Bank distress during the recent financial crisis caused significant damage to the real economy. But simply improving corporate governance does not guarantee that banks will be safer in the future. Other measures are also needed
Corporate governance is the practice of shareholders exercising control over managers so that they act in shareholders’ interests. In non-financial firms, this maximises firm efficiency. Such efficiency effects also exist in banks. For example, banks that face more active takeover markets are more cost-efficient.
Unlike non-financial firms, bank operations have another relevant dimension besides efficiency: risk. Banks are prone to risk-taking, due to:
■their high leverage,
■limited creditor market discipline (a consequence of deposit insurance and too-big-to-fail guarantees), and
■the ability to rapidly and opaquely increase the risk of assets by changing the investment strategy.
Moreover, large-scale (‘systemic’) bank failures may negatively affect the real economy. So we are interested in whether corporate governance can not only improve bank efficiency, but also reduce bank risk.
On this, the literature warns that better corporate governance, by itself, is unlikely to make banks safer. In fact, the opposite may occur. Bank shareholders benefit from greater risk, because their payoffs are protected on the downside by limited liability. Improving corporate governance (in the Anglo-Saxon sense of giving shareholders more power) may make banks riskier. For example, there is evidence that higher institutional ownership (a proxy for shareholder power) increases bank risk, and banks with more independent directors on boards suffered larger losses during the crisis.
So how to align bank corporate governance with financial stability?
Higher bank capital increases shareholders’ skin-in-the-game, making them more risk averse. While higher capital is undoubtedly useful, it is important to realise its limitations. First, while the cost of bank capital appears modest in steady state, the cost of building up capital quickly may be significant. This makes bank capital scarce during and after recessions. Second, banks often obtain new capital from passive shareholders (such as preferred equity holders or institutional investors), who are unable to influence bank risk taking. Then, higher bank capital provides risk absorption capacity, but does not correct risk attitudes.
Higher capital can be supplemented by CoCos (contingent capital): debt securities that convert into capital at the early stages of bank distress. CoCos are appealing because of their lower cost (similar to that of junior debt), ‘prepackaged’ bail-in of shareholders and CoCos holders, and the possibility of a punishing dilution of bank shareholders upon conversion, which may give them powerful incentives to contain risk. However, CoCos have not been tested in crises, have unsettled design features (e.g. market-based vs. regulatory conversion triggers). Most critically, it is unclear whether the market for CoCos is deep enough to accommodate their wide use.
There are indications that better risk management (measured by the ratio of CRO to CEO compensation, but understood broadly as appropriate budgeting and organisational centrality of the risk management function) reduces bank risk. However, the strength of risk management differs substantially and persistently across banks. This prompts the question of how to improve bank risk management.
Unfortunately, it is notoriously difficult to ‘impose’ better risk management through regulation. Direct tools, such as requiring banks to spend a given fraction of assets on risk management, may only lead to box-checking. A better approach may be to strengthen market discipline by requiring banks to undertake enhanced and timely disclosure of risk strategies and positions, including assessments of tail risk. This should include clearer risk communication within the bank and to the regulators, and to the public and the counterparties (Basel Committee 2014). Still, given the complexity of modern banks, one can be skeptical whether it is possible to design binding disclosure structures that would be sufficiently comprehensive and meaningful to have a material impact on market discipline.
Another avenue to improve risk management is to make banks risks ‘simpler’ by restricting the types of risk that banks can take. For example, imposing loan-to-value (LTV) limits on mortgages would remove hard-to-measure risks related to tail events in housing markets. Similarly, narrowing down banks’ financial market activities alleviates the challenges of responding to boom-bust dynamics in financial markets.
High-powered remuneration aligns the incentives of shareholders and managers, but may have unintended effects in banks. It may transmit shareholders’ high risk preferences to managers, and give managers incentives to take ‘tail risk’ to boost short-term return, or to herd to ensure on-par performance with peers. The literature confirms that high-powered remuneration led banks to take more risk and then suffer losses during the recent crisis.
The regulation of pay within firms is a contentious issue. The traditional policy approach is not to meddle within firms, but to ensure conservative shareholder incentives, allowing shareholders to put in place appropriate internal incentives themselves. However, for banks, it may be impossible to make shareholders sufficiently conservative. Moreover, even when shareholders are conservative, additional distortions may prevent them from putting in place low-risk internal incentives.
