The relationship between firm performance and board characteristics in Ireland

European Management Journal (2010) 28, 387– 399

journal homepage: www.elsevier.com/locate/emj

The relationship between firm performance and board characteristics in Ireland Vincent OÕConnell a b

a,*

, Nicole Cramer

b

Department of Accountancy, University College Dublin, UCD School of Business, Belfield, Dublin 4, Ireland ¨sseldorf, Germany Group Finance, West LB AG, Du

KEYWORDS Corporate governance; Board size; Board composition; Firm performance; Irish Stock Market

Summary In this study we explore the association between firm performance and both board size and board composition for companies quoted on the Irish Stock Market. We also investigate the impact of firm size on the relationship between firm performance and the aforementioned board characteristics. We find evidence that: (i) board size exhibits a significant negative association with firm performance, (ii) the relationship between board size and firm performance is significantly less negative for smaller firms, and (iii) a positive and significant association between firm performance and the percentage of non-executives on the board is apparent. While the latter finding is entirely consistent with a priori theoretical predictions, studies in a number of other countries generally fail to report any significant association between board composition and firm performance and potential reasons for this contrast are considered. ª 2009 Elsevier Ltd. All rights reserved.

Introduction Contemporary boards of directors are charged with the task of monitoring the performance and activities of top management to ensure that the latter act in the best interests of the owners (Jensen and Meckling, 1976). From this perspective boards have a potentially critical role to play in mitigating agency problems arising from the ubiquitous separation of firm ownership from control (Fama, 1980; Jensen, 1993; Shleifer and Vishny, 1997). In addition, Ruigrok et al. (2006) point out that boards also have important roles with respect to activities such as designing and implementing strategy and fostering links between the firm and its exter* Corresponding author. Tel.: +353 (0)1 7164735. E-mail address: [email protected] (V. OÕConnell).

nal environment. Given their multifaceted tasks it seems plausible that boards may impact firm performance and, if so, questions naturally arise as to what types of board structures are optimal from the perspective of maximizing stockholdersÕ wealth. Hence, it is hardly surprising that questions about the impact of board characteristics on firm performance have attracted significant research attention across a range of countries in recent years (Denis and McConnell, 2003). One important topic within the aforementioned research agenda is the potential influence of board size on firm performance. While larger boards may result in a wider pool of expertise (Zahra and Pearce, 1989) and greater external linkages (Goodstein et al., 1994), larger

0263-2373/$ - see front matter ª 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.emj.2009.11.002

388 boards may also lead to lower group cohesion (Evans and Dion, 1991) and greater levels of conflict (Goodstein et al., 1994). A second and related research question is the potential effect of board composition – in particular, the proportion of independent non-executive (i.e., outside) directors – on firm performance. Contemporary guidelines on corporate governance practices consistently emphasize the critically important role of non-executive directors in mitigating manager–shareholder conflicts. For example, both Hampel (1998) and Higgs (2003) recommend that independent non-executive directors should comprise at least 50% of UK boards. Meanwhile, from an agency perspective, greater board independence should result in more effective monitoring (Stiles and Taylor, 2001). However, many corporate governance researchers have questioned the true impact of board independence on firm performance (Dulewicz and Herbert, 2004). The empirical literature on the relationship between firm performance and board characteristics – such as size and composition – is quite extensive and while generalizations of such an impressive body of work are inevitably flawed, two overarching findings are apparent. First, prior research provides mixed evidence with regard to the impact of board size on firm performance (e.g., Yermack, 1996; Dalton et al., 1999). Second, prior work has largely failed to establish a convincing link between the proportion of outside directors and firm performance (e.g., de Andres et al., 2005). However, as Brennan (2006) points out, the impact of corporate governance characteristics on firm performance is likely to vary across jurisdictions and from this perspective, cross-country research can provide valuable incremental insights. Motivated by the need for additional comparative work, the present study presents the first empirical evidence on the relation between firm performance and board characteristics in Ireland. Our specific objective is to contribute to extant governance research in four ways. First, we explore the question of whether the negative relation between board size and firm performance observed in many other countries is apparent in the Irish context. Second, we test whether firm size has a moderating influence on the relation between firm performance and board size. This issue has been largely ignored in prior research. More specifically, we explore whether larger boards may be comparatively less disadvantageous in the context of smaller Irish firms. Third, we investigate whether having a higher proportion of non-executive directors on the board is positively associated with firm performance. While research in many other countries reports no association between firm performance and the proportion of outside directors we believe that certain characteristics of the Irish corporate environment – such as the comparative absence of interlocking directors – may lead to significant contrasts between the results for Ireland and other jurisdictions. Fourth, since we use a range of alternative measures of firm performance in our empirical analysis our study highlights the potential sensitivity of empirical work in this strand of the governance literature to alternative conceptualizations and definitions of performance. The remainder of the paper is structured as follows. The three hypotheses are developed in Hypothesis development section while the research design is outlined in Research design section. Corporate governance for listed Irish firms

V. OÕConnell, N. Cramer section provides some background information about corporate governance in the Irish context and this discussion is followed in Sample section by a summary of the sample selection procedures. The core findings from the empirical work are outlined in Empirical findings section. In Conclusions section we present a summary of our main findings, an overview of the key managerial implications of our results as well as an outline of potential avenues for future research and the limitations of our study.

Hypothesis development The impact of board size on firm performance The nature of the relationship between board size and firm performance has come under increased scrutiny in recent years. For example, Yermack (1996) investigates the impact of board size on firm value for a sample of large US industrial corporations between 1984 and 1991 and finds an inverse relation between firm value (measured by TobinÕs Q) and the number of directors. Yermack (1996) also shows that financial measures, such as return on assets and return on sales, are negatively related to board size. In contrast, the meta-analysis presented in Dalton et al. (1999), which draws on the results from a number of prior US studies, indicates a positive relationship between board size and firm performance. In a sense, these contrasting findings suggest that there may be advantages as well as disadvantages to larger boards. The results discussed above are based exclusively on US data. From an international perspective, Conyon and Peck (1998) find a negative relationship between return on equity and board size for a sample of European firms although their results with respect to market-based measures of performance are less clear-cut. More recently, de Andres et al. (2005) report a negative association between firm value and board size (controlling for a number of additional factors) in 10 OECD countries. Taken as a whole, these internationallybased results are consistent with JensenÕs (1993) argument that that the benefits resulting from larger boards are outweighed by the incremental costs of the potentially poorer communication and decision-making processes associated with larger groups. Hence, in line with the core findings from prior international research, we predict that board size is negatively associated with firm performance in Ireland: Hypothesis 1. Firm performance exhibits a negative association with board size.1 1

At the outset, it is important to emphasize that we test for ÔassociationÕ rather than ÔcausalityÕ. Freedman (1999) presents a comprehensive overview of the difficulties involved in drawing definitive conclusions about causality from regression analysis. Further, Listokin (2008, p. 100) points out that some of the most influential papers in corporate governance research ‘‘emphasize that the cross-sectional results prove associations rather than causation’’. ÔCausationÕ has fundamentally stronger conations than ÔassociationÕ and as Gujarati (2006, p. 23) states ‘‘a statistical relationship in itself cannot logically imply causation’’. ÔAssociationÕ in the present paper means that the dependent and independent variables exhibit a statistically significant relationship with one another within the context of the model presented in Eq. (1). The terms ÔrelationshipÕ and ÔassociationÕ are treated as interchangeable throughout the study.

The relationship between firm performance and board characteristics in Ireland

389

The moderating influence of firm size on the firm performance-board size relationship

Hypothesis 2. The association between firm performance and board size is significantly less negative for smaller firms.

