## Paul Tucker: Resolution of large and complex financial institutions - the big issues
Remarks by Mr Paul Tucker, Deputy Governor for Financial Stability at the Bank of England, at the European Commission Conference 'Building a Crisis Management Framework for the Internal Market', Brussels, 19 March 2010.
* * *
Many thanks for input to Mark Adams, Andrew Bailey, Peter Brierley, Geoffrey Davies and Andrew Hewitt, and to Financial Stability Board colleagues for discussions on resolution issues. For background work to Michael Grady and for secretarial support to Sandra Bannister and Cheryl Feeney.
I am delighted to be at today's conference, the first major conference under Commissioner Barnier.
As many colleagues here will know, I chair the Financial Stability Board's Working Group on Cross-Border Crisis Management. But my remarks today are offered in a personal capacity, albeit drawing on the international dialogue.
The conference is very well timed given the Commission's work on a possible EU directive on resolution regimes. An increasing number of national authorities have already established or are prospectively establishing Special Resolution Regimes for the resolution of commercial banks, enabling the transfer of deposits and good assets to another bank, with bad assets and other creditor claims going into an insolvency process. At a high level of generality, a distressed bank is split into a continuing good bank, typically acquired by a healthy bank; and a gone-concern bad bank, which goes into administration. Insured deposits travel with the 'good bank'. All other creditors typically travel with the 'bad bank' and become claimants in the insolvency.
We can be confident that this will help to buttress stability. Encouragement from the EU to all members states to establish such regimes would be really useful.
But we also know that, of itself, this is insufficient. It cannot cope with distressed LCFIs ('large and complex financial institutions') without injections of public money.
In the first place, every country having its own Special Resolution Regime does not solve the problem of resolving cross-border banking groups.
Second, such regimes cannot cope with investment banking, whether national or global, if a buyer cannot be found. Even for a (hypothetical) LCFI operating entirely within a single jurisdiction, with all of its counterparties, customers and contracts domiciled there, a standard resolution regime could not effect an orderly run down of, for example, a massive, complex derivatives portfolio. Putting over-the-counter derivatives though central counterparties would help. But it would not securely solve the problem.
Today I want to surface those big issues, and to encourage the EU and the international community more generally to explore possible solutions. In doing so, I take our objectives to be twofold:
- (i) First, to avoid a disorderly disruption to the provision of financial services to the economy in the event of firm distress;
- (ii) And second, to put market discipline back into the system, by having losses fall to uninsured creditors as well as to equity holders, and without public sector solvency support.
Bluntly, the challenge is how to sustain the provision of financial services during a crisis without the state bailing out uninsured creditors.
## Handling LCFI complex derivative/trading portfolios
Some types of business activity, particularly complex trading, do not lend themselves at all well to the standard resolution tools I summarised just now. Not least, unwinding a derivatives book is especially complicated; it typically requires dynamic hedging to preserve value, and that would be very hard indeed in a liquidation. Counterparties flee. Funding flees. The Lehman and AIG experiences demonstrate this. So, in a different context, does the time it took Berkshire Hathaway to wind down the complex derivatives portfolio of its Gen Re subsidiary; years, not months.
What can be done? One route, suggested by those examples, would be for the authorities to finance the wind down of complex trading positions, in what would be in effect an analogue to US-style debtor-in-possession financing. The authorities would be preferred creditors.
That kind of approach could be used either briefly, in order to bridge to an insolvency procedure, which would amount to putting a framework around the kind of operation the NYFed mounted for a week or so to contain disruption from the failure of Lehman's US broker-dealer. Or, alternatively, it could in principle be used over a prolonged period, to finance the run off of a whole book. In either case, however, it would expose the authorities to risk of loss, and so would not always deliver the Objectives I think we share. It would not be assured to preserve orderly markets in all cases. And it would be insufficient if a firm was actually insolvent.
A quite different, and rather more profound approach would be to deploy a super special resolution framework that permitted the authorities, on a rapid timetable, to haircut uninsured creditors in a going concern.
By haircutting creditors in a going concern, there would be no need to close the business immediately (or, if regulators wished, even at all). That is a very high-level description. Variants obviously exist. Instead of haircuts, uninsured creditors could have part of their claim converted into equity. And it would need to be clear how the regime applied to the normal ranking in the capital structure, with common equity and subordinated debt being wiped out before any senior unsecured debt was haircut (or converted) at all.