We can point to two reasons for such a failure of internal governance. The first is externalities across banks. Hiring talent is like an arms race, where each bank has to outbid others, including by offering employees more capacity to take risk. In equilibrium, employees may be given too much scope to take risk. The second reason is time inconsistency problems. When everyone expects shareholders to be conservative, a bank can attract cheap debt and capture its customers in long-term banking relationships. But this capture of creditors and customers is self-defeating, as it may give banks irresistible incentives to renege on them and increase risk.
In designing the regulation for managerial pay, some have proposed linking compensation to measures of bank risk. But such schemes are complex and open to the ‘Lucas critique’ that they may distort the pricing of risk, making banks appear safer than they are. Others have proposed pay claw-backs from managers that took too much risk. But claw-backs are hard to implement. When risk cannot be measured objectively, a manager can always claim that, at the time, the risk appeared not as acute as it has turned out to be. A blunt tool to regulate managerial pay would be to restrict performance-related elements of bank compensation (especially short-term performance-related elements, such as bonuses). Beyond affecting incentives within banks, such a restriction may also direct risk-loving talent away from banks to employers that have lower systemic importance and can better absorb losses (e.g. hedge funds). However, such regulation risks reducing risk to suboptimal levels.
Alternatively, one can increase the penalty for misbehaving managers by forcing bank managers to invest a stake in the bank without the protection of limited liability, putting the manager’s personal wealth at stake in the case of bank failure. While such a stake is unlikely to be large enough to prevent the bank failure (many CEOs of failed banks had significant stakes in their banks prior to the crisis), it would – to some extent – increase managerial risk aversion.
Besides shareholders, bank stakeholders include creditors and taxpayers (who may need to fund a bailout, or suffer from the impact of a financial crisis on the real economy). Bank shareholders are relatively minor (in terms of notional exposure) stakeholders, who nevertheless exercise most control over the bank, and can impose on a bank their high risk preferences that are at odds with the interests of other stakeholders.
To deal with this disparity, some suggest extending bank directors’ and managers’ fiduciary duty beyond shareholders. To some extent, stakeholder control of banks is already used in the real world:
■The transfer of control to the deposit insurer underpins the prompt corrective action regime of bank intervention in the US and similar bank resolution regimes in other jurisdictions. The limitation of such schemes is that the transfer of control may occur too late – when the bank is already in distress – making it difficult to apply corrective action to large, systemically important institutions, when too big to fail considerations get in the way.
■In Europe, supervisory boards of banks often include employee representatives (such as in Germany’s co-determination system), or bank managers may have fiduciary duties to stakeholders other than shareholders. Such arrangements are often ineffective in reducing bank risk, because creditors and other stakeholders may lack the grasp of the operations of a complex modern bank, while fiduciary duties may be insufficiently well defined to hold up in court.
A way to make stakeholder control more effective may be to make it state-dependent: increasing as the bank approaches distress, but becoming substantial well ahead of distress. This would focus stakeholder control on situations when a de-risking of bank strategy is both needed and still possible. To avoid government ownership of banks (which is associated with inefficiencies such as directed lending), the control may primarily be exercised by bank creditors.
Clearly, implementing creditor control of distressed banks has challenges. Bank creditors may experience a collective action problem. A divided board (shareholders vs. other stakeholders) may be indecisive. Bank creditors may reduce risk too rapidly, damaging the bank’s long-term value, or be over cautious in lending, compromising the provision of credit to the real economy. Still, creditor control may be preferred over a situation where shareholders ‘gamble for resurrection’ to extract value out of a distressed bank.
Traditional Anglo-Saxon corporate governance mechanisms make banks more efficient, but not necessarily safer. Higher bank capital, better risk management and pay regulation, and some stakeholder control may help steer corporate governance towards financial stability objectives. But as our analysis has indicated, the design of policies that affect bank governance leaves many open questions: there are possible unintended effects and the ultimate effectiveness of such policies remains uncertain. Moreover, the power of corporate governance in reducing bank risk may depend greatly on the strength of supervision.
Sound corporate governance, by improving market discipline, can send informative signals to supervisors that can speed up and improve the precision of regulatory intervention. However, ultimately such action depends on the resoluteness of the supervisor. Risk control in banks helps improve financial stability. But it cannot depend on market signals alone, because of the existence of market failures, which often cannot be foreseen, and can be especially large during financial crises when corrective action is needed the most. For these reasons, corporate governance cannot substitute for strong supervision. It can at best provide a helping hand.
Luc Laeven is deputy division chief and Lev Ratnovski is an economist in the International Monetary Fund research department. The full version of this post originally appeared on the Vox website (Disclaimer: the views expressed are those of the authors and do not represent those of the IMF)