Prior research on the relationship between corporate governance and firm size provides mixed evidence on whether or not larger firms tend to have better corporate governance structures (e.g., Drobetz et al., 2004; Ariff et al., 2007). Meanwhile, there are strong reasons to suspect that identical corporate governance practices and structures do not work equally well for both large and small firms (Klapper and Love, 2003). For example, Holmstrom and Kaplan (2003) argue that the costs of complying with the Sarbanes–Oxley act of 2002 in the US are proportionately greater for smaller firms and Chhaochharia and Grinstein (2007) report empirical evidence consistent with this view. Although the impact of firm size on the firm performance-board structure association is frequently ignored in empirical research, there are a number of reasons why size may be an important moderating variable.2 First, because of their reduced scale and complexity smaller firms are less likely to have coalitions of directors pursing a diverse range of interests. In particular, as Dalton et al. (1999, p. 679) point out, ‘‘It may be that smaller firms do not have as many competing coalitions as their larger counterparts. If so, stability and cohesiveness among constituent groups may be higher . . . and goal consistency stronger’’. However, it appears worthwhile to add the caveat that these potential effects may be attributable to smaller firms having smaller boards rather than always being present in smaller firms regardless of the size of the board. Second, work by Finkelstein and Hambrick (1996) suggests that directors are more likely to impact on the strategic direction of the organisation in smaller firms – from this perspective, larger boards may be less harmful in smaller firms. Third, it is well established that a firmÕs Ôinformation environmentÕ is directly related to its size (Collins et al., 1987; Ryan, 2005).3 For example, prior work by OÕConnell (1995) demonstrates that the information environment is significantly weaker for smaller Irish firms. Since larger firms typically operate in richer information environments, these companies are generally subject to greater levels of scrutiny from media, politicians and the general public. In contrast, because smaller firms operate in far more limited information environments the monitoring role of the board may well be even more critical. Since prior research shows that the monitoring role of the board is an increasing function of board size (e.g., John and Senbet, 1998; Kiel and Nicholson, 2003) larger boards may have a comparatively less negative impact for smaller firms. This discussion leads to the second hypothesis.

The impact of board composition on firm performance

2

While firm size is often used as a control variable, few studies directly test its impact on the firm performance-board size association. 3 Collins and Kothari (1989, p. 145) define the information environment as incorporating ‘‘all sources of information relevant to assessing firm value. It includes government reports on macroeconomic conditions, industry reports and trade association publications, firm-specific news in the financial press and reports issued by analysts and brokerage houses in addition to accounting reports, and vertical and intra-industry information transfers via sales and industry reports’’.

Extant work in the analytical agency tradition (e.g., Stiles and Taylor, 2001) suggests that a higher proportion of outside directors should be associated with stronger financial performance. Given the importance attributed to the role of independent non-executive directors in both the Hampel (1998) and Higgs (2003) reports, it is hardly surprising that UK research points to an increasing proportion of outside directors on corporate boards in recent years (Pye, 2000). Interestingly, it appears that the US stock market responds positively to the announcement of the appointment of nonexecutive directors (Rosenstein and Wyatt, 1990).4 From a strategic perspective recent work by Yawson (2006) also suggests that when facing performance declines, firms with a higher proportion of outside directors are more likely to sanction staff layoffs. Nonetheless, notwithstanding these findings, there is a relative dearth of empirical evidence pointing to a significant positive association between firm performance and board independence. For example, Hermalin and Weisbach (1991) conclude that there is no relation between the proportion of non-executive directors and firm performance in the US; Vafeas and Theodorou (1998) and Dulewicz and Herbert (2004) report similar findings for the UK. Recent work by de Andres et al. (2005) also fails to establish a statistically significant association between firm performance and board composition across a sample of OECD countries. Other research (Agarwal and Knoeber, 1996; Klein, 1998) suggests that US boards may in fact have an excessive proportion of non-executive directors. Furthermore, many academics and commentators have questioned the validity of the notion that a board should be comprised of at least 50% non-executives (e.g., Dulewicz and Herbert, 2004). Nonetheless, empirical work also reveals that firms with a higher proportion of outside directors have a smaller likelihood of experiencing financial distress (Elloumi and Gueyie, 2001). In addition, financially distressed firms with independent boards have a lower incidence of bankruptcy filings (Daily et al., 2003). The preceding discussion leads to our third hypothesis stated in the alternative form: Hypothesis 3. Firm performance exhibits a positive association with the proportion of non-executive directors on the board.

Research design As discussed earlier, the objective of this study is to explore the nature of the relationship between corporate performance and both board size and board composition for quoted Irish firms. In common with international studies in 4 Other evidence using event study methodologies points to higher announcement period returns for targets of tender offer bids when the boards of those targets are controlled by non-executive directors (e.g., Byrd and Hickman, 1992; Cotter et al., 1997).

390 Table 1

V. OÕConnell, N. Cramer Variable definitions.

Variable

Definition

BRDSIZE NED% OWNERSHIP RET

The number of executive and non-executive directors on the board The percentage of non-executive directors on the board The percentage of total directorsÕ stock ownership The one-year raw stock market return. The prices at the beginning and end of the fiscal year are adjusted for dividends Profit before interest and tax over total assets Sum of market capitalization plus long and short-term debt over the book value of total assets

ROA FINANCIAL Q

the field, board size is measured as the sum of the number of executive and non-executive directors while board composition is measured as the proportion of non-executive directors. These variables are referred to as BRDSIZE and NED%, respectively, and are summarised in Table 1. Since prior work (e.g., Faccio and Lasfer, 1999) has shown that directorsÕ shareholdings may also be an important influence when considering the relationship between firm performance and board size, the proportion of the company owned by the directors (OWNERSHIP) is included as a control variable in our empirical analysis. Three different measures of firm performance are used in the present work and summary definitions of each of these measures are outlined in Table 1. RET, the marketbased measure, is calculated in the usual fashion (Klein, 1998; Vafeas and Theodorou, 1998) as the change in stock price plus dividend for the period. Return on assets (ROA) which is defined as the ratio of earnings before interest and tax over total assets has also been widely used in prior corporate governance research (Yermack, 1996; Klein, 1998; Vafeas and Theodorou, 1998). Finally, in common with other studies (for example, Yermack, 1996; Hermalin and Weisbach, 1991; Faccio and Lasfer, 1999; de Andres et al., 2005) we also utilise a version of TobinÕs Q as a measure of corporate performance. Similar to de Andres et al. (2005) and Faccio and Lasfer (1999) we define our proxy for TobinÕs Q as the market value of equity plus the book value of debt divided by the book value of total assets. This measure is typically referred to as FINANCIAL Q and prior work (e.g., Lewellen and Badrinath, 1997) shows that this measure is a very useful practical proxy for TobinÕs Q. The three performance measures which we employ all capture different aspects of firm performance. Like any accounting metric, ROA is essentially a retrospective or historic measure which is independently audited and is freely available (Watts and Zimmerman, 1986). However, accounting numbers are prone to manipulation (Healy and Wahlen, 1999) and the accountantÕs tendency towards conservatism means that bad news is reflected more quickly in accounting numbers than good news (Basu, 1997; Watts, 2003; Beekes et al., 2004) While stock returns are an unbiased and independent measure, an important drawback is that returns tend to reflect a substantial element of anticipated as opposed to realized performance (Leone et al., 2006; OÕConnell, 2006). In contrast, accounting metrics such as ROA largely reflect realized or delivered performance (Barclay et al., 2005). Finally, FINANCIAL Q is a ÔmixedÕ measure in that it reflects both market-based and accounting-based elements of performance. Higher values of FINANCIAL Q indicate more efficient

asset utilisation in the sense that, as Enriques and Volpin (2007, p. 122) point out, the assets are ‘‘worth more within the firm than in alternatives uses’’. However, a potential weakness of FINANCIAL Q is that it may be a noisy measure from the standpoint of detecting the impact of governance variables on performance (Himmelberg et al., 1999). Overall, RET, FINANCIAL Q and ROA may be viewed as complimentary rather than competing metrics which capture different aspects of firm performance. The empirical specification utilised to test the three hypotheses outlined earlier is summarised in Eq. (1) below Performance measure ¼ b0 þ b1 Dt þ b2 BRDSIZEt þ b3 BRDSIZEt  Dt þ b4 NED%t þ b5 OWNERSHIPt þ et

