Comments prepared for the debate sponsored by the South Centre on Revolution Required The Ticking Time Bombs of the G7 Model, book authored by Hervé Hannoun and Peter Dittus, Palais des Nations, United Nations, Geneva, 13 April 2018
For reasons explained at length by the authors the principal focus of Revolution Required (RR) is the monetary policy in the Advanced Economies (AE), which has been the main response to the Global Financial Crisis (GFC). This response the authors view as leading to an unsustainable increase in debt levels in the medium term and to investment which may not be viable in the event of a reversal of interest rates and a return of the cost of finance to more normal levels. While the macroeconomic causes and consequences of the GFC have not been properly addressed by the AEs, the weaknesses in regulation revealed by the GFC have been the subject of an extensive regulatory programme which has targeted increases in capital and better risk management supported by enhanced incentives. However, there continue to be questions concerning the adequacy of the proposals included in this programme owing to limitations of multilateral agreements on regulation, which must overcome divergences in national financial systems, and resistance to many of the reform proposals from the powerful banking lobby.
RR deals concisely with the achievements and lacunae of the regulatory agenda so far. The central element of this agenda is the revised version of Basel III developed by the Basel Committee on Banking Supervision (BCBS) intended for Emerging Market Economies (EMEs) as well as AEs. Banks’ capital buffers, based on a stricter definition of capital, have been increased, the capital requirement for common equity now amounting to 7 per cent of risk-weighted assets. For global systemically important banks (G-SIBs) a capital surcharge has been prescribed with the objective of increasing the resources available for meeting losses in cases of insolvency (Total Loss Absorption Capacity), and resolution rules and procedures (“living wills”) to be followed in such cases are now in place.
An aggregate minimum leverage ratio – capital in relation to a figure for total assets with limited allowance for manipulation –is now specified for banks but only at a level of 3 per cent. For large, globally significant banks there will be a floor for equity capital in the ratio’s numerator of 50 per cent of their risk-weighted capital requirements.
Little attention in RR is paid to revision of the rules for capital requirements for market or trading risks – understandably since the rules had not been finalised at the time of RR’s completion. I shall spend some time on a summary of outstanding issues under this heading, which indicate the complexity of Basel III and reasons for the slowness of progress towards completion.
The proposals on market risk, which had already been strengthened in Basel II.5 in response to experience during the early part of the GFC, underwent in January 2016 a thorough revision to deal with still unmet weaknesses (BCBS, 2016a). The deficiencies which this revision flagged included the following: the definition of the boundary between the banking and the trading book, long the subject of regulatory arbitrage by banks seeking to lower their capital requirements; and methods for risk measurement which, relying heavily on banks own models, were insufficiently robust and led to the provision of inadequate capital for banking systems as a whole.
More specifically in the reforms of Basel II.5 a key determinant of the boundary between the trading and banking book was banks’ intent to trade, which was inherently subjective and subject to misinterpretation capable of accommodating capital arbitrage. The reforms under the heading of the internal models approach were dependent on the framework of Value at Risk (VaR). This measure failed to take account of the substantial exposures to credit as well as market risk of trading exposures. The restriction of VaR to protection against risks beyond the 99th percentile was shown to leave banks vulnerable to “tail risks” which were indeed rare in non-crisis situations but led to unexpected large trading losses for banks during the GFC. Allowance under the internal models approach for market illiquidity – allowance for exiting or hedging trading-book exposures over a 10-day period –was not realistic for stressed conditions. Moreover recognition of the potential for risk reduction through hedging and diversification was too generous, based as it was on estimates of correlations from historical data during normal conditions.
The standardised approach to market risk of Basel II.5 was not a satisfactory fall-back for the internal models approach. Shortcomings of the standardised approach revealed by the GFC included the following: lack of sensitivity to different risk exposures; inadequate rules for recognition of hedging and diversification; and failure to capture risks associated with more complex – but increasingly common – complex trading instruments. Weaknesses of the standardised approach meant that it was not a credible threat to banks facing withdrawal of approval of their use of the internal models approach and was thus of little help to supervisors’ efforts to improve banks’ risk management practices.
The 2016 revisions included revisions of the boundary between trading and banking books through the following: additional guidance on instruments and exposures appropriate for inclusion in the trading book; reductions in banks’ ability to arbitrage this boundary with unjustified transfers becoming potentially eligible for additional capital charges; enhanced powers for supervisors to use their increased authority in the case of instruments improperly designated; and clearer treatment in the rules concerning internal risk transfers of trading instruments between risk classes.
