Regulatory failures during the crisis resulted at least in part from a lack of sufficient incentives for bank examiners to act aggressively to prevent excessive risk taking by banks. While banker pay may have been too high-powered—too focused on shareholder value and insufficiently sensitive to potential losses, which would ultimately be borne by taxpayers—bank regulators’ pay was not high‐​powered enough and therefore, ironically, also insufficiently sensitive to potential losses to taxpayers.

Bank regulators are not paid for performance. They are civil servants paid a fixed salary that does not depend on whether their actions improve banks’ performance, protect banks from failure, or increase social welfare. In fact, trying to curb risk taking at a bank may be personally very costly for a bank regulator. Without a larger upside than what civil service compensation offers, regulators too often do the rational thing and play it safe, shying away from confrontation over potentially ill‐​advised bank policies.

To create better incentives, we propose that regulators, specifically bank examiners, be compensated in part with periodic bonuses tied to the value of bank debt and equity, as well as a separate bonus linked to the timing of the decision to take over a failing bank. Giving examiners a stake in bank performance, both upside and downside, will improve their incentives to act in the public interest. In contrast to most other types of bureaucratic functions, objective metrics exist to help measure the alignment of bank regulators’ activities with the public interest. With the right mix of banks’ debt and equity securities, public trading prices for these securities serve this purpose. A pay‐​for‐​performance culture therefore offers special promise for banking regulation as compared to other areas of regulation or government bureaucracy.

Because examiners have shown a bias toward nonintervention, we propose a debt‐​heavy mix of bank securities so that regulators bear the downside risk of nonintervention. To address insider trading and government ownership issues, we propose that regulators hold “phantom” securities whose payout is linked with actual bank debt and equity prices.

Though we do not discount the value of public spiritedness and reputation as inducements toward conscientious regulation, regulators’ dismal performance in the recent financial crisis makes us skeptical that public‐​spirited motivations are sufficient incentive. At scores of banks, examiners and other regulators were well aware of operational deficiencies and excessive risk taking several years before those banks failed. But regulators stood still in the face of this information. Instead of demanding corrective action by banks, examiners continued to rate risky institutions as “fundamentally sound.” Washington Mutual (WaMu), the largest bank to fail in U.S. history at the time, enjoyed a “fundamentally sound” rating until six days before its collapse. This regulatory failure was not a result of insufficient information or attention on the part of regulators. In the three years before WaMu’s collapse, examiners spent over 100,000 hours over 400 days inspecting its assets and operations. Despite the wealth of information, examiners failed to do the heavy lifting required to stop the excessive risk taking. The WaMu example is hardly unique.

Adding or subtracting examiner pay based on bank capital costs incentivizes regulators toward striking a socially optimal balance between increasing bank values and credit and reducing the costs of bank failure. This could mean more or less regulation, depending on the bank and the circumstances. For instance, examiners with pay linked to bank debt may pursue a more interventionist approach in some cases, since they bear some of the losses arising from the socially inefficient risk that exists on their watch. Regulators incentivized to worry about losses to taxpayers may be more diligent in their supervision of bank assets and management, may be more aggressive in assuring that corrective recommendations are implemented, may encourage or require changes to bank balance sheets, and so on.

Similarly, examiners who gain from increases in bank values (for example, by holding phantom bank stock) may take steps to make the examination process more efficient, to get the amount and type of disclosures right, and to encourage valuable lending and risk taking.

While we leave it to agency heads to develop optimal compensation practices over time, even small steps in the direction of our proposal could have large effects on the efficiency of banking regulation. The need to incentivize regulators is especially important after the 2010 Dodd‐​Frank financial reform legislation, whose say‐​on‐​pay provision is likely to generate even higher‐​powered incentives for managers to maximize shareholders’ private interests. High‐​powered bank CEO incentives require a corresponding impetus for regulators to proactively constrain bank risk taking.

Regulators’ Pay and Its Discontents

We are not the first to point out the problems with the standard pay structure for bureaucrats. Four decades ago, Gary Becker and George Stigler published a seminal article arguing for incentive pay for the enforcement of laws. Their suggestions were perhaps ahead of their time. Not until nearly 20 years later did performance pay for CEOs become common practice. It would likely have been something of a stretch to adopt performance pay for government agents before private sector actors.