Precisely because investment banking is highly levered, relatively small percentage haircuts (or conversions to equity) could materially recapitalise a distressed firm. Creditors would recapitalise distressed LCFIs, not the state. In that, it shares some features of contingentcapital bonds, but would be on a statutory basis and could apply to all unsecured creditors in the extreme circumstances of systemic distress.
The crucial feature is that haircuts/conversion would be applied up front to creditors of a going concern . By contrast, under standard commercial-banking resolution regimes, the 'bad bank' goes into an insolvency process as a gone concern, with de facto haircuts applied to its creditors ex post, after realisation of the assets by the liquidator - a process that almost always destroys value. A 'going concern' approach would in effect combine features of standard regimes for resolving commercial banks (rapidity, public policy objectives) with some features of the US Chapter 11 for non-financial companies (haircuts for creditors in a continuing business). Of course, the kind of super-special resolution regime I am airing could not be an exact replica of Chapter 11 - if only because the capital structure of a distressed LCFI would have to be restructured very quickly, over a weekend if not overnight, and so could not rely on a negotiated settlement with creditors.
For that reason alone, in common with the existing commercial-bank resolution regimes, there would plainly need to be safeguards . 1 Perhaps the most important might be that such
1 For safeguards in the UK SRR, see the Banking Act 2009 (Restrictions of Partial Property Transfers) Order 2009 (SI2009/322).
BIS Review 33/2010
super special resolution powers could be applied only where the authorities made a formal determination that the firm's distress would otherwise seriously threaten systemic stability.
The modalities would, I suspect, end up varying according to countries' particular circumstances. In some, the financial authorities might decide the scale of the haircuts/conversions; perhaps in other countries, it might be thought that a senior judge should decide on the basis of an application by the financial authorities.
It is also worth saying that maintaining an LCFI in business through such means would neither preclude nor of itself require regulatory action against management or, once the crisis had passed, to wind-down parts or all of the business over time.
Nor would a regime of the broad kind I have sketched exclude restructuring banking groups into functional subsidiaries, with different resolution tools applied to, say, deposittaking/payments; custody and global cash management; the business of illiquid lending to small and medium-sized enterprises and households; trading/derivatives business or wholesale banking more generally. That would not be inconsistent with reducing the number of legal entities in LCFI groups, perhaps by an order of magnitude. The current FSBsponsored work to develop 'recovery and resolution plans' for the top 25 or so LCFIs is designed to identify where those kind of steps are warranted. Resolution could, therefore, take its place alongside structure and regulation as one of the key elements of a reform package to strengthen the global financial system.
A super-special resolution regime along those lines would, on the face of it, go quite a long way towards reintroducing market discipline - by putting uninsured creditors at risk even in systemic situations. I am not innocent enough in the ways of the market to think that it wouldn't affect behaviour. Of course, it would lead to changes, over time, in the capital structure of investment banking operations. Proper transitional periods would be needed.
I suggest that it is worth debating. Maybe there is a fatal flaw. We should certainly search for the flaws. But if we do not find a way of resolving massive and complex trading books, the regulation of the business would probably have to be fiercer than otherwise in order to crush the probability of failure. Resolution is a way of coping with failure. Key voices are the nonbank financial sector and the corporate treasurers who finance the banking system. What would they get out of it? They might benefit from avoiding the risk that the authorities choose to liquidate a LCFI they had thought was too big to fail, with the depletion of value that entails; think of Lehman. They would, in effect, be trading off losses in one part of their portfolio against the prospect of protecting the rest of their investments from the fallout of a credit crunch that would follow the disorderly demise of a distressed LCFI.
## The cross-border dimension
The other big issue is how to handle the resolution of internationally active banks. This is a challenge even for fairly vanilla commercial banking groups.
There has been a Basel Concordat on how home and host authorities should co-operate in supervising international banks since 1975. It was updated in 1983 and 1992 to take account of group-level consolidated supervision following the failures of Banco Ambrosiano and then BCCI. But was the Concordat based on a false premiss: that the division of supervisory-labour during peacetime can carry smoothly over to the only period that really matters, distress?
Because, as it happens, neither standard resolution regimes for commercial banks nor any super-SRR of the kind I have just outlined for going-concern restructurings could be applied easily to the cross-border activities of LCFIs. SRR tools cannot automatically be applied extraterritorially.
It seems to me that policymakers need to face up to, and decide whether or not to fix, the really big issues about how insolvency and resolution laws are applied to internationally active financial companies.
