Editorial: Is There Still Something to Learn about the Association of Corporate Governance with Firm Performance?

The association of Corporate Governance (CG) with Firm Performance (FP) has always been an issue relevant to management literature, despite the notable heterogeneity of studies and their mixed results highlighting the opportuneness of continuing to investigate the association of CG with FP. The article aims to contribute to this lively debate by pointing out the main causes of the mixed findings and the contributions of three papers which, in the authors’ opinion, add knowledge to the CG-FP research stream, which deserves more empirical research, as there is still something to learn about this association.

In assuring that the firm is efficiently managed, it can be argued, at least theoretically, that an efficient and effective board can lead to improved financial performance of the firm (Conheady et al., 2015;Leong et al., 2015;Khansalar et al., 2015;Uadiade, 2010;Puni et al., 2014).
However, the huge number of studies in the CG literature addressed to investigate the association of CG with FP under different theoretical perspectives have not successfully provided unequivocal evidence of the positive association of CG with FP (Mazzotta, 2007;Mazzotta & Veltri, 2014).Some studies provide evidence of a link between well-governed firms and firm value/performance (for example, Bebchuk et al., 2009;Gompers et al., 2003), but others have been unable to suggest that good governance actually positively affects firm value/performance (for example, Core et al., 2006;Bhatt & Bhattacharya, 2015) (Note 1).

