Does Combining The CEO And Chair Roles Cause Poor Firm Performance?

Narayanan Jayaraman
Georgia Institute of Technology – Scheller College of Business

Vikram K. Nanda
University of Texas at Dallas – School of Management – Department of Finance & Managerial Economics

Harley E. Ryan Jr.
Georgia State University – Department of Finance

September 18, 2015

Georgia Tech Scheller College of Business Research Paper No. 2015-11

Abstract:

We use over 22,000 firm-year observations from 1995-2010 to investigate whether combining roles of CEO and board-chair causes poor performance. Our research design allows us to reconcile disagreement in the literature about whether CEO-chair duality impacts shareholder value. CEOs are awarded the additional title of board-chair following superior firm performance. A naïve analysis indicates a drop in firm performance following CEO promotion to chair. However, a research design that controls for the propensity to combine roles and performance mean-reversion reveals no post-appointment underperformance. Consistent with a learning explanation, investors react positively to combining both roles early in CEO’s tenure, but exhibit no reaction to combinations later in CEO’s tenure. Increases in post-combination compensation are unrelated to proxies for managerial power. Overall, there is no evidence that combining the CEO-chair positions hurts shareholder interests.

Does Combining The CEO And Chair Roles Cause Poor Firm Performance? – Introduction

In the aftermath of the governance failures of the early 20003 and the financial crisis of 2008, governance activists and, policy makers have increasingly called for separating the roles of CEO and chairman of the board. They base their demands on the premise that combining the two roles exacerbates agency problems since there is no arms-length monitoring of the CEO by an independent chairman. Empirical evidence is inconclusive, however, and clouded by a variety of endogeneity concerns. Suggestive of self-selection, some firms persist in combining the roles over time, while others never combine the roles or do so selectively. Motivated, by the increasing focus of governance reformers on separating these roles and a lack of clear evidence of causality in the literature, we use a large sample of over 22.000 firm-year observations for the period l995~2010 to investigate whether awarding the CEO the additional role of board chair causes poor performance. Our study provides little if any evidence to support the premise that combining the roles systematically causes poor performance. Moreover, our analysis explains why a research design that fails to control for selection issues can easily lead to the opposite – and biased conclusion that combining the roles destroys firm value.

An initial examination of the data reveals that many firms, at least over our l6-year time period, never combine both roles or always combine both roles. These two groups of dichotomous firms have strikingly different firm characteristics, which. suggest the presence of a, selection bias that make it difficult to attribute causality in cross-sectional regressions. Thus, we focus on a third group of firms that initially separate the roles of CEO and chair, and combine them only after a probationary period in which the board of directors can observe the new CEO’S actions and the firm’s performance. Brickley, Jairell, and Coles (1997) and, Vancil (1987) refer to this practice as “passing the baton” (PTB). We hypothesize that the PTB process can serve as a way to learn about the ability of the CEO before conferring additional power as chairman of the board. It can also serve as an incentive device that motivates the CEO to strive to receive the award. We label these as the learning and incentive hypotheses, respectively.

CEO, Chair Roles, Firm Performance

As the first step in our analysis, we examine differences between firms that always combine both roles, pass the baton, or always separate the roles. Of particular note, we find that PTB firms are more likely to be present in industries that are less homogenous than {inns that always combine or never combine the two roles. In less homogenous industries, CEO performance is difficult to benchmark to industry peers (Parrino (1997)),  suggesting that there may be a greater benefit to learning about CEO ability in these industries, before awarding the chair. We also use a hazard model to analyze the determinants of awarding both titles as opposed to keeping them separate. Supporting the premise that firms award both titles after a, probationary period in which the CEO proves his ability. we find that CEOs that exhibit superior industry-adjusted performance receive the chair title more quickly. However, we also find that good industry performance hastens the award of both titles, which suggests that, firms combine the two roles to retain CEOS when industry conditions create outside employment opportunities.2 Ceteris paribus, older firms take longer to award the title of Chair, while firms with multiple segments combine both titles more quickly. The latter result suggests that more complex organizations may be better served by combining the roles of the CEO and the chair, which is consistent with, the conclusions of other studies (Faleye (2007), Dey, Engel, and Lin {2009), and Palmon and Wald (2002)).

To study the consequences of combining the two roles on firm performance, we estimate CEO-firm pair fixed effect regressions in which the effect of the decision to combine the roles is captured by dummy variables for the year of the combination and the subsequent years. Our results indicate positive abnormal returns prior to the award. A naive 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.

CEO, Chair Roles, Firm Performance

See full PDF below.