## Andrew Bailey: Recovery and resolution plans
Remarks by Mr Andrew Bailey, Executive Director for Banking Services and Chief Cashier of the Bank of England, at the Santander International Banking Conference, Madrid, 17 November 2009.
I am grateful to Victoria Cleland, Chris Daniels, Iain De Weymarn, Andrew Hauser, Andrew Hewitt, Tony Lomas, Orla May, Maggie Mills, Sally Reid, Paul Tucker and Andrew Wardlow for their comments and advice.
* * *
It is a pleasure to be in Madrid today, not least because finding time to reflect on the future of the banking system means that we are at some distance from the urgent fire-fighting that has characterised the past two autumns. But I would not wish thereby to downplay the importance and urgency of the rebuilding task. The past two autumns were in many ways taken up with doing what had to be done in crisis conditions. And what had to be done involved the use of public money on a scale that has been unprecedented and was previously unimaginable. There was nothing easy about that task, and from the perspective of resolving the problems of the system too often we have had to grit our teeth, accept that 'we are where we are' and get on with the job in hand. Fire-fighters don't get to choose the fires they fight, or even sometimes how they fight them.
But fire prevention is better than fire-fighting. We cannot justify having a banking system that depends on the use of public money to douse the fire when the crisis comes. And we also cannot allow conditions to exist where risks are taken on the basis that this backstop exists. 'Too big' or 'Too important to fail' are shorthand for institutions that exist in their chosen form as a result of underwriting with public money. I don't believe that there is anything controversial about a statement that no industry can take on a form which is a consequence of such a massive potential, and of course recently actual, call on public money.
The big issue is therefore what to do to correct the situation. There are three elements to the debate: regulation, structure and resolution. We can re-design and re-build the fire prevention systems in the existing landscape. This is the regulation element. At the heart of this is action on capital requirements, including a role for contingent capital and larger liquidity buffers.
But we may also conclude that we want to prevent fires by having a different landscape of buildings - a town planning solution in which the structure of the industry is re-drawn. And in the past, cities were rebuilt with that objective in mind. This is the structure element.
But, whichever combination of these courses we choose to take, we cannot guarantee that there will be no fires in the future, and we would be extremely foolish if we made that promise. And, the odds are that the cause of the next one will not be the same. I have been involved in mopping up problems in banks long enough to have been involved in Barings, which was an idiosyncratic fraud, not a sector-wide problem. This is the resolution element.
In short, we need better fire prevention and better fire-fighting. In the UK we took a major step earlier this year by creating a Special Resolution Regime. But that is not the end of the response to the crisis. Rather, it is where 'living wills' or what we now prefer to call Recovery and Resolution Plans, or RRPs, come in. I want to spend the rest of my time setting out what I think should be the role and form of these plans, and why they should be critical tools both for banks themselves, banking supervisors and resolution authorities.
I have been struck by the strength of support for RRPs. In April, the G20 leaders endorsed a set of principles from the Financial Stability Board that include strong encouragement for firms to maintain contingency plans and procedures for use in a wind down, with regular reviews to ensure that they remain accurate and adequate. The Financial Stability Board cross-border Crisis Management Working Group, chaired by my colleague, Paul Tucker, has
published principles aimed at strengthening the ability of authorities to deal with a crisis in certain cross-border financial groups. In the UK, the FSA has recently issued a consultation paper on RRPs, and an important initial pilot exercise will involve reviewing the contingency plans and procedures of the major firms as they draw them up over the coming months. The timetable for this work is ambitious with the plans being drawn up in the first quarter of next year, but that is necessarily so. The FSA and the Bank of England, working together, will review and challenge the submissions to assess their credibility.
RRPs need to be owned by financial institutions at Board level. They should be of such importance that Boards need to understand that they are responsible for them.
At the heart of the role of these plans is taking a forward-looking view of the threats that should be tackled. This is a clear change of approach. It builds from the work already done to embed stress testing into the toolkit. Stress tests are useful but limited tools. They provide a forecast of what we think is most likely to happen in the future conditional on a set of assumptions on what the future will bring. There are three obvious limitations to stress tests used on their own. First, typically they are a single point forecast of what could happen rather than a distribution of outcomes conditional on a range of stresses. Second, because stress tests tend to take the form of point forecasts, they assume that we can predict the form of the next crisis, and typically they will place too much weight on the recent past to predict what will happen next. And third, they don't tell us what we should do about the stress. So stress tests are useful but nowhere near sufficient.