A major change in the methods of risk capture under the internal models approach was the replacement of VaR with an Expected Shortfall (ES) metric. VaR is intended to measure the maximum loss at a specified degree of probability during a specified period, i.e. how bad are things likely to get. By contrast ES is intended to answer the question of what is the expected loss if things do get bad, likewise within a specified time horizon. The initial popularity of VaR in risk management was due less to its superiority as a measure than to its advantages in comprehensibility and facility in exercises of back-testing (Hull, 2007: 198-200). ES was to be calibrated on the basis of periods of significant market stress. Account was to be taken of market illiquidity through the introduction in the ES measure of the concept of varying liquidity horizons, such an horizon being defined as the time required to exit or hedge a an exposure without materially affecting market prices in stressed conditions.
The process for supervisory approval of a bank’s models was to become more granular and was to apply at the level of each of a bank’s trading desks rather than, as previously, at a bank-wide level. Approval was to depend on a desk’s proficiency in modelling the dependence of profit and loss on risk factors with an appropriate degree of accuracy. This would include a proper classification of risk factors into those which are “modellable” and those which are “non-modellable”, with the latter subject to a separate stressed capital add-on under the ES approach. Potential advantages to a bank from hedging and diversification were constrained by rules concerning the classification of risk categories and correlations eligible for inclusion in risk mitigation through diversification.
The objectives of the revised standardised approach of the January 2016 proposals included keeping it suitable for banks with limited participation in trading activity and providing a fall back for banks whose standards of internal modelling did not meet those required for eligibility for the internal models approach. The revised standardised approach is also designed to measure the risks of securitisation exposures in the trading book.
Closer calibration between the revised standardised and internal model approaches was to be achieved through a sensitivities-based method involving the use of standardised “bucket” risk weights reflecting stressed market conditions under an ES framework, which also incorporated varying liquidity horizons as in the internal models approach. The approach was now to reflect more fully risk sensitivities which were already an integral part of the models used for risk pricing and management by banks with a more extensive involvement in trading activity.
The revised standardised approach included a standardised Default Risk Charge calibrated to reduce potential discrepancies between capital requirements for similar risk exposures in the banking and trading books. There was also a Residual Risk Add-on designed to capture risks not already covered by the sensitivities-based method or the standardised Default Risk Charge.
But even the 2016 proposals are now being reconsidered. A new consultative paper issued by the BCBS in March of this year proposes revisions in both the standardised and the internal modelling approaches to setting capital requirements for market risk (BCBS, 2018). An Annex sets out proposals for a simplified standardised approach to capital requirements for market risk for a bank meeting the following criteria: it should not be a G-SIB; it should not make any use of the internal models approach; and it should not hold any correlation trading positions, i.e. securitisation positions not subject to strict conditions as to the underlying instruments and as to the sharing of the proceeds of the securitisation tranches. Further clarifications are included as to the boundary line between exposures eligible for inclusion in the trading as opposed to the banking book.
The revised standardised approach of March 2018 follows broadly the rules of that proposed in January 2016. But there are revisions to the measurement and risk calibration of exposure categories. Most of these revisions appear likely to reduce rather than increase capital requirements. This would appear to be the result , for example, of allowances for the liquidity of exchange rates greater than that specified in the January 2016 proposals and for certain over-conservative correlation scenarios. Moreover revisions are proposed to estimates of capital requirements for non-linear financial instruments in response to certain shocks. Reductions in risk weights are also proposed for whole classes of exposure to market risk.
Under the simplified alternative to the standardised approach to capital requirements for market risk the requirements consist of the sum of four risk classes (interest-rate risk, equity risk, foreign exchange risk and commodities risk), each multiplied by a scaling factor greater than one.
The new draft of the internal models approach proposes a more rigorous test for estimates of profit and loss to be met by the bank’s own risk management model if the bank is to be eligible for use of this approach. Trading desks failing this test would be ineligible, thereby becoming subject to the less risk-sensitive standardised approach or to formula-based add-ons. As mentioned above, the proposals of January 2016 also covered the treatment of “non-modellable risk factors” (NMRF), whose risks would not be eligible for inclusion in the internal models approach but would be capitalised on the basis of a stress scenario or the maximum possible loss if a stress test acceptable to the bank’s supervisor was not available.
