1 Introduction

“Staggered boards are the most powerful antitakeover device in the current arsenal of takeover defense weapons” (Bebchuk et al. 2002, p. 950). This opinion is widely shared: staggered boards protect incumbent management teams from removal, can prevent hostile bidders from gaining control, and ultimately decrease the likelihood of a hostile takeover. In 2020, 31% of firms in the S&P 1500 and 53% of non-S&P 1500 firms had a staggered board, indicating the continuing popularity of this corporate governance provision (Guernsey et al. 2022a).

However, whether staggered boards are value-enhancing or value-destroying is subject to an ongoing academic debate. Proponents argue that staggered boards provide stability to a firm’s corporate governance and promote long-term strategic planning by reducing short-term shareholder pressure (Bebchuk & Stole 1993; Stein 1988, 1989). They also decrease the likelihood of costly disruptions to stakeholder relations and protect against hostile takeovers harming shareholder wealth (Stulz 1988). However, opponents argue that staggered boards hinder the free market for corporate control, limit accountability, and may encourage shirking and empire-building by board members (Jensen 1988, 1993; Manne 1965). They also raise the question of self-serving managers blocking beneficial takeovers or optimizing the terms of a takeover for their own benefit rather than maximizing shareholder value (Grossman and Hart 1980; Hartzell et al. 2004). Recent studies suggest a heterogeneous impact (Cremers et al. 2017; Daines et al. 2021), which is why it is particularly interesting to examine the circumstances that influence the effect of staggered boards on company value.

One factor that might mitigate the effect of staggered boards on firm value is a market shock. During market shocks, lower valuations place more value on anti-takeover provisions as they shield firms from the increased threat of hostile takeovers (Araujo et al. 2020). Managers can focus on managing the crisis instead of being distracted by activism threats (Shleifer and Summers 1988). Therefore, staggered boards should help preserve value in the face of market shocks. This paper aims to analyze this issue and empirically examine the effects of staggered boards on firm value during market shocks, contributing to the ongoing debate.

To this end, this study utilizes multiple databases to construct a broad sample of 354,517 firm-quarter observations, covering most US-listed firms over a period between 1994 and 2021. The chosen methodology draws on previous research and employs several alterations of an ordinary least squares regression model. The dependent variable is firm value, measured using the logarithmic value of Tobin’s Q. The key independent variables are staggered boards and market shocks. I obtain the staggered board variable from Guernsey et al. (2022a) and define market shocks as calendar quarters with a cumulative return on the CRSP value-weighted index of less than or equal to   − 10%. Some regressions include firm, industry, and/or time fixed effects to control for unobserved heterogeneity across firms, time-invariant industry-level omitted variables, and seasonality, respectively. Following prior work, I control for firm-level characteristics that could correlate with the staggered board variable.

I start the analysis by examining the relationship between firm value and staggered boards during market shocks for the entire sample and subsequently for different sub-periods. The results indicate a time-invariant effect of staggered boards: While they appear to be value-increasing during market shocks up to and including 2006, the effect disappears from 2007 onward. To check the robustness of the results, I conduct a second set of analyses in which I replicate the studies for the two most significant crises of the observation period: the dot-com crisis and the global financial crisis. The dot-com crisis took place before 2006, but the financial crisis falls in the period from 2007 onward. While a value-preserving effect of staggered boards can be observed in the dot-com crisis, thus confirming the previous results, the findings for the financial crisis are much less clear but indicate that staggered boards were rather value-decreasing.

These findings raise the question of which factors mitigate the effect of staggered boards during market shocks. To shed more light on this, I conduct a third set of analyses, splitting the observations sequentially by appearance in the ISS sample, R&D intensity, and size. The results suggest that the impact of staggered boards is influenced by all three factors, with the effect being strongest for non-ISS, R&D-intensive, and small firms. For non-ISS and small firms, the economically and statistically significant effect continues even after 2007. To examine this effect more closely, I combine the subdivision by appearance in the ISS sample with the subdivision by size. The findings suggest that for small non-ISS firms, staggered boards have a consistently positive impact during market shocks. The effect loses economic and statistical significance for large and/or ISS firms. These results can be replicated for both the dot-com crisis and the global financial crisis.

This study provides important insights into the heterogeneous nature of this relationship and sheds light on the specific circumstances under which staggered boards can be value-adding. The results suggest that the impact of staggered boards depends not only on the market environment but also on certain firm characteristics that mitigate their effect. During market shocks, the association between staggered boards and firm value becomes more positive and significant for R&D-intensive, small, and non-S&P 1500 firms. These findings support the argument that staggered boards benefit firms facing a high takeover threat due to low valuations, small size, and potentially greater information asymmetry between management and shareholders. Ultimately, this study challenges the notion of mandatory staggered boards and suggests that firms should have the autonomy to adopt them based on their individual circumstances.

