Bank resolution, the need to recognize reality in order to prepare for the next crisis

The financial crisis has led to a near universal cry that never again shall a banker land on a taxpayer. In the EU, this objective has been enshrined in The Bank Recovery and Resolution Directive (BRRD), which member countries were supposed to implement as of January 1, 2015. The BRRD is an impressive piece of legislation that prescribes bail in of creditors rather than bail out by taxpayers as a preferred resolution strategy. However, it is obvious today that full implementation at this moment, including bail-in of senior creditors, not to mention unguaranteed depositors, in many countries is neither politically nor economically feasible. Rather than compromising on the end objectives, we suggest that the differences in the situation of individual countries and banks today is recognized, and each country develop individual plans for every bank in their country in order to maximize the likelihood that the EU will be prepared, when the next crisis comes along. This also reflects a more general view that the EU is more likely to progress in a harmonious manner, if there is due consideration of individual circumstances, as opposed to a process where standards are imposed without such consideration. The revision to BRRD, the so-called BRRD-2, unfortunately represent steps in the wrong direction in a number of areas.

Lessons from the past and present

There are no easy ways out, when resolving a failed bank. Somebody will have to bear the immediate costs of resolution and on top of that, the local economy will suffer because of the shock to credit supply. There is also a spill over to the public’s perception of financial stability with potential consequences for other banks, including the risk of a run. Resolution is thus a task of searching for the least bad option. Thus far, the EU’s implementation of resolution policies have not been an unqualified success cf. IMF (2018).

The public will end up footing the bill as either investor, senior creditor or taxpayer. One may therefore ask if it matters from what pocket the public ends up paying. The answer is that it does matter. There are at least three arguments that speak in favor of bail in. One, bail in gives incentives to shareholders and subordinated creditors to monitor the banks that they have invested in and push these banks not to run an excessive risky operation. Bail out removes these incentives. Heads up, investors win. Tail up, the taxpayer pays. Two, the capacity of the tax system in most countries is limited. A bail out will crowd out other more needed expenditures, e.g. on hospitals, for the elderly and for others with worthy needs. In a worst case scenario, a bail out will lead to governments defaulting. This is the first leg in the so-called doom loop, where bank bailout leads to government default that leads to more bank defaults and so on. Three, fairness suggests that those who have earned a rate of return above the risk free rate should also bear the costs, when risk materializes.

There are also arguments against bail in. The most prominent is the risk of contagion. Bailing creditors in can lead creditors in other banks to run in fear of their bank is the next. This can lead to a systemic crisis. The risk of such a development is higher, when the banking system is already under stress than when a bank is failing because of idiosyncratic risk. Another argument against bail-in is that politically it is more difficult to impose losses on specific individuals rather than letting the taxpayer pay. This applies even more so, when creditors invested prior to the enactment of the bail in strategy and have not been educated as to the risks of holding claims on banks.

The BRRD is broadly similar in scope1 to FDICIA in the US, Federal Deposit Insurance Corporation Improvement ACT, which the US implemented in 1991 following the savings and loan crisis that had large cost for US taxpayers. FDICIA requires the FDIC to resolve banks using the least costly method available. However, US authorities de facto suspended FDICIA during the financial crisis, as they were afraid of the risk of contagion2.

Some may legitimately ask why authorities do not handle failing banks through bankruptcy procedures like any other failing company. Bankruptcy procedures takes out excess capacity in an industry and helps restore profitability to the industry. This is the reason that state aid authorities favor bankruptcy procedures as the resolution mechanism.

However, banks play a special role in the economy – many industries unjustifiable claim special treatment, but here is a valid case. Banks specialness relate both to their liabilities and to their assets. The liabilities of banks are for a large part deposits, and depositors have little chance to assess the quality of a bank. Furthermore, these liabilities are crucial for the depositors’ daily economy. If you close depositor’s accounts Monday morning, the ordinary depositor will not be able to make payments on anything from their daily needs to mortgage bills. Bankruptcies typically take years to resolve, and during those years, creditors may have no access to their money. The ordinary depositor cannot wait years to access the deposits that they need for their daily payments. Deposit insurance help, but deposit insurance covers only up to EUR 100.000. Small and medium size enterprises have even more need than households for transaction accounts, and many hold in excess of EUR 100.000. On the asset side, banks provide credit to the local economy. Credit is not always easily transferable as customers’ credit history is key and often rests with their bank. A failure of a bank that plays a large role in the local community can thus cause a credit crunch in the community.

The key insight behind both the BRRD and FDICIA is therefore to ensure that most of the banking activities of a failing bank can continue, even when the bank is not restructured but the authorities close the bank. The authorities will in the latter case transfer most of the assets and liabilities of the failing bank to a solid bank. They will cover the deficit on the balance sheet by bail in. Bail in involves a haircut on the claims of creditors, but allows the remaining claims, in particular insured deposits and potentially a substantial part of uninsured deposits, to be available Monday morning. It is thus much less disruptive than a bankruptcy, where the courts may lock up all uninsured claims for many years. On the asset side, the transfer of many assets to a good bank will also allow a continuation of credit. However, even on a good day authorities are likely to leave bad credits behind and wind them down in the shell of the bank that is resolved.