Our incentive compensation proposal borrows not only from the neglected economics literature of the past. It also finds hope in changed pay practices for government officials implemented in the last few years. The Obama administration has dramatically increased regulators’ salaries. According to public records, the number of federal government officials earning six‐​figure salaries has skyrocketed. In addition, bank regulatory agencies have begun using bonuses ostensibly tied to performance. During the period 2003–2006, three regulator agencies—the Federal Deposit Insurance Corporation, Office of Thrift Supervision, and Office of the Comptroller of the Currency—paid out nearly $20 million in retention and performance bonuses to bank examiners and other regulators. In 2006 alone, the FDIC gave bonuses to 2,000 bank examiners.

While perhaps a step in the right direction, those modest moves toward performance pay for regulators are less than ideal. Ex post bonuses are not likely to yield incentives as high‐​powered as our approach, which relies on ex ante incentive contracts tied to outside metrics. Prior to the 1990s, CEOs routinely received cash bonuses and yet pay and performance were not as tightly linked as when stock and stock options came into use. To the extent that ex post bonus payments are discretionary, they allow for the intrusion of nonperformance‐​based criteria, such as favoritism, political affiliation, and so on. The linkage between bonuses and conduct that maximizes social welfare may therefore be tenuous. Ex post bonuses are also likely to be one-sided—that is, paid in good times but not recouped in bad times. This is likely to bias regulation in a particular direction.

Regulator Pay and the 2008 Crisis

There is widespread agreement that regulators failed to act aggressively enough during the recent financial crisis. The problem was not one primarily of access to information, lack of expertise, or resource constraints. Reviewing regulators’ performance following bank failures, regulatory agencies’ inspectors general reached the same conclusion: regulators did a satisfactory job of identifying problems well in advance of failure, but they failed to act aggressively enough to remedy the identified problems. The problem, in our view, was that regulators did not have the right incentives to turn their recommendations into actual reforms of bank policies.

Regulatory failure | The examination process has two broad goals: review of the quality of bank assets, with special focus on the bank’s most important assets, its loans; and analysis of the bank’s financial condition and the quality of its management and operations. Examiners enjoy wide discretion as to the volume of loans reviewed, the nature of the examination, the time spent on each analysis, and the consequences of the examination results. Examiners make local judgments about the credit quality of each asset. After discussion with loan officers and bank managers, examiners make final determinations (effectively unreviewable) about how to classify particular loans for input into a final supervisory rating. Examiners also review loan portfolios as a whole for issues such as concentration risk, violations of legal rules, and deviations from bank loan and underwriting policies. They assess the behavior and impact of subsidiaries and affiliates, risks from litigation, the costs and benefits of off‐​balance‐​sheet activities, and the activities of insiders. Based on this process, examiners determine the bank’s CAMELS rating, which is the single metric used by regulators to capture bank safety and soundness. (“CAMELS” is an acronym for Capital adequacy, Asset quality, Management, Earnings, Liquidity, and Sensitivity to market risk.).

Regulators have tremendous power to influence bank decision‐​making. Much of the actual power resides with bank examiners in their conduct of bank examinations. For example, the decision to change a bank’s CAMELS rating from 2 to 3 (moving the bank from “fundamentally sound” to indicating “some degree of supervisory concern”) is largely if not entirely within the discretion of the bank examiner. Because any regulatory intervention depends on examiners to identify problems and pursue initial ratings downgrades, effective incentives for examiners to act are crucial for optimal regulation. Examiner passivity, by contrast, effectively insulates a troubled bank from higher‐​level scrutiny and corrective sanctions.

Reports by inspectors general of the Treasury Department conclude that regulators did not do enough to prevent multiple banks from taking excessive risk and failing. Although acute funding constraints were a precipitating factor for many bank failures, this shock was not sufficient to explain bank failures. One report explains, “Although the deterioration in the bank’s financial condition was severe in 2008, the underlying risks were evident in the preceding years.” The consensus seems to be that if regulators had been more aggressive, hundreds of billions in losses could have been avoided.

