## Ten years on: Lessons from Northern Rock
Speech given by Sir Jon Cunliffe, Deputy Governor Financial Stability, Member of the Monetary Policy Committee, Member of the Financial Policy Committee and Member of the Prudential Regulation Committee
Single Resolution Board Annual Conference, Brussels
Friday 29 September 2017
1
This month marks the 10 th anniversary of the failure of Northern Rock.
The failure of a middle size British mortgage bank was followed by the collapse of much bigger and globally systemic banks.
But the picture of the queues forming outside the branches of Northern Rock remains for many the picture of the start of the financial crisis.
Prior to September 2007, I knew of the existence of bank runs from economic history books, the experience of developing and emerging economies and the film of Mary Poppins.
Seeing one happen in London brought home that it is never different 'this time'. The fundamental basis of the financial system is trust. And trust, if not properly managed and protected, can disappear instantaneously.
And once that happened, the UK had no effective way of managing the failure of a bank. And no way to avoid the taxpayer having to step in to stop such a failure leading to the loss of critical services to the economy and contagion to other banks.
Anniversaries, even those of difficult events, offer a chance to step back and reflect - to ask whether we have learned the lessons of Northern Rock and the bank failures that followed?
I want today to look at why we were unable safely to wind up a failing bank without taxpayer intervention, at the progress we have made in the UK towards rectifying those failings, and at some of the challenges that remain.
## The pre-crisis regime
The failure of Northern Rock in 2007 and later, of RBS and Lloyds exposed brutally that the UK lacked the tools needed to manage the failure of a bank.
Depositors were expected to take comfort from a depositor compensation scheme based on the principle of 'co-insurance'. They were fully covered only for the first £2,000 of their deposit. Any depositor with more had a strong financial incentive to run on hearing rumours that the bank was failing - which is what they did on hearing that Northern Rock was in receipt of emergency liquidity assistance from the Bank of England.
When that happened, there was no public authority formally and clearly responsible for dealing with failing banks and with the powers to match. The UK relied wholly on its standard corporate insolvency regime to handle the fall-out from a bankrupt bank.
2
This required an insolvency practitioner to protect the interests of creditors as a whole. The public authorities had no means of directing the insolvency process to prioritise the protection of depositors, the continuation of the bank's critical functions or wider financial stability concerns. There was no way to recapitalise a failing bank by bailing in its creditors.
As a result, when faced with the failure of a bank of any size, the UK authorities had only a stark choice between a very disruptive insolvency, putting financial stability at risk, or a taxpayer bailout.
The UK was not of course the only jurisdiction to discover in the financial crisis that it did not have an effective regime for dealing with the failure of banks.
And, as we discovered when Lehman Brothers failed, the authorities in different jurisdictions had no preagreed means of coordinating with each other on how to deal with the sudden failure of a major cross-border bank. Nor had they any confidence that the actions taken by authorities in another jurisdiction would align to their own national interest.
With hindsight it is clear we had been lulled into a false sense of security. In Britain, Northern Rock represented the first major run on a bank since the failure of Overend Gurney in 1866. Unlike, for example, the US, the UK had not experienced bank runs in the 20 th century.
Banks in the UK had failed. But insolvency had generally been avoided by the Bank of England twisting arms to 'encourage' other banks, to support a rescue 'lifeboat', as in the secondary banking crisis of 1973 or to take over the failing bank as with Barings in 1995.
The failure of BCCI in 1991 did lead to a protracted and difficult insolvency. But the bank's activities in the UK were not of a scale that posed risks to the UK financial system and economy
The Bank of England's approach to bank failure, an approach shared by many central banks, was one of 'constructive ambiguity'. The aim was to guard against moral hazard by maintaining uncertainty about whether a bank in trouble would get liquidity support with the threat of insolvency in the background.
In the event the threat of insolvency proved neither credible nor effective. When Northern Rock got into trouble with £23bn in customer deposits and a balance sheet of £100bn, the UK financial authorities were left with no way of rapidly transferring parts of the failed bank's business to another bank; or of recapitalising the bank by imposing losses on shareholders and creditors without serious risks to financial stability.
