## Paul Tucker: The reform of international banking - some remaining challenges
Speech by Mr Paul Tucker, Deputy Governor for Financial Stability at the Bank of England, Member of the Monetary Policy Committee, Member of the Financial Policy Committee and Member of the Prudential Regulation Authority Board, at the Oliver Wyman Institute Conference, London, 1 October 2013.
* * *
Over the next few years the international regulatory authorities will move towards completing the programme of reforming the global financial system. Domestically, that will mean implementing those reforms, as discussed in the Record of the Financial Policy Committee's latest meeting published this morning.
Today, I plan to sketch a few of the issues in banking regulation that I think will confront policymakers as they take this forward. I should stress at the outset that these are my views and do not necessarily represent the views of any of the domestic or international bodies of which I am a member.
Just as a scene setter, let me recap the central components of the international reform programme, led by the 'G20' Financial Stability Board. They are essentially as follows. 1
To make banks more resilient to stress - the essential steps are, of course, new regulatory requirements for capital and liquidity. Second, to ensure that banks and other financial firms can fail in an orderly way without taxpayer solvency support - that's the resolution agenda. Third, to simplify the network of counterparty-credit exposures amongst banks and others that arises from derivatives transactions - central counterparties (CCPs). Fourth, to ensure that those CCPs are not themselves Too Important To Fail - resolution again, plus rules to allocate losses amongst bank clearing members and others so that a distressed CCP can recover without going into resolution. Fifth, to ensure that the reregulation of banks does not simply cause the systemic risks associated with excess leverage and liquidity mismatches to be reinvented outside banks - the shadow banking reforms. And sixth, to make securitisation safer and more efficient - a complicated raft of reforms, including overdue steps to remove mechanistic reliance on Credit Rating Agency ratings.
Cumulatively, those reforms will change the face of global finance - for the better. But no doubt there will also be effects that have not been anticipated, resulting in a need for running repairs - a point to which I will return. The work to produce the reforms has also revealed conceptual gaps and flaws in the inherited regulatory framework that will, sooner or later, need to be addressed. That is where I shall begin.
## The capital framework: going concern v. gone concern loss-absorbency
Perhaps most important, policymakers have focused more closely than ever before on what happens when a bank or dealer fails. Regulatory policy needs to work backwards from that, since banks are regulated in order to reduce the costs of failure. Policymakers are, therefore, now engaging with the crucial distinction between those parts of a bank's capital structure that absorb losses in a going concern - namely, equity - and those that can smoothly absorb losses in resolution or liquidation. The recent G20 Leaders Summit called on the Financial Stability Board to produce plans over the coming year for the level and location of goneconcern loss-absorbing capacity (or GCLAC) in global banks and dealers. In a nutshell, this
1 See Financial Stability Board (2013), 'Narrative progress report on financial reform'; Tucker (2012), 'Banking in a market economy - the international agenda', published in AFME (2012), Investing in Change: A book of essays on financial reform in Europe ; and Tucker (2013), 'Competition, the pressure for returns and stability', published in Dombret and Luciuis, Stability of the Financial System; Illusion or Feasible Concept .
will be a policy for the amount of term bonded debt issued by banks, and where in the group structure it is issued from.
I believe that in time the Basel Capital Accord could usefully be recast so that it has distinct components for going-concern and gone-concern requirements. That would replace what to my mind is the fuzzy distinction between what are termed 'Common equity tier 1', 'Additional tier 1' and 'Tier 2' capital - not all of which is capital in the ordinary sense of the term that it can absorb losses outside of liquidation. 2
The other element of this policy - which, again, the FSB will be addressing - is where in a group GCLAC should be issued from, ie whether from the group holding company, from intermediate regional or national holding companies, or from operating banks and dealers. This turns on the type of resolution strategy most appropriate for a particular group; ie topdown or Single Point of Entry (SPE) where a whole-group resolution is executed from the topco, or Multiple Point of Entry (MPE) where a group is split into parts which, as necessary, are resolved separately.