In general, regulatory failures fell into two broad but discrete categories that correspond to the supervisory functions. The first category is the failure to adequately inspect and supervise bank risk taking during “good” times—that is, periods without financial stress. We might think of this as a failure to do adequate preventive medicine. The failure reports describe many instances in which the regulators did not ensure compliance with basic risk policies and/​or restrict certain types of risk taking. For instance, regarding the failure of IndyMac in 2008, the inspector general of the Treasury Department concluded that “examiners did not identify or sufficiently address the core weaknesses that ultimately caused the thrift to fail until it was too late.” As noted above, problems often resulted from the failure to deploy regulatory tools as banks took increasingly large and risky positions.

The second category is the failure to react to signs of distress and intervene quickly enough to prevent further damage. The $2.5 billion collapse of NetBank illustrates. According to the Treasury Department inspector general, the Office of Thrift Supervision “did not react in a timely and forceful manner to certain repeated indications of problems.” A similar lapse preceded the $2 billion failure of ANB Financial. The regulator—the Office of the Comptroller of Currency—“did not issue a formal enforcement action in a timely manner” after the bank began to suffer losses and experience distress.

The failure reports show that, in both categories of supervision, regulators engaged in more box‐​checking and paperwork than aggressive oversight. Bank examiners did the important work of assessing bank assets and risk. They saw deficiencies and recommended changes, but then never followed up to see if those changes were implemented. For instance, WaMu’s regulator did not “formally track the status of examiner recommendations and [required] corrective actions.” Another (typical) report concluded:

We found that bank management did not effectively implement key examiner recommendations over several examination cycles regarding such controls as loan‐​to‐​value limits, interest reserve policies, stress testing and establishing meaningful concentration limits, and maintenance of a sufficient [allowance for losses] and adequate capital structure. (Emphasis added.)

This same phenomenon recurred with shocking frequency in the recent bank failures.

Regulators’ incentives | Why would examiners, who repeatedly identified problem areas, continue to rate WaMu and other banks so highly in the face of obvious shortcomings in their business models and practices? Why did examiners err so egregiously on the side of nonintervention, in the face of specific policy guidance to the contrary? The answer is incentives.

Like everyone else, bank examiners maximize according to the incentive structure in which they find themselves. Bank examiners are paid almost entirely in fixed salary that varies primarily by seniority. Examiners also cannot easily be terminated. They enjoy the special job security fashioned by the civil service rules. This job security may make some sense. With their fixed salaries, if examiners could be terminated for poor performance, they might be extremely risk averse. For example, if a bank failure on an examiner’s watch significantly increased her risk of termination, the regulator’s incentive would be to ensure that the bank was not taking much risk. Though good for the regulator, the social cost from reduced credit availability and lost bank profits might be quite high. Reduced job security might also subject examiners to political pressure for doing their jobs too well. Regulated banks might be able to bring political pressure to bear on conscientious regulators unwilling, say, to allow a failing bank to continue operating or to permit a bank’s excessive risk taking. Job security therefore reduces counterproductive risk aversion and the risk of political capture, giving examiners discretion in applying regulation, perhaps in ways that improve social welfare.

But without additional incentives, the civil service rules may also create perverse incentives by insulating regulators too well from the consequences of their job performance. With pay delinked from an objective performance metric, regulators may naturally focus on bureaucratic tasks with observable outcomes, rather than on more aggressive and costly actions with more complex and less transparent cause‐​and‐​effect relationships.

Performing the examination and filling out examination reports is entirely within the examiners’ control. This output is subject to objective performance metrics (e.g., is the report completed on time and in a competent manner?). And reports alone are unlikely to generate collateral costs for examiners. In contrast, aggressive follow‐​up enforcement is likely to raise the personal costs to examiners significantly, with little or no personal benefit. As the work moves from investigation to persuasion—both of higher‐​ups and the regulated party, each of which may push back strongly—costs for regulators will rise. Regulators interested in not appearing before congressional committees, defending budgets, and being forced to testify in court would likely err on the side of regulatory restraint, especially when they do not capture the upside from aggressive regulation and do not bear much of the downside cost of laxity.

