The Financial Crisis and the Future of Financial Regulation

This is a transcript of The Economist’s Inaugural City Lecture, which was delivered by Lord Adair Turner in London on January 21, 2009.

It is stating the obvious to say that over the last 18 months, and even more so the last four, the world financial system – and particularly but not exclusively the world banking system – has suffered a crisis as bad as any since the stock market crashes of 1929 and the various banking crises that followed. As a result, banking systems in many countries are suffering from an impaired ability to play their vital role in credit extension to the real economy and a process of deleveraging threatens severe adverse effects on real economic prospects. The crisis therefore presents the financial authorities – central banks, regulators and finance ministries – with two challenges:

• The first and most urgent is to design short-term policies so as to at least limit the adverse impact of deleveraging and deflation on the real economy. We cannot make that impact nil, but we do know how to avoid the policy mistakes which turned the initial problems of 1929-30 into the Great Depression. Fiscal and monetary policies need to be carefully designed, and – as we approach a zero interest rate and consider quantitative easing options – need to be increasingly coordinated. And there are a wide range of policies which can be taken to free up financial markets – measures which Ben Bernanke last week labeled “Credit Easing” – funding guarantees, liquidity provision, tail risk insurance, direct central bank purchases of assets, and regulatory approaches to capital regulation which avoid unnecessary pro cyclicality in capital adequacy requirements. The measures announced by the Chancellor of Exchequer on Monday were designed as an integrated package, which will have a significant impact. And if more measures are acquired they can and will be taken.

• It is not, however, on this challenge of short-term economic management – where the lead must be with the fiscal and monetary authorities – that I’m going to talk tonight. But instead on the second challenge: how to design the future regulation and supervision of financial services so that we significantly reduce the probability and severity of future financial crises? Last September, when I took over as Chairman of the FSA, the Chancellor asked me to conduct a review of our regulation and supervision of the banking system, and I will deliver that Review in March, alongside the publication of a comprehensive FSA Discussion Paper. That paper will set out the changes the FSA has already made, those where we have proposals in principle but need to consult on details, and those where we have defined our objectives but now need to play our role in achieving international agreement.

Those proposals for regulatory change need to be grounded in analysis of what happened – why this crisis occurred. Tonight therefore I will concentrate on that analysis. I will then draw out some issues and possible implications relating to the future shape and size of the banking and credit mediation markets. I will finally and briefly outline three changes which we know are in principle essential.

What happened and why?

So what happened? Why did this extreme crisis occur? I think with hindsight – and it is only with hindsight – a fairly compelling and broadly agreed explanation of what has occurred can be set out. At the core of the crisis was an interplay between macroeconomic imbalances which have become particularly prevalent over the last 10-15 years, and financial market developments which have been going on for 30 years but which accelerated over the last ten under the influence of the macro imbalances.

Macro-imbalances. First, the macro side. The last decade [Exhibit 1] has seen an explosion of world macro-imbalances, with very large current account surpluses piling up in the oil exporting countries, China, Japan and some other east Asian developing nations, and large current account deficits piling up in the USA, but also in the UK, in Ireland, Spain and some other countries. A key driver of those imbalances has been very high savings rates in countries like China; since these high savings are in excess of domestic investment, China and other countries must accumulate claims on the rest of the world. But since, in addition, those countries are committed to fixed or significantly managed exchange rates, these rising claims take the form of central bank reserves, typically invested not in a wide array of equity, property or fixed income assets – but almost exclusively in apparently risk-free or close to risk-free government bonds or government guaranteed bonds.

This in turn has driven a reduction in real risk free rates of interest to historically low levels [Exhibit 2]. In 1990 you could invest in the UK or the US in risk-free index-linked government bonds at a yield to maturity of over 3% real; for the last five years the yield has been less than 2% and at times as low as 1%.

These very low medium- and long-term real interest rates have in turn driven two effects:

• First, they have helped drive rapid growth of credit extension in some developed countries, particularly in the US and the UK – and particularly but not exclusively for residential mortgages [Exhibit 3] – with this growth accompanied by a degradation of credit standards, and fuelling property price booms which for a time made those lower credit standards appear costless.

