Clawback provisions authorize firms to recoup compensation from executives upon the occurrence of financial restatements or executive misbehavior. The first clawback provision in U.S. federal law was Section 304 of the Sarbanes-Oxley Act of 2002 (SOX 304). SOX 304 requires CEOs and CFOs to return any earned incentive compensation following a financial restatement due to misconduct and puts the burden of enforcement on the Securities and Exchange Commission (SEC). After SOX 304, Section 954 of the Dodd-Frank Wall Street Reform and Consumer Protection Act introduced clawback rules in 2010 requiring firms to adopt and enforce clawback provisions themselves. While Section 954 is not yet law, many firms have voluntarily adopted clawbacks. As of 2013, about 45 percent of Russell 3000 non-financial firms had such voluntary provisions in place.

Most studies report strong evidence that firm-initiated clawbacks are effective in improving financial reporting quality and that investors respond positively to clawback adoption. These studies treat clawbacks as a binary choice: A firm either does or does not have a clawback. The common interpretation of these results is that the mere adoption of the clawback is the primary cause of the observed benefits.

The strength and uniformity of the effects attributed to clawback adoption is striking for two reasons. First, firms have considerable freedom to design clawbacks, likely leading to a huge variety of clawbacks. One would expect that variety alone to result in different economic outcomes. Hence, the improvements in reporting quality may not result from adopting a clawback per se but rather from adopting a clawback with a specific design.

Second, prior studies focus mainly on the role of the CEO’s decisions in creating a risk of inaccurate reporting rather than on factors beyond the CEO’s immediate control. Firms that truly cared about their reporting quality would consider all such factors. Improvements in financial reporting may be interpreted as resulting not only from the adoption of a clawback but also from the combined effect of adopting a clawback of a certain design and other simultaneous actions.

We seek to address these two concerns by analyzing clawback design and distinguishing between two types of risk that can result in inaccurate financial reporting. In contrast to earlier studies on clawbacks, all of which focus on whether a firm adopts a clawback, we analyze the cross-sectional variation in clawback design. We capture differences in clawback design by performing a comprehensive linguistic analysis of 4,464 clawbacks adopted voluntarily between 2007 and 2013. Drawing on the linguistic analysis, we create a Clawback Strength Index that captures the strength of clawbacks. The index is based on five components: (1) the compensation types that are subject to a potential recoupment, (2) the employee groups that are covered by the clawback, (3) the extent to which the clawback mandates its enforcement, (4) the time period for the clawback, and (5) the events that can trigger a clawback.

We find a substantial variation in both the focus and scope of clawback design. We then distinguish between clawbacks according to whether they are designed to be enforced (“strong clawbacks”) or not (“weak clawbacks”).

We argue that overall reporting risk consists of controllable and non-controllable reporting risk. Controllable reporting risk is the risk of a restatement caused by the CEO’s decision to misreport. Non-controllable reporting risk is the risk of a restatement caused by factors beyond the CEO’s immediate control, such as the reporting decisions of lower level managers or the strength of monitoring by boards or external auditors. We assume that by initiating a clawback, the firm may change the CEO’s reporting decision, thereby reducing the controllable reporting risk but not the non-controllable reporting risk. Moreover, we posit that controllable reporting risk decreases when clawbacks are strong. Yet firms that seriously care about their reporting may not rely solely on the strong clawback but also on measures to reduce non-controllable reporting risk.

We therefore advance two explanations for our findings and those of earlier studies: one that determines causal effects based on clawback design (the isolated clawback explanation), not clawback adoption, and another that includes clawback adoption as part of a firm’s efforts to reduce reporting risk (the broader reform explanation).

Because the CEO’s expected costs from misreporting increase with clawback strength, we predict improvements in reporting quality for strong clawback adopters. Furthermore, we also expect improvements in reporting quality when a firm simultaneously decreases non-controllable reporting risk.

Given that restating firms often change executives following a restatement, we expect to observe less CEO turnover when there are fewer financial restatements due to the reduction of controllable or non-controllable reporting risk or both. Thus, we predict a decrease in the likelihood of CEO turnover for strong adopters.

The distinction between controllable and non-controllable risk also has implications for CEO pay. If firms adopt a clawback but do not simultaneously reduce non-controllable reporting risk, we predict an increase in CEO incentive pay: Strong clawbacks penalize the CEO for both controllable and non-controllable reporting risk. Because the CEO’s expected costs from non-controllable reporting risk increase after the adoption of a strong clawback, the board may have to increase CEO incentive pay. Thus observing an increase in pay is consistent with the isolated clawback explanation. However, if firms adopt the strong clawback as part of a broader reform, they probably reduce the CEO’s expected costs associated with non-controllable reporting risk, and thereby the need for increasing incentive pay. Therefore, observing a decrease in pay would refute the isolated clawback explanation while remaining consistent with the broader reform explanation.

Using an econometric design that compares adopters of strong clawbacks and weak clawbacks, we find that adopters of strong clawbacks experience significant improvements in financial reporting quality, a decrease in the likelihood of CEO turnover, and a decrease in both total and incentive-based CEO pay. Moreover, we find similar results when comparing adopters of strong clawbacks with a sample of firms that did not adopt clawbacks, whereas we find that adopters of weak clawbacks experienced no such improvements over firms without clawbacks.

The changes in reporting quality and CEO turnover can be reconciled with both the isolated clawback and the broader reform explanations. Strong clawbacks, either alone or as part of a broader reform, are responsible for the documented benefits. However, only the broader reform explains the decrease in CEO pay. This explanation implies that our findings reflect the effects of a broader reform rather than just clawback strength.

Our study provides the first evidence of the different economic consequences for firms that voluntarily adopt clawbacks. Not all clawbacks are the same: Some lead to significant economic benefits, while others do not. An important implication of our study is that one may interpret the results of prior clawback studies with more caution, as they attribute all beneficial effects of clawbacks to their mere adoption. This overlooks the variation in clawback design and the notion that governance mechanisms interact in potentially important ways, and thereby ignores the benefits of broader governance reform.

Our findings also have implications for shareholders and regulators. Given that the mere adoption of even strong clawbacks does not necessarily improve financial reporting, the impact of mandating clawback provisions under Section 954 of the Dodd-Frank Act may be overstated.

This post comes to us from professors Michael H.R. Erkens and Ying Gan at ERASMUS School of Economics and B. Burcin Yurtoglu at WHU – Otto Beisheim School of Management. It is based on their recent article, “Not All Clawbacks Are the Same: Consequences of Strong vs. Weak Clawback Provisions,” available here.