There is also the revolving door problem. Some regulators are bound to get some of their expected compensation from future employment with regulated banks. These banks may prefer as future employees those examiners who show diligence in their work but passivity in the face of bank interests or pressure.

Examiners may also fear making a mistake by restricting the lending of a seemingly successful bank. This problem may be exacerbated by the fact that examiners routinely work with the same bank for extended periods. They often go to work every day at the bank they are examining. While it is possible that familiarity breeds contempt, the opposite effect—akin to the Stockholm syndrome—may also skew regulatory decisions, especially where actions require confrontation.

Moreover, the relative secrecy surrounding bank examinations may also encourage regulatory inertia. Secrecy no doubt plays a useful role in encouraging bankers to be forthcoming with their examiners. Secrecy also insulates banks from the possibility of public overreaction to negative assessments from bank examinations, thereby avoiding the runs that deposit insurance and banking regulation were meant to cure. At the same time, however, secrecy also insulates examiners and the examination process from public accountability. When the Securities and Exchange Commission or the Environmental Protection Agency issues an order or takes other regulatory action against a violator, that action attracts public scrutiny. Failure to act in the face of egregious circumstances similarly attracts public attention. While public perception may not always be a useful metric for evaluating regulatory action, at the least it forces regulators to explain their actions—or inactions. Bank supervision, by contrast, is largely free from this accountability because of the secrecy of bank examinations.

Examiner passivity could be deterred by the loss of reputation, money, or other benefit because of the examiner’s failure to act against a bank that later collapses. But no such deterrent currently exists. If a bank fails, there are multiple causes to which blame can be assigned. In contrast, there is only the examiner to blame if reports are not accurately completed and done well. Under existing incentives, examiners might naturally conclude that their job is well done simply by accurately describing problems and bringing them to the attention of management and senior regulators. They have no stake in doing more. Job and salary security reduce incentives to do “good” work, however defined, since the consequences of “bad” work are reduced.

To be sure, many regulators value doing the right thing and serving the public interest. But given the ambiguity of these terms and the potential for rationalization, the absence of monetary or reputational rewards or sanctions means examiners care less than they would in the presence of more high‐​powered incentives. We do not doubt the honesty or good faith of the regulators who felt that they were doing the best they could do. We simply observe that regulators are influenced in ways beyond their ken, just as we all are. They respond rationally to the incentives they face, and can rationalize their conduct to fit to those incentives. We propose to change the incentives.

The Structure of Regulator Pay for Performance

The regulatory laxity preceding the recent crisis involved two distinct types of regulatory failure: the failure to apply preventive medicine when times were good and the failure to act aggressively when a bank showed signs of distress. Our incentive pay proposal has two distinct components to address those separate problems.

Bank debt‐​equity portfolio | First, offering a bonus linked to a bank debt‐​equity portfolio would offer real‐​time market feedback and long‐​run incentives to the examiner regarding the bank’s risk taking and its potential rewards. This component would matter primarily during good times, while the bank is operating in the ordinary course. Its purpose is to incentivize preventive and remedial measures well before a bank approaches distress.

We consider two potential debt benchmarks:

• a subordinated debt security issued by the bank, and
• a credit default swap contract (CDS) referencing a junior debt obligation of the bank holding company parent of the regulated bank (BHC).

The prices of publicly traded subordinated debt securities and CDS contracts reflect the market’s best estimate of the risk of the bank’s default on its debt. Holding this risk‐​sensitive instrument gives the examiner a personal financial incentive to curb excessive risk at the bank. (Some amount of BHC equity should be included as well in order to guard against undue examiner risk aversion.) With both the debt and equity components, we suggest a relative performance approach, which would filter out the effect of industry‐​wide or market‐​wide price movements. As noted, the lion’s share of this “preventive medicine” component of incentive pay should be debt‐​based.

To encourage a medium‐ to long‐​term regulatory perspective, each periodic phantom debt‐​equity allocation would have a specified medium‐ to long‐​term maturity. At maturity, say three to five years after the initial award, the allocation would be cashed out at the then–market values of its underlying debt and equity components. With regular periodic allocations, the examiner would hold multiple tranches of phantom securities with staggered payouts, giving the examiner incentive to consider the long‐​term as well as short‐​term consequences of her regulatory decisions, and making short‐​term manipulations of securities prices an unattractive strategy.