• And secondly, they had driven among investors a ferocious search for yield – a desire among any investor who wishes to invest in bond-like instruments to gain as much as possible spread above the risk-free rate, to offset at least partially the declining risk-free rate. Twenty years ago a pension fund or insurance company selling annuities could invest at 3.5% real yield to maturity on an entirely risk-free basis; now only 1.5%: any products which appear to add 10, 20 or 30 basis points to that yield without adding too much risk look very attractive.

Financial sector innovation. The fundamental macro economic imbalances have thus stimulated demands which have been met by a wave of financial innovation, focused on the origination, packaging, trading and distribution of securitised credit instruments. Simple forms of securitised credit – corporate bonds – have of course existed for almost as long as modern banking. In the US, securitised credit has also played a major role in mortgage lending since the creation of Fannie Mae in the 1930s; and securitisation had been playing a steadily increasing role in the global financial system and in particular in the American financial system for a decade and a half before the mid-1990s. But it was from the mid-1990s that the system entered explosive growth in both scale and complexity:

• With huge growth in the value of the total stock of credit securities [Exhibit 4]

• An explosion in the complexity of the securities sold, with the growth of the alphabet soup of structured credit products.

• And with the related explosion of the volume of credit derivatives, enabling investors and traders to hedge underlying credit exposures, or to create synthetic credit exposures. [Exhibit 5]

All of these developments, in different ways, seeking to satisfy the demand for yield uplift, and all predicated on the belief that by slicing and dicing, structuring and hedging, using sophisticated mathematical models to understand and manage risk, we can “create value” by offering investors combinations of risk and return which are more attractive than those available from direct purchase of the underlying credit exposures.

This explosion was supported by and in itself drove big increases in the leverage of major financial institutions – in particular investment banks and the investment banking activities of some large universal banks. [Exhibit 6]

And as it developed the rapid growth began to drive and to be driven by one of those self-fulfilling cycles of falling risk aversion and rising irrational exuberance to which all liquid traded markets seem at times to be susceptible:

• Credit spreads on a wide range of securities and loans falling to clearly inadequate levels. [Exhibit 7]

• The price charged for the absorption of volatility risk falling because volatility seemed to have declined. [Exhibit 8]

• And these falling spreads and volatility prices driving up the current value of a range of instruments, marked to market value on the books of banks, investment banks and hedge funds – fuelling in turn higher apparent profits and higher bonuses, and as a result reinforcing management and traders certainty that they must be doing the right thing.

Until we reached the point where people began to fear that the music was about to stop – but where others felt, in Chuck Prince’s words, that they had to keep dancing till the band stopped, which it did in summer and autumn 2007.

A cycle therefore of irrational boom and then bust; and therefore in some ways no different from other cycles which we have seen in markets in the past: in equities, in property, in South Sea project participations, in tulips. But what makes this one different – and potentially more economically destructive to the real economy – is that it is the first major global boom and bust of securitised credit instruments. Because at the core of this story is the development of a new model for delivering credit intermediation – the originate and distribute model of securitised credit. And one of the crucial questions we therefore have to ask is whether this originate and distribute model is inherently riskier than the one that it has partially replaced – or whether, provided we regulate it more effectively, it is capable of being a more stable system, or indeed of delivering the positive benefits of increased financial stability which its advocates originally proposed.

So before talking about the response to the crisis, I will make four observations relating to the growth and the implications of the securitised credit intermediation model:

Securitised credit: initial proposition and subsequent evolution

First, as already said, securitised credit has a history going back many decades. But it really began to take off in the 1980s, and it is interesting to revisit the arguments made in its favour at that time. One argument was greater liquidity for end investors (an issue I’ll come back to), but another crucial argument was that securitisation would reduce risks for individual banks by passing credit risk to end investors, reducing the need for unnecessary and expensive bank capital [Exhibit 9]. Rather than a regional bank in the US holding a dangerously undiversified holding of credit exposures in that particular region, which created the danger of a self-reinforcing cycle between the decline in a regional economy and the decline in the capital capacity of regional banks – much better to package up the loans and sell them through to a diversified set of end investors. And indeed it was argued that securitised credit intermediation could reduce risks for the whole banking system, since while some of the credit risk would be held by the originating bank and some by other banks acting as investors, much would be passed through to end non-bank investors. Credit losses would therefore be less likely to produce banking system failure.

