4.5.3 Studies from 2005 and later

Andres et al. (2005) conducted an analysis on a dataset of companies from North America and Europe.1 They conjectured the hypothesis that board independence2 has a positive effect on the value of the company, measured in this study by the market-to-book ratio. The percentages of NEDs and executive directors are not reported for the Netherlands, as a dual board structure is obligatory by law. However, the report does mention a value of the percentage of NEDs for Germany, which has a comparable situation to the Netherlands. The article does not give any explanation for this difference. The results do not show robust relationships between the percentage of NEDs and the market-to-book ratio. Results derived from an ordinary least squares regression show a negative coefficient, whereas a three-stage least squares regression found positive results for the impact of board independence on market value (Andres et al. 2005: 203-205). It must be noted in this respect that the roles of NEDs are not equal for all the countries in the analysis, which might be a reason for the lack of results.

Fich and Shivdasani (2006) researched the effectiveness of independent directors, but made a distinction between busy directors and directors who are able to spend sufficient time on their board position. Directors with three or more directorships were considered to be ‘busy’. Only positions in publicly traded companies were considered; all other directorships – such as in nonprofit organisations – were excluded in this study. Besides the percentage of busy directors on a board, the variable ‘busy board’ was introduced. If more than half of the outside directors were considered to be busy, the entire board received the label ‘busy’. The results of a sample of 508 United States companies over the period 1989 to 1995 (3,366 observations) show that board independence3 has a positive, but not significant, influence on performance, measured by the market-to-book value, ROA and return on sales (Fich and Shivdasani 2006: 700-706). But these same regression analyses show that the percentage of busy outside directors and the dichotomous variable busy board have a significant negative impact on all these performance factors. This means that independent directors may have a positive influence on performance. However, if they are not able to spend sufficient time on their directorship the performance of the company might decrease.

Gani and Jermias (2006) distinguished two types of strategies for a company: a strategy of pursuing cost efficiency and a strategy of innovation. For companies with the former strategy the authors hypothesised that board independence4 has more positive effects than for companies with the latter strategy (Gani and Jermias 2006: 297-299). When pursuing cost efficiency, board independence is expected to be beneficial, because independent directors limit the opportunities of management to engage in opportunistic behaviour, which might increase costs. When pursuing a strategy of innovation, creativity is required and strict monitoring by independent directors might hinder management in taking decisions that are important for the innovative processes within the company. The hypothesis was tested over the period 1997 until 2001 on a sample of 109 S&P 500 companies and 436 company year observations. The decision whether a company has a strategy of cost efficiency or innovation was based on a principal components analysis, which takes several characteristics into account such as research and development intensity. The results in several regression analyses show a positive influence of board independence on performance, measured by ROE and return on investment. They also confirm the hypothesis that this particular relationship is stronger for companies with a cost efficiency strategy than for companies with an innovation strategy (Gani and Jermias 2006: 306-308).

Chan and Li (2008) related the percentage of expert independent directors in the total board and audit committees to performance in a sample of 200 publicly traded companies of the Fortune 500 over the year 2000. Besides their position as NED on the board concerned, expert independent directors hold a top management position in another publicly traded company. This study used independence as a dichotomous variable; the board or audit committee is either independent or not. An audit committee is considered to be independent if more than 50 per cent of the members is an expert independent director. For the determination of board independence they used two cut-off points: boards with an expert independent directors percentage ranging between 35 and 50, and boards with an expert independent directors percentage of more than 50. The results show that the 50 per cent cut-off point for board independence yields a significant, positive relationship between board independence and performance measured by Tobin’s Q (Chan and Li 2008: 22-23). The 35 per cent cut-off point for the variable for board independence results in an insignificant negative relationship. The result that a 50 per cent cut-off point yields significant results and 35 per cent does not can be explained by the fact that the expert independent directors need control of the board in order to improve company value. Furthermore, if the percentage of expert independent directors on the audit committee is more than 50, this appears to have a significant positive influence on Tobin’s Q.