Of course, it is in operating companies - banks and dealers - that life-threatening losses are generated. Where those losses exceed the equity base of an operating subsidiary, they need either to be absorbed by debt issued into the external market or to be passed up the group structure via intra-group debt or guarantees. The first - GCLAC, ie term debt, issued to the market - is associated with Multiple-Point-of-Entry resolutions. The latter, where losses flow to the top of a financial group, are associated with Single-Point-of-Entry resolutions. 3 The choice of how equity and gone-concern loss-absorbing capacity should be distributed across a group is, therefore, part and parcel of deciding its preferred resolution strategy. That was not understood as international banking policy was developed over recent decades, and sets up my next issue.
## Reconciling the Basel Concordat, the Core Principles of Banking Supervision, the Capital Accord, and resolution policy
The development of international policy on resolution is, arguably, exposing an underlying need to reconcile the provisions of various international agreements in banking regulation and supervision. There are three that are relevant: the Concordat, the Core Principles and the Capital Accord. Bear with me for a moment while I set up the point I want to make.
The Concordat 4 establishes the division of labour amongst home and host authorities - the foundation stone of international co-operation amongst bank supervisors. First agreed in 1975 in the wake of the failure Herstatt, updated after Banco Ambrosiano in 1983 and again in 1992 following the closure of BCCI, the Concordat was most recently spring cleaned in 1996. It is premised upon 'consolidated supervision' of a banking group as a whole. What that means is that so long as the host supervisors of local outlets of an international banking group are satisfied that the home supervisor is capable of supervising the consolidated group as a whole, the distribution of capital resources across the group, both within and across jurisdictions, can be unfettered. In principle, legal entities within the group can be lightly capitalised so long there is adequate capital in the group as a whole.
2 See Tucker (2013), 'Banking reform and macroprudential regulation: implications for banks' capital structure and credit conditions.'
3 See Financial Stability Board (2013), 'Recovery and Resolution Planning for Systemically Important Financial Institutions: Guidance on Developing Effective Resolution Strategies'; and Tucker (2013), 'Resolution and the future of finance'.
4 For the most recent version see BIS (1996), 'The Supervision of Cross-Border Banking', available online at: http://www.bis.org/publ/bcbs27.htm
Meanwhile, the (1997/2006/2012) Core Principles for Effective Banking Supervision 5 set out minimum standards for prudential supervision. Supervisors are required to look at a bank on a solo basis; on a consolidated basis covering all related banking entities in a group; and on a group-wide basis, ie taking into account non-banking related entities as well. Under Principle 12, a home supervisor should supervise a bank on a consolidated basis, but 'in addition ... the responsible supervisor supervises individual banks in the group. The responsible supervisor supervises each bank on a stand-alone basis and understands its relationship with other members of the group.' Under Principle 16, jurisdictions may - I stress, may - also opt to be assessed by the IMF for compliance with the Principles against the 'additional' criterion 6 that a supervisor should ensure an adequate distribution of capital within a banking group. But, to repeat, that is not an obligation. So summing up, as with the Concordat, a country's prudential regime could be judged as complying with the Core Principles even though it does not ensure that the distribution of capital across a banking group leaves its individual entities adequately capitalised.
Finally, the Capital Accord 7 - now in its third incarnation - applies to banking groups on a consolidated basis (or at a sub-consolidated basis for different tiers of a group). The Accord gives jurisdictions an option to apply capital requirements on a solo-basis to individual legal entities as an alternative to applying the requirements to consolidated groups of legal entities 8 , but there is no requirement to exercise that option. Again, then, there is no hard rule that ensures a sensible distribution of capital across individual entities within a group.
If you cannot fit all that together, you can be excused!