Broadly speaking, countries' laws apply either territorial principles to international insolvencies, consistent with separate procedures for each of a bank's entities in different jurisdictions (so-called 'separate-entity resolution'); or universal principles, consistent with a single procedure applying to the resolution of an international bank and its branches abroad, wherever they are located ('single-entity resolution'). I believe that top-level policymakers have to choose what kind of regime they want. This issue needs to be elevated above the level of middle-ranking technocrats where, through no great fault of their own, it has wallowed for too long. The issues were articulated as long as 20 years ago, following BCCI's failure. 2 Nothing much has been done globally.
So in the second half of my remarks today, I just want to sketch why this problem arises.
## (i) Territoriality
Territorial approaches to resolution are adopted due to potential conflicts of interest between home and host-state authorities. These may reflect, for example, host countries not wanting to rely on home countries to act in the interests of financial stability in the host country; or different or conflicting triggers and objectives for Special Resolution Regime powers.
There are a number of problems with territoriality. It undermines coordinated and cooperative resolutions, and creates uncertainty by encouraging a race for assets and collateral across jurisdictions. Related to that, it reduces the effectiveness of Special Resolution Regimes by allowing host authorities to ring-fence local assets for the benefit of local creditors. And, it can even render resolution infeasible if the legal structure of an LCFI differs substantially from its operational structure, with local legal entities dependent on other parts of the group for vital services.
Greater harmonisation of SRRs across the major jurisdictions might help to reduce distrust in other countries' approaches. But harmonisation does not remove the potential conflict of interests between home and host authorities.
The current debate around subsidiarisation of global commercial banking groups stems in large part from these problems. The thought is that if policymakers cannot solve the territorial problem, then they should embrace it - banking would go local, with a degree of ring fencing. That course may conceivably have wider advantages - for example, in the area of macroprudential regulation 3 it might give authorities a greater degree of influence over domestic credit conditions. But we do also need to consider whether there might be costs to such an approach. For those policy issues to reach closure, we need to know the way forward on resolution.
## (ii) Pure universalism
What of the other end of the spectrum? In principle, an alternative approach would be to shift to a more universal approach. A pure form of universalism would rule out separate resolution or insolvency proceedings, ring-fencing and, therefore, a race for assets. It might encourage
2 Basel committee on Banking Supervision (1992). 'The insolvency liquidation of a multinational bank', Bank for International Settlements. Its key conclusions were: (1) when closing a multinational bank, supervisors should pay attention to the nature and timing of communications among themselves and of their communications with creditors, shareholders and management; (2) the nature of liquidation rules may be relevant to the manner in which multinational banks are supervised; (3) differences in liquidation rules across jurisdictions in a windingup can affect returns to depositors and other creditors and the operations of deposit protection schemes; and (4) coordination and cooperation between liquidators can affect the returns to creditors in a liquidation and can be affected by the role of supervisors in a liquidation.
3 See the Bank's recently published paper: 'The role of macroprudential policy: A discussion paper' and Tucker P M W (2009e), 'The debate on financial system resilience: Macroprudential instruments' at the Barclays Annual Lecture in London.
BIS Review 33/2010
- a more cooperative approach to cross-border resolution. There would be one, global resolution, led by the home authority.
But pure universalism also faces major obstacles. First, it would imply a surrender of national sovereignty - replacing it with pre-defined and enforceable rules that determined the roles of different authorities in different jurisdictions. Second, unless steps were taken to rule out the need for sovereign Capital of Last Resort operations 4 , it might well need to be accompanied by an ex ante burden-sharing mechanism, to facilitate a sharing of the costs of resolution across the jurisdictions in which a bank had activities.
In theory , that could be addressed by the main home and host authorities of each LCFI agreeing a set of binding principles to govern the distribution of the costs of a universal resolution across jurisdictions. An alternative would be to design a single resolution regime for all LCFIs, to be implemented along universal lines by a supranational resolution authority. But would countries be prepared to tolerate the loss of fiscal independence and the overturning of national insolvency laws that this model would entail, especially if it required governments to commit to recompense the agency for unlimited amounts in the future? That is most definitely not a question for central bankers.
## (iii) Modified universalism
Pure territoriality and pure universalism are corner solutions. We live in a world where some countries are closer to one corner, others closer to the other.
