SETTLING THE STAGGERED BOARD DEBATE.

• Author: Amihud, Yakov

INTRODUCTION 1476 I. BACKGROUND 1480 A. The Theoretical Effect of the Staggered Board 1480 B. The Policy Dispute 1484 C. Our Empirical Strategy 1487 II. EMPIRICAL FINDINGS 1487 A. Dataset 1487 B. Examining the Effect of a Staggered Board on Firm Value 1489 1. The Staggered Board and Omitted Variables 1489 2. The Determinants of a Staggered Board 1495 3. The Effect of a Staggered Board on Firm Value Using Firm-Fixed Effects 1498 4. A Note on Endogeneity 1501 III. POLICY IMPLICATIONS 1505 CONCLUSION 1507 APPENDIX 1508 Professor Lucian Bebchuk has engaged in two rounds of law-review-article duels with Professor Martijn Cremers and Professor Simone Sepe over classified boards. The weapons were statistics (and common sense). Cremers and Sepe wore the classified-board-stakeholder colors; Bebchuk, the agency-model-shareholder-democracy colors. Cremers' and Sepe's riposte was decisive. --Martin Lipton & Daniel Bulaevsky (1) INTRODUCTION

The staggered board debate has been both heated and confrontational. On the one side are those who argue, based in part on work by Professors Lucian Bebchuk and Alma Cohen, that the staggered board is value-decreasing and entrenches directors and managements (2) On the other side is the exact opposite argument--based in part on work by Professors Martijn Cremers, Lubomir Litov, and Simone Sepe--that the staggered board instead allows directors to bargain for higher takeover premiums and hence increases firm value. (3) In recent years, this debate has devolved into polemical statements from both sides often (but not always) citing key empirical studies on the issue. (4) These studies and this debate have driven recent law review policy proposals calling for either banning the staggered board or making the staggered board mandatory for all companies. (5) Studies finding negative wealth effects of a staggered board have also undergirded a campaign by the Harvard Law School Shareholder Rights Project to push publicly traded companies in the S&P 500 to eliminate their staggered boards. (6)

This Article sorts through this debate, gives clarity to the policy arguments, and provides an assessment of these empirical studies. We do so by analyzing the empirical and theoretical issues with studies both supporting and disparaging the staggered board. We then conduct our own empirical analysis of prior studies to determine their validity. We show that contrary to the prior major studies, a staggered board has no significant effect on firm value.

We begin by theorizing that prior studies are not robust to different estimation models. Specifically, when a regression is performed, explanatory variables are included to identify their effect on the dependent variable. For example, consider a study to ascertain the effect of irrigation on plant growth. Plant growth would be our dependent variable and irrigation an explanatory variable, along with other variables that affect plant growth such as fertilization, cultivation and weather conditions. The goal would be to isolate the effect of irrigation by considering all the relevant factors that make plants grow, because farmers who are careful about irrigation also diligently fertilize and cultivate the plants. Omitting these variables may incorrectly attribute their effects on plant growth to irrigation alone.

In the case of the staggered board, the dependent variable is the firm's market value (relative to its assets), and the main explanatory variable is the presence or absence of a staggered board, controlling for other firm characteristics that affect value. To obtain a good assessment of the relationship of firm value and the staggered board, it is important to include the main variables that effect the presence or absence of a staggered board to tease out the full relationship between the variables.

We show that prior studies, including the study by Professors Bebchuk and Cohen, do not include important explanatory variables that affect firm value and are correlated with the presence or absence of a staggered board. (7) The result is that these studies have inappropriately attributed a lower firm value to the presence of the staggered board instead of to these omitted variables. To illustrate with another example, while the birth of baby lambs is positively correlated with the arrival of storks, it is not that storks bring baby lambs. Rather, both are positively affected by the spring weather that causes the arrival of both. If we omit the spring weather from a model we may infer erroneously from the lambs-storks correlation that the former is affected by the latter. In the case of the staggered board, once we include our identified, omitted variables--firm characteristics--there is no longer a significant relation between a staggered board and firm value.

