Distracted Directors

The following post comes to us from Antonio Falato, Economist at Federal Reserve Board; Dalida Kadyrzhanova of the Department of Finance at the University of Maryland; and Ugur Lel of the Department of Finance at Virginia Tech.

In our paper, Distracted Directors: Does Board Busyness Hurt Shareholder Value?, which was recently accepted for publication in the Journal of Financial Economics, we examine the impact of independent director busyness on firm value in a setting that addresses a key challenge that the board of directors is an endogenously determined institution. A large number of publicly-traded firms in the U.S. have recently limited the number of multiple directorships held by their board members. For example, a recent survey shows that 74 percent of S&P 500 firms impose restrictions on the number of corporate directorships held by their independent directors, up from 27 percent in 2006, and the Institutional Shareholder Services recommends restrictions on the number of multiple directorships. Although such shareholder initiatives are consistent with standard theoretical considerations (e.g., Holmstrom and Milgrom, 1992), the empirical evidence on whether director busyness has any effect on the firm is thus far mixed. While several studies find that busy directors are associated with lower firm valuations and less effective monitoring (e.g., Fich and Shivdasani, 2006; Core, Holthausen and Larcker, 1999) others either do not, or provide mixed evidence (e.g., Ferris, Jagannathan and Pritchard, 2003; Field, Lowry, and Mkrtchyan, 2013).

This mixed evidence in the literature is perhaps not surprising, as Adams, Hermalin, and Weisbach (2010) point out that it is challenging to disentangle busyness costs from potentially offsetting selection effects of holding multiple directorships: busyness costs arise since individuals who hold several outside directorships will likely have relatively less time available to devote effort to each of them, while selection effects arise because more talented or reputable directors are more likely to hold multiple directorships and, hence, be considered “busy.” In this paper, we overcome this selection issue by examining the shareholder wealth effects of an exogenous increase in the demand for outside directors’ time while holding their talent constant. By doing so, we hope to provide endogeneity-free evidence that independent director busyness matters for firm value.

The main identification of the paper accounts for the selection concerns by exploiting a quasi-experiment to generate plausibly exogenous variation in independent directors’ workload. The ideal natural experiment would be to generate exogenous variation in the director’s workload by randomly assigning similar types of directors into different workloads. Our empirical research design is geared toward generating this random assignment. We consider directors that hold an outside board seat on at least two firms, firm A and firm B, and are subject to shocks originating from the death of either the CEO or a colleague on the board of, say, firm A. We offer evidence validating the notion that these shocks lead to an increase in board committee workload for some of firm A’s independent directors, which takes away time and resources from their board activities at interlocked firms (firm B). We label this effect as the “attention shock hypothesis.” For example, suppose that an independent director on the board of Apple Inc. passes away. Our sample would include all the other firms at which Apple’s non-deceased independent directors hold an outside board seat—i.e., all director-interlocked firms.

Our experiment constructs two groups of director-interlocked firms: a group of firms whose independent directors’ committee workload increased—which is our “treatment group,” and a group of firms whose independent directors’ workload did not increase—our “control group.” These two groups comprise a sample where we claim that the assignment of a firm into high versus low director workloads is random, in the sense that it is due to reasons that are plausibly unrelated to the interlocking directors’ type or productivity. Under this identifying assumption, we can difference out any selection bias by comparing the changes in firm value around the increase in directors’ committee workload (the death event) for firms in the treatment group with those in the control group.

For example, in the case of director deaths, the treatment group comprises director-interlocked firms that also have a committee interlock with the deceased director, and our testable hypothesis is that the interlocking directors of firms in this group will have to take over some of their deceased colleague’s committee responsibilities, which will take away time and resources from their board activities at the interlocked firms. Thus, for director-interlocked firms in the treatment group, we should see a negative stock market reaction to attention shocks, which is what our empirical tests focus on analyzing. The identification of our estimate comes from the control group of director-interlocked firms that do not have a committee interlock, which serves as a counterfactual for how the committee-interlocked firms would have performed had there not been a death.

In order to test the attention shock hypothesis, we hand-collect information on the deaths of directors and CEOs over the 1988 to 2007 period. We are able to find 633 independent director deaths, which result in a sample of 2,551 director-interlocked firms. 1,084 of these observations are due to sudden deaths and 843 are from directors who are interlocked through the same committee—i.e., the treatment group. An analysis of the likelihood of replacement of the deceased director by the next board meeting and within one or two years after it indicates that director deaths are material events and their effect on board committees is long lasting, which validates our “attention shock” interpretation. In the second part of our analysis, we use 189 CEO deaths, yielding a sample of 592 firms that have a director-interlocked relation with the firm that loses its CEO, out of which 338 are due to sudden deaths and 323 are in the treatment group.

