Corporate‐Governance Ratings and Company Performance: A Cross‐European Study

• Author: Annelies Renders,Ann Gaeremynck,Piet Sercu
• DOI: http://doi.org/10.1111/j.1467-8683.2010.00791.x
• Published date: 01 March 2010
• Date: 01 March 2010

Corporate-Governance Ratings and Company

Performance: A Cross-European Studycorg_79187..106

Annelies Renders,* Ann Gaeremynck, and Piet Sercu

ABSTRACT

Manuscript Type: Empirical

Research Question/Issue: Prior studies have failed to unequivocally establish a positive relationship between corporate-

governance ratings and company performance, although theoretically, we would expect to f‌ind one. In this paper, we try to

establish whether a positive relationship exists through modeling the relationship more carefully.

Research Findings/Insights: After controlling for selection bias and endogeneity simultaneously, we f‌ind a signif‌icant

positive relationship between corporate-governance ratings and performance. However, the strength of this relationship

seems to depend on the quality of the institutionalenvironment. Finally, we f‌ind that improvementsin corporate-governance

ratings over time result in decreasing marginal benef‌its in terms of performance.

Theoretical/Academic Implications: Our paper contributes to the literature by showing that improved corporate-

governance ratingslead to better performance, but that econometric problems might obscure this relationship. We also show

that for a sample of developed countries the institutional environment affects the relationship between governance ratings

and performance. Finally, this paper contributes to the literature on the impact, regarding compliance and effectiveness, of

codes of good governance.

Practitioner/Policy Implications: Our results are relevant for both companies and policy makers. They indicate

that companies can improve performance by adhering to good corporate-governance practices. For policy makers,

the f‌indings suggest that soft laws and the invisible hand of the market lead to companies improving their corporate

governance.

Keywords: Corporate Governance, Performance, Europe, Institutions, Econometrics

INTRODUCTION

Based on agency theory (Jensen & Meckling, 1976), a

positive relationship between corporate-governance

ratings and company performance should exist. To the

extent that higher corporate-governance ratings proxy for

better actual corporate-governance practices,1higher

corporate-governance ratings should translate into

improved operating performance and a higher market

value. Better monitoring forces insiders to invest in projects

with a positive net present value and to reduce perks and

waste, so that more of the benef‌its f‌low back to outside

investors (Shleifer & Vishny, 1997). Even though prior

empirical literature is extensive, many studies have thus far

failed to establish f‌irm evidence that corporate-governance

ratings, devised either by rating agencies or by researchers,

positively affect companyperformance or value.2There are a

number of reasons that may explain why this is the case. The

relationship may be obscured by econometric problems,

such as endogeneity, selection bias, or lack of statistical

power. Secondly, there may simply be no relationship either

because corporate-governance ratings do not measure what

they claim to measure or because a company chooses gov-

ernance practices based on its characteristics, thereby maxi-

mizing shareholder value.3This paper aims to avoid the

above mentioned econometric problems by: (1) controlling

for both sample-selection bias and endogeneity; (2) adding

statistical power by introducingmore variation in the regres-

sor (the corporate-governance ratings), across institutional

settings, companies, and time; and (3) ref‌ining the modeling

of the relationship. In doing so, we add to the corporate-

governance literature by showing how a modeling tech-

nique can be used to improve empirical results. If we still do

not f‌ind a link between corporate-governance ratings and

*Address for correspondence: KU Leuven, Facultyof Business and Economics, Naam-

sestraat 69, B-3000 Leuven, Belgium. Tel:+32-16-326-768; Fax. +32-16-326-683; E-mail:

annelies.renders@econ.kuleuven.be

87

Corporate Governance: An International Review, 2010, 18(2): 87–106

© 2010 Blackwell Publishing Ltd

doi:10.1111/j.1467-8683.2010.00791.x

company performance, then corporate-governance practices

are either value-maximizing or corporate-governance

ratings have no relevance.

We control for sample-selection and endogeneity bias in

the design of the model. The sample-selection issue arises

because often only the largest listed companies are rated,

which means that most studies analyze a sample that is

truncated in size. Size is related to both performance and

corporate governance,so without addressing this size-based

selection effect the link between corporate-governance

ratings and performance is diff‌icult to infer correctly (Vella,

1998). Theendogeneity of the relationship between corporate

governance and performance arises because performance,

and especially poor performance, maycause changes in gov-

ernance, which results in a two-way causality problem.

