Fixing bankers’ pay: punish bad risk management, not bad risk outcomes

Misa Tanaka and John Thanassoulis.

Post-crisis, a number of jurisdictions have introduced remuneration regulations in order to reduce bankers’ incentives to take excessive risks.  The UK is pioneering the use of bonus clawbacks under which bankers are asked to pay back their bonuses if certain circumstances materialise at a future date.  In our latest paper, we show that clawback can encourage better incentives as long as bankers believe that they will be held liable for failures of risk management, and not simply for poor outcomes.  Having a transparent mechanism in place to apply clawbacks is therefore critical.  If bankers fear that clawback will be wielded too generally upon bad business outcomes, then it could end up making them excessively risk averse.

The idea of bonus clawback for bankers has been mooted soon after the onset of the global financial crisis (for example, by Raghuram Rajan’s FT article in 2008).  But what’s actually wrong with privately-agreed pay packages, and why might they incentivise excessive risk-taking from society’s perspective?  There are at least three different types of agency problems – or conflicts of interest between a firm’s management and its stakeholders – which might lead to remuneration contracts that incentivise excessive risk-taking:

  1. First, the bank’s executives may not adequately take the long-term interests of its shareholders into account. This agency problem between bank executives and shareholders can be solved quite simply through deferred equity-linked pay, ie by making sure that part of the bonus is paid in equity that vests at some pre-specified future date.
  1. But the executives rewarded in equity-linked bonuses may still have incentives to take excessive risks at the expense of debtholders if the debt market cannot observe the risks and hence cannot price them. This agency problem between bank executives and debt holders can, to some extent, be mitigated if banks regularly issue debt and so constantly return to the judgement of the debt market.  Indeed, large banks borrow on average $2.7 billion every time they return to the debt market, and they do so more than twice a quarter.  Even so, to prevent the executives from taking excessive risks at the expense of debt holders, it has been proposed that the executives should be exposed to the price of debt and credit default swap (CDS) premia, or be paid in Contingent Convertible bonds (Cocos) (for example, see Edmans and Liu (2011); Bolton et al. (forthcoming); Dudley (2014); and Haldane (2015)).
  1. Finally, even if the shareholders’ and debt holders’ interests are protected by the executives, the interests of society may not be. In the presence of explicit deposit insurance and the implicit possibility of government bailouts, there is also an agency problem between the bank executives and taxpayers and the deposit insurance fund.  In this case, higher risk taking does not lead to a proportionally higher cost of debt funding, and so risk-taking is effectively subsidised (see for example, Sowerbutts and Zimmerman 2015).

We argue that remuneration policies designed to correct for the first two agency problems – such as deferred equity-linked bonus and paying in debt – may still incentivise excessive risk-taking from society’s point of view if the third agency problem is present.  For example, if equity prices are themselves inflated by the implicit and explicit guarantees on debt and deposits, then linking pay to long-term equity value would still incentivise the executives to take excessive risks.   Similarly, if debt prices are distorted by such guarantees, then incorporating standard debt that is pari passu with other creditors in bank executives’ pay package would still incentivise the executives to take too much risk.

So, what could be done about the excessive risk-taking caused by implicit and explicit guarantees on bank debt and deposits, short of eliminating these guarantees entirely?

The first, ex ante route is to shift the performance metric used for determining the bonus away from measures that are purely linked to returns on equity (RoE), which could be artificially inflated by government guarantees.   In theory, the excessive risk-taking incentives could be corrected if the executives’ bonuses are linked to equity returns minus the value of implicit and explicit guarantees on debt and deposits, rather than the raw RoE measure.  Our analysis suggests the adjustment required could well be around 10% of the bank’s equity value.  However, it is not easy to implement this in practice in the absence of reliable measures for the value of implicit and explicit government guarantees for individual banks.

The second, ex post route is to expose the executives to the possibility of a financial loss through bonus malus and clawback: malus adjusts variable pay that has not yet vested, while clawback adjusts variable pay that has already vested.  Importantly, the effectiveness of malus and clawbacks depends on how bankers expect them to be implemented.  Good risk management ex ante doesn’t mean zero risk, and bad outcomes at times occur despite good risk management.  So a well-designed remuneration regulation should lead the executives to believe that they will be penalised proportionately for losses that occurred because of their poor risk management, but not for losses that occurred in spite of good risk management.  This is one reason why deferred-equity linked pay is not fully effective, as it links pay to risk outcome rather than risk management.  Indeed, we expect malus and clawback to induce appropriate risk taking incentives if bankers are made to believe that these will be applied proportionately according to the degree of failure in risk management.

But malus and clawbacks could work imperfectly on incentives if bankers believe that these could be applied in the event their bank suffers large losses, even if they themselves have conducted appropriate risk management.  On the one hand, if the executives expect to suffer large losses when their firm performs poorly, then they end up being excessively risk averse.  Examples of this effect would be bankers investing in safe assets when riskier projects that yield higher expected returns are available.  On the other hand, if bankers expect to suffer only small losses in the event of bad outcomes, then they end up being insufficiently risk averse.   Examples of this effect would be bankers investing in risky but relatively low return projects in order to exploit the benefits from government guarantees, even though they are inferior to safer alternatives from society’s perspective.

Thus, for remuneration regulations to induce the right risk-taking incentives, they need to link financial losses imposed on bank executives through bonus malus and clawback to ex ante risk management, rather than making them depend on ex post risk outcomes alone.  In practice, assessing the quality of risk management several years later will be no easy task, and requires at a minimum a clear mechanism for identifying individuals responsible for specific risk decisions, and transparent criteria for judging these decisions.  This requires documenting and recording the processes, models and analyses used for taking important risk decisions for several years to ensure that individuals are aware that they can be held accountable for their parts in specific risk decisions; and clear standards against which their decisions will be judged, such that they will not be held liable for poor outcomes that may occur in spite of good prior risk management.   If this cannot be done, then the use of clawback would need to be tempered to induce sound risk-taking incentives, and thus efficient investment allocation for society.

Misa Tanaka works in the Bank’s Policy Strategy and Implementation Division and John Thanassoulis works at Warwick Business School, University of Warwick.

If you want to get in touch, please email us at

Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.

Comments Off on Fixing bankers’ pay: punish bad risk management, not bad risk outcomes

Filed under Banking, Financial Stability, Microprudential Regulation

Comments are closed.