VoxEU Column Financial Regulation and Banking Global crisis

Leverage and limited liability: A toxic cocktail

There is little doubt that one of the main causes of the Global Crisis was excessive risk-taking by large international financial institutions. This column argues that the combination of very high leverage and limited liability continues to incentivise risky behaviour by bankers. Dealing with this problem requires the alignment of bankers’ incentives with those of society, rather than of shareholders. Deferred compensation in the form of contingent convertibles presents one promising strategy.

There is little debate that one of the main causes of the Global Crisis was excessive risk-taking by large international financial institutions. Most observers would also agree that much has been accomplished under Basel III to address the problem. Banks today are required to have more and better equity capital, they are required to prepare ‘recovery and resolution plans’ (RRPs), and they must finance themselves through debt instruments that are bail-in-able or can be converted into equity (contingent convertibles, or Cocos). Bank owners and their creditors thus have significantly more ‘skin in the game’ than before the crisis. But is it enough? Many are answering in the negative in view of the still historically low capital requirements imposed on banks (e.g. Baldwin et al. 2017). But another issue merits consideration – the combination of very high leverage and limited liability, uniquely typical of modern banking. This constitutes a toxic cocktail that continues to be a source of excessive risk-taking which needs to be explicitly confronted. That convexity in remuneration patterns encourages risk taking is well-known and, for the most part, intended. The circumstances and the numbers involved in the financial sector, however, pervert the purpose of limited liability and generate highly problematic incentives that persistently undermine regulatory efforts and endanger financial stability. This issue justifies generalising remuneration contracts with long claw-back clauses and bonuses paid in the form of vested high trigger Cocos.

The limited purpose of limited liability

If there is one lesson in economics and finance that has not been invalidated by the Global Crisis, it is that incentives matter. The excessive risk-taking that was observed prior to the crisis must have been in large part the result of key decision-makers not being provided with the right incentives to carefully analyse and balance the possible consequences of the risks they agreed to take. They therefore wilfully accepted more risk on behalf of their institutions than would have been privately or socially desirable. The post-crisis resistance of bankers – often the same individuals, it must be said – to regulatory efforts designed to make their institutions more resilient is indicative that distorted incentives are still largely present and at work.1

The purpose of limited liability is to encourage risk-taking in situations where the prospect of moderately positive rewards for the decision-maker can be swamped by the possibility of extremely large losses that are unlikely to be bearable by a single individual. Limited liability is necessary to promote the level of risk-taking required for a growing economy. However, it leads to excessive risk-taking and can hardly be socially justified in the opposite configuration, that is, when the principal secures – for ‘good’ – huge personal gains when a gamble is successful, but bears no or only little personal financial losses if it is not. This is particularly damaging if the risks he bears increases the probability of large losses for his institution (in the current or subsequent periods).

It can be demonstrated that in the presence of limited liability, a levered cash flow (first-order) stochastically dominates the unlevered underlying cash flow (Danthine and Donaldson 2014: Section 4.9.3). This necessarily implies that little restraint can be expected in a situation where highly levered gambles are presented to (even highly risk-averse) decision-makers. Admittedly, a losing gamble does impose significant non-financial costs, such as the loss of prestige or job loss. These non-financial considerations probably suffice when the stakes are reasonable, but not when the degree of leverage and the amounts involved are very large, as is often the case in the financial arena.

Even with a (highly contested) 5% leverage ratio, the leverage level in banking would only be a fraction of what is typically seen in the rest of the corporate world (Admati and Hellwig 2013).  And bankers’ bonuses regularly amount to multiples of the lifetime remuneration of individuals contributing to society in more obvious ways (Zingales 2015). What restraint in risk-taking can one expect from a trader when his gamble (with other people’s money) brings him a lifetime-level remuneration in case of success and, at the very worst, a job change in case of a negative outcome? And isn’t it too tempting for a CEO to initiate large gambles, even if it means putting his firm’s survival on the line, if the odds are not too unfavourable and success brings him a comfortable lifetime income, but only benign consequences follow failure (in particular, if the state is expected to come to the rescue).

Distorted incentives

The problem of limited liability biasing the decision-maker’s perspective is present in the case of shareholders. They benefit from the upside of risk-taking; they are not symmetrically penalised on the downside. This asymmetry is particularly acute in the case of highly levered institutions. For these institutions, the return on equity in good times is high, say, above 20%. The trade-off between high returns (if the risky gamble pays off) and a zero outcome (if it doesn’t) is particularly lopsided. The combination of limited liability and the highly leveraged nature of banking thus come with a natural propensity for high, and most likely socially excessive, risk-taking on the part of bank owners themselves[2]. Moreover, the external monitoring of complex and often opaque institutions such as large banks, by their shareholders as well as their creditors, is a difficult, if not impossible, challenge. For this reason, the task of properly balancing risks should not be left to outsiders. It is crucial that proper incentives are set at all levels of an institution.

But the ‘limited liability with high leverage’ problem arises even more forcefully in the case of bank managers and other internal stakeholders entitled to bonuses, notably traders. Clearly, a managerial remuneration scheme that depends exclusively on a bank’s current performance, and traders’ bonuses that are related to a single or a few decisions taken over a short horizon, constitute inappropriate incentives.

