Does Combining the CEO and Chair Roles Cause Poor Firm Performance?

Harley Ryan is Associate Professor of Finance at Georgia State University. This post is based on a paper authored by Professor Ryan; Narayanan Jayaraman, Professor of Finance at Georgia Institute of Technology; and Vikram Nanda, Professor of Finance at the University of Texas at Dallas. Related research from the Program on Corporate Governance includes Learning and the Disappearing Association between Governance and Returns by Lucian Bebchuk, Alma Cohen, and Charles C. Y. Wang (discussed on the Forum here) and The Case for Increasing Shareholder Power by Lucian Bebchuk.

Considerable disagreement exists on the merits of CEO-Chair duality. In recent years, there has been growing regulatory and investor pressure to split the titles of CEO and Chairman of the Board. In fact, there is a significant trend towards separation of the two titles. However, the empirical evidence in the literature is inconclusive on the impact of separating these roles. We argue that the inconclusive evidence arises from endogenous self-selection that complicates empirical identification strategies and the ability to recognize the correct counterfactual firms.

In our paper, Does Combining the CEO and Chair Roles Cause Poor Firm Performance?, which was recently made publicly available on SSRN, we propose a learning model of CEO-chair duality and implement an identification strategy to address sample selection issues. Our model and identification is based on “passing the baton” (PTB) firms that award the chair position after a probationary period during which the board of directors learns about the ability of the CEO. In the model, the board optimally awards the additional position of board chair if the CEO demonstrates sufficient talent. The increase in CEO power improves the retention of high-quality CEOs by mitigating concerns about the board reneging on compensation contracts. The model delivers several implications that we test in our empirical analysis.

Using a very large sample of over 22,000 firm-year observations for the period 1995-2010, we explore the determinants and consequences of the combining the two roles. Firms that always combine the two roles, always separate the roles, or award the additional title following a period of evaluation exhibit significantly different firm characteristics, which suggest self-selection. We find that PTB firms are more likely to be from industries that are less homogenous. This is consistent with a learning rationale underlying PTB strategies: CEO performance is harder to benchmark in such industries and reneging on contracts may be of greater concern to CEOs. We also find that firms with more business segments are more likely to combine the two roles. These findings suggest that more complex organizations are better served by combining the roles of the CEO and the Chairman.

Overall, CEOs that receive the additional title of board chair outperform their industry benchmark before receiving both titles. In firms that combine the roles after observing the CEO’s performance under a separate board chair, the combination is positively related to both firm and industry performance in the two years prior to the combination. As predicted by our model, a naïve analysis of the post-chair appointment performance, one that fails to control for selection issues and mean reversion in performance data, indicates a significant drop in firm performance relative to the pre-chair period. However, in a matched sample of firms where the matching criteria includes the pre-appointment performance and firm attributes that predict a high propensity for using a PTB succession strategy, there is no evidence of post-appointment underperformance. These results suggest that the pass-the-baton succession process appears to be an equilibrium mechanism in which some firms optimally use the PTB structure to learn about the CEO and then award the additional title of board chair to increase the odds of retaining talented CEOS. Thus, the evidence is broadly consistent with the learning hypothesis that the additional title is awarded by the board after evaluating the ability of the CEO.

Our model suggests that, ceteris paribus, talented CEOs in a weaker bargaining position relative to the board will tend to be promoted to chair more quickly. The reason is that more vulnerable CEOs are more likely to pursue outside opportunities. Supportive of the prediction, we find that when the board is more independent, is not coopted and the CEO is externally sourced—the promotion to chair occurs more quickly. These findings are also counter to the notion that agency considerations and influence are central to the CEO being appointed chairman. We also show that stockholders react positively to combinations that occur early in the CEO’s tenure, which suggests that early promotions reveal directors’ private information about the quality of the CEO to the market. This is inconsistent with alternative explanations such as an incentive rationale for PTB or agency problem, since both of these alternatives would suggest a negative market reaction to such promotions.

A major implication of our analysis for researchers is that one should consider learning mechanisms and retention objectives when evaluating various board structures. Structures that are seemingly incompatible with effective monitoring may in fact be optimal when one considers the impact of learning on retention. For governance activists and policy makers, the implications of our analysis are straightforward: the results call into question the prevailing wisdom that suggests that shareholders will always be better served by separating the two roles. Thus, those who seek to reform governance should be cautious in proposing to unambiguously separate the roles of CEO and board chair. Forcing separation by fiat is likely not an ideal policy. Overall our evidence suggests that having one type of executive and board leadership structure is not optimal for all firms.

The full paper is available for download here.

Both comments and trackbacks are currently closed.