ð1Þ

With this model, the two key board structure variables (BRDSIZE and NED%) and the control variable (OWNERSHIP) are regressed on each of the three alternate performance measures (RET, FINANCIAL Q and ROA). To test for the impact of firm size, we rank firms on the basis of total market capitalization at the year-end (Chhaochharia and Grinstein, 2007). We then assign a dummy variable D which takes a value of 1 if a firm has a market capitalization lower than the median and zero otherwise. Hence, the top 50% of the sample firms (in terms of stock market value) are classified as ÔlargeÕ while the remaining companies comprise the ÔsmallÕ firm grouping. D is also included in the regression model and, in addition, we interact D with BRDSIZE to test whether there is any firm size-related differential with respect to the association between corporate performance and board size.5 Finally et in (1) represents the error term. Tests of the three hypotheses are based on OLS estimates of Eq. (1) and may be summarised in mathematical terms as follows: 5

We also explore an alternative to the dummy variable approach described here. In particular, we drop the dummy variable (D) and the dummy variable interaction term (BRDSIZE * D) in Eq. (1) and instead include a new interaction variable (BRDSIZE * MV) – where MV is the natural log of market capitalization. MV is also included in the alternative model as an independent variable so that the revised equation is: Performance measure = b0 + b1 MVt + b2 BRDSIZEt + b3 BRDSIZEt * MVt + b4 NED%t + b5 OWNERSHIPt + et. However, we do not persist with this alternative since preliminary analysis reveals that OLS estimates suffer from severe multicollinearity problems with individual Variance Inflation Factors (VIF) as high as 244 apparent (the mean VIF is 91). These VIF levels are far higher than the recommended maximum of 10 in Belsley et al. (1980). For estimates of Eq. (1) the mean VIF is 5.7 which is well below the Belsley et al. (1980) recommended level.

The relationship between firm performance and board characteristics in Ireland • Hypothesis 1 (firm performance is negatively related to board size) predicts BRDSIZE < 0. • Hypothesis 2 (the association between performance and board size is less negative for smaller as compared with larger firms) predicts BRDSIZE * D > 0. • Hypothesis 3 (firm performance is positively related to the proportion of non-executive directors on the board) predicts NED% > 0.

Corporate governance for listed Irish firms The Irish Stock Exchange is a comparatively small market by international standards. For example, the total market capitalization of the Irish market (in US$) at the end of June 2008 was $102,112 million while comparative figures for the UK and US were $3,154,234 million and $15,257,115 million, respectively (Standard and Poors, 2008). Firms listed on the Irish market follow the conventional Anglo-American style of corporate governance with a unitary board structure. Irish listed companies are required to ÔbenchmarkÕ their governance practices against the best practice guidelines contained in the Combined Code (Financial Reporting Council, 2006) and while compliance with the Code is not compulsory, firms are required to disclose whether or not they have complied and explain their reasons for any noncompliance (Donnelly and Kelly, 2005). This Ôcomply or explainÕ approach has been the subject of ongoing debate. For example, in their July 2009 report entitled ÔReview of the Effectiveness of the Combined CodeÕ, the Financial Reporting Council (2009, p.4) in the UK stated that ‘‘both companies and investors have expressed reservations about the way in which Ôcomply or explainÕ works in practice and it is clear that more needs to be done to encourage all parties to apply it in the intended manner’’. The version of the Combined Code in operation for the period covered by the present study was that issued by the Hampel Committee on Corporate Governance in June 1998 (London Stock Exchange, 1998). An important aspect of that particular version was that, while it highlighted the important role of independent non-executive directors, it did not present a formal definition of independence.6 The subsequent Higgs report on the role and effectiveness of NEDs was published in January 2003. This report presented guidelines for ascertaining director independence and recommended that the board (excluding the chairman) should be comprised of at least 50% independent directors. These points were formally incorporated in the July 2003 version of the Combined Code which was implemented for reporting years commencing 1 November 2003. However, it is important to note that our data is from the Ôpre-HiggsÕ period and our findings have to be interpreted in this context.7

6

Page 17 of the Combined Code issued by the Hampel Committee (1998) states that ‘‘it is important that there should be a sufficient number of non-executive directors, a majority of them independent and seen to be independent; and that these individuals should be able both to work co-operatively with their executive colleagues and to demonstrate objectivity and robust independence of judgement when necessary’’. 7 The Combined Code has more recently been revised in 2006 and 2008.

391

When considering the issue of board appointments, one pervasive aspect of the Irish governance regime is that, as OÕHiggins (2002) points out, non-executive directors are largely co-opted through Ôthe old boysÕ networkÕ. For example, she concludes (p. 27) that the ‘‘selection of non-executive directors is an opaque one that tends to perpetuate an inner homogenous clique in the corporate life of Ireland’’. Nonetheless, while non-executive directors are drawn from a relatively homogenous pool of potential candidates, MacCanna et al. (1999) found few interlocking directorates in Ireland compared with other countries. In addition, Brennan and McDermott (2004, p. 331) report ‘‘only six situations of interlocking directorates of various degrees’’ in their comprehensive analysis of the 751 directors of all but one of the 81 firms listed on the Irish Stock Exchange in mid 2002.8 Turning to the issue of board composition, Brennan and McDermott (2004) provide some interesting insights in their study which applied the seven NED independence criteria of the just-published Higgs report retrospectively to Irish listed companies. Their work shows that Irish companies were generally compliant with the applicable stock exchange independence requirements of the time. In particular, most firms in their sample complied with recommendations in the Combined Code in terms of having a balanced board structure although only approximately 60% of firms had majority independent boards. However, the Brennan and McDermott study showed that many Irish companies would have to make substantial changes in order to comply with the recommendations of the Higgs report which were soon to become the required benchmark. In particular, Brennan and McDermottÕs (2004, p. 325) state that their research reveals ‘‘a lack of consistency in interpreting the definition of ÔindependenceÕ, a lack of disclosure of information and, by applying criteria generally regarded as prerequisite to independence of non-executive directors, certain situations which imposed upon their independence’’.

Sample Our initial potential pool of sample firms consists of the 77 firms listed on the following markets of the Irish Stock Exchange at the end of December 2001: the Official List (65 firms), ITEQ – the market for Irish technology stocks (8 firms) and the Developing Companies Market (4 firms).9 In common with other studies (e.g., Vafeas and Theodorou, 1998) we exclude 8 financial companies because their accounting reports are different to those for other sectors. In addition, to avoid empirical problems associated with 8 One firm is excluded from the Brennan and McDermott (2004) sample because of difficulties in obtaining the relevant information. Readers should note that the MacCanna et al. (1999) findings are impacted by the fact that there is a far higher proportion of nonlisted firms in Ireland compared with other countries employing the Anglo-Saxon governance model. 9 We ignore firms quoted on the Exploration Securities Market since, in addition to being thinly traded, these 8 firms tend to have more limited information requirements and different types of accounting characteristics. On April 12 2005 the Irish Stock Exchange launched a new market, the Irish Enterprise Exchange, which replaced both the Exploration Securities Market and the Developing Companies Market. The ITEQ market was also discontinued in 2005.