It may reasonably be asked why so much space in my discussion should be devoted to a part of Basel III largely ignored by RR. My answer would be that the series of Basel III documents on capital requirements for market risk issued as part of the post-GFC reform agenda illustrate a major underlying point of RR, that the post-GFC reform agenda after a promising start has become increasingly detached from what should be its priorities.
Moreover the revisions of the rules for market risk concern a relatively minor part of what is anyway only a part of the reform agenda – though admittedly an important part. Moreover these revisions reflect what many would consider the questionable assumption that banking risk can be properly dealt with by adding further quantitative bells and whistles to an already elaborate set of measures of market risk. These of course may serve their purpose. But is it realistic to assume that they will provide protection in situations where market participants are subject to severe stress or panic – and may even conceal the intentions of positions taken in the interest of limiting their losses ? Might not simply good supervision and well designed incentives for bank operatives serve this purpose more effectively ? This question applies to AEs but perhaps still more strongly to EMEs.
As already mentioned, owing to the date of RR’s publication the authors could not comment on the March 2018 proposals. The continuing revision of these proposals does not seem to me consistent with the official statements at the end of last year applauding the finalisation of Basel III. The Basel Accord after all is supposed to include capital requirements for market as well as credit and operational risk.
Generally the authors of RR commend the changes to Basel III as well as other regulatory reforms introduced as part of the agenda up to the end of 2012. They draw attention to the increases in banks’ capital reported by the BCBS’s monitoring mechanism, though to levels which, especially for some large European banks, still involved shortfalls vis-à-vis Basel III’s final rules. Although the authors do not explicitly mention the reduced reliance in Basel III on reliance on banks’ own internal models, a subject of lengthy contention between European banks and their regulators, from the general tone of their comments I should expect them to approve. But their commendation of Basel III is accompanied by qualifications. One senses a certain scepticism on their part that even the new capital levels will be adequate in conditions of severe stress or crisis, in which credit and liquidity difficulties can be mutually reinforcing. Moreover the Basel III minimum leverage ratio of 3 per cent is hardly proving a binding constraint – with the 2017 figures for major banks already at 5 per cent and more distant historical levels for banks in AEs comfortably exceeding 10 per cent.
RR draws attention to the effects of lobbying by banks for a relaxation of the new rules in both the Eurozone and the United States. In the former there are pressures for a weakening rather than a strengthening of regulatory standards, even for banks still undercapitalised. This reflects partly erosion of banks’ profitability – and consequent pressure on banks’ capital – owing to the low or negative interest rates associated with the current stance of monetary policy. Moreover in the United States there is talk of a review of the Dodd-Frank Act, the comprehensive vehicle for the country’s post-crisis regulatory reform. This review likely to result in deregulatory changes in the Act’s framework.
RR draws special attention to four weaknesses of the post-crisis regulatory agenda: inadequate reforms of the rules for exposures to structured finance; the lack of a capital charge for interest-rate risk in the banking book; the continuing attribution in many cases of a zero risk weight to sovereigns; and unsystematic and inadequate reshaping of industrial structure in the banking sector.
The technology of structured finance has included – notoriously – creating tranches of securitised assets with credit ratings often higher than those of the different instruments in the underlying collateral. In the period before the GFC the tranches, frequently enhanced by guarantees from insurance companies, were distributed through specially created entities to investors (including banks) in the form of financial instruments designed to correspond to different degrees of investor risk preference or aversion. This technology played a major role in the development of the so-called shadow banking system which consisted of entities with balance sheets resembling in many respects those of more traditional financial institutions but which was not subject to appropriate regulation. The exposures created by the shadow banking system contributed to levels of leverage and thus vulnerability in the financial system in ways which rendered external scrutiny and measurement extremely difficult.
Despite the exceptionally large losses associated with these practices during the GFC the so-called “super senior tranches” of securitisation were not subject to special treatment in the rules of Basel III a “notorious failure” for RR. Moreover the BCBS has actually been subject to pressures to relax capital requirements on at least some forms of securitisation in the interest of generating a revival of such financing. In the United States financial engineering of structured products, partly based on different forms of securitised collateral, remains active.
On interest-rate risk in the banking book the BCBS has opted for an approach in the form of supervisory review of this risk under Pillar 2 of Basel III (BCBS, 2016) rather than a standardised measure for the potential impact of such risk on banks’ equity. This more cautious approach was adopted in response to feedback from consultations. These emphasised the complexities of a standardised measure which, given the heterogeneous nature of interest-rate risk in the banking book, would not be sufficiently accurate and risk-sensitive to allow it to serve as the basis for appropriate regulatory capital requirements. The authors of RR clearly believe that this approach is likely to be an inadequate response to the dangers to banks and their customers which could result from a sharp return (“snap-back”) to more normal levels of interest rates – a return which could result in a crash in bond values of “supernova” dimensions.