3
And so, on 17 September 2007, the Chancellor was compelled to guarantee all Northern Rock deposits. Subsequently the bank was nationalised to avoid insolvency and to allow the bank's critical functions to continue while a buyer was found for its deposits.
The lessons of Northern Rock were dramatically underscored a year later by the failure of RBS and Lloyds. Faced again with the choice of insolvency or taxpayer intervention, the government was forced to inject £37bn into the two banks. The banks were not put into insolvency and it proved impossible fully to write down the existing shareholders or impose losses on the bond holders.
The first lesson from this was that constructive ambiguity simply didn't work. The market had always suspected this to be the case. The implicit subsidy enjoyed by the largest banks before the crisis illustrated the market's assumption that the state would always intervene to prevent their bankruptcy. And the market was generally proved right.
The exception that proved the rule was the Fed's decision to let Lehman Brothers enter insolvency. The fallout globally from Lehmans demonstrated dramatically, why, for large banks that perform critical functions, insolvency is not a viable option.
The second lesson, which follows from the first, was that if we want, when a bank fails, to have better options than a disruptive insolvency or a taxpayer bailout, those options have to be put in place well beforehand. Doing so comes at a cost to banks and to public authorities. But if at the point of failure, there are no better options than were available in 2008, moral hazard cannot be avoided.
Rather than constructive ambiguity, we need credible clarity that when a bank gets into trouble, the losses will be made to fall on shareholders and creditors and not the taxpayer. And if the bank provides critical services to the economy, that these can continue while the bank is resolved in an orderly way.
## The Purple Book
It is in order to achieve such credible clarity, that next week the Bank will publish an update of its approach to resolution. This document - known by the colour of its cover as the Purple Book - was first published in October 2014. Its purpose is to set out very clearly the options that the Bank has to deal with a failure of a bank and the way in which we would use our powers.
Explaining how resolution is designed to work in practice and what is needed to remove barriers to resolvability are necessary steps to ensure that resolution regimes are credible.
4
The Purple Book illustrates the scale of progress that has been made in the UK towards putting in place a credible and effective way of dealing with bank failures. I would pick out three crucial areas of reform:
First, there is now a comprehensive statutory framework to deal with failing banks, for which the Bank of England is formally responsible. We have statutory powers to match this responsibility and there are a wide range of options available to us.
Unlike in 2008, there, is now a special bank insolvency procedure which requires that the insolvency practitioner prioritise pay-out of insured depositors or the speedy transfer of these deposits to a purchaser. Eligible depositors are protected up to £85,000. 'Co-insurance' has been consigned to history.
For the large number of small building societies and banks, this procedure should mean that their failure can be managed and their depositors paid out or transferred quickly following entry into insolvency - leaving shareholders and creditors to take the losses.
But the larger banks that hold the majority of deposits in the UK provide critical functions for the economy. These would be disrupted by insolvency, even in the new regime. Were a large bank to fail, the Bank of England can now trigger the use of 'stabilisation' powers outside insolvency. These powers include 'bail-in' to recapitalise the bank by imposing losses on shareholders and creditors, so that its critical operations can continue. This provides time for the firm to be safely restructured to address the causes of failure.
Second, we are well on the way to ensuring that if a bank fails and is taken into resolution, there will be sufficient, private sector, financial resources, in the form of debt and equity, that can be bailed in to absorb losses and recapitalise the bank so that it can continue to operate.
To achieve this the Bank has set every UK bank (and building society) a requirement for the minimum amount of such loss absorbing resources - known as MREL - it needs to hold. This requirement will need to be met in full by 2022. MREL covers both the capital a bank holds in going concern and the capital and debt that can be bailed in if it fails and enters resolution.
The biggest UK banks already have going and gone concern resources sufficient to absorb losses of almost a quarter of their risk-weighted assets, and are well on their way towards meeting their full MREL requirements. Current levels of loss absorbing resources mean that even if the major UK banks saw losses six times the losses they incurred over 2008 and 2009, there would be sufficient private sector resources that could be bailed in to recapitalise the bank and stabilise it without taxpayer support.
And next week, alongside the Purple Book, we will publish for consultation the Bank of England's proposals on how these loss absorbing resources should be distributed within banking groups.