392 Table 2

V. OÕConnell, N. Cramer Descriptive statistics.

Variable column no. BRDSIZE

NED%

OWNERSHIP

RET

ROA

FINANCIAL Q

All (N = 44) (1)

Large (N = 22) (2)

Small (N = 22) (3)

p-Value for difference (4)

Mean Std dev. Median

9.295 4.157 8

11.409 4.738 10.5

7.182 1.893 7

0.0004

Mean Std dev. Median

0.536 0.152 0.5

0.549 0.164 0.5

0.523 0.142 0.5

0.5693

Mean Std dev. Median

0.136 0.166 0.062

0.126 0.152 0.057

0.145 0.182 0.062

0.7115

Mean Std dev. Median

0.012 0.379 0.009

0.182 0.349 0.134

0.159 0.331 0.123

0.0018

Mean Std dev. Median

0.088 0.106 0.079

0.085 0.087 0.093

0.092 0.123 0.065

0.8356

Mean Std dev. Median

1.285 1.082 0.895

1.778 1.358 1.244

0.794 0.206 0.789

0.0017

0.002

0.7681

0.6221

0.0049

0.1456

0.0002

Notes: The sample consists of all Irish non-financial firms with the relevant data listed on the Irish Stock Exchange in 2001. Large (small) firms are those with a market value above (below) the average at the year-end. The p-value in column (4) are for t-(Wilcoxon-) tests of equality of means (medians) across the large and small firm groupings. Exact definitions of the variables are presented in Table 1.

thinly traded stocks (McKillop and Hutchinson, 1988) we also exclude 8 firms which had a total trading volume on the Irish Stock Exchange for the entire calendar year of 2001 of less than 0.1 million shares. In order to be included in the final sample each firm selected from the remaining 61 had to: (i) have its accounting information and stock market data for fiscal 2001 available on Datastream and (ii) have all of the required information about the board of directors available from either the annual financial report and/or Primark Global Access. The 17 firms which failed to meet these criteria were excluded from our study so our final sample consists of 44 firms broken down as follows: 37 out of a possible 49 Official List firms (76%), 5 out of a possible 8 ITEQ firms (63%) and 2 out of a possible 4 Developing Companies Market (DCM) firms (50%). Overall, although our final sample is small by international standards, the unique features of the Irish market discussed earlier mean that it is a potentially interesting research setting. From the perspective of our subsequent analysis, it is important to note that the regulatory regime for ITEQ firms is similar to those for Official List firms and 4 out of the 5 ITEQ companies in our sample were included on the Official List prior to moving to the technology market. However, the regulations for the DCM are slightly less stringent than those for the Official List and ITEQ markets and later in the study we test whether our empirical findings are impacted by the exclusion of the 2 DCM firms in our sample. However, this analysis reveals that the results with respect to the three hypotheses outlined earlier remain unaltered when these firms are excluded. As outlined earlier, one of the goals of our study is to explore the impact of firm size on the firm performance-board

size association. The mean (median) market capitalization for all firms is €755 (€172) million while that for large and small firms is €1,430 (€608) million and €79 (€69) million, respectively. In addition, t-(Wilcoxon) tests reveal that the difference between the mean (median) market capitalizations for both groups is statistically significant at the 1% level. Summary statistics for the sample are presented in Table 2. Table 2 reveals that the mean (median) board size is 9.29 (8) directors for the overall sample. Empirical research by Lipton and Lorsch (1992) suggests that a board composed of around 8 or 9 members is ideal from a monitoring perspective so the figures reported for the present sample would appear to be within the range of their suggested target.10 The mean board size for Irish firms is below the 11.67 reported by de Andres et al. (2005) for 10 OECD countries but is larger than the 8.07 reported by Vafeas and Theodorou (1998) for the UK. In addition, the board size of Irish companies appears to be generally smaller than that of US firms (e.g., Yermack (1996) shows that the median board size for his sample is 12). Another notable aspect of the statistics in Table 2 is the significant difference in the average size of boards for small and large firms (11.40 versus 9.29). Turning to the NED% variable, Table 2 reveals that the mean percentage of non-executive directors on the board 10 The summary data for the present sample is roughly comparable to that presented in Brennan and McDermott (2004) who report a mean board size and proportion of non-executive directors of 9.4% and 61%, respectively. However, in contrast to our study, the Brennan and McDermott (2004) sample includes financial stocks and firms listed on the Exploration Securities Market.

The relationship between firm performance and board characteristics in Ireland

393

Empirical findings

current study offers some support for the notion that a negative relation between firm performance and board size is also apparent in the Irish setting. Hypothesis 2 predicts that the relationship between firm performance and board size is significantly less negative for smaller firms. The results in columns 1 and 2 of Table 3 (i.e., when RET and FINANCIAL Q are the respective dependent variables) support this prediction since the BRDSIZE * D coefficient for these regressions is positive and significant at the 1% and 5% levels, respectively. However, when ROA is used as the dependent variable, BRDSIZE * D is negative although statistically insignificant. Taken as a whole, the results offer reasonably strong support for H2 and hence we conclude that the relationship between performance and board size is significantly less negative for smaller firms.12 Hypothesis 3 predicts a positive association between firm performance and the percentage of non-executive directors (NED%). The results in Table 3 suggest that while NED% is positive for all three regressions, the association is statistically significant (at the 10% level) when RET and ROA are used as the dependent variables (columns 1 and 3). Hence, in line with a priori theoretical predictions but in contrast to results for many other countries (e.g., Dalton et al., 1998) the findings in Table 3 provide some support for the assertion that firm performance is positively and significantly associated with the proportion of NEDs.13 Finally, two other aspects of the findings in Table 3 are noteworthy. First, the dummy for firm size categories (D) is statistically significant for two of the three regressions. Taken in conjunction with the results for BRDSIZE * D, this finding suggests that empirical work in the field needs to carefully account for firm size effects in order to avoid potential specification error.14 Second, it appears that the OWNERSHIP variable is statistically insignificant for all three performance measures. This outcome mirrors evidence in Daily et al. (2003) highlighting the absence of a relationship between corporate performance and insider or outsider equity ownership.

Core findings

The potential impact of endogeneity

The coefficients and their associated standard errors from estimates of Eq. (1) using the three alternate measures of firm performance (i.e., RET, FINANCIAL Q and ROA) are presented in columns 1–3, respectively, of Table 3.11 Hypothesis 1 predicts a negative relationship between firm performance and board size (BRDSIZE) and the results in Table 3 support this prediction since the b2 coefficient is negative in each of the three regression estimates. When FINANCIAL Q and ROA are the dependent variables (columns 2 and 3, respectively), the relationship between the performance metrics and board size is statistically significant at p < 0.1. Hence, in line with international research in the field (e.g., Yermack, 1996; de Andres et al., 2005) the

In any tests of the relationship between firm performance and board characteristics, it is critical to explore the

is 53.62%. This is considerably higher than the 39% reported by Vafeas and Theodorou (1998) for the UK; however, their sample is for 1994 and, as pointed out earlier, the proportion of NEDs in the UK has risen sharply in recent years (Pye, 2000) – for example, de Andres et al. (2005) reveal a mean proportion of outsiders of 50% for UK firms. Similar to the Irish experience, Yermack (1996) reports that his sample of US companies includes an average of 54% of outside directors. The statistics in Table 2 also show that there is no statistically significant difference in the NED% variable across company size categories (with means of 54.9% and 52.3% for large and small firms, respectively). The OWNERSHIP variable (with an overall mean of 13.6%) is also relatively similar across large and small firms. Because of data availability issues, we do not distinguish between independent and ÔgreyÕ NEDs (where the latter are defined as those non-executives with substantial business, family or other connections to the firm). Our decision to measure board independence with reference to the proportion of NEDs is similar to the approach adopted in many related studies (e.g., de Andres et al., 2005) and does not appear to be a major limitation since prior work (e.g., Vafeas and Theodorou, 1998) reports equivalent findings with respect to the association between firm performance and the proportion of NEDs regardless of whether NEDs are split into the aforementioned subgroupings. In any case, John and Senbet (1998) and Dalton et al. (1999) argue that the degree of a boardÕs independence is closely related to its composition. Turning to the performance measures, Table 2 shows that the average FINANCIAL Q and ROA values are 1.28% and 8.8%, respectively, while the mean RET is barely positive at 1.2% reflecting the downturn in the Irish stock market in the early part of the decade. Large firms significantly outperform their smaller counterparts on the RET and FINANCIAL Q measures while there is no significant difference in ROA across the two firm size categories.