The credit risk weight for sovereigns under Basel III’s existing rules is not automatically zero: under the Standardised Approach the risk weight depends on a country’s credit rating; and under the Internal Ratings Based Approach banks make their own estimates of the probability of a sovereign’s default. However RR notes that in the Eurozone there is widespread resistance to abandoning a generalised risk weight of zero for sovereigns regardless of their creditworthiness. The BCBS now also appears to be moving away from the choice of a zero risk weight for sovereign exposures. Its new ideas on the subject after further consultations include the following: more rigorous definition of different categories of sovereign; abolition of reliance on banks’ internal models for most sovereign exposures for which risk weights would henceforth follow a standardised approach; risk-weight add-ons to mitigate concentration risk for most sovereign exposures; and more rigorous supervision and disclosure (Pugsley, 2018; BCBS, 2017).
Structural changes affecting big banks have actually been undertaken by the governments of some countries (FSB, 2014). In the United Kingdom, for example, retail banking is now ring-fenced in a separate legal entity from investment banking and related activities. Moreover there is an EU proposal for the prohibition of proprietary trading by banks at group level and the possible separation of other trading activities into separate legal entities. In the United States Dodd-Frank includes restrictions on commercial banks’ engagement in proprietary trading (the Volcker Rule) and in derivatives trading (a measure which originated in an amendment introduced by Senator Blanche Lincoln) as well a cap on the size of banks and other systemically important financial institutions (waivable in situations characterised by the danger of default).
Nevertheless, like many other commentators, the authors of RR are concerned that, despite the new rules imposed on systemically significant institutions, there is still a real possibility that in periods of severe stress big banks will be protected from bankruptcy, with the costs of such protection borne by the population as a whole. Here they draw sympathetic attention to the proposal of Neel Kashkari, President of the Federal Reserve Bank of Minneapolis, that consideration should be given to the possibility of “breaking up large systemic banks into smaller, less connected, less important entities” or “turning large banks into public utilities by forcing them to hold so much capital and so little leverage that they virtually can’t fail” (Corkery, 2016). Under Kashkari’s proposal banks with assets of USD 250 billion or more would have to increase their common equity to 23.5 per cent. According to Kashkari’s analysis of IMF data his proposal would reduce the probability of another banking bailout to 9 per cent from the 67 per cent which he views as the probability under current reforms. Another proposal from the upper reaches of the United States regulatory community has come from the perennial gadfly, Tom Hoenig, vice-chairman of the Federal Deposit Insurance Corporation, who has proposed that the largest banks set up separately capitalised banking and investment units, receiving in exchange relief from Dodd-Frank stress tests (Orol, 2017).
The crisis has generated many other proposals concerning size and corporate governance in the financial sector. These proposals have undoubtedly been influenced by several factors: the choice of officially sponsored and arranged mergers in many AEs as an instrument for avoiding more widespread bankruptcy during the GFC – which in the United States has led to increases in the size of major banks and in the level of concentration in the sector – despite doubts that adequate management and internal control of large financial conglomerates is still possible.
A fuller review of the many ideas and proposals on banking structure and governance provoked by the GFC would take me far beyond the scope of my comments today. I should none the less like to describe briefly a comprehensive, if utopian, scheme put forward by the well known financial commentator, David Shirreff, because of the way it brings together under a single heading so many of the dimensions of possible reform of bank structures and governance (Shirreff, 2016).
Shirreff’s underlying vision of the financial system is as a utility – non-financial analogues which he cites being the sewage system and electricity distribution. The key functions of banks for ordinary people in his view are to facilitate payments and to safeguard and keep track of cash balances. These functions are performed primarily by retail banks, which should thus be in separate institutions with capital levels and deposit insurance rendering failure almost impossible.
The second of the three major institutional groupings in Shirreff’s scheme is corporate/wholesale banking. The core activities of such banking would be lending, cash management, and standard foreign-exchange and interest-rate hedging services. These could be accompanied by limited and carefully defined support to investment banks with restrictions designed to ensure avoidance of the interconnectedness which made the collapses of Bear Stearns and Lehman Brothers so potentially dangerous. Corporate/wholesale banks would be subject to capital and liquidity rules but not necessarily the same ones as retail banks.