We make this finding by first estimating the effect of a staggered board on firm value across firms and over time, employing the explanatory variables used in prior studies. Using data compiled by Institutional Shareholder Services (ISS) and its predecessors, we examine the effect of a staggered board on up to 2961 firms over a timespan of twenty-three years for a total of up to 27,016 firm-years. (8) Our initial results show that firm value is negatively affected by a staggered board, which is consistent with the prominent study of Bebchuk and Cohen. (9) This is notable since the Bebchuk and Cohen study ended in 2002; we extend it through 2013 with a similar result. However, the effect of a staggered board becomes insignificant once related explanatory variables are included in our analysis. Putting this result another way, Bebchuk and Cohen's analysis does not include important variables related to both firm value and the incidence of the firm having a staggered board. The inclusion of these variables renders the effect of a staggered board on firm value insignificant. In particular, we find that the negative firm value-staggered board relation becomes insignificant once we include in the model an entrenchment index developed by Bebchuk, Cohen, and Ferrell that quantifies other corporate governance measures. (10) This leads to the conclusion that the firm-value effects of a staggered board are driven by other variables and not by the staggered board itself.

We next turn to studies that have shown that the staggered board enhances value, examining the most prominent study in this area, that of Cremers, Litov, and Sepe. (11) Their empirical strategy was to add firm-fixed effects to the models of Bebchuk and Cohen. These effects control for unobserved individual firm characteristics that are time invariant and purportedly account for the idiosyncratic nature of the firm, thus attempting to address the omitted variable problem of the Bebchuk and Cohen study.

We replicate the Cremers, Litov, and Sepe methodology and find similar results. However, given that our sample spans twenty-three years (1991-2013) and firm-fixed effects are by definition constant over this period, it is unlikely that all firms have invariable characteristics over this long period of time. Rather, we know that some firms choose to stagger their boards and other firms decide to destagger their boards. It is likely that these decisions are undertaken because for some firms there are changes in unobserved characteristics and conditions over time. We therefore split the sample into two subperiods and estimate the model for each subperiod. When we make this adjustment to account for the varying effect of firm-fixed effects, we find that a staggered board has no significant effect on firm value in any of the two subperiods.

Our results thus find shortcomings in both sides of the debate. They also highlight the sensitivity of any analysis of the staggered board and firm value to the choice of variables and models. To the extent that variable selection produces such different results, it highlights the unreliability of prior studies concerning the wealth effects of staggered boards.

We also address the fact that having a staggered board--its adoption, retention, or removal--is a result of a decision made by the firm. In other words, the estimated effect of a staggered board may reflect the consequences of the factors that led to the decision rather than resulting from the presence or absence of a staggered board. We correct for this problem by employing an instrumental variable estimation method, which is detailed in subsection II.B.4. In this analysis, we find the value effect of a staggered board becomes insignificant.

We conclude by examining the policy implications of our findings. Our analysis means no definitive conclusion can be made at this time as to the positive or negative wealth effects of a staggered board. In terms of wholesale policy efforts to adopt or repeal staggered boards, our results suggest caution. We find that staggered boards appear to be affected by firm characteristics that account for decisions on adoption, retention, and removal of the staggered board provision, which are partially unobserved or cannot be quantified for the purpose of research. Our findings ultimately highlight the theoretical proposition that the staggered board is endogenous, and that the decision to adopt (or not) the staggered board is unique to each firm and its characteristics. (12) In some firms the staggered board may be value-enhancing, in others value-destroying. Therefore, the battle over the staggered board must be waged at the individual firm level.

More generally, our results provide evidence for measured skepticism of prior corporate governance studies and their implications for the structure of the board of directors generally. We conclude by analyzing various legal proposals related to the staggered board, and find that our results suggest caution about each of them. The rhetoric unfortunately does not match reality and the staggered board is neither value-decreasing nor value-enhancing overall. Instead, it...