11

Initial tests suggest that while heteroscedasticity is not a problem when RET is used as the dependent variables, this is not the case for ROA and FINANCIAL Q. Hence, the White (1980) approach is used to estimate the standard errors when ROA and FINANCIAL Q are utilised as the dependent variable. It should also be noted that multicollinearity is not a problem with respect to Eq. (1) since the mean variance inflation factor is less than 5.7.

12 When D = 0 (i.e., for large firms) we are effectively estimating the following equation: Performance measure = b0 + b2 BRDSIZEt + b4 NED%t + b5 OWNERSHIPt + et. When D = 1 (i.e., small firms) we estimate the full Eq. (1). Hence, from an econometric perspective, this means that b1 in Eq. (1) reflects the difference in the intercept between small and large firms while b3 in Eq. (1) reflects the difference in the BRDSIZE slope coefficient between small and large firms. 13 On this point Daily et al. (2003, p. 375) observe that ‘‘rather than focusing predominantly on directorsÕ willingness or ability to control executives, in future research scholars may yield more productive results by focusing on the assistance directors provide in bringing valued resources to the firm and in serving as a source of advice and counsel for CEOs’’. 14 In other (unreported) tests we also included the interaction between the size dummy and NED% as an additional independent variable in (1). However, for all three regressions this variable is statistically insignificant and its inclusion has no impact on the findings reported in Table 3.

394 Table 3

V. OÕConnell, N. Cramer The relationship between alternative measures of performance and board characteristics in Ireland. (1) RET OLS estimation

Intercept

Prediction ?

D

?

BRDSIZE



BRDSIZE*D

+

NED% OWNERSHIP N F-ratio R2 Adjusted R2

0.063 (0.27)

(2) FINANCIAL Q OLS estimation

(3) ROA OLS estimation

(4) ROA 2SLS estimation

1.935 (2.76)***

0.078 (1.40)

0.385 (1.48)

1.121 (3.18)***

2.149 (2.90)***

0.003 (0.03)

0.590 (1.25) *^

0.007 (0.43)

0.080 (2.10)**

0.007 (1.73)**

0.055 (1.53)*

0.106 (2.50)***

0.114 (1.77)**

0.002 (0.13)

0.053 (1.14) *^

+

0.570 (1.55)*

0.985 (0.87)

0.187 (1.43)*

0.610 (1.64)**

?

0.099 (0.31)

1.693 (1.54)

0.141 (1.41)

0.041 (0.21)

44 4.16*** 0.35 0.28

44 3.45** 0.31 0.22

44 1.44 0.16 0.05

44 0.79 NA NA

at 1%; **significant at 5%; *significant at 10%; *^significant at 12% for one-tailed p-values for variables with signs as predicted, two tailed otherwise. Notes: The sample consists of Irish non-financial firms with the relevant data listed on the Irish Stock Exchange in 2001. The regression analysis is carried out using three alternative dependent variables (RET, FINANCIAL Q and ROA). Standard t-statistics are reported when RET is the dependent variable while White adjusted t-statistics are reported in parenthesis when FINANCIAL Q and ROA are employed as the dependent variable (columns (2) and (3), respectively). Exact definitions of the variables are presented in Table 1. D is a dummy variable taking the value of 1 when a firm has a market value at the year-end below the median and 0 otherwise. All estimates are based on Ordinary Least Squares (OLS) except those in column (4) which are two-stage least squares estimates (2SLS). ***Significant

potential impact of endogeneity on the empirical findings. In particular, while performance may be a function of board characteristics, there is also the possibility that certain board characteristics (such as board size) are actually determined by performance. Hence, building on de Andres et al. (2005) we assume that board size is an endogenous variable which may be modelled as function of performance, the proportion of outsiders (NED%), firm size (measured by the natural log of market value (MV)) and the size dummy (D). We then test for endogeneity using the approach described in Wooldridge (2006, p. 532). First we regress BRDSIZE on the full set of exogenous variables (i.e., NED%, OWNERSHIP, SIZE and D). We then include the residual from this regression (RESID) as an additional independent variable in estimates of Eq. (1). If RESID is statistically significant then this is evidence that BRDSIZE is in fact an endogenous variable. When we carry out this test using each of the three alternate performance measures we find that RESID is statistically insignificant when RET and FINANCIAL Q are the dependent variables (with p values of 0.858 and 0.549, respectively). These findings suggest that endogeneity is not a significant issue when RET and FINANCIAL Q are the performance measures and so we rely on conventional Ordinary Least Squares estimation for these performance metrics. However, when ROA is the dependent variable, RESID is significant with a p-value of 0.056. Since our findings with respect to ROA may be impacted by endogeneity we repeat

our earlier analysis using a two-stage least square (2SLS) estimation procedure with an adjustment for the small sample size. These results are presented in column (4) of Table 3. For the 2SLS estimation, ROA is modelled as per Eq. (1) while BRDSIZE is expressed as a function of ROA, NED%, MARKETCAP and D. When we estimate the regression parameters using the 2SLS approach, we find that the results with respect to Eq. (1) are similar to those reported in column (3) of Table 3 since BRDSIZE is negative and significant while NED% is positive and significant.15 However, in contrast to the ordinary Least Squares (OLS) results, the BRDSIZE * D interaction term is now positive and significant at the 12% level. Finally, both D and OWNERSHIP are statistically insignificant for the estimates of Eq. (1) using 2SLS. The findings for the second equation (i.e., when BRDSIZE is the dependent variable) reveal that none of the explanatory variables is statistically significant implying that performance does not significantly impact board size.16 Overall, the results 15

The only difference is that BRDSIZE is now significant at 10% but not at 5% while NED% is now significant at 5% rather than at the 10% level. 16 The results for the second equation using the two-stage least squares procedure when BRDSIZE is the dependent variable are as follows: INTERCEPT [Coefficient = 17.502 (t-statistic = 1.13)], ROA [1.499 (0.09)], NED% [6.228 (1.30)], MARKETCAP [1.238 (1.60)]), and D [0.645 (0.29]. Hence, none of the explanatory variables are statistically significant in the second equation.

