Yes, there is a good basis for concern that executive pay arrangements have contributed to excessive risk-taking during the run-up to the financial crisis. To be sure, other factors were clearly at work: the environment within which firms operated grew riskier due to asset bubbles generated by macro policies and global factors, and regulatory constraints on risk-taking and capital requirements were too lax. As financial economists generally recognize, however, for any given environment and outside constraints, the performance and risk choices of firms depend substantially on the incentives of firms’ executives. Unfortunately, rather than provide incentives to avoid excessive risk-taking, the design of pay arrangements in financial firms encouraged such risk-taking.
Of course, despite incentives to take excessive risks, some executives might have avoided doing so due to professional integrity, reputational concerns, or fiduciary duty norms. And some executives taking excessive risks might have done so due to their under-estimation of the risks taken. But economics and finance teach us that incentives often matter. Thus, to the extent that pay arrangements provided significant incentives to take excessive risks, the possibility that such incentives in fact contributed materially to the excessive risks taken in the run-up to the crisis should be seriously considered.
In fact, pay arrangements did provide substantial incentives for excessive risk-taking. Under the standard design of pay arrangements, executives were fully exposed to the upside of risks taken but enjoyed substantial insulation from part of the downside of such risks. As a result, executives had incentives to increase risk-taking beyond optimal levels.
The first feature of pay arrangements that generated excessive risk-taking incentives was the partial insulation of executives’ payoffs from effects on long-term shareholder value. Both bonus and equity compensation have had excessive focus on short-term results. Jesse Fried and I warned about this problem and its consequences in our 2004 book Pay without Performance: The Unfulfilled Promise of Executive Compensation. Under the standard design of pay arrangements, executives were able to pocket bonuses based on short-term results and were permitted to unload substantial parts of their equity incentives based on short-term stock prices. These arrangements provided executives with incentives to seek short-term increases in profits even when these came at the expense of piling up latent and excessive risks of an implosion later on.
Bebchuk, Cohen, and Spamann (2010) illustrate the problem through a case study of compensation at Bear Stearns and Lehman Brothers. We document that, notwithstanding the 2008 meltdown of the firms, the bottom lines for the period 2000-2008 were positive and substantial for the firms’ top five executives. These top executives regularly unloaded shares and options, and thus were able to cash out a lot of their equity before the stock price of their firm plummeted. The top executives’ payoffs were further increased by large bonus compensation during 2000-2007; while the earnings providing the basis for these bonuses evaporated in 2008, the firms’ pay arrangements did not contain any “claw-back” provisions that would have enabled recouping the bonuses that had already been paid. Altogether, while the long-term shareholders in these firms were largely decimated, the executives’ performance-based compensation kept them in decidedly positive territory. Bhagat and Bolton (2011) find a similar pattern – pre-crisis cashing out of large amounts of compensation by the CEO that exceeded losses suffered by the CEO from stock price declines during the crisis – for other large financial firms that had to be bailed out during the financial crisis.
The second relevant feature of pay arrangements, analyzed in detail in Bebchuk and Spamann (2010), was the insulation of executives’ payoffs from potential losses to contributors of capital other than the shareholders – such as bondholders, preferred shareholders, and the government as guarantor of deposits. Consequently, in considering choices that could impose large losses on such stakeholders, executives had insufficient incentive to internalize these potential losses. Thus, executives had insufficient incentives to avoid risk-taking that was beneficial for equityholders but whose potential consequences for bondholders and other stakeholders made them excessive on the whole. This second problem reinforced the incentives to take excessive risks produced by the first problem.
Going forward, these two problems can and should be addressed by improved design of pay arrangements. To address the first problem, pay arrangements should generally tie executives’ payoffs to long-term results (along the lines proposed in Bebchuk and Fried (2010) or otherwise). To address the second problem, executives’ payoffs should be tied not only to long-term results for shareholders but also (as Bebchuk and Spamann (2010) advocate) to long-term results for other contributors of capital to their financial firm. Had financial firms used such arrangements in the past, they would have avoided providing excessive risk-taking incentives.
While Fahlenbrach and Stulz (2009) (as well as follow-up work by Erel, Nadauld, and Stulz (2011)) did not find evidence supporting the hypothesis that pay structures have contributed to risk-taking, a substantial number of subsequent studies provide such evidence. Several studies find that risk-taking was associated with the sensitivity of the CEO’s wealth to return volatility (Chesney, Stromberg, and Wagner (2011), DeYoung, Peng, and Yan (2010), Gande and Kalpathy (2011) and Suntheim (2011)). Risk-taking was also identified to have been associated with the sensitivity of the CEO’s compensation to short-term earnings per share (Bhattacharyya and Purnanandam (2011)). Furthermore, consistent with the view that risk-taking incentives were generated by the tying of executive interest to results for equityholders but not for bondholders, risk-taking was found to be negatively correlated with inside debt holdings by bank CEOs (Tung and Wang (2011)) and positively correlated with CEOs’ equity-based compensation (Balachandran, Kogut, and Harnal (2010)). Finally, consistent with the view that executives did at least partly perceive the significant risks they were taking, Cziraki (2011) finds that insiders of banks with the highest exposure to subprime risk increased their equity selling starting in mid-2006.
While the above evidence is telling, it is worth stressing that the body of empirical work beginning to accumulate is unlikely to reveal the full extent of the contribution of incentives to risk-taking. While there is some variation that has enabled researchers to obtain the above results, the variation in the available data is insufficient to identify the full effects of flawed incentives. While some pay arrangements were marginally better than others in relevant respects, financial executives’ pay arrangements generally provided substantially more exposure to short-term results, and substantially less exposure to payoffs of bondholders and other stakeholders, than was necessary to provide optimal risk-taking incentives. Thus, it will be difficult for researchers to identify in pre-crisis data the full difference between actual risk-taking levels and those that would have been chosen under optimal pay arrangements.
The view that compensation incentives likely have an important effect on executives’ decisions has rightly led financial economists to provide strong support to providing executives with incentive compensation. This view also warrants serious concerns about the risk-taking incentives that financial firms provided (and to a substantial extent continue to provide) their executives. Giving these concerns the weight they deserve is important not only for understanding the recent past – which it is – but also for according adequate priority to the task of fixing pay arrangements. Fixing these arrangements to eliminate excessive risk-taking incentives would contribute significantly to maintaining financial stability in the future.