Here's an attempt to make sense of it. There appears to be a disconnect between the Concordat (consolidated supervision); the Core Principles (both consolidated and solo perspective recognised as important, but no mandated solo capital requirements); and the Capital Accord (consolidated and sub-consolidated, with solo subject to deductions allowed but not required).
Or, put another way, each of the three BCBS documents does an important job but none of them addresses the distribution of capital across a group. That is a gap.
Do not assume that this is a theoretical point. It is one reason there have been attempts by regulators to shore up the financial soundness of local subsidiaries during the crisis. In a crisis, sovereign countries care about the international distribution of losses. Standard setters should look into this.
They should do so in the light of the principles and policies that have been agreed on how to improve the resolvability of banks - at both the consolidated and solo level. Because, quite simply, it is legal entities that fail.
5 For the most recent version see BIS (2012), 'Core principles for effective banking supervision', available online at: http://www.bis.org/publ/bcbs230.htm.
6 Paragraph 33 of the foreword to the Core Principles states that 'countries undergoing assessments by the IMF and/or the World Bank can elect to be graded against the essential and additional criteria. It is anticipated that this will provide incentives to jurisdictions, particularly those that are important financial centres, to lead the way in the adoption of the highest supervisory standards. As with the essential criteria, any assessment against additional criteria should recognise the concept of proportionality.'
7 See BCBS (2013), 'Basel III: A global regulatory framework for more resilient banks and banking systems', available online at: http://www.bis.org/publ/bcbs189.pdf.
8 To understand the difference of applying capital requirements on a solo basis relative to a sub-consolidated basis, imagine Bank A owns Bank B which in turn owns Bank C. Under both approaches, capital requirements apply to the group on a consolidated basis. Under the solo-approach, capital requirements are also applied to each of Bank A, Bank B and Bank C individually, after having deducted from Bank A's capital, its investment in B, and from B's capital its investment in C. Under the sub-consolidation approach, capital requirements would be applied on the consolidated accounts of Bank B, ie after adding the assets and liabilities of B and C and eliminating any intra-group exposures.
For groups for which Single-Point-of-Entry - broadly, bail in of debt issued through a holding company - is the preferred resolution strategy, local subsidies need a combination of equity and intragroup GCLAC (ie intragroup bonds) that permits losses to flow smoothly to the topco. By transferring losses upwards, an ailing subsidiary is recapitalised. Whether that suffices for the host authorities will depend on their confidence in the capacity and willingness of the group's home authorities to execute an effective resolution of the topco. By contrast, for groups for which Multiple-Point-of-Entry - breaking up a group - is the preferred resolution strategy, the host authorities of a subsidiary identified for 'separation' must ensure that it has a capital structure and sufficient operational independence to enable them to resolve it themselves with their particular set of powers. Put simply, for firms within a group for which SPE is the preferred resolution strategy, the group as whole is a source of strength on which local supervisors can place weight. That is less true for groups for which MPE is the preferred resolution strategy.
Summing up, from a resolution perspective, the distribution of loss-absorbing capacity (the sum of going-concern and gone-concern loss absorbency) is crucial. Yet, from a regulatory and supervisory perspective, there is as yet no international standard that requires a banking group to ensure a sensible distribution of equity (and debt) within the group. The international standards on banking will need to catch up.
## The calibration of capital requirements: stress testing
But how much capital (equity and gone-concern loss-absorbing capacity) should a bank have, and how should regulators decide? I will conclude by saying something about how stress testing can help.
The minimum standard set in the Basel Capital Accord is just that: a minimum standard. There has to be provision for add-ons for individual banks' particular circumstances; and also for temporary across-the-board increases when the environment has become more threatening than anything envisaged when the Accord was calibrated, ie when the world has become extraordinarily risky. In the United Kingdom, the Financial Policy Committee (FPC) has the power to recommend, or direct, the micro-supervisors to vary capital requirements temporarily in order to protect against just such unusual threats to stability.