5
In setting the timetable for meeting full MREL requirements, the Bank has considered the balance of costs and benefits. Issuing MREL imposes costs on banks. These costs will be minimized if banks are able to build up their MREL loss absorbing debt to replace their existing debt as it matures. It could be counterproductive if, in seeking to impose requirements intended to address financial stability, we did so on a timetable that dislocated the banking system and made it more not less vulnerable during the process.
Third, statutory powers and loss absorbing resources are necessary but not sufficient conditions for an effective resolution regime. There are other barriers to resolution that also need to be addressed. Resolution must provide continuity, whether continuity of access to financial market infrastructure, continuity of contracts or operational continuity. For example, services such as IT that underpin critical functions will need to be set up in a way that enable them to continue in resolution. The Bank is working alongside the PRA in assessing firms' readiness to meet operational continuity in resolution.
We should however have no illusions about the resolution of a major bank. If it happens, even when the regime is fully in place, it will be a very painful exercise. Resolution is not a magic wand; losses will need to fall on creditors. Even if we are prepared in advance, stabilising a large failing bank will not be easy.
But taken together, these reforms mean that we would be able to handle a failing bank very differently today compared to 2008.
There are in place now credible options, other than insolvency or bailout, that ensure that bank shareholders and creditors will bear losses if a bank fails.
And we are much better able now than we were to ensure that a failing bank can if necessary be stabilised so it can continue to provide critical services to the economy.
## What remains to be done?
However more remains to be done.
First and foremost, we need to implement fully the reforms I have mentioned. Banks need to continue to build up the necessary loss absorbency and to restructure as necessary to ensure operational continuity in resolution.
This is perhaps an obvious point, but one that needs repeating. It is 10 years since Northern Rock failed and memories may be beginning to fade.
6
Ensuring that we have a better option than insolvency or a bailout, is not costless. And, as that cost becomes apparent in a number of jurisdictions, there are increasingly voices calling for the reforms to be watered down or abandoned. It is argued that they are too expensive for banks, especially small banks, to implement and will restrict lending to the real economy.
On the cost, I would emphasise the cost not only to the taxpayer but to the economy as a whole of disruptive bank failure. In the UK, the going concern capital regime is based on assurance of there being an effective way to resolve failing banks. Absent such assurance, if this risk of disruptive bank failure remained as in 2008, we would require banks to hold appreciably more capital to absorb losses.
In the UK, the Independent Commission on Banking suggested a capital surcharge of 3% of risk-weighted assets (RWA) for banks that could not be resolved. In its 2015 assessment of capital adequacy, the Financial Policy Committee estimated that capital would need to be around five percentage points higher if there was no resolution regime in place.
It is also argued that resolution will not work; that while it may be a way of dealing with idiosyncratic failures, in a systemic crisis authorities will be reluctant to bail in shareholders and bank creditors.
This binary distinction between idiosyncratic failure and fully blown systemic crisis seems to me oversimplistic. For sure, if we were suddenly to find ourselves pitched back into the middle of a systemic crisis, with a number of major institutions having failed or on the point of failure and a complete breakdown of trust and confidence, orderly resolution of individual banks in itself is unlikely to be able to stabilise the system as a whole.
But we did not suddenly arrive in late 2008 in the midst of full blown financial crisis. Had the authorities had better, less disruptive options available over the previous 18 months to deal with a series of failing institutions, the disruptive, explosive nature of the crisis might well have been minimised. There would still have been very major losses and failures. But we would, in my view, have had a much better chance of a more orderly, less damaging correction of an overleveraged banking system.
And it is also probable that the buildup of leverage and bad debt itself would have been significantly restrained in the years before the crisis by the discipline imposed by shareholders and bondholders aware that they stood first in line to bear losses if the banks failed.
We not only have to follow through domestically. There is further work to do to ensure that we can not only manage the failure of a bank large enough to be systemically significant in one jurisdiction, but that we can manage the failure of large, internationally active banks.
7
Again, if we want better options than bailout or insolvency, we need to continue working now, internationally, to put those options in place.
This of course requires trust and cooperation between the home authority for the group and the host authorities of the jurisdictions in which it has major operations. But trust and cooperation, while necessary conditions, are not in themselves sufficient.