Recent events and their agreed handling by politicians across the EU suggest that – despite all its’ virtues – we were not ready in most countries to implement the BRRD as scheduled. In many countries, the banking system is still struggling and there is a risk of contagion. Furthermore, the willingness to bail in senior creditors is not there yet. There is an acceptance that shareholders shall lose their shares. There is also some acceptance that you can bail in holders of subordinated debt. However, the willingness to bail in, what many consider nonconsenting senior creditors, not to mention unguaranteed depositors, is feeble in most countries.

Sweden in the early 1990s is often referred to as case that shows that bailout can be profitable for the government as the Swedish banks that were bailed out eventually could be sold at a profit for the government.3 The reference neglects three things. One, the Swedish government took an enormous risk by issuing a blanket guarantee for the banking system and the result does not reflect the possible distribution of outcomes at the time of the guarantee. Two, Sweden managed to change its economic policies in a manner that together with a large devaluation of the currency helped bring the economy back on a solid footing. Three, there are plenty of other examples of where government bailout has been extremely expensive. Ireland, Iceland and the UK intervention in relation to RBS are three recent examples.

It is also incredible tempting to avoid both bail-in and bailout, kick the can down the road and hope that the economy improves and in turn resolves the banking problems. The public at some stage turns their wrath against financial supervisors, when a bank fails. The longer a financial supervisor can pretend that the bank is not failing, the longer the financial supervisor avoids the wrath of the public. However, this can become very expensive. The savings and loan crisis in the US in the 1980s is a good example of the risks of allowing banks to gamble for resurrection. The Japanese banking crisis that began in the early 1990s is an example of how zombie banks allow zombie firms to continue and thereby hamper the creative destruction process in an economy. Most observers also agree that the US did better than many European countries by tackling weak banks early, rather than late, during the financial crisis.

The experience from Denmark during the financial crisis is instructive4. Like in the US, the Danish taxpayers did not lose money on the various support mechanisms (the Bankpackages). From early October 2008 virtually all senior, unsecured claims on Danish banks were covered by a 2-year government guarantee for timely payment. The participating banks paid fees for the guarantee and had several obligations. In case of regulatory failure, there was an obligation to accept a transfer of their assets and senior, unsecured claims to a government bridge bank. The government’s operator of most of the Bankpackages (Finansiel Stabilitet) would then auction off viable activities and do a controlled liquidation of the remaining parts. In early 2009 viable banks were offered an injection of government AT1-capital to avoid a credit crunch. At the same time viable banks were offered 3-year individual government guarantees on senior bonds as an exit mechanism for the general government guarantee. Non-viable banks were – after the expiration of the general government guarantee – offered a choice between liquidation or a transfer of assets and senior claims of an equivalent value to a government bridge bank. To the extent that Finansiel Stabilitet only partially transferred senior claims, the outcome was similar to a bail in of the non-transferred senior claims. Non-transferred senior claims covered by deposit-insurance or government guarantee were paid-off. Apart from that there was no change in the handling of the bridge bank. If the failing bank chose liquidation, the liquidator would make an equivalent transfer to the government bridge bank. From 2008 to 2012 the activities of 12 banks were transferred to a government bridge bank. The direct exposure of the government to the banking sector is today close to zero as before the financial crisis. A very broad majority in parliament enacted the Bankpackages. They did not affect Denmark’s AAA sovereign credit rating at any time.5

A strategy for the future

So, what should we do now? The first step is to recognize, where we are. The next is figuring out, where we eventually want to be. We will not get to where we want to be, if we fudge where we are.

The situation right now is that the banks across the EU are in very different stages of readiness in terms of applying the BRRD and so are the politicians. There are a few places, mainly in some countries in Northwestern Europe, where the banks are both sufficiently robust and there is a political willingness to apply bail-in to senior creditors. There are many other places, where this is not the case. In the EU, there is a strong desire to move ahead in unison. This has many merits, but when the situation is as different as in this case, we have to recognize it and act accordingly. Otherwise, some countries will have to move faster than they can and others to backtrack.

The objective should be to prepare for the next banking crisis, not only to repair the remnant problems from the recent crisis. Each country should set out a realistic but ambitious plan for how they will achieve the objective for each bank in the country. The plan should include how the bank should build bail-in-able liabilities in light of the possibility to retain earnings as well as issuing debt. Banks that cannot meet reasonable objectives should not pay dividends or buy back shares. Furthermore, the plan should include steps to eliminate impediments for resolution. Finally, countries should be obliged to educate their public and key stakeholders that claims on banks are subject to bail-in. If the plans are sufficiently ambitious, the EU state aid authorities should allow for a time some leeway in applying bail-in and in relation to state aid rules.

The EU has just revised the BRRD, but despite a lot of progress reflecting efforts by member states and the Commission, BRRD2 papers over the different situations across banks in the EU and does not reflect all the recent lessons6. In addition to taking into account the differences across the union, the EU should address six issues.