But that is not what happened. Because when the music stopped – as these figures from the IMF Global Financial Stability Report of April, 2008 make clear [Exhibit 10] – the majority of the holdings of the securitised credit, and the vast majority of the losses which arose, did not lie in the books of end investors intending to hold the assets to maturity, but on the books of highly leveraged banks and bank-like institutions.

Because what increasingly happened [Exhibit 11] was that the credit securitised and taken off one bank’s balance sheet, rather than being simply sold through to an end investor, was:

• bought by the propriety trading desk of another bank;

• or sold by the first bank but with part of the risk retained via the use of credit derivatives; or

• used as collateral to raise short-term liquidity – creating a complex chain of multiple relationships between multiple institutions, each performing a different small size of the credit intermediation and maturity transformation process, and each with a leveraged balance sheet requiring a small slice of capital to support that function.

Some banks were truly doing “originate and distribute”: but the trading operations of other banks (and sometimes of the same bank) were doing “acquire and arbitrage [1]. The new model left most of the risk still somewhere on the balance sheets of banks and bank-like institutions but in a much more complex and less transparent fashion.

Changing forms of maturity transformation

My second point is that in this story, what happened to maturity transformation and to assumptions about liquidity was particularly important. One of the key socially valuable functions of the banking system is to deliver maturity transformation, holding longer term assets than liabilities, and thus enabling the non-bank sector to hold shorter term assets than liabilities. This absorbs the risks arising from uncertainties in the cash flows of households and corporates, and results in a term structure of interest rates more favourable to long-term capital investment than would pertain if no such maturity transformation were being performed.

It is a very important function of undeniable social value, but also one which creates risks. If everybody wants their money back on the contractual date, no bank could repay them all. Therefore we have insurance via lines from other banks, liquidity policies to measure and limit the extent of maturity transformation, and ‘lender of last resort’ facilities provided by the central bank. A complex balancing act of individual bank practices, regulatory policies and central bank facilities and discretion, but at least one we know is of central importance.

But one of the striking developments of the last several decades has been that a growing part of this maturity transformation has been occurring not on the books of regulated banks with central bank access, but on the off-balance sheets of banks, and on the balance sheets of shadow banks or near banks. SIVs and conduits performed large-scale maturity transformation between short-term promises to noteholders and much longer term instruments held on the asset side. Investment banks funded holdings of long-term to maturity assets with much shorter term liabilities. And while mutual funds with long-term assets and immediately available redemption were not banks since their liabilities to investors did not have certain capital value, the implicit promise not to “break the buck” meant that their behaviour in a liquidity crisis – selling assets rapidly to meet redemptions – could reinforce the liquidity crisis elsewhere.

While it is difficult to get the aggregate figures, it therefore seems highly likely that the aggregate degree of maturity transformation being performed by the financial system in total has increased substantially over the decades. And it is certainly clear that a wide range of institutions – both banks and near banks – have been relying on “liquidity through marketability” to assure themselves that their maturity transformation activity is safe. “Liquidity through marketability” – i.e. I can count this as a one-day asset because I can sell it within a day in a liquid market – has always been an important concept. Instruments which have long contractual tenor but which can be sold or discounted to generate immediately funds have been a key element in bank liquidity management since the days of Bagehot. But the extent to which the bank and near-bank system in total has relied on “liquidity through marketability” has increased dramatically over the last three decades and particularly in the years running up to the crash. The system in total has become significantly more reliant on the assumption that a very wide range of assets could be counted as liquid because they would always be sellable in liquid markets [2].

And while some of these developments – in particular the growth of SIVs, and investment bank balance sheets and mutual funds – were most prevalent in the US and less important elsewhere, the impact in a global funds market was felt throughout the world. Northern Rock and Bradford & Bingley were directly or indirectly dependent on the maturity transformation function of US mutual funds and SIVs, enabling them to access the funds of short-term US investors to provide long-term UK mortgages , with the macro-imbalances I mentioned earlier, including the feature that while the US was a huge net recipient of Asian central bank investment, it was simultaneously on the private-sector side, a large net investor in, among other things, British residential mortgage backed securities [Exhibit 12]

Irrational exuberance in credit prices more harmful than in equities?