Some have suggested that an intermediate solution might be feasible. This would involve 'territorial' countries deciding to resolve financial institutions headquartered in their own jurisdictions on a universal basis. (A possible example of this approach in the recent crisis was the FDIC's resolution of United Commercial Bank, which involved a 'purchase and assumption' 5 that extended to the Chinese subsidiary and the Hong Kong branch following coordination amongst the authorities in the US, China and Hong Kong.)
A 'modified' form of universalism would involve host authorities choosing to defer to and cooperate with a resolution brought by the home country authorities, provided that certain conditions held. Separate insolvency proceedings might be employed by host countries, but allowing the proceeds to be remitted to the home country, 'lead' resolution. Variations might be needed for subsidiaries and branches. Subsidiaries are distinct legal entities, but they too can have worldwide creditors and claims. So it would all need to be within some kind of joined-up whole.
At a high level, the preconditions for something like this might include:
- (1) equitable treatment of all creditors regardless of their jurisdiction, because if creditors in host countries were likely to be penalised in a resolution brought by the home country authorities, host authorities would have a strong incentive to ring-fence;
- (2) broad harmonisation of resolution regimes across the major jurisdictions.
Also, if governments do continue to contemplate the possibility of fiscal support in a crisis, then there might need to be some kind of agreement - conceivably case by case - on
4 For a discussion of capital of last resort see, See Tucker P M W (2009f) 'The Crisis Management Menu', SUERF, CEPS and Belgian Financial Forum Conference: Crisis Management at the Cross-Roads, pp. 13 and Tucker P M W (2009b), 'The repertoire of official sector interventions in the financial system: last resort lending, market-making, and capital', Bank of Japan's 2009 International Conference, Tokyo, May 2009, pp. 17.
5 A 'purchase and assumption' is the term used by FDIC to describe when a healthy institution purchases some or all of the assets of a failed deposit taker.
principles governing the contributions imposed on each country to meeting the costs of the resolution. But the ideas I aired in the first part of the remarks were aimed at helping to take that off the table.
A principle of equitable treatment of worldwide creditors and depositors would be consistent with the imposition of losses on unsecured creditors or uninsured depositors, either in a going-concern restructuring of the kind I outlined earlier or through the application of more standard SRR tools to gone-concern bad banks. But it would require losses to be imposed on an equitable basis, respecting the existing priority ranking regardless of the jurisdictionallocation of creditors. That might require recognition across the main jurisdictions of certain safeguards. 6 And, such an approach would require binding agreements amongst home and host resolution authorities. Vague assurances would not do the trick, as domestic authorities could change their minds when it really mattered. Host and home authorities would need 'skin in the game' to underpin co-operation. And co-operation in distressed circumstances could helpfully flow back to how supervisors work together in peacetime.
An obstacle to such an approach might be that in some countries, legislation mandates preferential treatment of creditors and depositors of local entities of cross-border banks. But that might be weighed by legislatures against the current reality that when the fiscal authority finds itself injecting equity to save a bank and preserve stability, it protects worldwide creditors not just its home creditors.
To be clear, I am not advocating any particular approach here. I just want the authorities to articulate the issues clearly, so that our leaders can make a choice and think through the consequences of that choice. I'm not sure that it has really been put to them up to now. And, of course, whatever is done, within Europe we would need to find an EU version that sat between national and global regimes.
## Summary
My purpose today is not to advocate solutions. It is to identify the big issues and to float some possible avenues for discussion in the EU and internationally.
The big question is clear enough. Can we resolve distressed LCFIs in an orderly way without injecting public money?
One possible way to cut through Too-Big-To-Fail is to adjust our sense of what 'Fail' involves. Perhaps it does not have to involve liquidation or administration. Perhaps it does not have to involve a binary shift from 'going concern' to 'gone concern'. But it does have to involve loss for equity holders and uninsured creditors. It does have to rekindle market discipline. It does have to preserve the flow of financial services. Those are, surely, the essential components of the resolution leg of the regulation, structure, resolution triangle that make up the reform debate.
Individual countries or economic areas such as the EU can get only so far on their own. To cope with distress in global banking, we need to decide whether or not we want international collaboration in the resolution of cross-border banking groups. That is a question which, one way or another, the highest reaches of the authorities should answer as part of the reform debate. Once that is clear, some of the questions about structure and regulation might become clearer.
A lot rests on whether we can resolve distressed LCFIs. We need to answer the big questions.
6 Examples of such safeguards include: creditors left no worse of than in liquidation; netting and set-off agreements respected; secured creditor claims and collateral kept together; and capital market arrangements, such as covered bond programmes and securitisations, protected.