Does Board Size Matter?

Dirk Jenter is a Professor of Finance at the London School of Economics and Political Science, Thomas Schmid is an Associate Professor of Finance at Hong Kong University, and Daniel Urban is an Assistant Professor of Finance at Erasmus University Rotterdam. This post is based on their recent paper.

Our study (Does Board Size Matter?) exploits a minimum board size requirement in Germany to show that excessively large boards of directors reduce firm performance and value. Boards play a crucial role in corporate governance, and regulators in many countries have tried to improve their effectiveness by discouraging large boards. The academic literature, however, provides little causal evidence on the effects of board size. Empirical studies, going back to at least Yermack (1996), have mostly found negative correlations between board size and firm performance. However, because large boards are a choice, and because this choice is likely to be correlated with other drivers of performance, these correlations are difficult to interpret.

German corporations have a two-tier board, with a management board that runs the firm and a supervisory board that hires, advises, and monitors the management board. Since 1976, the supervisory board’s legally mandated minimum size depends on the size of the firm – it is 12 directors for firms with 2,000 to 10,000 domestic employees (DE), 16 directors for 10,000 to 20,000 DE, and 20 directors above 20,000 DE. These requirements are binding for many firms: most firms just below 10,000 DE have exactly 12 directors, and most firms just above have exactly 16. This suggests that many firms just above this threshold would have chosen a smaller board if permitted, which allows us to test whether forcing them to adopt a larger one is harmful.

We find robust evidence that forcing firms to have large boards is detrimental. Our main analysis uses a panel of German firms from 1987 to 2016 and compares firms just below 10,000 DE to firms just above. Performance declines at the threshold, with a drop in return on assets (ROA) of 2-3 percentage points and a decline in Tobin’s Q (the ratio of the market value to the book value of assets) of around 0.20. Accounting for the higher probability of firms above the threshold having a large board, large boards are associated with a (noisily estimated) 5-6 percentage point drop in ROA.

Even though this setting is useful and unique, it does not create an ideal experiment. The number of DE is at least in part under the control of management and, to avoid a larger board, some firms might strategically choose to remain below 10,000 DE. Even though we find no evidence of such behavior, we cannot rule it out. We therefore analyze a second setting in which strategic behavior is less likely.

This second setting is the introduction of the board size requirement in 1976. We compare changes in firm performance and value from before to after the law’s introduction of treated (>10,000 employees) to control firms (≤10,000). We minimize concerns about firms strategically choosing to remain below 10,000 employees by classifying firms shortly after the threshold was announced. Depending on the length of the measurement period, the law’s introduction reduces treated firms’ ROA by 1.3-2.0 percentage points and their Tobin’s Q by 0.04-0.10. Thus, despite different time periods and research designs, the two analyses show similar reductions in operating performance. The smaller effect on Tobin’s Q might be explained by stock prices anticipating the new law’s effect.

We explore several mechanisms that might explain why large boards underperform. Our results show worse performance to be a steady-state feature of large boards, not a transition effect, and large boards to be associated with lower profit margins (consistent with worse cost control) and lower acquisition announcement returns (consistent with worse M&A deals). We also find suggestive but insignificant evidence that directors added after crossing 10,000 DE are busier, less experienced, and less likely to have a doctorate. Notably, there is no evidence that large boards pay more for executives or employees, choose lower pay-for-performance sensitivities, or reduce CEO turnover. However, many of these results are noisily estimated, so more work is needed to understand why exactly large boards are less effective.

Our results support the prediction, made intuitively by Lipton and Lorsch (1992) and Jensen (1993) and more rigorously by the literature on optimal committee size (Sah and Stiglitz 1988, Persico 2004, Kakhbod et al. 2022), that effectiveness declines as boards grow too large. This might be because of frictions in group decision making, such as free-riding and coordination problems, or because the additional directors are of lower quality. Independently of the exact mechanism, our findings are a warning that ill-designed board regulations can be costly.

Both comments and trackbacks are currently closed.