The mixed findings of the Empirical Studies Investigating the Relationship between Corporate Governance and Firm Performance: Some Critical Considerations
This section would like to provide some critical reflections on the likely explanations for these mixed findings.We need to underline that we base our reflection on the empirical papers addressed to investigate the links between CG and FP.These papers make use of empirical-archival research methods, refer to wide databases and employ the statistical techniques of regression.For this reason, we concentrate on the problems that are typical of financial accounting research using econometrics, namely the measurement of dependent (CG) and independent (FP) variables, the inclusion of all other variables related to CG and supposed to affect FP, the choice of the suitable regression model.
In our opinion, there are four main reasons likely to affect the CG-FP mixed results: 1) the multidimensional character of CG and FP variables and the related difficulties for researchers to measure CG and FP; 2) the use of research frameworks unable to unveil the role played by CG in affecting FP because of overlooking variables also affecting CG and FP; 3) the use of econometric models not able to take into account properly the statistical problems related to the study of the CG-FP relationship; 4) the missing consideration of the specificities of the contexts investigated.
The multidimensional character of CG and FP and the correlated difficulties encountered in measuring these two variables have led researchers to use different indicators to measure the same issue (Alimehmeti and Paletta, 2014;Al-Matari et al., 2014;Hult et al., 2008;Richard et al., 2009).Even when the CG variable is circumscribed to the board of directors, there are several possible measures; some researchers focus upon some specific board of directors measures, such as the board size, the proportion of non-executive directors, in turn divided into non-executive and independent, i.e. nominated without the votes of the controlling shareholders and non-executive but non-independent directors, i.e. nominated with the votes of the controlling shareholders into independent and nominee (Veltri and Mazzotta, 2016), CEO duality, that is when the role of CEO and board chairperson are filled by same person (Tenuta and Cambrea, 2016;di Donato et al., 2016), CEO features, e.g.belonging to or not to the family-managed firm (Mannarino et al., 2016;Binacci et al., 2015), gender diversity (di Donato et al., 2016;Ferraro and Mazzotta, 2016), some others employing composite measures (CG indices), derived from the literature (Gompers et al., 2003;Bebchuck et al., 2009) or created by summing up different board dimensions (Reverte, 2009;Mazzotta and Veltri, 2014).Also FP is a difficult construct to measure, even though we refer only to objective measures of organizational performance, leaving aside subjective measures of FP (i.e.derived from surveys) (Note 2).Organizational performance is one of the most important constructs in management research (Richard et al., 2009).In the literature, briefly the objective FP ratios are categorized into accounting ratios (i.e.ROE, ROA, ROS); market ratios (i.e.EPS, price-to-earnings ratio) and mixed ratios (i.e.Tobin's Q).The literature has both pros and cons: accounting measures are simple to use, easy to understand, and based on audited figures, but they are historical and backward-looking, can be easily manipulated by changes in accounting policies and distorted by inflation, and do not consider IC, the risk and the cost of capital; on the other hand, market-based performance measures reflect value given by share prices thus the future performance, but they can reflect market expectation rather than true performance.The literature, while searching a relationship between accounting and market FP measures (Gentry and Shen, 2010;Masa'deh et al., 2015;Aliabadi et al., 2013), agree that FP is a complex, multidimensional construct (Richard et al., 2009;Hult et al., 2008;Gentry and Shen, 2010).The choice of the variables to proxy CG and FP affect the results of empirical regression models used to investigate that relation.
The mixed results of studies addressed to investigate the CG-FP relationship could also be due to overlooking several variables in the model used to analyze that relation that could cause biased results of the coefficients of the variables on interest, both in the sign (positive or negative) and in the magnitude.We are not speaking only of control variables, which researchers are used to considering in the light of previous studies, such as the size of the firm or its leverage (Note 3).we are speaking of variables that only recently have been included in the framework of researchers.The reference is, e.g. to the ownership structure (OS) and to the intellectual capital (IC).Only recently, the literature highlighted that OS is a variable that affects FP (Perrini et al., 2008;Barontini and Caprio, 2006), and also the board of directors (De Miguel et al., 2004) and, furthermore, that OS characterizes entire contexts in a different way, being therefore responsible for the differences in CG across countries (Acero Fraile and Alcade Fradejas, 2014).On the other hand, recent studies have focused on the relationship between IC and CG (Keenan and Aggestam, 2001;Muttakin et al., 2015: Mazzotta andVeltri, 2016), IC and OS (Saleh et al., 2009;Bohdanovich and Urbanek, 2013;Bohdanovich, 2014) (Note 4).
The use of a proper regression model could also be one of the key explanations for mixed results.There are two main issues; the first is related to the choice of econometric technique to employ.Researchers, in fact, have the possibility of using Ordinary Least Squares regression on poled data (POLS) or fixed-effects (FE) or random-effect (RE) panel models.POLS regressions estimate a single intercept for all the companies, omitting all those time invariant characteristics that could be peculiar to each company (individual effects).This estimation method can be used on longitudinal data (i.e.observations on each company are repeated over time) only if there are no individual effects and the statistical relevance of company-specific fixed effect should be verified by implementing a poolability test, e.g. the Breusch and Pagan Lagrange-Multiplier (LM) test for random effects.When individual effects are statistically relevant, researchers should use panel models, as the omission of unobservable, yet relevant, factors would make the model misspecified from an econometric point of view and would produce OLS-biased (or inconsistent) estimates.Panel models, instead, offer the opportunity to control properly for unobserved individual company heterogeneity by factoring out a different fixed effect for each company.The other relevant econometric issue is the endogeneity issue, which is caused by simultaneous causality between X and Y (CG and FP) and that can be addressed with sophisticated econometric models (Brown et al., 2011;Larcker and Rusticus, 2010) or by choosing measures of CG and FP variables determined exogenously.
Finally, another reason for the mixed findings could be linked to the consideration that few papers until now have been aware of the context and on how the context can affect the CG-FP relationship (Bauer et al., 2004).In their 2011 review, Brown et al. (2011) conclude their article suggesting researchers be more focused on measures that capture local conditions, as best governance practices are not impenetrable to local conditions.In order to advance our knowledge, we do not need one-size-fits-all measures of governance; instead, we need to take into consideration different contexts (i.e.countries like Continental Europe and East Asian economies are characterized by different ownership configuration with respect to countries like the UK and the United States, and this in turn affects the type of CG problem that companies face in each of these different contexts, Acero Fraile and Alcade Fradejas, 2014) and different industries (i.e.CG in financial institutions could be measured using a different proxy from CG in industrial companies and CG in family firms, also providing different results, Bachiller et al., 2016).
Summarizing, despite the breadth, and depth, of the literature on CG, room for improvement could be found in the form of better models, better empirical proxies, better data, better estimators, better analysis or better interpretations.Of course, because the literature already has a degree of maturity, most improvements will be incremental, but, consistently with Brown et al. (2011), we believe that "there is still much we can do".