So we have had the huge growth in securitised credit intermediation and a related increasing reliance of the total system on liquidity assured by marketability. That raises the question of whether a system of securitised credit intermediation is inherently more risky, at the systemic rather than the idiosyncratic level, than a system of on-balance sheet intermediation.

In a securitised system, credits become marketable instruments, tradeable in liquid markets. And we know from history that all liquid markets – markets where you can buy something one day in the hope of selling it the next day for profit – can be susceptible to swings in sentiment which produce significant divergence from rational equilibrium prices. The historical record of such irrational swings has been extensively documented by economists such as Kindlebeger, Minsky and Shiller: and the root causes of these swings in human psychology and in the incentives facing institutions and individual traders are increasingly well understood. Internet equity prices in 2000 were driven irrationally high by irrational exuberance, and subsequently fell. Bond yields were driven irrationally low and prices irrationally high by irrational exuberance between 2002 and early 2007, and the yields subsequently soared, the prices collapsed.

But while the former boom and bust in equity prices had surprisingly small consequences for the real economy, the latter boom and bust is likely to have a much bigger one. And that contrast may be inherent. It may well be that the world economy has greater ability to absorb without dire consequences severe cases of irrational exuberance and then depression in equity prices, than in the prices of a broad range of credit instruments, held to a significant extent on the trading books of banks, shadow banks or near banks. Banks are highly leveraged: they perform maturity transformation which exposes them to liquidity risk: and they are involved in a process of continual rollover of new credit supply to the real economy without which economies will contract. Irrational swings in the prices of credit securities held by banks, and thus in their capital resources, are therefore likely to be far more economically significant than irrational swings in the prices of equity investments held by end investors.

It is therefore possible that the growth of the securitised credits intermediation model has increased some aspects of systemic risk in ways which are not just the result of poor execution – bad remuneration practices, inadequate risk management or disclosure, failures in the credit-rating process – but absolutely inherent.

But if that is true, it would be precisely the opposite of what many clever, hardworking and well-meaning people believed about the securitised credit markets only two years ago. The IMF’s Global Financial Stability Report of April 2006 stated that [Exhibit 13] ‘the dispersion of credit risk by banks to a broader and more diverse set of investors, rather than warehousing such risk on their balance sheets, has helped to make the banking and overall financial system more resilient’. It noted that this dispersion would help to “mitigate and absorb shocks to the financial system” with the result that “improved resilience may be seen in fewer bank failures and more consistent credit provision”. And many other economists argued that in addition to increasing financial stability, the development of securitised credit, structured credit, and of credit derivatives, by creating or completing markets which had not previously existed, must have increased economic allocative efficiency.

The core issue which we now need to face is whether in that analysis we significantly overstated the allocative efficiency benefits which could possibly have arisen from this completion of markets, even if they had operated rationally, and significantly underestimated the inherent dangers that any liquid-traded market will at times be susceptible to irrational exuberance followed by irrational despair.

The growth of the financial sector: fundamental benign effects, illusory profits and rent extraction.

My fourth and final observation about the growth of the securitised credit markets is that it is striking that it has been accompanied by a quite remarkable growth in the relative size of wholesale financial services within the overall economy. If, for instance, we look at debt as a percent of GDP – an income measure of leverage – [Exhibit 15] we do indeed see a growth of household debt as a percent of GDP, and to a smaller extent of corporate debt as a percent of GDP, but what is really striking is the extent to which the debt of financial companies as a percent of GDP has grown, both in the US and in the UK.

On a consolidated basis of course – i.e. stripping out claims between financial institutions – financial sector assets and liabilities can only grow pari passu with non-financial sector liabilities and assets. So what this disproportionate growth represents is an explosion of claims within the financial system, between banks and investment banks and hedge funds – that multiplication of balance sheets involved in the credit intermediation process which I suggested earlier has accompanied the increasing complexity of securitization.