The third of Shirreff’s groupings –which is the subject of his most radical proposed treatment – is “pure” merchant and investment banks. These would provide advice and transaction services, underwriting the placement of shares and bonds, and taking equity and lending stakes in new and existing ventures. Their activities would include trading in financial instruments. Especially notable is Shirreff’s proposal that the corporate form for this grouping should be partnerships, long common amongst investment banks. The partnership in his view achieves a better alignment of interests between a bank and its top management. Guidelines for employees are less likely to accommodate or encourage risky trading since losses will be met by banks’ owners, i.e. the partners. There would be no legal capital requirements for this group, and regulation would be light, consisting principally of conduct-of-business and stock-exchange rules. This would not exclude ad hoc regulatory measures considered necessary, for example, to restrain overheated property markets which are frequently harbingers of crisis situations.
I should add that the regime reforms sketched by Shirreff are not limited to his proposals on corporate groupings. They include caps on bankers’ remuneration and the abolition of bonus pools, and forbidding the purchase and sale of credit derivatives in regulated banks unless there is also an offsetting position in the underlying credit.
RR’s coverage includes interesting comments and ideas on several subjects other than macroeconomic policy and regulation. I particularly liked the brief chapter on G7 foreign and defence policy. The authors’ observations seemed especially apt on the way in which NATO has been surrepticiously transformed from a defensive alliance into a broader military and political grouping with objectives including enlargement eastwards and military intervention in the Middle East and Asia.
Appropriately RR devotes considerable attention to environmental problems, which the authors view as requiring radical policy responses based on a new concept of the common good. Steps taken so far as part of the financial regulatory agenda indicate growing awareness of the need for action under major headings including the following: globally agreed carbon pricing, carbon caps, and carbon taxation as well as large transfers of resources to developing countries to facilitate their control of emissions and management of other problems caused by climate change. On the financial front initiatives under the auspices of the Financial Stability Board (FSB) have so far concerned mainly climate-related disclosures by financial institutions. Much publicity has also been given to issuance of green bonds.
Less well reported are initiatives at the level of banking contracts and transactions. Many of these initiatives involve soft commodities and deforestation (International Chamber of Commerce, 2016). The initiatives reflect the realisation that continuation of current trends will have disastrous environmental consequences and, by obvious extension, disastrous consequences for business – and for supporting finance – in the activities involved. An example of these initiatives is the resolution of the Consumer Goods Forum (CGF), a network of about 400 consumer goods companies with combined sales of EUR 2.5 trillion, to achieve “zero net deforestation” in their supply chains by 2020. A group of international banks has worked with the CGF to achieve agreement on a “Soft Commodities Compact” to align the banks’ services with the Forum’s goal. Other actions involve appropriate incorporation in risk management and pricing at the level of individual firms of risks due to unsustainable production and trade of soft commodities. All this no doubt seems painfully slow in the face of the pressing character of environmental and climate-change problems. But it does represent a start.
I am delighted to have been a member of the panel discussing the book of Hannoun and Dittus – from which I learnt much – and I am grateful for your attention.
BCBS (2016a), Explanatory note on the revised minimum capital requirements for market risk, BIS, January;
BCBS (2016b), Interest rate risk in the banking book, Standards, BIS, April;
BCBS (2017), “The regulatory treatment of sovereign exposures”, Discussion paper, BIS, December;
BCBS (2018), “Revisions to the minimum capital requirements for market risk”, Consultative document, BIS, March;
Corkery M (2016), “Minneapolis Fed chief proposes eliminating ‘Too Big To Fail’ banks”, New York Times, 16 November 2016;
FSB (2014), Structural banking reforms: cross-border consistencies and global financial stability implications, Report to G20 leaders for the November 2014 Summit, 27 October;
Hull J (2007), Risk Management and Financial Institutions, New Jersey, Pearson/Prentice Hall];
International Chamber of Commerce Banking Commission (2016), 2016 Rethinking Trade and Finance an ICC Private Sector Development Perspective, ICC;
Orol R (2017),”FDIC’s Hoenig suggests a bank rule that even the White House may love”, The Street, 13 March;
Pugsley J (2018), “Has Basel made banks safe ?”, The Banker, March;
Shirreff (2016), Break Up The Banks a Practical Guide to Stopping the Next Global Financial Meltdown, Brooklyn and London, Melville House Publishing.