The Contribution of Empirical Research to the Debate
All the articles included in this special issue have been presented to the Sidrea (Società Italiana Docenti di Ragioneria ed Economia Aziendale) International Workshop (SIW) "Innovations in corporate governance and performance management", organized by UNINT University in Rome on 21 and 22 of April, 2016.
The articles have been selected by the Editors of this issue because all of them deal with the relationship between CG (namely the board of directors) and FP in a creative and innovative way, trying to overcome the problems underlined in the first section of this article and using different yet rigorous and sophisticated econometric models.
The contributions share some common points (all the articles employ quantitative econometric models and are based on the Italian context, a country which presents relevant peculiarities to consider from a CG point of view) and differ for some other points (the choice of variables to proxy CG and FP) that we are going to briefly illustrate.

Italian Context
Italy is a European Continental country characterized by a weak equity capital and a low legal enforcement system, thus belonging to the code-law European countries (Nobes and Parker, 2012).One main characteristic of the code law countries, including that of Italy, is the predominance of small-or medium-sized companies and ownership being mostly concentrated in the hands of few family members (Melis et al., 2012).A recent survey conducted by the Italian Chamber of Commerce in 2013 reveals that 86% of 6,061,960 Italian companies are family businesses, out of which almost 99% are small-and medium-size (Jaggi et al., 2016).
As regards the stock market, Italy is a small equity market (a quarter of the UK market) (Note 5), in which few medium-and large-size companies form pyramidal groups, in which the holding company controls other companies through voting trusts or by issuing nonvoting shares (Bianchi et al., 2001;Di Pietra et al., 2008).The Italian stock exchange is also characterized by its having large controlling shareholders: in 2012, the average percentage ownership held by the first stakeholder was 46.8%, with a low presence in terms of the average percentage of shares (less than 7%) belonging to international and national institutional.Instead, despite the privatization process, the government-owned companies made up approximately 34% of the total market capitalization of Italian companies listed on the Italian Stock Exchange in 2013.Three different classes of major shareholders of Italian companies are commonly identified: (i) active family members, (ii) non-family members, and (iii) groups of shareholders with venture capital or entrepreneurial backgrounds.This high ownership concentration means the stock market plays a limited role because the predominant source of capital for these Italian companies is a bank loan (Note 6), which makes Italy a credit-oriented financing systems (Veltri and Ferraro, 2012).Banks, as main capital providers, play a relevant role in affecting firm financial disclosure policies, but are not involved in the CG mechanisms (Manes Rossi et al., 2015).
The Italian firms mainly have a two-tier corporate governance structure, which consists of the Board of Directors and the Board of Statutory Auditors (Bianchi et al., 2001).The Board of Directors is entrusted with the responsibility of managing the company, with the primary goal to create value for shareholders over a medium or long-term period.Shareholders are allowed to appoint (and remove) directors.In such a context, the controlling families dominate the governance mechanism, where family members are appointed as directors on corporate boards and also as CEOs of the companies (Enriques and Volpin, 2007;Prencipe et al., 2008).The weak distinction between owners and managers causes a Type 2 agency problem, which result in expropriation of minority interests by the majority shareholders (Melis et al., 2012) and an inadequate system of protection of outside investors.As a result of this high ownership concentration, the effectiveness of corporate boards in monitoring managerial decisions and activities appears considerably reduced (Di Pietra et al., 2008).

