Recent corporate governance reforms have emphasized increasing the boardroom influence of independent directors, and thus developments contrary to this trend would seem to be important to investors. The authors examine the disadvantages of a related reversal—namely, the surprise departures of independent directors and the unfavorable financial consequences. Not only is the announcement effect of such events shown to be negative, suggesting some market inference of subsequent bad news, but also further evidence is offered regarding the direction of that effect’s causality. Directors appear to step down in anticipation of bad events, as opposed to the departures themselves leading to unfavorable corporate results.

Modeling the supply side of the director labor market and the director-specific factors and firm characteristics that explain departures, the authors find that most exits can be anticipated because they are often due to retirement. The significant minority of exits described as unexpected exhibit persistently (for 12 months afterward) bad associated stock performance. For a baseline comparison to assess causation, the authors evaluate turnover due to deaths of independent directors and do not find a consequently increased prevalence of adverse developments at affected firms.

Several research implications are spotlighted:

• The governance benefits connected to independent directors seem limited, because such directors have incentives to quit ahead of bad news and thus are gone (and difficult to replace) when their former firms subsequently struggle.

• Reputational concerns can have a perverse effect on the behavior of independent directors. Instead of being incentives to greater effort, such concerns can motivate departures as a response of self-preservation against trouble.

• Clustered surprise departures, though infrequent, would presage a firm’s deteriorating future performance.

The authors’ primary data sources are the RiskMetrics Directors database (now ISS) and Standard & Poor’s Compustat for accounting information and CRSP for stock returns. Their final sample covers the tenures of 16,497 directors at 2,282 distinct firms over 1999–2010.

The authors’ method contains a number of steps. The determinants of independent director departures are established via a Cox proportional hazard regression to predict expected tenure, with age (69 years or older) being critical. Instead of relying on published reports of turnover, which are seen as biased, the regression model’s results are used to construct measures of unexpected departures, which are defined as instances when directors left despite the model’s predicting a likelihood of more than 75% that they would stay.

Both expected and surprise departures are related to a firm’s future stock and operating performance. Adverse corporate events (e.g., earnings restatements, shareholder lawsuits, extremely poor stock returns) are checked for increased frequency after surprise departures. The issue of direction of causation is addressed by using directors’ deaths as an instrumental variable for regression tests of surprise departures, with corroboration for a preemptive motivation to depart before an adverse event.

The authors complete the analysis by examining the special issues of clustered departures of directors, the quality of replacement directors after surprise departures, and investor reaction to surprise departures.

It seems safe to assume that both proponents and skeptics of market efficiency are aware of market anomalies and their potential profit opportunities. Both could possibly exploit the type of economic development that the authors’ study of independent directors examines. Its findings are explicitly qualified: The authors note that adverse events follow surprise departures merely “on average,” with no certain direct correspondence. They also acknowledge other plausible explanations—for example, turnover arising from fundamental disagreements about company policy.

Continuing in this cautious vein, the authors could have strengthened the study’s reliability by various means:

• Using a model to designate unexpected departures while rejecting written evidence available from subsequent reporting on turnover (or even from directly interviewing or surveying former directors) places great confidence in a researcher’s ability to come up with the right specifications.

• Postdeparture activities of former directors are overlooked. Yet given that reputational concerns are highlighted, the taint of adversity could have an ongoing and extensive reach for directors who depart right before crashes.

• Similarly, the nature of real independence is unstated, with outsider status apparently seen as sufficient. Long tenure could effectively make a director a constructive insider, so the impact of surprise exits along a seniority continuum would be informative.

Nonetheless, the authors offer a substantial weight of evidence that surprise departures convey a significant signal of bad times ahead. As a further incentive for a closer look, it shows that, at least to date, the prospects for rent seeking have not been fully arbitraged away given that the related negative abnormal stock returns extend for one year into the postannouncement future.