In the next section, I briefly discuss the most common theories and arguments of the proponents and opponents of staggered boards as well as key empirical studies. Then, I present the procedure used to construct the datasets of this study and describe the data. Section 4 continues by introducing the empirical methodology used in this paper. The findings of this study are described in Sect. 5. Finally, I offer concluding remarks.

2 Literature review

2.1 Staggered boards and firm value

Staggered boards, also known as classified boards, refer to a corporate governance structure in which the board of directors is divided into several classes, each having staggered terms of office. Typically, staggered boards are divided into three classes, each serving a three-year term. This means that only one-third of the board is up for election and can be replaced, while two-thirds remain in their positions. In contrast, a non-staggered board of directors allows the entire board to be replaced at once.

The purpose of a staggered board is to provide continuity and stability to a firm’s corporate governance. A core task of the board is to oversee the company and to discipline the management. Board members whose retention in the company is secured over a longer period naturally have more power and are therefore better able to fulfill this task (Amihud et al. 2018). Additionally, when a board member’s retention in the company is secure, it decreases the likelihood of costly disruption to stakeholder relations (Shleifer and Summers 1988). Staggered boards could bond a firm’s commitment to its stakeholders, serving as a “commitment device toward more stable stakeholder relationships” (Cremers et al. 2017, p. 439).

Staggered boards also promote long-term strategic planning by reducing the risk of short-term pressure from shareholders (Bebchuk and Stole 1993; Stein 1988, 1989). Information asymmetry between board members and shareholders may lead to shareholders perceiving value-increasing long-term projects as poor performance in the short term, therefore wanting to replace management or accept suboptimal takeover offers (Goshen and Squire 2017). Boards may anticipate this and forgo profitable long-term projects to avoid appearing unprofitable in the short term. Staggering the board may provide a solution to this problem as it shifts the time perspective of its members.

Proponents also perceive a positive effect in connection with hostile takeovers. Staggered boards can protect against hostile takeovers, due to only a minority of the board being open for replacement at any given time. This improves the bargaining power of acquisition targets and enables them to increase acquisition premia (Koppers et al. 1998; Lipton et al. 2012; Stulz 1988).

Thus, following the proponents’ argumentation, the adoption of staggered boards should have a positive effect on firm value.

However, critics argue that staggered boards hinder the free market for corporate control and limit accountability. A free market for corporate control implies a constant threat of takeovers and replacement, which should serve as a disciplining force for boards (Manne 1965). Staggering a board would reduce the threat of removal in the context of a takeover and make it more difficult for shareholders to replace poorly performing board members, which in turn could encourage shirking and empire-building (Jensen 1988, 1993).

Another way how staggered boards could destroy value is by enabling self-serving managers to block hostile takeovers that would benefit shareholders (Easterbrook and Fischel 1981; Grossman and Hart 1980). Or they could utilize their power to optimize takeover terms, prioritizing their own benefit over maximizing the premium paid to shareholders (Bebchuk 2002; Hartzell et al. 2004).

Hence, the critics’ argumentation suggests a value-decreasing effect of the adoption of staggered boards.

A broad set of empirical studies deals with the question of whether staggered boards have an overall positive or negative effect on firm value. The results, however, are mixed.

A set of studies by Bebchuk et al. (2002, 2003) focuses on the effect of effective staggered boards on shareholder returns after a hostile takeover bid is made. The authors find a statistically significant and economically relevant negative effect in both the short and long terms.

These papers are complemented by Bebchuk and Cohen (2005) and Faleye (2007). Both investigate the overall impact of staggered boards on firm value, controlling for other governance provisions. They unanimously concluded that staggered boards are associated with a significant reduction in firm value.

Later, Cohen and Wang (2013) aim to identify causality, using a natural experiment involving two Delaware court rulings in 2010. The authors look at the stock price reaction of Delaware-incorporated firms to the rulings in the Airgas, Inc. case, finding a negative association.Footnote 1 Thus, they conclude that the adoption of staggered boards causes a reduction in firm value.

In contrast, some more recent studies have identified no association or even a positive association between staggered boards and firm value.

Amihud and Stoyanov (2017) examine the Cohen and Wang (2013) study, finding that the results become insignificant when excluding penny and OTC stocks or when using an alternative sample.

Cremers et al. (2017) revisit the debate and conduct several studies, using a longer sample period and focusing on decisions to stagger or de-stagger. Also, they use firm fixed effects to address the omitted variable problem. Unlike previous studies, they find that the decision to stagger (de-stagger) has a positive (negative) effect on firm value under certain conditions. The authors conclude that staggered boards have heterogeneous effects.

Amihud et al. (2018) take the debate to the next level by showing that previous studies, such as those by Bebchuk, fail to account for important variables affecting both value and the incidence of staggered boards. By controlling for these, the effect of staggered boards on firm value becomes statistically insignificant. The same results occur for their replication of Cremers et al. (2017) when accounting for the varying effect of firm fixed effects. The authors conclude that the effects of staggered boards are idiosyncratic.