Home and host authorities needs to agree, in advance, on the resolution strategy for a major cross border banking group. Hosts need to be confident that the chosen strategy is viable and that it will respect their own financial stability needs.
In other words, confidence that local operations will not be cut loose in resolution and that, if necessary, hosts have the ability to draw down loss absorbing resources through the parent.
We have made very significant progress in agreeing international standards on resolution. And we have established crisis management groups, and, in the EU, resolution colleges, that bring together the regional supervisors for the major cross border banks. These provide the mechanism for agreeing, in advance, the strategy for managing the failure of a major bank with a presence in a number of jurisdictions and for monitoring progress in making that strategy possible.
This requires the continued build up, in line with the agreed international standard, of loss absorbing resources that can be used to recapitalise the group in resolution. And hosts' confidence in the resolution strategy will depend on the contractually enforceable allocation of those resources across the groups major operating entities, again in line with international standards.
Recapitalising a bank in resolution by bailing in the private sector restores solvency. But a solvent bank in resolution will still have liquidity needs. Our first preference - and that of the resolved firm - would be for the bank to meet those needs from some combination of its own liquid assets and private funding sources. But we cannot guarantee resolved firms will be in this position, even once they have been restored to solvency and are continuing to meet the requirements for authorisation.
Ensuring solvent firms in resolution have access to public sources of liquidity is therefore a critical part of an effective resolution strategy and an area where there is great merit in clarity. The clearer it is to the bank's creditors, counterparties and financial market infrastructure firms that a resolution comes with the expectation of access to public liquidity, if needed, the smaller the amount that may end up needing to be drawn-down. In contrast, doubts over whether and how liquidity will be made available to a bank upon its entry into resolution risk undermining a resolution, and leaving the authorities with a potentially far bigger problem.
8
So it is important to ensure resolution and liquidity strategies are aligned - whether they are delivered by a single institution, as in the UK, or by separate ones.
The Purple Book, to be published next week will set out the Bank of England's approach to providing a liquidity backstop in resolution, where required.
To be eligible, a bank will need to be restored to solvency by the bail in of shareholders and debt holders. It will need to meet the PRA's authorisation conditions, including capital requirements, so that it can continue to operate while in resolution. Solvent but illiquid banks in such a situation would have access to the Bank's published facilities subject to meeting the necessary eligibility criteria.
To supplement those arrangements, the Bank also has put in place a new, flexible Resolution Liquidity Framework providing the tools to lend to banks which are in a Bank of England led resolution. 1
Such liquidity may be secured against a wide range of collateral, building on the collateral eligible in Sterling Monetary Framework operations. 2 The Bank's objective would be to provide liquidity in sterling or foreign currency as required, in the necessary scale and for a sufficient period of time to allow the firm to make the transition to market-based funding. The terms would be set in a way designed to support the effectiveness of the resolution regime, incentivise the transition of the firm back to market-based funding, and protect public money.
Confidence in the regime for the resolution of international banks is of crucial importance to the UK. We are home to a number of major international banking groups. But equally, if not more important, we are host to a very large number of foreign banks, many of which have sizeable wholesale market operations in the UK. As the leading international financial centre, we import considerable risks from other jurisdictions. It is therefore crucial to financial stability in the UK that we can rely on foreign banks operating in our jurisdiction having viable resolution strategies in line with international standards. Absent such assurance, we would need to ensure the entities operating here have greater resilience locally.
## Non bank resolution
It was the failure of systemic banks in the crisis that exposed the lack of an effective resolution regime. But the lesson that if you want better options to deal with a failing financial institution they need to be put in place beforehand applies more widely than banks.
1 RLF would not be available to any firm subject to an insolvency or administration procedure.
2 See http://www.bankofengland.co.uk/markets/Pages/money/eligiblecollateral.aspx
9
So we need also to think about whether resolution is necessary in other parts of the financial sector, particularly systemically important insurance companies and CCPs. These did not fail in the last crisis. But they may pose similar problems in a future episode of stress.
Insurance companies are very different animals to banks. Insurer failure is more likely to be in much slower motion and solvent run-off may present a credible solution.