Corporate Governance Variables
All the articles included in the special issue focus on the board of directors as the key governance tool in monitoring executives on behalf of shareholders, so ensuring executives pursue firm's instead of their own interests (Hermalin and Weisbach, 1991).An efficient and effective board of directors is supposed to lead to improved firm performance (Conheady et al., 2015), anyway board characteristics and roles cannot be defined in a general way, because they differ widely across various countries and company types (Charkham, 1995).Following this assumption, each article stress one or more different peculiarities of the board of directors in the Italian context.Veltri and Mazzotta (2016) in their article focus on the board independence and ownership concentration, di Donato et al. (2016) on board diversity and interlocking directorship, while Mannarino et al. (2016) on the family management.

Board Independence
The board of directors is defined by law.In Italy, the regulations make a requirement that "at least one third" of directors on the Board should be independent (Jaggi et al., 2016).Two are the definitions of independence provided by the Italian legal system: the first is provided by the Law in Finance No. 58/1998 (revised over time) for listed companies (Note 7), the second is provided by the Code of Corporate Governance, a voluntary Code published in 1999 and revised over time by the Stock Exchange Committee for Italian listed companies (Note 8).Generally, the members of the board of directors are distinguished into executive (i.e.non-independent) and non-executive (independent), but in a context of high ownership concentration, like Italy, it is necessary to distinguish among independent board members the members elected with the controlling shareholder votes (but that are independent because of the lack of formal family or other ties to the controlling shareholder) from those elected by minority shareholders (Lefort and Urzúa, 2008).To this last category is assigned the protection of minority shareholders' interests.Veltri and Mazzotta (2016) in their research framework measure the board of independence exogenously from the Italian Code of Corporate Governance to address endogeneity issue.

Ownership Structure
With regard to corporate ownership structure, a conventional classification distinguishes between two broad categories.In the first, ownership is more dispersed and cross-shareholdings are rare.This ownership configuration characterizes countries like the UK and the United States (Vintilă et al., 2015).In the second, firm ownership is often concentrated within a small number of other firms, banks, and families.This ownership configuration characterizes the countries of Continental Europe and East Asian economies.The problem of corporate governance that companies face in each of these contexts is different: in countries where ownership is dispersed (first category), the predominant problem is the agency problem between shareholders (principals) and managers (agents), as a result of the separation of ownership and control (type I agency problem).On the other hand, in contexts such as Italy, where equity ownership is highly concentrated, the corporate governance problem is focused on the relationship between small and large shareholders (type II agency problem).In the first case, the board of directors addresses mainly supervision tasks, whilst in the second case (Italy) the board of directors is called to broaden the scope of its supervisory function to safeguard the interests of small shareholders (Acero and Alcade, 2016;Baglioni and Colombo, 2013).Veltri and Mazzotta (2016) measure the ownership structure exogenously from the Italian Stock Exchange to address endogeneity issue.

Interlocking Directorship
The interlocking directorship is a typical feature of the Italian Board of Directors, where executives of one company can be appointed as directors of another company and that company can in turn have its executives as directors of the first company.The Italian legislative body has made significantly changes to corporate governance issuing new sets of law related to financial market to discourage interlocking (Note 9).The data evidence a decrease in multiple directorship in Italian listed companies from 2011 to 2013 in all sectors (Jaggi et al., 2016).In the literature there are two competing positions, one assuming a negative effect of interlocking directorship on firm performance, the second assuming instead a negative effect of interlocking directorship on firm performance.This second position is grounded on the resource dependence theory, according to which the board of directors must become the main link with the environment, ensuring control of company resources through social and professional networks, as well as through the mechanism of interlocking directorship (Bachiller et al., 2015).Based on resource dependence theory, di Donato et al. (2016) include this board variable in their research framework, theory hypothesizing a positive relationship with the firm performance.