This huge growth of intra-financial system leverage has a relevance to the urgent issue of short-term macro-economic management. The more that we can ensure that bank deleveraging takes the form of the stripping out of inter-trader complexity, and the less it takes the form of leveraging vis-á-vis the non-bank real economy, the better. But for this evening my focus is on why this growth has occurred, why indeed many other measures of financial system importance – output as a percent of total GDP, profits as a percent of total corporate profits, as financial sector market cap as a percent of total equity market capitalisation [Exhibit 16], show a similar long-term trend, with a strong acceleration in the last five years up to 2007.

For this growth appears to be at variance with one of the other arguments made for securitisation, that it would be a more cost efficient system – delivering the service of credit intermediation to the real economy at a lower total cost. So why has the wholesale financial services industry instead had to grow so significantly?

Well, there are some underlying and entirely benign factors which do tend to increase the relative importance of financial services (retail and wholesale combined) as incomes grow. The wealthier people become, the more lifecycle consumption smoothing that occurs, and the more diverse they become in their preferences for consumption at different points in their life cycle; as a result there is a simultaneous increase in demand for both savings and borrowing products. And the more complex and globalised the world economy becomes, the more complex are the functions which the world’s banks have to perform in intermediating credit and other flows, and in themselves managing and helping corporates manage, the risks that arise from global operations, and fluctuating exchange rates, interest rates and commodity prices. In general, income per capita and financial sector value added as a percent of GDP are somewhat correlated, across at least a range of income per capita levels, for inherent and benign reasons.

But it is also possible that the importance of financial services as a percent of GDP has been swollen by two other factors – one of which is illusory and short term, and the other harmful and longer term.

• The illusory one arises from mark to market profits in a rising market. If the bank and near-bank system in total holds a net long position in those assets which we mark to market – which it does – and if irrational exuberance can push the price of those assets to irrationally high levels (which I think it clearly did in the years running up to early 2007) then mark to market accounting will swell declared profit in an unsustainable way, but in a way which, reflected in bonuses, may reinforce management and traders’ determination to do more of the clever stuff, which is delivering those profits.

• The possible long-term and harmful possibility is rent extraction. For there must be a suspicion that some and perhaps much of the structuring and trading activity involved in the complex version of securitised credit, was not required to deliver credit intermediation efficiently, but achieved an economic rent extraction made possible by the opacity of margins and the asymmetry of information and knowledge between end users of financial services and producers. Simply put, wholesale financial services, and in particular that element devoted to securitised credit intermediation and to the trading of securitised credit instruments – grew to a size unjustified by the value of its service to the real economy, and is now going through a downsizing, part of which is cyclical, but part a permanent one-off adjustment to a more economically efficient size.

Now of course, if you are an extreme Chicago school economic liberal, what I have said cannot be the case. If the industry grew dramatically in the decade to 2007 that must be because it was performing value added services: if complex product innovations were able to sustain themselves economically, they must have been socially useful innovations. But after what has happened, I think we know that that is not the case. I think we know that imperfections and irrationality in financial markets which are not fixable just by disclosure, but are inherent, mean that financial innovation which delivers no fundamental economic benefit, can for a time flourish and earn for the individuals and institutions which innovated, very large returns.

Not all innovation is equally useful. If by some terrible accident the world lost the knowledge required to manufacture one of our major drugs or vaccines, human welfare would be seriously harmed. If the instructions for creating a CDO squared have now been mislaid, we will I think get along quite well without. And in the years running up to 2007, too much of the developed world’s intellectual talent was devoted to ever more complex financial innovations, whose maximum possible benefit in terms of allocative efficiency was at best marginal, and which in their complexity and opacity created large financial stability risks.

Implications and what to do next

So if that’s what happened, and some of the reasons why it happened, what are the implications for the possible future shape of the financial system, and what are the implications for how we should manage the future financial system?

One implication is that there is a very important macro element to the required response.

Major macro-economic imbalances – large surpluses and deficits – were an important underlying driver of what occurred, and their more effective management is important not just to a more stable global system in the long term, but to the challenge of limiting the severity of the immediate economic downturn. The fundamental problem of potentially asymmetries between the policy responses of deficit countries and surplus countries, which lay at the root of the mismanagement of the gold standard in the 1920s and early 30s, and which Keynes warned of ahead of Bretton Woods, remain a crucial issue today. Without more Chinese consumption to balance more Americans saving, the deflationary impact of the crisis could be prolonged.