Our tests employ a specification, which is akin to the difference-in-differences (DID) methodology. We derive estimates of the treatment effect of a plausibly exogenous increase in directors’ committee workload on firm value by comparing the change in director-interlocked firm value within the treatment and control groups before and after the death event dates (first-difference) and then taking the difference across the two groups (second-difference). The latter is the key step needed to difference out potential biases arising from obvious differences between busy and non-busy directors, such as talent, reputation, and access to a wider network of corporate relations. In our large sample of director-interlocked firms, we document robust evidence that relative to the control group, firms in the treatment group experience a significant negative stock market reaction to director attention shocks. On average, for interlocked firms in the treatment group the market reaction is as substantial as -1.55% and statistically significant at the one percent level when the death event is sudden. By contrast, in our control group of observations that are subject to the same sudden shocks but whose interlocked relation does not involve serving in the same committee, there is no statistically significant market reaction (0.19%, t = 0.354). The difference in the market reaction between the treatment and control groups is -1.74%, which is statistically significant at the one percent level and robust to a battery of tests that address potential outliers and sample composition issues.

Further, we investigate the drivers of the treatment effect by examining its variation with alternative measures of the degree of constraints on independent directors’ time such as board size, the number of board meetings required by the firm, the number of directorships of the board and the interlocked director, and an overall measure of director busyness obtained through a principal component analysis of all these measures. We robustly document that different measures of director busyness significantly magnify the treatment effect. The results are robust to a multivariate regression analysis where we control for several interlocked firm characteristics such as firm size, growth opportunities and profitability, interlocking director age and tenure, and year, industry, and deceased individuals’ fixed effects.

We also show that there is significant heterogeneity in the market reaction associated with outside directors’ busyness by type of committee membership, strength of the interlocked firms’ governance, and to some extent, interlocked firms’ advisory needs and director expertise. The average cumulative abnormal returns (CARs) on attention shocks are more negative for interlocks through an audit committee, which is likely the most demanding assignment in terms of duties and time among all committee memberships, and for firms with stronger corporate governance, which is where the impact and degree of actual independence of interlocked directors is likely to be higher. In addition, CARs are more negative for younger firms and for directors with greater industry experience, consistent with recent theories that emphasize the board’s advisory role (Adams and Ferreira, 2007).

When we examine the long-run stock price impact of attention shocks using a calendar-time portfolio return approach, we find that the valuation effect of attention shocks is not purely temporary and tends to persist over time. In fact, the DID analysis of interlocked firms’ operating, financing, accounting, and CEO pay policies shows that the effect is at least in part attributable to a deterioration in earnings quality, a decrease in leverage, and higher CEO rent extraction. These findings further reinforce the notion that investors attach a negative value to shocks that increase outside directors’ workload and that investors’ concerns are indeed warranted based on the evidence from corporate policies that are consistent with a weakened monitoring of the CEO.

In our final battery of tests, we repeat this analysis using 189 deaths of CEOs as the source of our attention shock. We identify 592 interlocked firms that are linked through an independent director to such corporate events during the sample period. Given CEOs’ prominent role within firms, their deaths likely create greater shocks than directors’ deaths, which should improve the power of our tests. In addition, they offer an alternative validation of our identification strategy since they allow us to show the robustness of our previous results to an alternative set of treatment and control groups. However, the sample size is much smaller in the case of CEO deaths. Keeping this tradeoff in mind, we find that the market reaction is also negative and statistically significant for attention shocks generated by CEO deaths, and the treatment effect is somewhat more pronounced in these cases.

The main contribution of our paper is to the literature on the impact of outside director busyness on shareholder value by documenting endogeneity-free evidence of significant costs of director busyness on shareholder wealth and corporate policies. We provide direct evidence that independent directors’ workload matters for investors. Our natural experiment allows us to overcome the notoriously difficult sample selection challenge that busy directors can increase firm value because they can be of superior quality, as well as the potential omitted variables issue that any negative relationship between firm value and director busyness could be the result of an underlying corporate governance problem that director busyness may be proxying for (Adams et al., 2010).

Our results provide direct evidence supporting the too-busy-to-mind-the-business view of studies such as Fich and Shivdasani (2006). As such, our evidence that additional demands on directors’ time have adverse consequences for board monitoring quality and firm value suggests that multiple directorships can be detrimental to shareholder value especially when the independent director and the board are already busy. These results also suggest that the recommendations of Institutional Shareholder Services to limit the number of seats directors can hold in publicly traded firms may benefit shareholders by making boards more resilient to adverse shocks that increase demands on directors’ time.

The full paper is available for download here.

Both comments and trackbacks are currently closed.