As argued by the authors of the Squam Lake Report (SLR 2010), what is primarily at issue is less the level than the structure of pay – that is, the link between the remuneration and the medium- to long-run performance of the institution. While I also agree with the SLR that the level of remuneration should not be directly regulated, I am somewhat less confident that the extreme remuneration levels seen in banking are innocuous and necessarily the result of productivity or skill differences.

Risk is intimately linked with luck. In a risky world, heroes are as likely to be lucky as smart. In asset management, Barras et al. (2010) find that around 8% of mutual funds display a significant positive alpha, but of them only about 0.5% deliver a positive alpha that is not driven by luck. I find it presumptuous to assume that the associated statistics are radically different in the population of successful bank managers and traders. Bonus payments could as often be a reward for luck as they are compensation for actual skill or effort. We need to improve our ability to distinguish between skill and luck, and must be prudent before concluding on the former. Simultaneously, we need to draw the adequate conclusions from the difficulties that will always exist in signal extraction on this issue. This means, notably, that compensation should be geared to the medium- to long-term performance of the individual and the institution, and not to a single trade or the current financial result.

Deferred compensations in the form of vested Cocos

Getting rid of limited liability in banking is probably unfeasible and is likely to be attached to significant negative collateral damage (notably through regulatory arbitrage). The implementable partial substitute is deferred compensation. In the words of the Squam Lake Report, “systematically important financial institutions should withhold a significant share of each senior manager’s total annual compensation for several years”. Noting that the goal is to align the incentives not to the interests of shareholders but to those of society, the report adds that “the withheld compensation should not take the form of stocks or stock options. Rather, each holdback should be for a fixed dollar amount, and employees would forfeit their holdbacks if their firm goes bankrupt or receives extraordinary government assistance.”

This recommendation has not found its place in explicit regulation (Basel III) but has remained a recommendation only (endorsed by the FSB). As a result, compliance is not monitored. This is highly regrettable as constructive behaviour from bankers in regulatory matters and, ultimately, financial stability are unlikely to be attained if the deep incentive problems highlighted here are not squarely confronted.

The excessive risk-taking behaviour engendered by limited liability, combined with very high remuneration, can be fully corrected only with very long clawback periods that may be hard to implement legally. The SLR suggests a period of five years; this has to be viewed as a minimum. As the Global Crisis has shown, the negative consequences of ill-guided risky investments can take many years to unfold, in particular when the sector in its entirety is supported by public policies. The case of zombie institutions hobbling along ten years after the start of the crisis provides a vivid illustration. Unless one finds a way to secure long-duration clawbacks as a default rule for large remunerations, one should have the courage to question the very notion of limited liability itself, at least for compensations exceeding a certain threshold.

It is important to add that the problem does not exclusively concern top managers of a financial institution, but all individuals entitled to high-level compensation, particularly in the form of bonuses. The many incidents that have occurred in the last ten years – most conspicuously outright frauds on LIBOR and in forex markets – have highlighted the intuitive fact that large and complex institutions cannot be controlled from the top in the absence of a perfect alignment of the incentive structure with the long-run interests of the institutions. This requires that all highly paid individuals, notably traders, are subject to very strict and long-maturity clawback clauses.

Finally, an extra step can be taken by awarding bonuses in the form of vested high-trigger Cocos. As Elliott (2015) notes, “paying banker bonuses in bonds makes sense. Cash bonuses are hard to claw back, while stock awards have unlimited upside.” UBS announced in January 2015 that 40% of the bonuses of its employees earning more than $300,000 would be deferred, and a proportion (40% again) of these deferred bonuses would be paid in Cocos that vest after five years and would be wiped out if, in the meantime, UBS’s common equity Tier 1 ratio falls below 7% (or even 10% in the case of executive board members). This innovation is commendable; it should be firmly endorsed and promoted by regulators. A universal adoption of similar principles would go a long way in restoring individual incentives conducive to financial stability.

Author’s note: This is a condensed and revised version of an article prepared for the special volume that the Reinventing Bretton Woods Committee (RBWC) will be releasing for the 10-year anniversary of the 2007 financial crisis. I thank John B. Donaldson for his very useful comments.


Admati A R and M Hellwig (2013), The Bankers’ New Clothes: What’s Wrong with Banking  and What To Do about It, Princeton University Press.

Baldwin R, T Huertas and T Ogden (2017), “Ten years after the crisis: looking back, looking forward”, VoxEU.org, October.

Barras L, O Scaillet and R Wermers (2010), “False discoveries in mutual fund performance: Measuring luck in estimated alphas”, Journal of Finance 65: 179-216.

Cochrane J H, D Duffie, D Diamond, A Kashyap, J Y Campbell, K R French, M N Baily (2010), The Squam Lake Report: Fixing the Financial System, Princeton University Press.

Danthine, J P and J B Donaldson (2014), Intermediate Financial Theory, 3rd edition, Academic Press.

Elliott, D (2015), “They should coco”, Reuters Breaking Views, 23 January.

Stulz, R and R Fahlenbrach (2011), “Bank CEO incentives and the credit crisis”, Journal of Financial Economics 99(1): 11-26.

Zingales, L (2015), “Presidential address: Does finance benefit society?” The Journal of Finance 70: 1327–1363.


[1] The hypothesis of distorted incentives does not exclude the possibility that involuntary mistakes also play a role in the crisis. Incomplete information and competence issues were certainly relevant as well (Fahlenbrach and Stulz 2011). 

[2] Which is why aligning the incentives of managers to those of limited liability bank owners is not sufficient to ensure socially optimal decisions.

2,310 Reads