Daines et al. (2021) conduct a quasi-experiment, using a 1990 adopted state law that imposed staggered boards on Massachusetts-incorporated firms as an exogenous event. They find causal evidence for a positive value effect for early-life-cycle firms facing high information asymmetry, again showing that the effect of staggered boards might depend on the firm and the specific situation they are facing.

In summary, previous studies have found mixed results on the effect of staggered boards on firm value with some showing a negative association and others finding no or even positive effects. Most recent studies have suggested that the effect is heterogeneous and idiosyncratic, depending on the firm and its situation.

The purpose of this paper is to contribute to the exploration of the circumstances under which a staggered board can be value-increasing.

2.2 Staggered boards and market shocks

According to the theories described above, one situation in which staggered boards could have a particularly positive effect on firm value would be market shocks. During such times, the threat of a takeover increases due to depressed share prices (Araujo et al. 2020; Gottfried and Donahue 2020). By implication, this places more value on the increased bargaining power arising from staggered boards, which could help management extract a higher takeover premium (Shleifer and Vishny 2003). Conversely, staggered boards could partially shield firms from the increased threat of hostile takeovers during market crises as it is more difficult and costly for the acquirer to gain the board majority. This should increase the benefits of staggered boards as a commitment device toward stakeholder relationships (Shleifer and Summers 1988). Also, managers are less distracted by activism threats arising from the takeover. Instead, they can focus their efforts on managing the crisis (Eldar and Wittry 2021).

In summary, staggered boards could help preserve value in the face of market shocks, as the benefits increase in value due to the increase in takeover threat.

There is a small literature base that empirically examines the effect of anti-takeover provisions during market shocks.

Ding et al. (2021) study the relationship between stock returns and the number of anti-takeover provisions a firm had during the COVID-19 crisis. These anti-takeover provisions include staggered boards. However, contrary to the theory described above, the authors find lower stock returns for firms with more anti-takeover provisions.

On the other hand, Guernsey et al. (2022b) examine the effect of state-level anti-takeover provisions on firm value during crises over a longer period. During 14 crises, they find no evidence for a value-decreasing effect of such state-level provisions. Instead, the findings suggest a positive association between the number of anti-takeover provisions and firm value. In a robustness test, the authors control for staggered boards. And while the variable for staggered boards itself remains insignificant and does not support either side of the argument, its interaction with the shock variable supports the argumentation above in favor of staggered boards during market shocks.

Due to the poor empirical basis and the divided nature of the results, further research is needed on this topic. This paper aims to further previous scientific efforts and examine the effects of staggered boards on firm value during market crises.

3 Data

This study combines several databases, with data availability varying between sources. For the staggered board variable, which is the key independent variable of this study, I use data provided by Guernsey et al. (2022a) who utilized machine learning algorithms to compile the staggered board status for all U.S. public firms.Footnote 2 Shock data is obtained from CRSP.

For the base sample, I use the quarterly CRSP-Compustat merged database between the first quarter of 1994 and the fourth quarter of 2021.

The augmented sample is extended to include control variables related to insider ownership, S&P 500 membership, and corporate governance. For insider ownership and S&P 500 membership, I use S&P’s ExecuComp database. For corporate governance data, I use the ISS Governance Legacy database (up until 2006) and the ISS Governance database (from 2007 onward). Since the ExecuComp and ISS data are on an annual basis and this study looks at fiscal quarters, I assume that the annual variables remain the same across all four fiscal quarters.

I exclude utility firms (SIC codes 4900 to 4999), REITs (SIC code 6798), and public administration firms (SIC codes equal to or greater than 9100). All observations are required to have non-negative equity and non-missing data for all control variables, excluding R&D and CapEx. If data are available, I also exclude firms with a dual-class structure. In addition, I conduct a 99% winsorization on all continuous variables before merging the datasets to limit extreme values and the effect of outliers.Footnote 3

Table 1 reports the descriptive statistics for all variables of the base sample. Tables 2 and 3 tabulate the descriptive statistics for all variables of the augmented sample for the period up until 2006 and from 2007 onward, respectively.

Table 1 Descriptive statistics base sample
Full size table
Table 2 Descriptive statistics augmented sample up until 2006
Full size table
Table 3 Descriptive statistics augmented sample 2007 and after
Full size table

The base sample includes 354,517 observations of which 50.3% have a staggered board. The average Q in this sample is 1.911. In the augmented sample, the first half contains 23,683 observations of which 61.2% have a staggered board. That number declines to 42.0% in the second half which includes 60,220 observations. The average Qs are 2.013 and 1.972.Footnote 4