And looking to the long-term, as we think about what is needed to avoid future crises, it is clear that better analysis of and response to macro-prudential problems – problems which lie at the interface between macro-economic policy and financial system regulation – will be vital. The FSA has been more open than I think any institution involved in this crisis in admitting that it made mistakes in the institution specific supervision of northern rock. But I think the best judgement is that better institution specific supervision of Northern Rock would have made only a very small difference to the shape and impact of this global crisis.

The far bigger failure – shared by bankers, regulators, central banks, finance ministers and academics across the world – was the failure to identify that the whole system was fraught with market-wide, systemic risk. The key problem was not that the supervision of Northern Rock was insufficient, but that we failed to piece together the jigsaw puzzle of a large UK current account deficit, rapid credit extension and house price rises, the purchase of UK mortgage-backed securities by institutions in the US performing a new form of maturity transformation, and the potential for irrational exuberance in the market price of credit. We failed to realize that there was an increase in total system risk to which financial regulators overall – authorities, central banks and fiscal authorities – needed to respond.

Regulators were too focused on the institution-by-institution supervision of idiosyncratic risk: central banks too focused on monetary policy tightly defined, meeting inflation targets. And reports which did look at the overall picture, for instance the IMF Global Financial Stability Report which I quoted earlier, sometimes simply got it wrong, and when they did get it right, for instance in their warnings about over rapid credit growth in the UK and the US, were largely ignored.

In future, regulators need to do more sectoral analysis and be more willing to make judgements about the sustainability of whole business models, not just the quality of their execution. Central banks and regulators between them need to integrate macro-economic analysis with macro-prudential analysis, and to identify the combination of measures which can take away the punch bowl before the party gets out of hand. We also need to create deliberate mechanisms to increase the likelihood that major analytical institutions such as the IMF challenge the conventional wisdom rather than go along with it. And we need to ensure that when the IMF or other international surveillance bodies do issue warnings, that big powerful countries, and not just weaker developing countries potentially dependent on IMF support, feel that they have to respond.

Alongside that more effective macro-prudential analysis, however, we also need more effective approaches to the regulations which govern the financial system. That regulation needs to be designed in the light of how the system we are regulating is likely to evolve. Two questions can help frame our thinking about the future: what is going to happen, and what should happen, to the originate and distribute securitised credit model: and what will and should be the institutional relationship between “narrow banking” – deposit taking, extending loans, and providing payment services – and more complex treasury and trading activities?

On the originate and distribute, securitised credit model, I argued earlier that, especially if it involved a substantial holding and trading of securitised credit instruments on the balance sheets of banks involved in maturity transformation, it created significant and inherent risks. But it does not follow that the originate and distribute model will now or should now largely disappear. Some of the arguments which were advanced in favour of this model can be good ones: taking regionally or sectorally concentrated credit risk off bank balance sheets and distributing it to diversified investors, could be beneficial. Many forms of credit, for instance residential mortgages, are best credit assessed via quantitative scoring techniques, rather than by individual bank officer judgement, and can therefore be turned into securities, the risk of which can be well captured in credit ratings. And while we are now facing a crisis of the securitised credit model, we must remember that the past has had many examples of credit crises in good old fashioned on balance sheet banking – the British secondary banking crisis of 1973 to 74, the US savings and loans debacle of the 1980s, the Japanese and Swedish banking crises of the 1990s.

It is therefore, I believe, highly likely that the future system will and should involve a combination of traditional on balance sheet credit intermediation and securitised intermediation, and that a combination of better regulation and market response to the crisis, should and will produce a safer version of the originate and distribute model – less complex, more transparent to end investors, with less exclusive reliance on credit ratings and more independent judgement, and with less packaging and trading of securitised credit through multiple balance sheets. The securitised originate and distribute model will change significantly but it will still play an important role in national and global credit intermediation.