Board Diversity
Arguments in favor of a stronger presence of woman on corporate boards to enhance financial performance rely both on agency theory, according to which gender diversity increases board independence, which is known to improve monitoring (Carter et al. 2003) and resource dependence theory, according to which a diverse board brings more resources to the firm, which will result in better firm performance and value (Huilman and Daziel, 2003).Empirically speaking, despite a large body of literature examining the relationship between women on boards and firm financial performance, the evidence is mixed (Isidro and Sobral, 2015).Some studies have found that female directors add value (Nguyen and Faff, 2012); other studies have suggested that female directors decrease firm performance (Darmadi, 2011).A recent meta-analysis (Post and Byron, 2015) found out that that female board representation is positively related to accounting returns, but that this relationship is lower in countries with weaker shareholder protections like Italy.In Italy, female representation within companies is still weak (Bianco et al., 2011); furthermore, although the women on corporate boards serve as independent directors in the majority of cases (nearly 60% of women directorships), women CEOs account for 3.2% of total female directorships (Jaggi et al., 2016).However, in 2011 Law no.120 was issued, addressed to support the presence of the less represented gender in the corporate boards of listed companies (Note 10).The research of an association between gender diversity in boards in the Italian context deserves more investigation.Based on the above considerations, di Donato et al. (2016) includes this variable in their research framework.

Family Firms
If the articles of Veltri and Mazzotta (2016) and di Donato et al. (2016), focused on Italian listed companies, where the "family effect" is not clearly distinguishable from the "ownership effect", the article of Mannarino et al. (2016) stress the peculiarities of family firms, deriving from the strong interconnections between company and family.In this context, the board, especially when non-family members are present, play a relevant role in balancing both entities (Bachiller et al., 2016).In their article, Mannarino et al. (2016) identify family firms on the basis of management type, including in the family subsample both family firms run by a family members and family firms run by a professional manager (Note 11).According to EFIGE estimates (Note 12), Italy has far more family-owned firms where the whole management is from the same family, and this is a peculiarity with respect the other European countries, in that these firms make up two-thirds of Italian firms, against a quarter in France and Germany and one-tenth in the UK (Cucculelli and Bettinelli, 2016).In the literature there are two competing positions about the relationship between CG and FP (Rossi, 2012).The first hypothesizes a positive effect of family firms on FP, owing to features such the long-term orientation, the tighter control of management by family, the alignment between family and management (Corbetta, 2012).A second position instead stresses negative characteristics such as the pursuit by family members of their own benefits despite the firm benefits, the risk aversion, the unwillingness to grow dimensionally, the ineffective selection of CEOs (Tiscini and Raoli, 2013;Barontini and Bozzi, 2011).Empirical evidence provided mixed results (Baschieri, 2014).In order to overcome the difficulties of existing research, Mannarino et al. (2016), coherently with Cucculelli and Bettinelli (2016) introduced an external variable, namely the institutional quality, to proxy the role of the institutional environment in affecting the performance difference between family firms run by a family member and firms run by a professional manager.

Firm Performance
As regards the choice of the proxy of the dependent variable (i.e.FP), the three articles choose a different variable, thereby providing evidence of the multidimensionality of FP.In detail, Veltri and Mazzotta (2016) and di Donato et al. (2016) chose to proxy the financial performance outcome with an accounting ratio (Veltri and Mazzotta, 2016), and with an accounting ratio and a mixed ratio (di Donato et al., 2016).On the contrary, Mannarino et al. (2016) use the total factor productivity as measure of performance, a measure typical of economic studies, so addressing the call of Brown et al. (2011) for a multidisciplinary approach in CG research.

Contribution of the Papers
In our opinion, all the articles included in the special issue add knowledge to the literature on the CG-FP relationship: they employ an original research framework (2016), innovative techniques to process data (i.e.di Donato et al.,2016), novel techniques to measure variables (i.e.Mannarino et al., 2016).The results provide evidence that research frameworks built up to capture context and local conditions are more likely to provide evidence of the CG-FP relationship.Veltri and Mazotta (2016) in fact, provide evidence that Italian board features (i.e. the ownership concentration and the two different figures of nonexecutive board members) affect firm value and also di Donato et al. (2016) provide evidence that interlocking directorship affects firm performance.Finally, Mannarino et al. (2016) provide evidence that the institutional environment affects the CG-FP relationship in family firms and that family firms will more significantly outperform non-family firms in an underdeveloped institutional environment than in a developed institutional environment.