4 Empirical methodology

The approach I am using draws on the methodology developed by Guernsey et al. (2022b). To examine the effect of Staggered Boards during market shocks, I employ several alterations of the following ordinary least squares regression model:

$${y}_{it}=\;{\beta }_{1}{\mathrm{SB}}_{it}+{\beta }_{2}{\mathrm{SB}}_{it}\times {\mathrm{Shock}}_{t}+{\beta }_{3}{\mathrm{Shock}}_{t}+{\beta }_{4}{\mathrm{MEI}}_{it}+{\beta }_{5}{\mathrm{MEI}}_{it}\times {\mathrm{Shock}}_{t}+{\gamma }_{1}^{\mathrm{^{\prime}}}{{\varvec{X}}}_{it}+{\gamma }_{2}^{\mathrm{^{\prime}}}{{\varvec{Y}}}_{it}+{\psi }_{j}\times {\omega }_{k}+{\nu }_{i}+{\eta }_{i}+{\varepsilon }_{it}$$

The dependent variable \({y}_{it}\) is a measure of firm value. This study uses the logarithmic value of Tobin’s \(\mathrm{Q}\), \(\mathrm{log}\left({Q}_{it}\right)\).Footnote 5 This is in line with prior research (Amihud et al. 2018; Bebchuk and Cohen 2005; Cremers et al. 2017; Daines 2001; Daines et al. 2021; Gompers et al. 2003; Guernsey et al. 2022a, b).

Similar to Guernsey et al. (2022b), I treat market shocks as exogenous events and define the dummy variable \({\mathrm{Shock}}_{t}\) as a quarter with a cumulative return on the CRSP value-weighted index less than or equal to  − 10%. By this definition, twelve quarters qualify as shocks between Q2 1994 and Q3 2021.

In all regressions, I control for the following firm-level characteristics \({{\varvec{X}}}_{it}\): the logarithm of total assets, Delaware incorporation, CapEx, CapEx missing, R&D, R&D missing, tangible assets, return on assets, assets to equity, leverage, asset growth, sales growth, and industry sales share. These characteristics could correlate with a firm’s decision to stagger or de-stagger its board, thus biasing the effect. Their selection is guided by the current literature (cf. Amihud et al. 2018; Cremers et al. 2017; Daines et al. 2021; Guernsey et al. 2022a, b).

Regressions on the augmented sample also control for \({\mathrm{MEI}}_{it}\), \({\mathrm{MEI}}_{it}\times {\mathrm{Shock}}_{t}\), and \({{\varvec{Y}}}_{it}\). \({\mathrm{MEI}}_{it}\) stands for the company’s modified E-index, which can take on values between 0 and 5, depending on how many of the five governance-related provisions, excluding staggered boards, are in place.Footnote 6 I include this to ensure that the value effects of \({\mathrm{SB}}_{it}\) are not driven by other corporate governance provisions which correlate to \({\mathrm{SB}}_{it}\) (Amihud et al. 2018). \({{\varvec{Y}}}_{it}\) includes the following additional firm-level characteristics: S&P 500 membership, insider ownership, and insider ownership squared.

Some regressions include fiscal-quarter fixed effects \({\psi }_{j}\) \(\times \) fiscal-year fixed effects \({\omega }_{k}\) to control for seasonality and other time-varying factors which affect all firms similarly.Footnote 7

Some regressions include firm fixed effects \({\nu }_{i}\) to control for unobserved heterogeneity across firms and mitigate the omitted variable bias. Those regressions that do not include firm fixed effects include industry fixed effects \({\eta }_{i}\) based on two-digit SIC codes to account for time-invariant industry-level omitted variables.

The main focus of this study is to analyze the effect of staggered boards on firm value during market shocks. For this purpose, the following section of the regression model is of particular importance: \({\beta }_{1}{\mathrm{SB}}_{it}+{\beta }_{2}{\mathrm{SB}}_{it}\times {\mathrm{Shock}}_{t}+{\beta }_{3}{\mathrm{Shock}}_{t}\). The model makes it possible to consider the effect of \({\mathrm{SB}}_{it}\) (\({\beta }_{1}\)), that of \({\mathrm{Shock}}_{t}\) (\({\beta }_{3}\)), and that of \({\mathrm{SB}}_{it}\times {\mathrm{Shock}}_{t}\) (\({\beta }_{2}\)) separately, to subsequently compute the total effect of staggered boards during market shocks (\({\beta }_{1}+{\beta }_{2})\).

5 Empirical findings

5.1 Main set of regressions

Table 4 presents the results of the main regressions, estimating the effect of staggered boards during market shocks on the logarithm of Tobin’s Q.

Table 4 Main regressions (I/II)
Full size table

Columns 1–3 tabulate the results from the base sample, only controlling for the base controls, while columns 4–9 report the results from the augmented sample and include the extended set of controls. Columns 4–9 are subdivided into columns 4–6, which cover the period 1995–2006, and columns 7–9, which look at the period 2007–2021. I make this separation because the Institutional Shareholder Services data is taken from two different datasets.Footnote 8 Columns 1, 4, and 7 include industry fixed effects based on two-digit SIC codes, columns 1, 3, 4, 6, 7, and 9 include fiscal-quarter fixed effects \(\times \) fiscal-year fixed effects, and columns 2, 3, 5, 6, 8 and 9 include firm fixed effects.