The relationship between firm performance and board characteristics in Ireland of the 2SLS estimation approach with respect to ROA reported in column (4) of Table 3 offer support for our three hypotheses.17

Sensitivity tests A number of separate sensitivity checks are carried out with respect to the empirical work. In particular, we extend the primary analysis by estimating the model outlined below18: Performance measure ¼ b0 þ b1 Dt þ b2 BRDSIZEt þ b3 BRDSIZEt  Dt þ b4 NED%t þ b5 OWNERSHIPt þ b6 DEBTt þ b7 SPLITt þ b8 NEDCHAIRt þ et

ð2Þ

First, a control for debt levels, (where ÔDEBTÕ is measured as the ratio of total debt to total assets) is included as an additional independent variable (e.g., Agarwal and Knoeber, 1996). The inclusion of this control has no impact on the reported results and DEBT is only significant when FINANCIAL Q is used as the dependent variable. Second, a dummy indicator (ÔSPLITÕ) taking the value of 1 if a director other than the CEO is the chairman and 0 otherwise is also included in Eq. (2) (e.g., Yermack, 1996). For our sample, the CEO is chairman for only 18% of the sample firms and the CEO indicator variable is not significant for any regression and its inclusion has no impact on the reported findings. Third, a dummy variable (ÔNEDCHAIRÕ) taking the value of 1 if a non-executive director is chairman and 0 otherwise is introduced into Eq. (2). Once again this indicator variable is insignificant in each of the regressions and its inclusion does not alter the reported findings. Overall, the core results in Table 3 are robust to the inclusion of the three additional variables in Eq. (2).

Analysis The empirical findings discussed above portray a remarkably consistent picture. In total, we undertake nine hypotheses tests (i.e., our study presents three hypotheses and utilises 17 Because of concerns about the small sample properties of three stage least squares (3SLS) estimator we do not rely on 3SLS in the current paper. Notwithstanding our additional work, an important caveat (Larcker and Rusticus, 2008) with respect to both 2SLS and 3SLS is that there are many circumstances in which OLS without any correction for endogeneity leads to more reliable statistical inference than these estimation approaches – even when endogeneity is present. Barnhart et al. (1994) also point out some of the critical drawbacks of alternative means of dealing with endogeneity in corporate governance research. 18 Multicollinearity problems preclude the joint inclusion of the three additional control variables. Following suggestions from an anonymous reviewer, in other tests we attempt to ascertain whether the smaller board size in Ireland has a moderating effect on our results with respect to NED%. To pursue this question, we include a new variable which is the interaction between the proportion of non-executives (NED%) and board size (BRDSIZE). However, including this new variable (NED% * BRDSIZE) in the regression analysis leads to extremely serious multicollinearity problems (the VIF rises to 51). Unfortunately, then, we are not in a position to further explore this issue.

395

three alternate measures of performance when testing these hypotheses) and our findings are of the predicted sign for all nine tests and are statistically significant in seven of the nine tests.19 Hence, our overall findings offer a clear and consistent story with respect to the relationship between firm performance and board characteristics in Ireland. As explained above, the results for two of the hypothesis tests are not statistically significant although the regression coefficients are of the predicted sign. Naturally the question arises as to why the results for these tests (i.e., Hypothesis 1 when RET is the performance measure and Hypothesis 3 when FINANCIAL Q is the performance measure) are not significant. In this regard it is worth noting that RET and FINANCIAL Q both reflect – albeit to different extents and in different ways – market-based performance and as de Andres et al. (2005, p. 206) point out, market measures ‘‘have been sometimes criticised since they can be affected by market ÔmoodsÕ, suffer from anticipation problems or deal with other issues such as market power in addition to firm performance’’. Perhaps the insignificant t-statistics for the aforementioned tests somehow reflect the weaknesses of market-based performance. However, in the absence of detailed additional work – which is beyond the scope of the present study – this is only one of a number of potential explanations. In any case, no performance measure is ÔperfectÕ since accounting metrics are impacted by other difficulties (such as potential managerial manipulation and a lack of timeliness). Kothair (2001) presents a comprehensive overview of prior research on the complex nature of the relationship between market- and accounting-based measures of performance.

Conclusions Core findings This study explores the relationship between firm performance and both board size and board composition for a sample of quoted Irish firms using three alternative performance measures: stock returns (RET), FINANCIAL Q and return on assets (ROA). The key statistically significant results are as follows. First, board size exhibits a significant negative association with firm performance (using ROA and FINANCIAL Q). Hence, the present work shows that the well established negative association between firm performance and board size is also apparent for the Irish market. Second, the relationship between board size and firm performance is significantly less negative for smaller firms (using all three alternative performance measures) and this result emphasizes the need for researchers to consider the potentially moderating impact of firm size on the company performance-board size association. Third, we find a positive and significant association between firm performance and 19 As discussed above, the 2SLS findings (which are more appropriate for ROA in light of the discussion above) offer support for all three hypotheses. However, as noted earlier, the p-value for the test of Hypothesis 2 for the 2SLS estimates is 0.12 which is only marginally insignificant at the ÔconventionalÕ 10% level and we refer to this as being significant at the 12% level. We believe that a pvalue of 0.12 is non-trivial in light of the small sample size (recall N = 44).

396 the percentage of non-executives on the board (using RET and ROA). Although this outcome is entirely consistent with predictions from agency theory, studies in other countries have not generally established a statistically significant link between these variables.20

Managerial implications The empirical findings reported above suggest that, as with many other countries, larger board sizes in Ireland are associated with weaker corporate performance. From this perspective, the Irish experience dovetails with that in other jurisdictions. However, a relatively unique aspect of the present study is that we explore whether firm size has a moderating impact on the overall negative association between firm performance and board size. Clearly, the results presented earlier do indicate that for smaller Irish firms, the potentially disadvantageous impact of larger boards is significantly reduced. However, it is difficult to comment on whether or not Ireland is unique in this regard since the impact of firm size on the firm performance-board size relationship has been somewhat ignored in prior research. One important exception is the work by de Andres et al. (2005) who use a similar approach to ours although they report no firm size-related differentials in the relationship between performance and board size.21 This general neglect of the potentially moderating impact of firm size on the performance-board size association in extant research has at least two important consequences. First, it means than an interesting and potentially important research question is ignored. Second, ignoring firm size can lead to econometric difficulties since the researcher is effectively constraining the relationship to be identical across size categories. This constraint may not always be appropriate and one of the strengths of this paper is that it highlights this point as an important consideration for future empirical research in the field. The findings with respect to our third hypothesis appear to be somewhat specific to the Irish setting. While agency theory and contemporary codes of corporate governance emphasize the key role of independent directors, prior research in many other countries has - perhaps surprisingly failed to establish a convincing link between firm performance and the proportion of outside directors on the board. However, while our finding of a significant positive association between performance and NED% is entirely consistent with theoretical expectations, it does represent a striking contrast with the results of equivalent studies for other 20 However, some marginal support for this conclusion is also apparent when we use ROA as the performance measure and 2SLS estimation. 21 Many empirical studies include firm size as a control variable but do not directly test the potential impact of firm size on the firm performance-board size association. In a similar fashion to our work, de Andres et al. (2005) test for size related differences by interacting a dummy variable for size categories with the board size variable. The type of dummy variable technique utilised in our study is also widely used in empirical research in a variety of other fields (e.g., Wooldridge, 2006). For example, Collins et al. (1987) employ a similar approach to assess the impact of firm size on the information content of security prices with respect to accounting earnings.

V. OÕConnell, N. Cramer markets. Hence, the question arises as to what characteristics of the Irish governance setting may be responsible for this contrasting outcome. One possible (albeit speculative) explanation is that, as discussed earlier, Ireland has significantly fewer interlocking directorates than other countries (Brennan and McDermott, 2004). From the perspective of the monitoring role of the board, a high level of interlocks may compromise director independence. Furthermore, a high degree of interlocks can also have critical implications from a strategic standpoint. For example, Ruigrok et al. (2006) report that boards in Switzerland are highly interlocked and they find that this phenomenon has a negative impact on the boardÕs contribution to strategic decision-making. As Ruigrok et al. (2006, p. 1219) point out: ‘‘The more board mandates an individual director has, the more limited the time and attention he or she can devote to a single company. Active involvement in strategic decision-making requires significant knowledge about the company and its industry. To build up such knowledge is time intensive’’. Hence, one possible reason for the contrast between our findings and those for other countries is that the low levels of board interlocks in Ireland actually help improve the quality of the contribution made by outside directors.22 A second possible – and far from mutually exclusive – explanation arises from the way in which Irish NEDs are selected. While Irish NEDs have few interlocks, they do tend to be drawn from a small homogenous pool (OÕHiggins, 2002). While this approach to selection is not optimal from the perspective of encouraging board diversity (Ruigrok et al., 2007), one advantage is that the chosen non-executives are likely to be highly experienced business people with the capacity to make an immediate positive contribution to board activities.