RRPs are a big step beyond that and one that should complement stress tests. The objective of a recovery plan should be to set out how a bank may react to a whole range of stresses and the steps that it could plausibly take to head off the impact of these stresses in order to avoid formal resolution actions in the event of failure. Recovery plans should encompass contingency funding plans and the use of contingent capital instruments as well as the sale of assets and/or business lines. They should not assume the use of extraordinary public support. A key element of the recovery plan is that it should make management more aware of the risks run in their businesses while they can still manage out of a stressed situation, thus reducing the likelihood of failure. This calls for a sharper use of risk management tools and systems. To be blunt, I have seen too many banks in the past two years where the systems simply did not support that sort of management under stress. I should also add that both internal and external auditors must play their full part in providing assurance that recovery plans are robust. Recovery plans must be ready to implement if it comes to it, but they should I hope and expect be used on a regular basis to test the robustness of business models, by having management rigorously ask the question, 'Can I recover from these stressed scenarios with my current business model?'.
The second leg of RRPs is the resolution plan. Resolution plans must be produced and owned by the authorities, since only we can determine how best to apply the tools of our regime. But firms have a vital role to play in these plans. They will have the best sense of the existence of impediments to a resolution plan, for instance the existence of negative pledges that could obstruct emergency lending and how to remove these impediments. Second, firms must produce and maintain the information that is necessary to enable a resolution plan to be enacted, and understand the problems and issues that would arise in a resolution from the way their firm operates. The last point is critical - it is not a data dump process.
There are two key objectives for resolution plans. First, to make it more straightforward and less costly for resolution authorities to resolve firms by creating a permanent state of readiness. Second, as a means to deal with the 'too complex to resolve' state by demonstrating where structures need to be changed. This is all about reducing the cost of resolution and thereby the risk to the stability of the system. Just like recovery plans, resolution plans need to be there to be used, and I can assure you that the Bank of England, in its role as a resolution authority, will be placing great emphasis on the existence of credible and useable resolution plans. They will be a tool for the authorities rigorously to ask
the question, 'With this structure and business model, could I achieve a resolution at an acceptable cost'. This will be true for domestic banks and for those operating across borders, and in the case of the latter there is a large amount of work to be done to ensure both complete coverage of resolution regimes and that they can work together.
I have emphasised that I see RRPs as a device to enable tough questioning of structures and business models. I want finally to illustrate the sorts of questions that should arise by referring briefly to a couple of well known cases of the last two years or so.
Let me start with Northern Rock, the first casualty in the UK and use this case to illustrate the role of a recovery plan. Had there been one, the Northern Rock recovery plan should have challenged the management and the Board on how its business model could adapt to and survive the closure of its key funding markets, given the sheer scale of its dependence on securitisation. Likewise, how could it have handled a serious stress when it had pursued a business model of aggressively taking market share by squeezing its net interest margin and accepting more borrower default risk? And then moving further into the guts of the bank, how would it deal with stress to a funding model where the bank was excessively dependent on cash deposits from the securitisation master trust but with a hard rating agency trigger linking that funding to the position of the bank, thus creating a severe amplification of any stress? Finally on Northern Rock, how would the bank deal with a funding stress when it was having to encumber such a high proportion of its good quality assets in order to over-collateralise its securitisation master trust?
A very good example of the need for resolution plans is the case of Lehmans in Europe. Two questions are relevant here: what should the resolution plan have contained; and what changes to its organisation and operations might have made the massive task of dealing with Lehmans more straightforward? You will, I hope, appreciate that I can only scratch the surface of the Lehmans case in the time available so I apologise for the laundry list approach. Starting with the useful content of a resolution plan: a detailed balance sheet for all relevant corporate entities in the group at the most recent month end; a clear mapping of financial and operational interdependencies between affiliates; wind-down plans for all business areas linked to a comprehensive information data room; and a contact plan for major stakeholders.
Now, listing the changes to Lehmans' organisation and operations that should have resulted from drawing up robust RRPs: corporate entity based accounts and management information to supplement the business-line version (because corporate entities fail); clear segregation of handling client asset activities; controls over depositing client money with affiliates; employees hired by the entity for whom they work; a clear record of each entity's title to business and associated intellectual property; robust risk systems that allow ready provenance of the balance sheet; contracts for the provision of key business services (eg banking systems) that allow continuity of provision of services in the event of a resolution; and arrangements for continuity of access to payment and settlement systems.
This has been a very brief run through some very relevant case studies. But I hope it has illustrated the role that RRPs should play and the pressing need to have them in place. RRPs have come a long way in a short period of time, which is both essential and gratifying. But, of course, we haven't actually produced any yet. This situation will not last much longer, certainly not in the UK, because the authorities are very keen to see real implementation of RRPs, alongside the continuing international work on them. One thing is for sure - you will hear a lot more about Recovery and Resolution Plans. Thank you.