The regressions on the base sample clearly show that the dependent variable \(\mathrm{log}\left(Q\right)\) decreases less during market shocks for companies with staggered boards. In all three columns, the coefficients of 0.017, 0.026, and 0.029 on \(\mathrm{SB}\times \mathrm{Shock}\) are positive and statistically significant at the 1% level. Also, the total effect of staggered boards during market shocks, computed as the sum of the coefficients on \(\mathrm{SB}\) and \(\mathrm{SB}\times \mathrm{Shock}\), is positive in all three regressions. This effect is independent of whether the regression includes industry fixed effects or firm fixed effects.

In columns 4–6, the statistically significant coefficients of 0.035, 0.030, and 0.030 on \(\mathrm{SB}\times \mathrm{Shock}\) confirm the findings of the base sample for the augmented sample in the period 1995–2006. The total effect of staggered boards during market shocks is even more positive in the three extended regressions.

It is noteworthy that the modified E-index in columns 4–6 is highly significant and negative, while its interaction with the market shock variable is statistically and economically insignificant. This indicates that other corporate governance provisions have a negative effect on firm value, independent of market shocks.

Columns 7–9 contrast the findings of the previous regression analyses. The coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) become both economically and statistically insignificant for the augmented sample in 2007–2021. This is contradictory and not in line with the theories discussed above.

An additional finding in Table 4 is that the coefficients on \(\mathrm{SB}\) are inconsistent in terms of their sign as well as their economic and statistical significance. In nine regression analyses, the coefficients of \(\mathrm{SB}\) are statistically significant in six and positive in only four. Moreover, no pattern is discernible that could explain this variation. These findings suggest that the effect of staggered boards during normal times is not conclusively clear.

The findings in Table 4 raise a central question: Why do the results for the second half of the augmented sample differ so much from the first half and the base sample?

Table 5 shows the same regressions as Table 4, with the base sample split into two parts according to the division of the augmented sample. The aim is to determine whether the difference is due to the observation period or other factors such as the composition of the augmented sample or the extended controls.

Table 5 Main regressions (II/II)
Full size table

For the first half of the base sample (columns 1–3), the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) remain positive and even gain economic significance. For the second half (columns 4–6), however, the coefficients become statistically and economically insignificant, except for column 6 where the coefficient drops to 0.010 at a 5% significance level. These findings suggest that the observation regarding the second half of the augmented sample is due to the observation period.

Overall, the results support the hypothesis that staggered boards can help preserve value in the face of market shocks, but only until the end of 2006. For the period from 2007 onward, the positive value effect of staggered boards on firm value during market shocks seems to have disappeared.

5.2 Case studies: dot-com crisis and global financial crisis

In the previous analyses, I looked at the coefficients across all market shocks. Now I want to test whether the findings can be replicated for individual crises. The two most prominent market shocks in the US during the period under review were the dot-com crisis and the global financial crisis. For this study, the exact quarters that are to be labeled as part of the crisis need to be defined.

In the period following the bursting of the dot-com bubble in March 2000, five quarters can be classified as market shocks according to the definition given in Sect. 4, and which can be considered part of the market correction following the unusually high valuations of the dot-com bubble. I call the totality of these five quarters “Dot-Com All Losing Quarters”. However, one of these quarters follows the attacks on September 11th, 2001. Although the dramatic loss of the quarter without a full recovery in the following quarters may well be due in part to the inflated valuations of the dot-com bubble, it makes sense to run a regression without it being part of the dot-com crisis. The remaining quarters include the bursting of the dot-com crisis itself as well as the following correction, I call them the “Dot-Com Burst and Correction”.

The global financial crisis is generally dated to the period from mid-2007 to early 2009. During that time, there is a negative value-weighted return in all six quarters between Q4 2007 and Q1 2009. I call this period the “Financial Crisis”. However, only the quarters Q3 and Q4 2008 have a value-weighted return of less than  − 10% and thus qualify as shock quarters. Those two quarters follow the Lehman Brothers filing for bankruptcy on September 15th, 2008, and are considered to be the climax of the financial crisis. Therefore, I call them the “Climax of Financial Crisis”.

I run selected regression analyses again, with the change that the shock variable now marks those quarters that fall under the new definitions of the respective crises.

Table 6 presents the results of the regressions for the dot-com bubble. Columns 1–4 use the “Dot-Com All Losing Quarters” definition for the shock variable, and columns 5–8 use the “Dot-Com Burst and Correction” definition. Columns 1, 2, 5, and 6 tabulate the results from the base sample, only controlling for the base controls, while columns 3, 4, 7, and 8 report the results from the augmented sample and include the extended set of controls. All regressions include firm fixed effects and columns 2, 4, 6, and 8 include fiscal-quarter fixed effects \(\times \) fiscal-year fixed effects.

Table 6 Case study: dot-com crisis
Full size table

The findings support the results of Sect. 5.1. In all eight columns, the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) are positive and statistically significant. Additionally, the total effect of staggered boards during the dot-com crisis, computed as the sum of the coefficients on \(\mathrm{SB}\) and \(\mathrm{SB}\times \mathrm{Shock}\), is positive in all eight regressions.