One, subordination is a necessity for MREL to work. At a minimum, the EU should allow countries to require full subordination. It is much easier politically to bail in creditors that are consenting adults. Furthermore, subordination makes it easier to transfer non-subordinated claims to a sound bank. There are many “No creditor worse off problems”, if there is not subordination and MREL rank pari-passu with other liabilities that the authorities do not want to bail in. The BRRD-2 puts limitations on the share of MREL for which authorities can require subordination. Different legislative initiatives across the EU show that there are different ways to achieve subordination, including through changes in the creditor hierarchy.

Two, authorities should impose MREL requirements not only for large banks, but for all banks7. Bail in is in many ways easier for smaller banks, e.g. the risk of contagion is less. In addition, as mentioned above, it is a fiction that any bank can be resolved thru the bankruptcy courts. However, MREL requirements can be lower for smaller banks, as there is not the same need to maintain the existence of the bank. You can normally sell the largest part of a small bank to a sound bank. The remaining part needs recapitalization, but much less than the whole bank. For some time the remaining part will need to carry out bank operations for the remaining customers. Thus, the recapitalization need is less than for a large bank. In Denmark almost, all failing banks are subject to resolution as the authorities see a public interest in avoiding bankruptcy. However, only SIFI’s have a full MREL-requirement.

Three, MREL needs to include a loss absorption amount beyond the solvency requirement. When a bank changes status from going to gone concern, the value of the bank’s assets, including the franchise value, declines. In addition, contingent liabilities pop up, e.g. litigation. This change in the value of a bank is not included in the way solvency requirements are calculated. The various forms of valuation exercises that resolution authorities have to undertake recognize this. The original BRRD allowed for inclusion of an additional loss absorption amount, but the BRRD-2 is less clear on this issue.

Four, MREL instruments should have a soft bullet structure that allows for the prolongation of the maturity of the instruments, if refinancing is not possible. Supervisors and resolution authorities will be caught between “a rock and a hard place”, if MREL instruments expire and cannot be refinanced resulting in the bank breaching its’ MREL requirement. When the MREL instruments expire, the authorities face the choice of restructuring the bank to avoid amortization or allowing a significant change in the creditor hierarchy by allowing amortization. The short maturity of most MREL instruments, around 5 years, makes it likely that such a situation will arise, when there is turbulence in financial markets. It is a mitigating factor that the legislation does not allow the use of instruments with less a one-year remaining maturity as MREL-instruments, even though they serve as such in case of resolution. However, this only gives authorities one year before they face a wall.

Five, there should be – subject to the above qualification – a requirement of prompt corrective action when a bank fails to meet MREL (once adequate levels of MREL has been reached). MREL can only protect a bank, when it is there. If a bank is allowed to go below its’ MREL, the protection of its’ other creditors/the tax payer will not be there. MREL is not only the ultimate line of defense in resolution. In the early intervention phase, MREL also makes it easier to avoid resolution. MREL is thus a buffer that can help in ensuring a sale of an institution or its’ activities. This should be the preferred strategy. However, it requires that there is no forbearance and that there are still MREL that can serve as a buffer. It should be the expectation that when MREL is breached, the bank in question will be asked to explore the possibility of a merger or a sale of assets and liabilities, if the bank cannot raise MREL. This would create less contagion than, if the resolution authorities have to intervene. Both the original BRRD and BRRD-2 could be clearer on the appropriate strategy.

Sixth, it is necessary to restrict the distribution of MREL instruments. MREL instruments is not a suitable instrument for retail customers. If retail customers hold MREL instruments, bail-in will be politically difficult. Furthermore, banks should not hold significant amounts of each other’s MREL instruments. Interbank holdings risk creating contagion. The BRRD-2 addresses G-SIFI holdings of other G-SIFI MREL instruments and to some extent constrains distribution to retail customers. We need to do more. All interbank holdings of MREL instruments are potentially contagious and any sale to retail customers causes political difficulties, when bail-in is applied. Restrictions on distribution will make it more difficult to raise MREL, but MREL that you cannot bail in is useless.

Final remarks

Bank resolution has proved a very expensive experience for taxpayers in many countries. It is ok to fail once, but there is no excuse, if you do not learn from your failures. The public will rightly blame supervisors, resolution authorities and policy makers, if we do not do better next time around. One of the few sure things is that banking crisis are a reoccurring phenomenon. However, with a bit of luck, we will have some years ahead of us, before the next crisis hits us. Therefore, it is less important to have the framework in place now than to build a resolution framework over the next years that the public believes is feasible and credible.

The BRRD framework is broadly the right framework, but it needs strengthening in a number of areas. The compromise between countries, where the banking system is still vulnerable and countries, where there is both a capacity and political willingness to adhere to the most stringent standards, should not be a BRRD that is too harsh for the former countries, and too lenient for the latter countries. The compromise should rather be one, where the different starting points are respected and where we in due time all will be in a position that is defendable, when a bank is about to fail. This should be a joint objective that everybody would support.