Bhagat and Bolton (2007) show that controlling for endoge-

neity is relevant– they f‌ind a signif‌icant positive relationship

between corporate governance and operating performance

only when they control for endogeneity. Nobody, to our

knowledge, controls for both biases simultaneously.4We are

convinced that controlling for both is necessary– some of our

conclusions depend on whether or not one takes into account

not just endogeneity, but also sample-selection bias.5

To obtain more variation in the cross-section of corporate-

governance ratings, we study a panel of European data,

instead of the US samples that most prior studies have ana-

lyzed. One reason why a European sample may offer more

statistical power is that legal and regulatory requirements

differ across countries, which in itself generates cross-

sectional variation in corporate-governance ratings. Further-

more, corporate-governance recommendations issued at

country level are largely voluntary, allowing companies to

signal their compliance with governance standards, which

again creates variation in the ratings.

By opting for a panel instead of a cross-sectionalapproach,

we add power to the design. Corporate-governance prac-

tices, even though they remain voluntary in Europe, have

somewhat improved during our sample period, namely the

period 1999–2003. These were the years when corporategov-

ernance came to the fore – big accounting scandals in a

number of countries stimulated more shareholder scrutiny

and led to the introduction of corporate-governance codes in

Europe, which were widely adopted by f‌irms even when not

mandatory (Conyon & Mallin, 1997; Weir & Laing, 2000).

This is also evident in the increase in governance ratings

over time (see Table 4).

Lastly, in terms of modeling, some progress can be made

by not imposing a uniform relation across f‌irms and over

time. We allow the relationship to vary with the extent of

shareholder-protection laws (Djankov, La Porta, Lopez-de-

Silanes, & Shleifer, 2008) and in the longitudinal dimension

we keep track of changes over time.

Our results indicate that, after controlling for sample-

selection bias and endogeneity, corporate-governance

ratings have a highly signif‌icant and positive impact on per-

formance, whereas they have only an insignif‌icant or even a

negative impactwhen we do not control for these economet-

ric problems. So controlling for sample-selection bias and

endogeneity simultaneously is relevant. With regard to the

degree of the bias, we f‌ind that, at least in our sample,

selection bias and endogeneity have a comparable inf‌luence

on the bias in the regression coeff‌icient for corporate gover-

nance. After controlling only for endogeneity, for instance,

we no longer f‌ind a relationship between corporate-

governance ratings and operating performance. Regarding

market measures of performance, in contrast, we do f‌ind a

positive relationship after controlling for either selection

bias or endogeneity. This relationship still becomes much

stronger after simultaneously controlling for both problems.

Next, we f‌ind that companies in countries with strong

shareholder-protection laws or extensive corporate-

governance recommendations have higher corporate-

governance ratings and the effect on performance is smaller

in comparison to countries with weak shareholder-

protection regulations. Up to now, the impact of the institu-

tional environment has only been documented with regard

to emerging countries (Durnev & Kim, 2005; Klapper &

Love, 2004).

Finally, our results show that the relationship between

corporate-governance ratings and performance has become

weaker over time. This may suggest that the governance of

companies does not have to be regulated by law, but can be

left to the “invisible hand” of the market (Chhaochharia &

Laeven, 2008; De Jong, De Jong, Mertens, & Wasley, 2005;

MacNeil & Li, 2006).

This paper is organized as follows. The following section

reviews the literature on the relation between corporategov-

ernance and performance. The research design is presented

in the third section. In the fourth section we describe the

sample and data. Sections f‌ive and six present the primary

f‌indings and the results from robustness tests. The f‌inal

section summarizes the f‌indings.

RELATED LITERATURE

In the literature on agency conf‌licts it is argued that manag-

ers and corporate insiders have different objectives than

outside investors and will act in their own best interest

whenever they have the opportunity, usually at the expense

of the outside investors (Jensen & Meckling, 1976). Such

opportunities are more likely to arise in companies with

poor governance, characterized by the absence of effective

monitoring and disciplining mechanisms. Company insid-

ers in these companies are more likely to adopt suboptimal

strategies, manipulate performance measures, resist take-

overs, and expropriate value (Shleifer & Vishny, 1997) and,

as a consequence, these f‌irms often exhibit signif‌icant under-

performance (e.g., Core, Guay, & Rusticus, 2006; Gompers,

Ishii, & Metrick, 2001). By installing good governance prac-

tices, companies can reduce the agency costs and curtail this

suboptimal behavior. This should result in improved

company performance. So on the basis of agency theory we

expect to f‌ind a positive relationship between company per-

formance and corporate-governance ratings, which we use

as a proxy for corporate-governance practices.

However, the f‌indings of prior empirical studies investi-

gating the relationship between corporate-governance

ratings and f‌irm value or performance are mixed. As far as

emerging countries are concerned, they show that

corporate-governance ratings have a signif‌icant positive

impact on market value (Black, 2001; Black, Jang, & Kim,

88 CORPORATE GOVERNANCE

Volume 18 Number 2 March 2010 © 2010 Blackwell Publishing Ltd