It is noteworthy that in most regressions, the coefficient on SB is statistically significant and positive, except for columns 1 and 5 where the coefficients 0.005 and 0.005 lack significance. Those two regressions only control for base controls and do not include fiscal-quarter fixed effects × fiscal-year fixed effects. They become significant after including the extended set or controls of controlling for the fiscal quarter and year. These results suggest that staggered boards may have had a positive effect on firm value independent of market shocks. The findings are consistent with Cremers et al. (2017) who also used firm fixed effects in their regressions.

Table 7 is structured similarly to Table 6. Columns 1–4 use the “Financial Crisis” definition for the shock variable, and columns 5–8 use the “Climax of Financial Crisis” definition. Because the financial crisis falls into the second half of the entire period analyzed, I use the subsamples covering the years 2007–2021 instead of 1994/1995–2006. The rest of the setup in terms of control variables and fixed effects is the same as in Table 6.

Table 7 Case study: global financial crisis
Full size table

The results are partially surprising. Based on the findings in Tables 4 and 5, one would expect the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) to be statistically insignificant as the financial crisis falls into the period after 2007. However, this is only the case in columns 2, 4, and 8. In the other five columns, the coefficient is positive and statistically significant which would be more in line with the findings for the period until 2006. Yet, in four out of five of those regressions, the coefficient on \(\mathrm{SB}\) is negative and larger than the one on \(\mathrm{SB}\times \mathrm{Shock}\), resulting in a negative total effect of staggered boards during the financial crisis. This in turn supports the conclusion that for the period from 2007 onward the hypothesis that staggered boards are value-increasing during market shocks no longer holds.

Conclusively, the two most prominent market shocks largely support the findings of the previous set of regressions. The results of the dot-com crisis, which took place before 2006, suggest a value-preserving effect of staggered boards during the crisis itself. The results of the financial crisis, which took place after 2007, are less clear but suggest an overall value-destroying effect of staggered boards during the crisis.

5.3 Firm characteristics mitigating the effect

The regressions so far show that there is an association between staggered boards and firm value during market shocks. However, it is unclear whether this relationship is homogeneous across firms or heterogeneous. Next, I want to look at different firm characteristics that could mitigate the effect.

5.3.1 S&P 1500 membership

Many studies in the field of corporate governance use the ISS Governance databases to obtain information on the staggered board status of companies (cf. Bebchuk et al. 2009, 2013; Gompers et al. 2003). ISS, however, mainly covers the S&P 1500, neglecting non-S&P 1500 firms. This is problematic as there is reason to believe that corporate governance provisions have a different effect on those firms.

The S&P 1500 is composed of all constituents of the S&P 500, the S&P MidCap 400, and the S&P SmallCap 600 (S&P Global 2023). Those firms tend to get much more attention from the media and are more closely observed, potentially helping to reduce information asymmetry. Additionally, Guernsey et al. (2022a) find that shareholders in S&P 1500 firms are more experienced, resourceful, and generally more effective in coordinating with other investors, compared to their counterparts in non-S&P 1500 firms. This has led to different dynamics in terms of staggering and destaggering: while the share of staggered boards in S&P 1500 firms has decreased over the past two decades, it has actually increased for non-S&P 1500 firms. Consequentially, it is uncertain whether staggered boards have the same effect on S&P 1500 and non-S&P 1500 firms.

These circumstances underline the necessity for this study to examine whether the effect of staggered boards in the event of market shocks differs for companies covered in the ISS governance databases from those that are not. Therefore, I will conduct the main regressions again, but distinguish between ISS and non-ISS firms. Since the augmented sample relies on ISS data, I am limited to the base sample.

Tables 13 and 14 report the results from the regressions for ISS firms and non-ISS firms, respectively.

For the period up until 2006, the significance of the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) for ISS firms decreases over all three estimates (Table 13, columns 1–3), compared to the main regressions in Table 5. For non-ISS firms, the coefficients remain highly significant at a 1% level and become even more positive (Table 14, columns 1–3). The findings suggest a stronger effect for non-ISS firms.

In the second period from 2007 onward, most coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) for ISS firms are economically and statistically insignificant, just like in the main regressions. In column 5 (Table 13), the negative coefficient becomes slightly significant at a 10% level. However, it turns insignificant after including fiscal-quarter fixed effects \(\times \) fiscal-year fixed effects (Table 13, column 6). These results are largely consistent with those of the main analysis.

For non-ISS firms from 2007 onward, however, the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) become highly significant and positive when including firm fixed effects (Table 14, columns 5–6). In all columns, the total effect on staggered boards during market shocks is positive.

The findings suggest that S&P 1500 membership influences the effect of staggered boards during market shocks. In contrast to the S&P 1500 firms, non-S&P 1500 firms seem to benefit from staggered boards even after 2007. This effect is further amplified during market shocks.