The appropriate debt‐​equity mix in the regulator’s portfolio will depend on a number of factors, some of which will be specific to the regulated bank, to the regulating agency, to the particular times, and perhaps even to the individual examiner. We therefore make no attempt to offer firm prescriptions for the optimal ratio. The mix should induce regulators to care about bank profits but not at the expense of risk shifting to creditors. In the face of excessive risk, the negative reaction from debt markets should reduce the value of the debt component of the portfolio by more than any positive reaction from equity markets would augment the value of the equity component.

Takeover bonus | Our second component, the takeover bonus, becomes important as a bank approaches distress. The examiner would be eligible for a cash bonus based on the timing of her decision to take over a failing bank. Regulators have a number of reasons to wait too long before effecting a takeover. This bonus would ameliorate the problem.

Bank regulators are by statute tasked with the specific goal of minimizing losses to the Deposit Insurance Fund (DIF). The bonus could therefore be tied specifically to the ultimate losses sustained by the DIF at the resolution of the FDIC’s receivership proceeding. We suggest several approaches to estimating these losses ex ante in order to incentivize improved timing of bank takeover decisions.

The takeover decision requires special treatment for a number of reasons. First, it is the most difficult and drastic decision the regulator must make in her supervision of the bank. The regulator has a number of reasons for being reluctant to pull the plug on a failing bank. Regulatory capture and the Stockholm effect may dissuade the regulator from taking over the bank. Pulling the plug might also highlight the regulator’s past mistakes in not intervening more forcefully. At any given point, the regulator might prefer to wait and see, hoping the bank will turn itself around. As the recent financial crisis illustrates—like all others before it—regulators tend to err on the side of taking over too late rather than too early. This delay in the crisis exacerbated banks’ losses and the ultimate costs to the DIF. A resolution bonus would offer a direct incentive to make the right timing decision at a critical juncture.

Moreover, although the FDIC is charged by statute with the specific goal of minimizing DIF losses in its dealings with troubled banks, the FDIC is typically not involved in the takeover decision, which rests with each bank’s chartering agency or primary federal regulator. The FDIC takes over only after a bank has been declared insolvent and put into resolution. Because the other agencies do not have their own money at stake in the timing of the takeover, the incentives to wait and see may be overwhelming.

Market discipline from the regulator’s bank debt‐​equity portfolio may not be useful in optimizing the timing of takeover because of information asymmetry. Market signals are likely to be noisy as a bank nears distress. Optimal timing will depend to a great extent on fine‐​grained private information, which is available only to the regulator and is constantly being updated once takeover becomes a real possibility. A one‐​time bonus distinct from any market assessment of the decision is therefore advisable.

Because of the importance of minimizing DIF losses, low losses should trigger high bonuses and vice‐​versa. In theory, then, if an examiner were to put a bank into resolution at time T1, and as a result the FDIC losses were 100, the examiner would get a larger bonus than if the decision were made at T2 when the FDIC losses would be 200.

Implementing this simple idea may not be straightforward. First, if we could discern the counterfactual losses that would have occurred at T2, then the calculation would be simple. But, of course, a takeover at T1 makes it impossible to know the counterfactual T2 outcome. Second, it is entirely possible that a conscientious examiner might decide, given the information available to her at T1, to wait until T2 for more information. Important developments concerning the bank’s prospects—the direction of certain asset or lending markets, for example—might be worth waiting for. Perhaps paradoxically, the more uncertain are the bank’s prospects, the more value there is in waiting. Third, even if a regulator made an (ex ante) optimally timed decision to resolve a failed bank, disposing of the bank’s assets may take several years. That process will affect the ultimate DIF loss figures and will not be under the examiner’s control. So a “final” resolution will be hard to predict at the time of takeover.