Notes
Note 1.An in-depth analysis of the huge literature addressed to investigate the association of CG with FP is outside the scope of the paper.See the two meta-analyses performed by Dalton et al. (1998); Dalton et al. (1999) among others.Note 2. In the paper we consider only one dimension of performance, that is we focus only on financial aspects, so leaving aside the concept of social performance and the studies that focused on it (Bachiller et al., 2016).Note 3. In the literature, both for firm size and leverage control variables we have alternative positions.Firm size is believed to have a positive effect on firm performance, as large companies can access funds more easily and are also able to create entry barriers (Short and Keasey, 1999), even though there are other theories that suggest the possibility of a negative relationship between firm size and FP, based on the considerations that the larger the firm, the more serious the bureaucracy and agency problems (García-Meca and Sánchez-Ballesta, 2011).On the other side, Leverage according to trade-off theory has a positive relationship with FP, as firms with higher profitability (ROA) tend to have higher debt to benefit from the tax shield (Miller, 1977).Anyway, according to the pecking order theory, in selecting the capital structure managers first recur to retained earnings, then to debts and finally to equity, thus the debt level in the capital structure of a firm is inversely related to financial performance (ROA).Empirical studies supporting the first view are those of Abor (2005), Berger and Bonaccorsi di Patti (2006), Margaritis and Psillaki (2009); empirical studies supporting the second view are those of Titman and Wessels (1988), Rajan and Zingales (1995), Zeitun and Tian (2007), Quang and Xin (2014).Note 4. Intellectual capital (IC) can be defined as the dynamic and firm-specific system of intangible resources and activities in gaining and sustaining firm competitive advantages (Reed et al., 2006;Martin-de-Castro et al., 2011).In the IC literature the search for an association of IC with FP has always been a key issue, even though empirical and case evidence was inconclusive (Guthrie et al., 2012).For a meta-analysis on the association of IC with FP, see Veltri (2012).For an analysis of this relationship within the Italian market, see (Pucci et al., 2015;Celenza & Rossi, 2014).Note 5.In June 2013, there were 247 companies listed on the Italian Stock Exchange (MTA) with an average market capitalization of €1,402 million.Source: Italian Stock Exchange Commission.Note 6.In June 2013, the Italian listed companies had an average leverage ratio (financial debts to equity) approximately of 47%.Source: Italian Stock Exchange Commission.
Note 7.The law 158/1998 in the art 148 so defines independent members: "(a) persons who are in the conditions referred to in Article 2382 of the Civil Code [mainly debarred, disqualified or bankrupt persons] (b) spouses, relatives and the like up to the fourth degree of kinship of the directors of the company, spouses, relatives and the like up to the fourth degree of kinship of the directors of the companies it controls, the companies it is controlled by and those subject to common control; (c) persons who are linked to the company, the companies it controls, the companies it is controlled by and those subject to common control or to directors of the company or persons referred to in paragraph (b), by self-employment or employee relationships or by other relationships of an economic or professional nature that might compromise their independence".Note 8.The code of Corporate Governance states that directors, in order to be considered independent, "do not maintain, directly or indirectly or on behalf of third parties, nor have recently maintained any business relationships with the issuer or persons linked to the issuer, of such a significance as to influence their autonomous judgment".Note 9.In particular, the so-called Rescue-Italy Law Decree (no.6/2011) has provided a new eligibility requirement, according to which "no member of management boards, supervisory boards and statutory board of auditors, as well as no executive officer, of undertakings or group of undertakings which are active on the markets for banking, insurance and finance shall, at the same time, serve in corresponding positions in competing undertakings or groups of undertakings" (Article No. 36).
Note 10.The law requires that the companies shall introduce gender quotas in their by-laws, equal to 1/5 of the board membership for the first mandate and 1/3 for the following two mandates.Note 11.Most of the literature identifies the family character considering both family ownership (control of the firm) and family involvement in management.
Note 12. EFIGE (European Firms in a Global Economy) is a database founded also by Banca d'Italia, which provides microdata related to single firms in a sample of European countries.