Avenues for future research and limitations Our study points to a number of potentially fruitful avenues for future research. First, we believe that future work using inter-temporal modelling (e.g., Davidson and Rowe, 2004) and/or a Granger causality estimation approach (e.g., Wooldridge, 2006) could build upon and extend the insights presented here. Second, we also believe that researchers in the field should endeavour to adopt a multidimensional approach to performance measurement when investigating the types of questions considered in our paper. As discussed earlier, different measures tend to capture different aspects of performance. A third potentially intriguing area for future work arises from those contrasts in board characteristics which are apparent across countries and which may be related to factors such as country size differentials. For example, in international terms, board size in Ireland is comparatively small; the median for Ireland (8) is less than the board size of all but one of the 10 countries covered in the de Andres et al. (2005) study – the exception being the 22

We are grateful to both anonymous reviewers for encouraging us to think more deeply about why the results for Ireland are different from those for other jurisdictions.

The relationship between firm performance and board characteristics in Ireland Netherlands (a median board size of 7) which is another relatively small country. Firm size is also comparatively lower in Ireland. Hence, our study suggests that the impact of country size – and firm size within individual countries – on corporate governance in general (and board characteristics in particular) is a hugely important area for future research. It is likely that these effects are inextricably linked to factors such as national culture, national legal systems, the degree of economic development and other related factors. While some important work on these themes has already begun to emerge (Doidge et al., 2007) this strand of research represent a critically important avenue for future investigation. Overall, comparative crosscountry work represents one of many potentially interesting and important opportunities for future corporate governance research (Brennan and Solomon, 2008). Fourth, while we have offered potential explanations for the observed positive association between firm performance and NED%, since these assertions are essentially speculative, future research work is clearly required. We hope that our findings highlight some of the intriguing possibilities for work on these issues in both the Irish and international settings. Fifth, the data for our study is from the Ôpre-HiggsÕ period. However, as discussed earlier, the Ôcomply or explainÕ approach outlined in Higgs has been the subject of much subsequent criticism. An interesting avenue for future research would be to explore the impact (if any) of the Higgs recommendations on the relationship between firm performance and board characteristics in Ireland. Finally, the major limitations of this study are as follows. First, the results are based on a one year sample period – hence, our findings may not generalise to other years. Second, the lack of publicly available information (highlighted by Brennan and McDermott, 2004) and the small number of quoted firms on the Irish market limits the sample size. Third, the small sample size means that we have little opportunity for meaningful industry or sector analysis. Fourth, the sole focus of our study is on firms quoted on the Irish stock market. Hence, it is not possible to accurately predict the impact on our findings of the fact that – when compared with many other countries – Irish firms are generally smaller and have fewer board members.

References Agarwal, A. and Knoeber, C. R. (1996) Firm performance and mechanisms to control agency problems between management and shareholders. Journal of Financial and Quantitative Analysis 31, 377–397. Ariff, A., Ibrahim, M. and Othman, R. (2007) Determinants of firm level governance. Malaysian evidence. Corporate Governance: An International Review 7(5), 562–573. Barclay, M., Gode, D. and Kothari, S. (2005) Matching delivered performance. Journal of Contemporary Accounting and Economics 1(1), 1–25. Barnhart, S. W., Marr, M. W. and Rosenstein, S. (1994) Firm performance and board composition: Some new evidence. Managerial and Decision Economics 15, 329–340. Basu, S. (1997) The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics 24, 3–37. Beekes, W., Pope, P. and Young, S. (2004) The link between earnings timeliness, earnings conservatism and board composi-

397

tion: Evidence from the UK. Corporate Governance: An International Review 12(1), 47–59. Belsley, D., Kuh, E. and Welsch, R. (1980) Regression diagnostics. John Wiley & Sons, Inc., New York, NY. Brennan, N. (2006) Boards of Directors and firm performance: Is there an expectations gap? Corporate Governance: An International Review 14(6), 577–593. Brennan, N. and McDermott, M. (2004) Alternative perspectives on independence of directors. Corporate Governance: An International Review 12(3), 325–336. Brennan, N. and Solomon, J. (2008) Corporate governance, accountability and mechanisms of accountability: An overview. Accounting, Auditing and Accountability Journal 21(7), 885–906. Byrd, J. W. and Hickman, K. A. (1992) Do outside directors monitor managers – evidence from tender offer bids. Journal of Financial Economics 32(2), 195–221. Chhaochharia, V. and Grinstein, Y. (2007) Corporate governance and firm value: The impact of the 2002 governance rules. The Journal of Finance 62(4), 1789–1825. Collins, D. and Kothari, S. (1989) An analysis of the intertemporal and cross-sectional determinants of earnings response coefficients. Journal of Accounting and Economics(11), 143–181. Collins, D., Kothari, S. and Rayburn, J. (1987) Firm size and the information content of prices with respect to earnings. Journal of Accounting and Economics 9, 111–138. Conyon, M. J. and Peck, S. I. (1998) Board size and corporate performance: Evidence from European countries. European Journal of Finance(4), 291–304. Cotter, J., Shivdasani, A. and Zenner, M. (1997) Do independent directors enhance target shareholder wealth during tender offers? Journal of Financial Economics 43, 195–218. Daily, C., Dalton, D. and Cannella, A. (2003) Corporate governance: Decades of dialogue and data. Academy of Management Review 28(3), 371–382. Dalton, D., Daily, C., Ellstrand, A. and Johnston, J. (1998) Metaanalytic reviews of board composition, leadership structure, and financial performance. Strategic Management Journal 19, 269–290. Dalton, D., Daily, C., Johnson, J. and Ellstrand, A. (1999) Number of directors and financial performance: A meta-analysis. Academy of Management Journal 42(6), 674–686. Davidson, W. and Rowe, W. (2004) Intertemporal endogeneity in board composition and financial performance. Corporate Ownership and Control 1(4), 49–60. de Andres, P., Azofra, V. and Lopez, F. (2005) Corporate boards in OECD countries: Size, composition, functioning and effectiveness. Corporate Governance: An International Review 13(2), 197–210. Denis, D. K. and McConnell, J. J. (2003) International corporate governance. Journal of Financial and Quantitative Analysis 38(1), 1–36. Doidge, C., Karolyi, A. and Stulz, R. (2007) Why do countries matter so much for corporate governance? Journal of Financial Economics 86, 1–39. Donnelly, R. and Kelly, P. (2005) Ownership and board structures in Irish plcs. European Management Journal 23(6), 730–740. Drobetz, W., Schillhofer, A. and Zimmerman, H. (2004) Corporate governance and expected stock returns: Evidence from Germany. European Financial Management 10(2), 267–293. Dulewicz, V. and Herbert, P. (2004) Does the composition and practice of boards of directors bear any relationship to the performance of their companies. Corporate Governance: An International Review 12(3), 263–280. Elloumi, F. and Gueyie, J. P. (2001) Financial distress and corporate governance. Corporate governance 1(1), 15–23.

