Ladies and Gentlemen ,
I would like to thank CAREFIN and the Department of Finance at Università Bocconi for inviting me to speak today in Milan. I would like to use this opportunity to focus on what is in my view a crucial aspect of bank stability, namely, the existence of effective resolution mechanisms.
Excessive risk-taking by banks is at the origin of the financial crisis. One of the many factors contributing to their risk-taking behaviour was the distorted incentive structure resulting from the lack of an effective resolution mechanism. Establishing a strong approach to bank resolution is an essential component of properly realigning incentives towards risk.
Why is this the case? As you know, bank resolutions are often costly, complex and time-consuming, in particular for large institutions. Without alternative ways to restore their viability, governments have too often found themselves compelled to bail-out entities using taxpayer money. This government support may be successful in avoiding financial contagion within closely interconnected banking system. But the expectation of assistance via publicly funded bailouts amplifies moral hazard, leading to excessive risk-taking.
Moreover, in the euro area it has nurtured what became the nexus of the crisis, the fatal feedback loop between bank and sovereign creditworthiness. So these are already two reasons to rethink the allocation of losses in bank resolution. The former is general and the latter is specific to the fragmented structure of the euro area.
A third reason, which I will not discuss here, relates to the political economy of bank resolution. Even if it were efficient for financial stability reasons to let taxpayers bear the burden of losses in resolution, it would be unpalatable to them, and it would be unfair unless they can exert effective control on the risk-taking behaviour of banks – which has not been the case so far.
Existing bank resolution mechanisms are being overhauled in many countries following the adoption in 2011 of the Financial Stability Board (FSB)’s Key Attributes for Effective Resolution Regimes . In Europe, where most countries did not have a resolution framework before the crisis, extraordinary efforts have been made to design a new regulatory framework, the Bank Recovery and Resolution Directive (BRRD). The BRRD, which has been agreed by the EU Council and is now being discussed by the European Parliament, will set harmonised standards for national resolution regimes across Europe in compliance with the FSB Key Attributes. It aims to reduce the costs of bank resolution, make banks resolvable without using taxpayer funds, and hence induce bank creditors to exert more market discipline.
I would like to use my remarks today to comment on this new Directive, and in particular the introduction of creditor-funded recapitalisation - otherwise known as bail-in. I want to explore, on the basis of the economic literature, how this is likely to affect bank funding and lending and hence, as suggested by the title of this conference, the financing of the recovery. I will not discuss here resolution strategies, which are at the heart of the FSB discussion and can be particularly challenging for large and internationally active banking groups.
The proposal for a Bank Recovery and Resolution Directive
Let me begin by briefly reviewing the main features of the new bail-in tool. Bail-in rules are envisaged to apply no later than four years after the entry into force of the new resolution Directive. If this is 1st January 2014, as is currently planned, then they will have to be incorporated in national laws by 1st January 2018 at the latest. I will comment this timeline later.
The scope of bail-in clauses is generally broad, encompassing all liabilities apart from those explicitly excluded. Non-excluded liabilities will be bailed-in following the order of their ranking in national insolvency laws.
The only exception to the seniority rankings in national insolvency laws relates to customer deposits. The latest Council compromise text introduces a “super priority” for insured deposits (up to € 100 000) and a ‘simple priority’ for retail deposits (deposits of natural persons, micro and SMEs above €100 000). The latter category will rank below insured deposits, but above all other senior unsecured claims. 
It is important to stress that the Directive does not exclude other funding sources to manage banking crises. Privately-funded national resolution funds may step in, but only up to a certain limit and only after some losses have been borne by shareholders and creditors.  In addition, the Directive does not completely exclude the possibility of a bail-out via extraordinary public financial support. However, the principle is that support has to be given exceptionally and temporary in nature. 
In order to prevent banks from circumventing the bail-in rules, the Directive also defines minimum requirements for own funds and eligible liabilities (MREL), ensuring that banks have sufficient loss-absorbing capacity, i.e. sufficient bail-inable liabilities. The requirements are currently based on the size, risk characteristics and the corresponding business model of banks. In 2016, the proposal aims at introducing harmonised MREL for all banks based on recommendations by the European Banking Authority (EBA).
Additional to the bail-in rules, the European Commission adapted temporary state aid rules, applicable as of the 1st August, 2013. These rules demand mandatory capital write-down and bail-in of subordinated debt before any public supports is granted. Therefore, these rules further clarify the interaction between bail-in and bail-out in the European Union, even before the BRRD becomes effective, and provide for a level playing field between similar banks located in different Member States.