However we need to think carefully about whether we can rely wholly on that and what, if any, systemic risks could arise from insolvency of a major insurance company. Given the relatively small amount of debt in insurance companies relative to policy liabilities, who should bear the losses in resolution? And what tools are needed to allocate them? The answers to these questions are not yet fully apparent, though in my view, the case is probably made for a resolution regime for insurers, if only as a precaution should it turn out that run-off is not enough.
## CCPs
CCPs are also very different to banks. They exist to manage and reduce the risks faced by their members to ensure financial contracts are reliably and transparently margined and collateralised.
The main prudential risks CCPs face is from the failure and consequently inability of clearing members to meet their obligations to the CCP.
The steps that have been taken since the financial crisis to increase the resilience and resolvability of their bank clearing members are therefore a key protection for CCPs. But CCPs also need the backstop of a credible resolution regime.
The principal challenge here is not solvency per se but rather the ability of a CCP to restore itself to a matched book if members default, and to do so in a way that does not undermine the stability of the system. CCP rule books provide for very substantial mutualised resources and a comprehensive series of recovery actions. These include, as a last resort, the cancellation or 'tear up' of contracts and the end of the clearing service. However, waiting until the mutualised resources of a CCP are exhausted and subjecting participants to the unpredictability of a full tear-up may well pose unacceptable risks to financial stability. Resolution allows the resolution authority to intervene, if necessary, to tear-up a subset of contracts earlier in the process than would be possible in CCP recovery, and before the mutualised resources in the CCP have run out. The losses can then be spread across the membership in the order set out in the rule-book. This is key to allowing clearing members to measure and manage their exposures at a time of stress.
It is therefore important that if the resolution authority intervenes it avoids disturbing the order of losses in the rule-book.
10
This requires a robust 'No Creditor Worse Off' (NCWO) safeguard which takes as its counter-factual the loss allocation rules in the CCP's rule book.
The issue of NCWO protection is the subject of current debate in the draft EU regulation on CCP Recovery and Resolution. One school of thought is that it is essential to have a weaker NCWO safeguard so that resolution authorities have flexibility to deviate from the way in which losses would fall under a CCP's rules.
This degree of flexibility puts particular focus on the objectives by which home authorities would exercise this discretion. Resolution aims principally to deliver financial stability. But in the case of a global CCP, whose financial stability will be given prominence?
This uncertainty may leave participants located outside of the home jurisdiction fearing that they will be exposed to disproportionately greater losses in order to protect the home jurisdiction's financial stability.
The UK does not support that approach. As with banks, it is entirely possible to establish resolution frameworks for CCPs that ensure that interests are mutually aligned and that do not permit or require the home authorities to protect national financial stability at the expense of participants outside of that jurisdiction.
The conduct of CCPs depends on clarity and certainty. A regime that does not provide the same certainty as that set out in CCP rules runs the risk of undermining the very reasons why international leaders have placed CCPs at the heart of the response to the financial crisis.
## Conclusion
To conclude. Anniversaries are not always milestones to celebrate - and the anniversary of the failure of Northern Rock is a case in point. But even where the lessons of the past were painful, their anniversary provide a chance to step back and consider whether they have been learned.
The lessons of the crisis, of course, go much wider than resolution. The first defence against bank failure is to ensure that banks are properly capitalised to withstand losses and continue to serve the real economy. A vast amount of work has been done over the past 10 years to put in place capital and liquidity standards and stress testing regimes to ensure a much safer and stronger banking system.
Resolution should be seen as an integral part of making the financial system safer and stronger. Credible resolution regimes that impose losses on shareholders and investors, rather than taxpayers, when things go wrong will incentivise banks to manage their risks properly.
1
And regimes that enable systemic banks to be stabilised if they fail, so that they can be resolved in an orderly way without disrupting critical economic functions will reduce the cost of such failures to the real economy.
There is still significant work to do to implement fully the resolution regime, domestically and internationally. And resolving a failing bank will never be a simple or painfree exercise. But, as the Purple Book and recent experiences in a number of jurisdictions show, we increasingly have available to us options that we did not have 10 years ago which will contribute to a safer and stronger financial
12