature review that criticizes the methodological challenges of this
procedure (Bozec and Bozec, 2011;Aguilera and Desender, 2012). For instance, there is
a theoretical gap among the variables (Florou and Galarniotis, 2007), the criteria
(Nerantzidis, 2015), the weights (Graafland et al., 2004;Bozec and Bozec, 2011) and the
norms that someone should use to construct a reliable and valid index. These theoretical
gaps support the idea of Bhagat et al. (2008, p. 1803), according to which, the CG indices
are highly imperfect instruments.
2.2 Main criteria in index construction
According to literature review, there are two main criteria that researchers should take into
account when constructing indices: reliability and validity (Krippendorff, 2013). However, a
prerequisite for the index construction is a prior investigation of the country’s environment,
which, in the case of CG, is the main characteristic of each country’s CG framework
(Figure 1).
Based on the first criterion, an index is considered as reliable if the results can be
replicated by another researcher (Marston and Shrives, 1991, p. 197). Thus, in case the
scores are extracted from CG statements (CGS) and remain constant over time, then
there is no obstacle to repetition. Moreover, the developments in CG reporting give the
opportunity to researchers to decide more objectively whether a variable is disclosed or
not. This means that the problem of “applicability” (that probably companies will not
provide any information regarding some provisions) is not such an important issue for our
analysis, as it is reflected in EU Directives (2006/46/EC and 2007/63/EC). However, it is
worth mentioning that the theory of measurement indicates that researchers have to pay
attention to statistical items analysis when constructing indices (Tsipouri and Xanthakis,
2004;Florou and Galarniotis, 2007). This means that researchers (in compliance studies
and CG indices) prefer not to rely on statistical methods for their analysis, namely, classical
test theory (CTT) (Allen and Yen, 2002;Crocker and Algina, 2006) and item response
theory (IRT) (Hambleton et al., 1991;Embretson and Reise, 2000) but to calculate ratings
using a simplistic procedure: 1 point if the provision is disclosed and 0 otherwise. In fact,
this practice would be excellent in studies that present compliance scores; however, it is
not “ideal” in constructing indices because it does not take into account the effects and
interrelationships between variables (Cox, 1972;Edelen and Reeve, 2007).
Referring to the second criterion, an index can be considered as valid if it measures what
its user claims it measures (Krippendorff, 2013, p. 329). But does this really happen in CG
index construction? As we can see in empirical studies (Black, 2001;Klein et al., 2005;
Black et al., 2006;Bebchuk et al., 2009;Cheung et al., 2010), there is no consensus on one
particular CG index or on a measuring approach of CG (Marston and Shrives, 1991,
p. 198). However, this is one facet of the validity problem. In particular, as Krippendorff
(2013, p. 334) mentioned, researchers have to consider a plethora of validity dimensions
such as construct validity, content validity and criterion-related validity. Nevertheless,
researchers in compliance studies and CG indices have the tendency to “overlook” the
criterion of validity either as a whole (Klapper and Love, 2004;Tsipouri and Xanthakis,
2004;Cheung et al., 2007;Lazarides and Drimpetas, 2011) or partially (Tsalavoutas et al.,
2010;Nerantzidis, 2015).
2.3 Prior CG indices
Several empirical studies have constructed CG indexes. As it can be seen in Table I, these
are classified in two categories. On the first panel (Panel A), the commercial CG indices
(GMI Ratings, 2013;Institutional Shareholder Services, 2014;Credit Lyonnais Securities

To continue reading

Request your trial