But in order to use those degrees of freedom, policymakers need a good framework for assessing banks' capital adequacy in different circumstances, both at the level of the system a whole and for individual firms. The Bank of England - Financial Policy Committee and Prudential Regulation Authority - believe that a regime of concurrent stress testing of banks is an important part of the answer.
Following a recommendation from the FPC earlier in the year, the Bank has today published a Discussion Paper setting out plans, to be implemented progressively over a number of years, for just such a stress-testing regime for the United Kingdom. In broad terms, each year the FPC would design a (set of) common scenario(s) that it judges the banking system needs to be able to survive; the PRA would apply those scenarios to the major banks, dealers and building societies, alongside more firm-specific scenarios; drawing on analysis from across the whole institution, Bank staff would analyse capital adequacy under stress drawing on a range of models, including firms' own models. The results would inform judgments by the FPC and the PRA Board on whether remedial measures were needed. In some degree, both the stress scenarios and the results would be published.
That very brief summary leaves a lot of questions unanswered, but suffice to say that we believe that a stress-testing regime can enhance the quality of the Bank's macro and micro prudential supervision and, over time, underpin confidence in the banking system.
In addition, I believe that a stress-testing regime will be very important for another reason: accountability to the public via Parliament. The Bank of England now has extensive powers as a micro supervisor of individual firms and a macro supervisor of the financial system as a whole. Parliament has set objectives for both - primarily, safety and soundness for the PRA,
and the resilience of the system for the FPC. But unlike in monetary policy, we do not have a quantified target - and, to be clear, we do not think that that would be possible. To have such power with limited transparency would be uncomfortable. If nothing else, it would leave the Bank in a position where supervision was of interest only when it obviously goes wrong. We need to engage society - public, Parliament - in debates about the resilience of the financial system and how we are supervising it in the good times as well as in bad times. The FPC and PRA have already taken various steps to help with this - the six-monthly Financial Stability Review, the published Record of FPC policy meetings, the Committee's statement on the use of its directive powers, the PRA's Approach document and annual report, and so on. But stress testing can provide a quantum leap in transparency and accountability.
For example, the FPC will need to be able to defend its choice of stress against questions and criticisms that the scenarios are too extreme or not tough enough. That will implicitly be a healthy debate about objectives. At the micro-supervisory level, the PRA Board will need to explain some of its remedial measures using the published stress tests. This can make for a richer public discussion of prudential supervision, as well as encouraging firms to address latent vulnerabilities promptly.
That brings me back to resolution. Occasionally a firm will be revealed to be so feeble that major surgery is needed. Sometimes recovery measures, overseen by the supervisors, will suffice. Sometimes they will not, with a firm having to go into resolution instead. A sensible stress-testing regime - indeed effective supervision in general - cannot work without a welldesigned resolution regime. That should not be a surprise, but is in fact one of the great lessons of the crisis.
## Conclusion
Over past decades, banking, and banking supervision, somehow got detached from the most essential truth: that banks can fail. Thus, we have banks using brands that do not represent any legal entity, as if banks were simply trying to sell something, not running a business in borrowing and lending. We have an international regime that encourages but does not require adequate loss-absorbency in each and every entity within a banking group. We have had a regime that is focused on the capital needed in a going concern, but not on the organisational and capital structures that would help the authorities to contain disorder in the face of chronic distress. All those issues are now, I believe, in the open. There is a clear way forward on resolution. But there will have to be some consequential recasting of the international regulatory regime.
Of course, none of the reforms will be fool proof. Mistakes will be made; mistakes we cannot see. In future, it has to be possible to make running repairs to the regime where faultlines are exposed.
Stress testing - not only of banks but of CCPs and, in time, of other intermediaries - can help to expose those faultlines. It can help to calibrate macro and micro prudential policy. And a good stress-testing regime will make it easier to explain and defend the standard of resilience that the Bank requires. And for that, the Bank can be held to account.