Florian Hoffmann is Associate Professor of Finance at KU Leuven; Roman Inderst is Chair of Finance and Economics at Goethe University Frankfurt, and Marcus Opp is Associate Professor of Finance at the Stockholm School of Economics. This post is based on their recent paper, forthcoming in the Journal of Finance. Related research from the Program on Corporate Governance includes Rationalizing the Dodd-Frank Clawback by Jesse Fried (discussed on the Forum here).
The 2007-08 financial crisis put compensation practices in the financial sector on the agenda of financial regulation. On a supra-national level, the Financial Stability Board (FSB) adopted its Principles for Sound Compensation Practices in 2009 “to reduce incentives towards excessive risk taking that may arise from the structure of compensation schemes.” In particular, short-term oriented bonus schemes have been identified as a key factor contributing to excessive risk-taking by financial institutions. This triggered regulatory initiatives around the world to intervene in the timing of bankers’ incentive compensation. For example, in the United Kingdom bankers’ variable pay compensation is now subject to minimum deferral periods of 3 to 7 years, and can be clawed back upon severe underperformance for 7 to 10 years. These regulatory restrictions do not only apply to top-level executives, but a broad set of banks’ “material risk-takers:” E.g., for Barclays alone, the compensation packages of 1746 employees are currently affected by this regulation (see Barclays 2020 Pillar 3 report).
Our study analyzes the positive and normative effects of such regulatory interventions in the timing dimension of bankers’ compensation packages. The punchline of our theoretical analysis is that even if bankers’ laissez-faire compensation contracts are socially suboptimal, such policy interventions do not robustly help mitigate risks in the financial sector, and, potentially, even backfire. At a very high level, the fallacy of targeting compensation packages is that “wrong” compensation contracts are merely a symptom of distortions in the financial sector, but not their root. That is, whichever distortion has led bank shareholders to write contracts incentivizing their key risk takers to take excessively risky actions in the first place, it is still present if they face regulatory constraints on compensation design. Capital regulation instead directly targets the root of the key distortion towards excessive risk tolerance in the financial sector, excessive leverage fueled by bailout expectations. Accordingly, our study suggests a “pecking order” of regulatory tools, in which compensation regulation should be considered only if capital regulation is restricted.
But, why doesn’t deferral regulation work robustly? Wouldn’t bankers be more prudent if their compensation is longer at risk? Such an agent-centric view of compensation regulation is misguided for a fundamental reason, the Lucas critique: Compensation contracts are not exogenously set, but endogenously determined. It is bank shareholders (the principal) who design compensation contracts to incentivize their employees (the agent) to take particular actions. And the issue with any structural constraint on pay, like deferral or clawback provisions on bonuses, is that they allow shareholders to restructure unregulated dimensions of the compensation contracts they offer to their key employees. Concretely, if deferral and clawback requirements pertain to bonus pay, shareholders are free to adjust, e.g., fixed pay components (wages) and/or perturb the bonus size, as has been documented empirically (Colonnello et al. 2018). So, to assess the overall effect of such regulatory intervention on risk-taking, we need to know how shareholders adjust unregulated dimensions of the compensation package in response to structural constraints on the timing of pay.
The net effect of any form of compensation regulation on the level of shareholders’ profits is clear-cut: Any additional restriction on contracts must raise the cost of incentive compensation, which acts like a tax on shareholders. However, exploiting the tax analogy, it is not the level of the tax that matters for which action is ultimately chosen in equilibrium, but the marginal tax: Risk-taking becomes less attractive to shareholders if and only if the regulatory restrictions make risk-taking relatively more expensive to incentivize than more prudent actions. This effect is subtle and unclear without a concrete model.
To this end, we develop a parsimonious continuous-time principal agent model that allows us to endogenize compensation contracts with and without regulatory constraints. The model captures three central aspects of compensation regulation in the financial sector: First, the agent is a “key risk-taker”—as targeted by real world regulation—able to affect the survival of the entire financial institution. Second, to capture the concern of regulators that “bad bets by financial-services firms take longer than three years to show up” (Wall Street Journal, 2015), we assume that the banker’s unobservable action has long-lasting effects on the bank’s failure rate, which generates a central role for the timing of incentive compensation (cf., Hoffmann et al. 2021). Perhaps surprisingly, we show that in such a setting observing short-term bonus payout dates need not be indicative of high risk-taking incentives. The third aspect we seek to capture relates to the scope for regulatory intervention in bankers’ pay: Bailout expectations allow bank shareholders to finance risky projects with subsidized debt leading them to incentivize excessively risky actions from the viewpoint of society (see, e.g., Dewatripont et al., 1994). Hence, compensation regulation is not motivated by corporate governance concerns, i.e., that bankers do not do what their shareholders want (see, Bebchuk and Fried, 2010). The reason for regulation is precisely that bankers do what shareholders want, namely to take on too many risks.
This model allows us to derive concrete conditions for when deferral regulation indeed reduces risk-taking and when such regulation backfires. In particular, the effectiveness of regulation is determined by (a) the stringency of the mandatory deferral period, (b) the outside options of the targeted bank employees, and, (c) the information environment determined by how risk-taking affects bank defaults over time. Our analysis reveals that deferral regulation has a robust risk-reducing effect if and only if deferral periods are moderate and targeted risk-takers have lucrative outside options, e.g., due to high competition for their talents. Instead, it always backfires when regulatory deferral periods are sufficiently stringent, i.e., far away from the laissez-faire choice. The case of low outside options—which formally corresponds to a slack participation constraint—is most subtle: Then, the effectiveness of deferral regulation hinges on the specifics of the information environment, i.e., on how the unobservable risk-taking action affects observable performance signals in the form of the occurrence or absence of bank defaults over time.
This dependency of the regulation’s effectiveness on economic fundamentals makes clear that “one-size-fits-all” compensation regulation applied to a highly heterogeneous group of risk takers including bank executives, traders and risk-managers cannot be a robust tool to curb risk-taking. A successful implementation requires a large amount of detailed information allowing the regulator to fine tune the intervention to each targeted group of financial sector employees. However, before trying to micromanage compensation contracts, regulators should resort to a more robust tool to address excessive risk-taking arising from excessive leverage: Capital regulation. Intuitively, if shareholders need to contribute more equity, they are themselves more concerned about downside risks and accordingly incentivize their employees to avoid or better manage these risks. Arguably, shareholders are much better informed about the relevant fundamentals governing optimal compensation design for their key risk takers than regulators and, thus, are also better equipped to adequately account for downside risks in compensation contracts, provided they care about these risks. Still, if capital regulation is restricted, our analysis suggests that moderate deferral regulation can be a useful supplementary tool if targeted to those risk-takers that have lucrative outside options.
The complete paper is available for download here.