The impact of mandatory bail-ins on bank funding
The implementation of bail-in rules can be expected to affect banks activity in several respects that matter not only for bank stability but more broadly for the funding of economic activity. Here I want to focus on two in particular: bank funding and bank lending.
Starting with bank funding, the bail-in rules transfer risk from taxpayers to unsecured bondholders. In this way, they counterbalance the perceived implicit subsidy from which some very large or highly interconnected institutions have benefited, leading to excessively low funding costs, particularly in good times. As a result of this increased risk, holders of bail-inable liabilities may be expected to ask, everything being equal, for a higher return. The cost of funding for such liabilities would reflect more accurately their actual risks.
How will this affect banks’ overall cost of funding? This is not easy to answer definitively. There are many factors that we have to take into account. Let me list four of them.
First, the systemic importance of the bank. The greatest effect of the new bail-in rules should in principle be on those banks that would in the past have benefited most from an implicit guarantee. In other words, it should affect more significantly systemically important financial institutions (SIFIs). And among the SIFIs, it should affect the riskiest banks the most.
Second, the bank’s liability structure. For those banks that fall short of the bail-inable debt requirements, the Directive will force them to increase their unsecured liabilities, since secured ones cannot be bailed-in. This could increase overall funding costs, as it would require banks to adapt their liability structure such that more costly debt plays a greater role. But equally, the higher share of senior unsecured liabilities could also result in a lower level of encumbered assets, possibly lowering the overall cost of funding. Indeed, asset encumbrance makes balance sheets more complex and poses significant challenges to investors in assessing banks’ riskiness, which may result in higher risk premia for unsecured debt, especially at a time where the overall demand for collateral assets is expected to increase. 
Moreover, we need to consider the effect on the overall cost of funding from the strengthening of depositor protection, which has a potential to reduce the relative interest rate paid on deposits compared to unsecured debt. While depositors have, in normal times, low sensitivity to risk and any decline in their cost is expected to be limited; deposits typically represent a large share of the overall bank debt. As a result, even small reductions in the remuneration of deposits could produce a significant impact on the overall cost of funding. On top of this, banks that already meet the minimum bail-inable debt requirement may adjust their liability structure to incorporate a larger fraction of deposits.
Third, to assess the overall impact of bail-in rules on bank funding, we have to consider their interaction with other initiatives in the field of regulation. Let me give me two examples. There is a possibility that the higher cost of senior unsecured long-term debt could lead banks to raise more short-term debt, resulting in a higher maturity mismatch and greater liquidity and interest rate risks. This could happen in particular for those banks that have already a high level of encumbered assets and so are unable to issue long-term secured liabilities. Such reshuffle of liabilities will be limited by the proposed Basel III requirements on the Liquidity Coverage Ratio (LCR) and the net stable funding ratio (NSFR). Similarly, recourse to the repo market, which we know is inherently fragile,  will be constrained by the envisaged FSB rules on shadow banking.
As another example, as there is at least in theory a risk that bail-in rules could trigger financial contagion if a significant proportion of banks’ liabilities are held by other banks or by systemically important non-banks such as insurance companies,  we need to understand and if necessary limit such cross-exposures.
Fourth, we have to take into account the effect of bail-in on banks investment and lending behaviour. Let’s not forget that one objective of an efficient resolution mechanism is to mitigate excessive risk-taking. As a result of the removal of an implicit bail-out guarantee, bank debt-holders should exert more efforts in monitoring banks, thus mitigating moral hazard phenomena. Such market discipline may mean that the expected rise in the cost of bank funding would be counterbalanced by an overall reduction in bank risk and hence the cost of funding may not rise.
By redistributing the risk between tax-payers, depositors and debt holders, bail-in rules can reduce the overall risk in the system. The corporate finance literature has pointed out the role of secured bank debt as an equilibrium response to the agency problems of equity financing.  This highlights the risk of a “race to the bottom” towards a liability structure that would leave bank managers unchecked. It is therefore crucial that all sources of bank financing are subject to adequate monitoring.