5.3.2 R&D

An important theory discussed above is that staggered boards reduce the risk of short-term pressure from shareholders and promote long-term strategic planning (Bebchuk and Stole 1993; Stein 1988, 1989). Some proponents of management entrenchment argue that takeover defense mechanisms such as staggered boards increase value by enabling management to focus on innovation (Manso 2011). Following this argumentation, staggered boards are likely to be particularly advantageous for those companies that are innovative and engage in long-term investments such as R&D. Again, this effect should be larger for firms experiencing a higher takeover threat due to lower valuations during market shocks (cf. Sect. 2.2.). I test this by dividing the samples used in the main regressions into innovative and non-innovative firm observations by using R&D intensity as a proxy and repeating the main regressions. Beforehand, I exclude all firm-quarters for which R&D is missing.

Tables 15 and 16 contain the results of the regressions with innovative firms, while Tables 17 and 18 contain the results of the regressions with non-innovative firms.

It is clear to see that the effect of staggered boards during market shocks is generally more positive and significant for innovative firms compared to non-innovative firms. For non-innovative firms, only columns 2 and 3 of Table 18 and none of Table 17 return positive and statistically significant coefficients on \(\mathrm{SB}\times \mathrm{Shock}\). For innovative firms, they are positive and statistically significant in columns 2, 3, 5, and 6 of Table 15 and columns 2 and 3 of Table 16.

It can be concluded that innovative firms benefit more from staggered boards during market shocks than those that are not innovative. This conclusion is consistent with the theoretical argumentation above.

It should be noted that the results are much less clear and consistent than in the previous parts of this study. The reason for this may be that R&D intensity is calculated based on the financial quarter and can fluctuate. Companies that have invested a lot in R&D are classified as non-innovative as soon as they stop their investments, even though the information asymmetry regarding the investment remains. This is problematic and may be the cause of the unclear results.

5.3.3 Size

Next, I examine how a firm’s size affects the impact of staggered boards during market shocks. Assuming that smaller companies become more frequent takeover targets, they are likely to benefit more from the advantages of staggered boards (cf. Sect. 2.1.). Therefore, I divide the samples used in the main regressions into small and large firm observations by market value and repeat the main regressions.

Tables 19 and 20 present the results of the regressions with small firms, and Tables 21 and 22 present the results of the regressions with large firms.

In most columns of Tables 21 and 22, the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) become economically and statistically insignificant, indicating that for large firms, staggered boards have no exceptional effect during market shocks.

For small firms, all coefficients on \(\mathrm{SB}\times \mathrm{Shock}\), which were statistically significant and positive in the main regressions, remain statistically significant and positive. Additionally, the coefficients in most other columns become statistically significant and positive, too. Column 8 in Table 19 is the only column in which the coefficient remains insignificant. However, this changes as soon as one adds fiscal-quarter fixed effects \(\times \) fiscal-year fixed effects. This suggests that for small firms the value-preserving effect of staggered boards during market shocks continues after 2007.

Overall, the regressions with small firms show much more positive and statistically significant coefficients on \(\mathrm{SB}\times \mathrm{Shock}\), compared to those with large firms. The total effect of staggered boards during market shocks, calculated as the sum of the coefficients on \(\mathrm{SB}\) and \(\mathrm{SB}\times \mathrm{Shock}\), is much higher for small firms, too. This supports the hypothesis that smaller firms are more likely to benefit from the advantages of staggered boards during market shocks, even after 2007.

5.3.4 S&P 1500 membership and size

In the previous three blocks, the effect of staggered boards during market shocks has been found to be more positive and significant for non-ISS, R&D-intensive and small firms. Moreover, there is evidence that the positive effect continues after 2007 for non-ISS and small firms. To examine this effect more closely, I combine the subdivision by ISS membership with the subdivision by size and repeat the main regressions. For this purpose, I first divide the base sample into small and large firms and then subdivide these subsamples according to ISS membership.

Tables 8 and 9 present the results for small non-ISS and ISS firms. Tables 10 and 11 contain the results for large non-ISS and ISS firms, respectively.

Table 8 Regressions with small non-ISS firms only
Full size table
Table 9 Regressions with small ISS firms only
Full size table
Table 10 Regressions with large non-ISS firms only
Full size table
Table 11 Regressions with large ISS firms only
Full size table

Noticeably, size and ISS membership are not independent of each other, but the vast majority of ISS companies belong to the group of large companies. However, due to a large number of remaining observations in the samples of small ISS and large non-ISS firms, this is not a problem.

Table 8 reports that for small non-ISS firms, the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) are positive and statistically significant at a 1% level for every single regression, even after 2007 (columns 4–6). Moreover, the total effect of staggered boards during market shocks is consistently positive in every column.

In Tables 9, 10, and 11, the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) lose economic and statistical significance. For large firms, the coefficients turn negative after 2007 (Tables 10 and 11, columns 4–6) and even become slightly significant if the firm is covered by ISS (Table 11, columns 4–5). The total effect of staggered boards during market shocks varies between columns.