398 Enriques, L. and Volpin, P. (2007) Corporate governance reforms in Continental Europe. Journal of Economic Perspectives 21(1), 117–140. Evans, C. R. and Dion, K. L. (1991) Group cohesion and performance: A meta-analysis. Small Group Research 22, 175–186. Faccio, M. and Lasfer, M. A. (1999) Managerial ownership, board structure and firm value: The UK evidence. Cass Business School Research Paper. Available at SSRN: http://ssrn.com/ abstract=179008. Fama, E. (1980) Agency problems and the theory of the firm. Journal of Political Economy 88, 288–307. Financial Reporting Council. (2006) The combined code on corporate governance, London. Financial Reporting Council. (2009) Review of the effectiveness of the combined code progress report and second consultation. Available from www.frc.org.uk. Finkelstein, S. and Hambrick, D. C. (1996) Strategic leadership: Top executives and their effects on organizations. West, Minneapolis/St.Paul. Freedman, D. (1999) From association to causation: Some remarks on the history of statistics. Statistical Science 14(3), 243–258. Goodstein, J., Gautam, K. and Boeker, W. (1994) The effect of board size and diversity on strategic change. Strategic Management Journal 15, 241–250. Gujarati, D. (2006) Essentials of econometrics. (third ed.). McGrawHill. Hampel, R. (1998) Committee on corporate governance: Final report. Gee Publishing, London. Healy, P. and Wahlen, J. (1999) A review of the earnings management literature and its implications for standard setting. Accounting Horizons 13(4), 365–383. Hermalin, B. E. and Weisbach, M. S. (1991) The effects of board composition and direct incentives on firm performance. Financial Management 20(4), 101–112. Higgs, D. (2003) Review of the effectiveness of non-executive directors. Department of Trade and Industry, London. Himmelberg, C., Hubbard, R. and Palia, D. (1999) Understanding the determinants of managerial ownership and the link between ownership and performance. Journal of Financial Economics 53, 353–384. Holmstrom, B. and Kaplan, S. (2003) The state of US corporate governance: WhatÕs right and whatÕs wrong. Journal of Applied Corporate Finance 15, 8–20. Jensen, M. (1993) The modern industrial revolution, exit and the failure of internal control systems. Journal of Finance 48, 831–880. Jensen, M. and Meckling, W. (1976) Theory of the firm: Managerial behaviour, agency costs and ownership structure. Journal of Financial Economics 3, 305–350. John, K. and Senbet, L. (1998) Corporate governance and board effectiveness. Journal of Banking and Finance 22, 371–403. Kiel, G. C. and Nicholson, G. J. (2003) Board composition and corporate performance: How the Australian experience informs contrasting theories of corporate governance. Corporate Governance: An International Review 11(3), 189–205. Klapper, L. and Love, I. (2003) Corporate governance, investor protection and performance in emerging markets. Journal of Corporate Finance 10(5), 703–728. Kothair, S. P. (2001) Capital markets research in accounting. Journal of Accounting and Economics 31(1–3), 105–231. Klein, A. (1998) Firm performance and board committee structure. Journal of Law and Economics 41(1), 275–303. Larcker, D. and Rusticus, T. (2008) On the use of instrumental variables in accounting research. Available at SSRN: http:// ssrn.com/abstract=694824. Lewellen, W. G. and Badrinath, S. G. (1997) On the measurement of TobinÕs Q. Journal of Financial Economics 44, 77–122.

V. OÕConnell, N. Cramer Leone, A., Wu, J. and Zimmerman, J. (2006) Asymmetric sensitivity of CEO cash compensation to stock returns. Journal of Accounting and Economics 42, 167–192. Lipton, M. and Lorsch, J. (1992) A modest proposal for improved corporate governance. Business Lawyer 48, 58–77. Listokin, Y. (2008) Interpreting empirical estimates of the effect of corporate governance. American Law and Economics Review 10(1), 90–109. London Stock Exchange (1998) Combined code: Principles of good governance and code of best practice. Gee Publishing, London. MacCanna, L., Brennan, N. and OÕHiggins, E. (1999) National networks of corporate power: An Irish perspective. Journal of Management and Governance 2, 355–377. McKillop, D. and Hutchinson, R. (1988) The small firm effect, seasonality and risk in Irish equities. Irish Business and Administrative Research 9, 21–29. OÕConnell, V. (1995) The information content of security prices on the Irish stock exchange. The British Accounting Review 27(4), 311–324. OÕConnell, V. (2006) The impact of accounting conservatism on the compensation relevance of UK earnings. European Accounting Review 15(4), 627–649. OÕHiggins, E. (2002) Non-executive directors on boards in Ireland: Co-option, characteristics and contributions. Corporate Governance: An International Review 10(1), 19–28. Pye, A. (2000) Changing scenes in, from and outside the boardroom: UK corporate governance in practice from 1989 to 1999. Corporate Governance: An International Review 8(4), 335–346. Rosenstein, S. and Wyatt, J. C. (1990) Outside directors, board effectiveness and shareholder wealth. Journal of Financial Economics 44, 229–250. Ruigrok, W., Peck, S. and Keller, P. (2006) Board characteristics and involvement in strategic decision making; evidence from Swiss companies. Journal of Management Studies 43(5), 1201–1226. Ruigrok, W., Peck, S. and Tacheva, S. (2007) Nationality and gender diversity on Swiss corporate boards. Corporate Governance: An International Review 14(4), 546–577. Ryan, P. (2005) The market impact of directorsÕ trades: Relationship to various measures of a firmÕs information environment. British Accounting Review 37, 319–337. Shleifer, A. and Vishny, R. W. (1997) A survey of corporate governance. Journal of Finance 52(2), 737–783. Standard and Poors. (2008) Monthly report: World by numbers, June. Available at www.standardandpoors.com. Stiles, P. and Taylor, B. (2001) Boards at work – how directors view their roles and responsibilities. Oxford University Press, Oxford. Vafeas, N. and Theodorou, E. (1998) The relationship between board structure and firm performance in the UK. British Accounting Review 30, 383–407. Watts, R. and Zimmerman, J. (1986) Positive accounting theory. Prentice-Hall. Watts, R. L. (2003) Conservatism in accounting part I: Explanations and implications. Accounting Horizons 17(3), 207–221. White, H. (1980) A heteroscedasticity-consistent covariance matrix estimator and a direct test for heteroscedasticity. Econometrica 48, 817–838. Wooldridge, J. (2006) Introductory econometrics: A modern approach. (3rd ed.). Thompson South-Western. Yawson, A. (2006) Evaluating the characteristics of corporate boards associated with layoff decisions. Corporate Governance: An International Review 14(2), 75–84. Yermack, D. (1996) Higher market valuations of companies with a small board of directors. Journal of Financial Economics 40, 185–211.

The relationship between firm performance and board characteristics in Ireland Zahra, S. and Pearce, A. (1989) Boards of directors and corporate financial performance: A review and integrative model. Journal of Management 15, 291–334. VINCENT OÕCONNELL is Senior Lecturer, Department of Accountancy, UCD School of Business. Previously, he was Associate Professor – Accounting at Korea University Business School. He received his Ph.D. from the London School of Economics & Political Science and holds B.Comm and MBS degrees from University College Cork. His current research interests include the impact of accounting information on stock prices and the interaction between marketing and accounting measures of performance. He work has been published in Accounting Horizons, the European Accounting Review, the International Journal of Research in Marketing, the British Accounting Review and the International Journal of Managerial Finance.

399 NICOLE CRAMER is Financial Adviser, Group Finance, with WestLB AG, Du ¨sseldorf, Germany since 2002. She graduated from the University of Applied Sciences, Trier, Germany in Business Management. She completed the MBS Corporate Finance and Accounting programme at University College Cork, Ireland in 2002.