The economic literature clearly points out that exposure to credit risk should be left primarily to those actors who have the best information on counterparty risk, namely, shareholders and debt-holders, hence the new capital requirements and the bail-in rules. Let me note however that inducing greater monitoring efforts is a necessary, but not a sufficient condition to mitigate excessive risk-taking. Bank debt-holders have to be able to monitor banks ( market monitoring), but by monitoring them and pricing risk correctly they are also effectively influencing banks’ behaviour ( market influence). Making market discipline work calls for a more comprehensive regulatory and supervisory framework of which bail-in rules are only a part, even if a significant one. And since bail-in rules have confirmed that insured depositors are protected by deposit guarantee schemes, i.e. ultimately by the taxpayer, the banking supervisor can be represented as the public authority to whom insured depositors delegate the monitoring of bank behaviour.
All in all, while a good resolution framework is needed, it can only be a complement, and not a substitute for market discipline and supervisory vigilance.
What does the empirical evidence of the effect of bail-in on funding costs say so far?
The evidence points to a possible increase in the cost of funding for senior unsecured debt. For instance, a recent survey  suggests that under a bail-in regime investors would demand around 90 basis points more for a senior bond issued by a single ‘A’ bank. The investors surveyed also expect that the implementation of a bail-in framework will lead to an increase in price differentials across issuers with diverse credit quality. However, all this has to be weighed against the offsetting considerations that I just described.
Indeed, econometric results undertaken at the ECB for a sample of European banks suggest that the announcement of the Directive had so far limited impact on the cost of senior unsecured debt. But for the financial sector, there are still two sources of uncertainty with regards to the bail-in rules. First, about the timing of the implementation of bail-in. The European Parliament, for example, wants to have bail-in from 2016. Second, about the degree of discretion national resolution authorities will have in applying the bail-in rules, which in my view should be as constrained as possible.
This would seem to suggest that, although likely to be priced in to a considerable extent, the ultimate effect of bail-in on funding costs, and its success as a market discipline device, will depend on the final design of bail-in and how markets see it being used in practice. This calls for an early implementation of bail-in as a way to reduce uncertainty for banks themselves and market participants, and allow for a once-for-all repricing of risk. .
The impact on lending
Turning now to the effect of bail-in on bank lending, my baseline assumption is that the higher cost of bail-inable liabilities will have a limited impact on bank lending. Let me explain.
Clearly, if overall funding costs were to rise, this could in theory have an effect on banks’ lending activity: banks may respond by providing less credit and charging higher loan interest rates. That said, I am not certain how realistic this scenario would be in practice. Building on the vast literature on the bank lending channel, one can argue that even if overall funding costs increase, there are several factors that can mitigate the transmission of funding shocks to lending. For instance, capital and liquidity buffers  allow banks to shield borrowers from funding shocks and these are in the process of being strengthened by the Basel III requirements. Moreover, the type of relationship between banks and borrowers may affect the way banks translate shocks into their lending activity.  A very recent paper, to be presented in this conference, suggests that banks more reliant on relationship lending are in a better position to shield borrowers from shocks.
Finally, any assessment of the effect of bail-in on bank lending and ultimately on economic welfare should take into account the general equilibrium benefits of reallocating risk to those who have the ability to monitor and eventually absorb it: shareholders and debt-holders. As noted by Paul Tucker, there is no place to hide risk: if the risk in banking is not incorporated in the yields of bonds issued by banks themselves, then it will be reflected in higher sovereign borrowing costs.  The euro area crisis has shown the negative and potentially highly non-linear consequences of such an increase in today’s situation of high public debt.
Let me conclude.
Creating an efficient mechanism to swiftly resolve insolvent banks is a crucial step towards making the costs of tackling bank crises less onerous for taxpayer. While there is a lot of work ahead, a well-designed, comprehensive and consistent regulatory framework for banking resolution, including creditor-funded recapitalisation schemes, will mitigate excessive risk-taking phenomena. By improving financial stability, it would also have a favourable long-term impact on the real economy.
While a good resolution framework is needed, it can only be a complement, and not a substitute for market discipline and supervisory vigilance. The Single Supervisory Mechanism (SSM) and its founding act, the balance sheet assessment and asset quality review of all directly supervised euro area banks that will take place in 2014, can therefore be seen as complementing and supporting the new resolution framework.
Effective resolution is not only about strong rules creating sound incentives, but also about quick judgment and decision-making. Given the inherent instability of financial markets, there is no time for a negotiation between bank creditors and shareholders leading to an equity write off and a conversion of debt into equity, as would be usual outside the financial sector – and in the case of a large systemic bank, such a process would have the potential to impose significant negative externalities on the economy. In this respect, the creation of a strong and independent Single Resolution Authority (SRA) is needed in Europe as a necessary complement to the SSM. The SRA should have the capacity to take swift decisions in the European interest and to tap a Single Resolution Fund financed by the industry.