However, after 2007, the coefficients on \(\mathrm{SB}\) are consistently negative and statistically significant for small ISS firms and positive and statistically significant for large non-ISS firms, while the interaction effect with \(\mathrm{Shock}\) remains economically and statistically insignificant (Tables 9 and 10, columns 4–6). Therefore, the total effect of staggered boards during market shocks is positive for large non-ISS firms and negative for small ISS firms for that period.

To test whether the findings regarding small non-ISS firms can be replicated for individual crises, I repeat regressions for the dot-com and financial crisis with the subsample of small non-ISS firms. Table 12 presents the results of the dot-com crisis and financial crisis. As expected, the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) are positive and highly significant in all eight columns. The coefficients on \(\mathrm{SB}\) are mostly economically and statistically insignificant so the total effect of staggered boards during market shocks is positive. This confirms the findings that small non-ISS firms can benefit from staggered boards during market shocks.

Table 12 Case studies with small non-ISS firms only
Full size table

Conclusively, the findings suggest that staggered boards can help to preserve value in the face of market shocks, especially for small non-ISS firms. They are rather value-increasing for other non-ISS firms and value-decreasing for ISS firms.

5.4 Further robustness tests and endogeneity concerns

To further test whether the findings are sensitive to alternative specifications, I conduct additional analyses.

First, I test whether different industry definitions have an impact on the results of the main regressions. For this purpose, I use the three-digit SIC code and the S&P Industry Sector Code instead of the two-digit SIC code to construct the industry fixed effect. The results are presented in Tables 23 and 24, respectively. Overall, the coefficients and their significance remain much the same. In the three-digit SIC code definition, the base sample 2007–2021 even becomes significant at the 1% level (Table 23, column 3), and in all regressions, the total effect of staggered boards during market shocks is positive. The results hold for different industry definitions.

Second, I use the ISS Governance database instead of Guernsey et al. (2022a) to determine the staggered board status of firms. It has many coding errors but is used frequently. Thus, I repeat the main regressions on the augmented sample, which already include the ISS Governance database. Table 25 reports the results. Despite all the incorrect entries, the analysis generates results similar to those using Guernsey et al. (2022a).

Third, I lag selected explanatory variables to address endogeneity concerns, as the market shock and lower valuations might cause changes in independent variables. Lagging the staggered board variable as well as the mod. E-index by one quarter helps to mitigate concerns about firms adopting corporate governance provisions in response to market shocks. Lagging all other firm-level controls by one quarter helps to mitigate the endogenous controls problem, too, as changes in the firm characteristics might be caused by different market conditions. Therefore, I conduct a series of analyses in which I first lag only corporate governance variables and then lag all firm-level controls by one quarter for both the base sample and the augmented sample. Table 26 presents the results. For the base sample (Table 26, columns 1 and 4) and the augmented sample up until 2006 (Table 26, columns 2 and 5), the coefficients on \(\mathrm{SB}\times \mathrm{Shock}\) are positive and highly significant and the total effect during market shocks is positive, too. For the augmented sample after 2007 (Table 26, columns 3 and 6), the coefficients turn economically and statistically insignificant. Thus, the findings are consistent with those of the main analysis.

In summary, the additional analyses confirm the main analysis.

6 Conclusion

This paper examines the effect of staggered boards on firm value during market shocks. The question of whether staggered boards are value-enhancing or value-destroying has long been debated. Most recent empirical studies suggest that the effect of staggered boards is homogeneous and idiosyncratic, depending on the firm and its situation. To further explore the circumstances under which staggered boards have a positive effect, this paper looks at market shocks and different firm characteristics that could influence the effect of corporate governance provisions.

Running various regressions, I find no evidence of a homogenous effect of staggered boards on firm value for all firms throughout the observation period. Instead, the total effect of staggered boards during market shocks depends on firm characteristics. I find that the association between staggered boards and firm value during market shocks becomes more positive and significant for R&D-intensive, small, and non-S&P 1500 firms. Especially for small non-S&P 1500 firms, the total effect is strictly positive and highly significant throughout the entire observation period. This effect does not occur to the same extent for large and S&P 1500 firms.

The results indicate that staggered boards are particularly value-enhancing for small non-S&P 1500 firms during market shocks. They are consistent with the argumentation of staggered boards’ proponents who argue that staggered boards benefit firms that are experiencing a high takeover threat. In this case, the takeover threat stems from low valuations, a small size, and an arguably greater information asymmetry between management and shareholders. Overall, the results are also consistent with the theory that the effect of staggered boards is heterogeneous.

Like other recent studies, this paper challenges proposals for mandatory staggered boards. Due to the heterogeneous effects, firms should instead be allowed to decide independently whether to stagger their boards.

Future work should explore further channels through which staggered boards affect firm value and performance. This could help provide clarity on the circumstances in which staggered boards increase or decrease value. It should also be examined whether this study can be replicated for other regions, provided data availability improves through the use of machine learning in addition to greater interest in broader geographic coverage of key databases.