In their analysis of board independence,5 Coles et al. (2008) made a distinction with respect to the kind of company: simple and complex companies. Their sample consists of 8,165 United States company year observations over the period from 1992 to 2001. By means of a factor analysis, based on the number of business segments, the size of the company and leverage, the authors determined whether a company is either simple or complex. High complex companies need much advising, whereas simple companies are less in need of advice. The research found that board independence has a negative impact on the performance of simple companies, but that this relationship is significantly positive for complex companies (Coles et al. 2008: 343-344). Highly complex companies, which are expected to need more advisory resources than simple companies, benefit from the presence of independent NEDs.

Henry (2008) investigated whether the adoption of corporate governance best practices – prior to the introduction of the Australian corporate governance code in 2003 – had positive effects on the value of Australian companies in the period from 1992 to 2002. By using this particular timeframe, Henry was able to investigate whether companies that voluntarily comply – although they are not obliged to do so – with corporate governance best practices have a better performance than companies that do not comply. A sample of 116 of the 300 largest – measured by market capitalisation – companies from Australia listed in June 1996, resulted in 1,127 company year observations over an eleven-year period. The Pearson correlation of board independence6 shows a negative relationship with before-tax ROA (Henry 2008: 928). In a fixed effects regression analysis, board independence appears to have a positive influence on company value, measured by Tobin’s Q and Tobin’s Q adjusted for industry means (Henry 2008: 927-930). The relationship with Tobin’s Q is significant; the industry means adjusted Tobin’s Q does not have a significant relationship. The results show limited impact of self-appliance of corporate governance attributes on the value of the company (Henry 2008: 938).

Adams and Ferreira (2009) researched the impact of women directors on governance and performance; additionally they focused on the influence of other governance variables on performance. Their dataset consists of 8,253 company year observations over the period 1998 until 2003 of 1,939 companies from the S&P 500, S&P MidCap and the S&P SmallCap indices. The results for the relationship between board independence7 and performance, measured by ROA and Tobin’s Q, do not show any consistent results (Adams and Ferreira 2009: 304-308). The authors used ordinary least squares, fixed effects regressions, instrumental variables, and industry dummies; the results for the coefficient of board independence range from significantly negative and negative to positive and significantly positive. Therefore, a conclusion about the impact of board independence on performance cannot be given based on this study.

Brick and Chidambaran’s (2010) main aim was to investigate the relationship between, on the one hand, board activity and committee structures and, on the other hand, performance. Board independence8 was used as a control variable in these regressions. In a dataset of 5,228 American company year observations from the period 1999 until 2005, they did not find any consistent relationship between board independence and financial performance, either measured by industry-adjusted ROA or Tobin’s Q (Brick and Chidambaran 2010: 542-547). The results vary from significantly negative to significantly positive, but no real conclusion can be drawn from this article.

Lückerath-Rovers and Smits (2010) conducted research on the relationship between board independence and performance in two countries: the Netherlands and United Kingdom. They split their dataset into two parts and calculated the correlations and executed the regressions separately for both countries. This resulted in data of 121 companies from the United Kingdom and 81 companies from the Netherlands over 2006. They reported mixed results for the relationship between board independence9 and performance, measured by ROA, return on investment, ROE and Tobin’s Q (Lückerath-Rovers and Smits 2010: 157-160). The correlation between board independence and ROE is significantly negative in the United Kingdom; the ordinary least squares regression confirms this negative relationship by reporting an insignificant negative coefficient. The relationship between board independence and ROE is significantly positive in the Netherlands, for the Pearson correlation as well as the least squares regression (Lückerath-Rovers and Smits 2010: 159-160). Regression analyses are only conducted for ROE; the relationship with the other measures of performance is only reported for the correlations. These correlations show small insignificant relationships between board independence and performance.