Despite these seeming hurdles, a resolution bonus may still offer important motivation for a regulator to act promptly in putting a bank into resolution, as compared to the current compensation system. The timing of the takeover will no doubt have an important effect on the severity of DIF losses, and warning signs in terms of bank characteristics and practices that lead to large resolution losses are not so mysterious. Researchers have identified factors that correlate with increased losses and estimated the economic magnitudes of these effects. The findings generally comport with common intuition. Bank asset composition and quality affect ultimate losses, as does liability structure. For example, brokered deposits—“hot money” aggregated by brokers seeking higher yields—are positively associated with high‐​cost bank failures and shorter time to failure. The same is true for real estate owned and loans past due. Uncollected income—basically, nonperforming loans—correlates with high‐​cost failures. Local economic conditions also matter. State personal income growth and the health of the local banking industry are negatively correlated with FDIC losses, while in‐​state bankruptcy growth and the unemployment rate are positively correlated with FDIC losses.

The depth of existing research strongly suggests that a resolution bonus algorithm could be constructed to both guide and cabin regulators’ discretion as to the timing of a bank takeover. Some trial and error would be involved in optimizing the bonus structure in pursuit of minimizing DIF losses, but with learning and experience it might be possible to design a fully automated system in which market and other data are incorporated into the bonus algorithm.

Short of that, agencies can develop a mechanism for estimating what losses would have been had the examiner not acted when she did. For example, post‐​mortem reports, like those described above, could be helpful. The inspector general of the FDIC could estimate losses at hypothetical future intervals had the examiner not taken over the bank when she did. These reports could deploy a mix of economic models, learning from past failures, and expert opinions from inside and outside the regulatory agency.

Could the takeover bonus induce the regulator to act too hastily in seizing a bank? Premature takeover is no more desirable than waiting too long; both destroy value. The regulator’s debt‐​equity portfolio would ameliorate the problem to a great extent because it gives the regulator a financial stake in the bank’s recovery. The key is to scale the takeover bonus relative to the potential value of the regulator’s debt‐​equity portfolio so that she neither permits futile gambles for solvency nor pulls the plug too early. Agencies might even implement non‐​trivial penalties for a takeover decision if the DIF ultimately suffers no losses because the examiner pulled the plug too soon.

Even if predicting ultimate FDIC losses is not an exact science for the regulator making the takeover decision, neither is the loss assessment inscrutable. A lack of research or analysis is unlikely to be the reason why regulators have been too slow to pull the plug on failed banks. Empirical studies—many done by the FDIC—and post‐​mortem reports offer regulators a wealth of research to support the goal of minimizing DIF losses. Instead, regulators may simply need better incentives to get it right. A well‐​structured bonus may help ameliorate this problem.

Qualifications and Objections

Incentive structures may sometimes generate not only desired outcomes but also some that are unintended and undesired. Our regulator incentive pay proposal is no exception. We do believe, however, that potential problems either can be overcome or are not sufficiently serious to preclude the experimentation along the lines we suggest.

Crowding out the public interest | Some might object that incentive pay is fundamentally inconsistent with public service. Financial rewards for “success” might change the public‐​regarding culture within regulatory agencies; financial incentives may crowd out the public spiritedness that would otherwise motivate employees. Instead of diligent altruistic service to the public, regulators and other agency employees might begin to view their roles in terms of market exchange. Regulators desiring higher compensation would pursue the proffered financial rewards, while those who value leisure might feel free to work less and forgo the rewards for diligence. Once diligence has been priced, perhaps some regulators will slack.

In addition, the type of person that chooses to be a bank examiner could change. Regulators have employment choices, and their choice to be regulators likely derives at least in part from their interest in public service. This public spirit is an important regulatory asset and should be husbanded. Public service motives might be displaced by financial motivations among new hires after implementation of an incentive compensation scheme. Eventually, the composition of the regulatory agency could change for the worse.

We do not discount these concerns. Social scientists have documented the crowding‐​out effect in experimental settings. We do not believe, however, the effect is necessarily universal or sufficiently well understood that experimentation with incentive compensation for regulators should be precluded. Moreover, as described above, the federal government has already begun experimenting with financial incentives for regulators. Enormous pay raises have been implemented at several executive agencies. Bank regulators have received bonuses for “good” performance during the crisis, although without any transparency or standards of which we are aware. These examples suggest that public spiritedness and financial reward are not mutually exclusive, at least up to a point. Our innovation is to rely on market pricing and specific observable outcomes to set bonus pay, instead of relying on fiat. Our approach makes incentive pay more transparent, more sensitive to performance, and less subject to political, class, gender, racial, or other biases.