If these steps are taken, bail-in rules will be conducive to a better monitoring of banks’ risk behaviour and will thereby lead to a better banking system – a system that fulfils its economic role without creating an excessive risk for society.
I thank you for your attention.
I wish to thank Paolo Emilio Mistrulli for his contributions to these remarks. I remain solely responsible for the opinions contained herein.
See Financial Stability Board, “Key Attributes for Effective Resolution Regimes for Financial Institutions”, October 2011.
This differs from the US where the depositor preference applies to all deposits, both insured and uninsured.
National resolution funds (RF) have to be set up in each Member State and financed by ex-ante contributions of the institutions. They can only be used to absorb losses and finance resolution after losses of at least 8% have been borne by shareholders and creditors. When the RF steps in its total contribution is limited to 5% of total liabilities.
Public support may be given as (i) State guarantee backing central bank liquidity; (ii) State guarantee for newly issued liabilities; or (iii) capital injection at market price, as long as the institution still fulfils the authorisation requirements and is solvent.
Committee on the Global Financial Stability (2013), “Asset encumbrance, financial reform and the demand for collateral assets”, CGFS Papers, n. 49
See G. Gorton and A. Metrick (2012), “Securitized Banking and the Run on the Repo”, Journal of Financial Economics, Vol. 104, No. 3, pp. 425-451.
See M. Dewatripont and X. Freixas (2012), “Bank Resolution: Lessons from the crisis” in M. Dewatripont and X. Freixas (eds.), The Crisis Aftermath: New Regulatory Paradigms, eReport, Centre for Economic Policy Research.
This literature can be traced back to D. Diamond (1984), “Financial intermediation and delegated monitoring”, Review of Economic Studies, Vol. 51, pp. 393-414. See also the discussion in A. Kashyap, R. Rajan and J. Stein (2008), “Rethinking Capital Regulation”, in Federal Reserve Bank of Kansas City, 2008 Economic Symposium, "Maintaining Stability in a Changing Financial System", pp. 431-471.
“European bank bail-in will cost +87 basis points”, Financial Times , October 25, 2010.
 To give a flavour of the issues markets will have to consider, we can look at the literature on contingent convertibles bonds tied to market evaluation, or CoCos - see among others, O. Hart and L. Zingales, 2011, “A New Capital Regulation for Large Financial Institutions”, Journal of Antitrust Enforcement, Vol. 13, pp. 453-490. The literature have shown that if bail-in takes place at too low share values, a “death spiral” can occur whereby the bail-in leads to a decrease in stock prices which triggers additional bail-in (P. Hillion and T. Vermaelen, 2004, “Death Spiral Convertibles”, Journal of Financial Economics, Vol. 71, pp. 381-415). At the same time, it is essential that the conversion is triggered early enough and leads to enough dilution of existing shareholders. If not, then the conversion of the debt into capital will increase the value of the firm for shareholders after losses arise, with the perverse effect of increasing risk-taking incentives (See C. Berg and T. Kaserer (2012), “Does contingent capital induce excessive risk-taking and prevent an efficient recapitalization of banks?”, New York University and Humboldt University.) A somewhat related issue is that banks may have a perverse incentive to increase the risk of its assets’ returns in order to transfer value from contingent capital investors to the original bank shareholders (G. Pennacchi, 2011, “A structural model of contingent capital”, Federal Reserve Bank of Cleveland, Working paper No. 10-04). In principle, this is less an issue in the case of mandatory bail-ins, since the conversion of debt into equity occurs at the point of non-viability. In practice, however, it is difficult to establish what this point of non-viability actually entails, so effectively these concerns could also be relevant in the case of bail-ins.
See A. Kashyap and J. Stein (2000), "What do a million observations on banks say about the transmission of monetary policy?" American Economic Review, 90(3), pp. 407-428. R. Kishan and T. Opiela (2000). " Bank size, bank capital, and the bank lending channel," Journal of Money, Credit and Banking, vol. 32(1), pages 121-41.
See P. Bolton, X. Freixas, L. Gambacorta and P.E. Mistrulli (2013), “Relationship and Transaction Lending in a Crisis”, CEPR Discussion Paper , No. 9662.
P. Tucker (2012), “Resolution – a progress report”, speech at the Institute for Law and Finance Conference, Frankfurt, 3 May.