As for selection effects, our incremental approach suggests that such effects from variability of pay are likely to be minor, at least in the early stages. More generally, the possibility that increased pay variability might change the mix of individuals opting to serve as examiners could be a good thing. Examiners screened by their commitment to the public interest were in fact insufficiently attentive to that interest during the recent crisis. Accordingly, attracting individuals interested in a variable pay‐​for‐​social‐​performance compensation structure may be a beneficial change.

Noisy proxies | A basic objection to our approach is that it simply won’t work. Our market‐​based incentives may be too blunt to be effective. Even after adjustments for relative performance, many important influences besides the regulator’s input will affect the market pricing of the bank’s debt and equity securities. Decisions by the CEO and senior officers, for example, will generally dwarf the regulator’s influence over the bank’s performance and the market price of its securities. If the regulator’s decisions have little impact on the bank or the price of its securities, the argument goes, then our scheme will have weak if any incentive effects on regulators.

Moreover, though private firms often extend option compensation to rank‐​and‐​file employees, and not just executives, there is some debate as to whether broad‐​based option plans create effective performance incentives. No matter how much harder they work, individual employees are not likely to be able to exert much influence on firm value. Given their individual small shares in their firms, they might rather free ride than increase their effort.

These potential obstacles to performance pay schemes in private firms should not deter us, however. Our situation is different. Regulators are not tasked with the general goal of increasing banks’ value. Their charge is far more specific and their incentive structure is more targeted. Regulators’ charge is to guard against excessive bank risk taking, and our debt‐​heavy portfolio of phantom bank securities focuses regulators on that task. In a well‐​run bank that does not incur excessive risk, it may be true—as with rank‐​and‐​file employees in private firms—that examiners’ ability to affect the value of the bank’s securities and their own debt‐​heavy portfolios is weak or non‐​existent. But that is as it should be. The regulator has only a minor role to play at a bank that is not pushing the risk envelope. Moreover, in that situation, the costs to the government of performance incentives are low since the market value of the bank’s debt will probably not move much. The bank’s debt will enjoy a consistently low risk premium.

However, in the opposite scenario, when a bank pushes the risk envelope and the market value of its debt declines, examiners have personal financial incentives to respond. This is the situation where performance incentives cause the regulator’s self‐​interest to correspond with social welfare interests, inducing the regulator’s vigilance. If that situation never comes to pass, all the better. It may be that, especially during good times, regulatory action has little effect on most banks’ value or the value of most banks’ securities. However, when a bank strays, prompt and effective regulatory action may be critical to avoiding large losses. For this bank and this regulator, the incentives will matter.

Conclusion

There is no reason we can think of why bank regulators should not be paid for performance. The crucial issues are whether one can identify what “good” and “bad” performance are, whether contracts can be written ex ante that operationalize these metrics, and whether the potential negative effects from introducing a pay‐​for‐​performance culture for regulators outweigh the potential social welfare gains. We have argued that bank regulation is an area where there are readily available metrics, where plausible contracts or payment schedules could be devised, and where the potential for crowd‐​out or other downsides from incentive pay are limited.

Accordingly, we propose that bank examiners be paid in part with a mix of debt‐​heavy incentives linked to bank equity and debt values. This pay should represent a substantial but not dominant part of examiner pay, should be paid out over a number of years, and should adjust in order to maintain incentives aligned with the regulatory mission of ensuring that bank risk taking is aligned with the social welfare. A separate takeover bonus would encourage examiners to make bank takeover decisions optimally to minimize DIF losses.

Although seemingly radical, our proposal is consistent with recent moves by regulators to pay bonuses for good work and to generally increase the quality and efficiency of regulation. It is also consistent with laws and academic proposals to alter bank CEO pay to take greater account of the social component of bank losses. Our contribution is to merely point out that regulator incentives are an overlooked but crucial factor affecting bank risk taking, and that improving the social performance of banks and the banking system requires a consideration of the incentives not only of bank CEOs but also of bank regulators. Insofar as we can improve the efficiency of government regulators, we need to worry less about the structure of private incentives, which are further from the control of government.