Narrow banking and investment banking. The somewhat related issue is then, should the different functions of classic on balance sheet banking – deposit taking, loan extension, and payment services provision – and more complex and risky investment banking activities be done in the same institutions or in separate firms. The actual trend of the last year has clearly been for these functions to be combined to a greater extent than before – as Bear Stearns has folded into JP Morgan, Merrill Lynch into Bank of America, and part of Lehmans into Barclays, and as Morgan Stanley and Goldman Sachs have become bank holding companies with access to the Fed discount window and covered by the implicit assumption that the US government would consider them too important to fail.

But even while this is been the de facto trend, several commentators have argued that regulation should be designed to produce the precisely opposite result – a separation of “narrow banking” from risky investment bank trading activities, a re-imposition in the US of the Glass Steagall separation of commercial and investment banking, and the introduction of that separation for the first time into European bank regulation.

At times this vision is expressed in terms close to a nostalgic elegy for a past age of innocence and stability: with Captain Mainwaring back behind the desk in the branch at Walmington-on-Sea casting a censorious eye over any householder or small-business man silly enough to want to take on too much credit, while the wide boys of the City and Wall Street are free to speculate but well away from sober middle England. But there is a very important issue here; we need to think carefully about how to insulate the vital functions of deposit taking, maturity transformation, and credit extension, from adverse impacts arising from the potential irrationality of liquid traded markets; we need to control the extent to which large universal banks can take the benefits of stable retail funding, deposit insurance and too big to fail status, and use them to fund activities of little economic value and/or of high risk, such as unnecessarily large proprietary trading.

I am not convinced, however, that this can or should take the form of any absolute separation between institutions which do to narrow banking and those which perform investment banking activities. The desire for this distinction fails, I think, to reflect the fact that the originate and distribute model can have some advantages, and the fact that in between narrow banking as performed by say a building society, and pure proprietary trading performed by say a hedge fund, there are a wide range of functions essential to the provision of finance to major corporates, to the lubrication of global capital flows, and to the management of risk naturally arising in a world of fluctuating exchange rates, interest rates and commodity prices, which mean that global banks involved in deposit taking, and credit extension are also involved in complex treasury and market making activities.

A crucial issue for regulators is therefore going to continue to be how we regulate the very large and very complex systemically important banks which are too big to fail and which are involved both in narrow banking and in complex treasury and trading activities. The Group of 30 Report Financial Reform: A Framework for financial stability, published last Thursday, suggested that “large systemically important banking institutions should be restricted in undertaking proprietary activities that present particularly high risks”. The precise ways in which we achieve this end need to be carefully considered, and the crucial change may be simply the better treatment of trading book capital, which I will come to shortly: but the issue certainly needs to be addressed.

So let me turn finally to the key elements of the regulatory response required to reduce the probability and severity of future financial crises, whether arising from classic on balance sheet banking or from the securitized model of credit intermediation. The response needs to be multifaceted, and there are many facets which I am not going to discuss tonight, but which will be covered in the Review and Discussion Paper to be published in March. These include actions relating to remuneration and incentives, to rating agencies, to counterparty risk in derivatives. They also involve consideration of the appropriate balance between mark to market and accrual accounting principles in published accounts. And, very importantly, we need to address issues relating to the regulation of large cross-border financial institutions, the realistic scope for global supervisory coordination, appropriate structures for local operations (subsidiaries or branches) and the appropriate balance of responsibility between home and host supervisors; the events of the last year – the Icelandic bank and Lehmans failures in particular – have taught us that we live in a world of global finance and global banks, but where, when disaster strikes, bankruptcy procedures and fiscal support are national, and we need to be clear about the implications of such a world.

But while all these issues are all important, there are three regulatory responses which I would like to highlight as the most fundamental and where what we need to do is in principle now clear.

New approaches to capital adequacy

The first is new approaches to the regulation of the capital adequacy of banks. These have of course been extensively revised by the introduction of Basel 2, which has aimed to achieve greater sensitivity of capital levels to the different risks which banks are running, and there are certainly benefits to the Basel 2 approach on which the future system should build. It is important to realize that the crisis developed under the Basel 1 regime not Basel 2, and that Basel 2 would have addressed some of the problems which led to it – for instance the failure to distinguish between the capital required to support mortgages of different credit quality. But it is also clear that we will need to adjust Basel 2 in a number of ways. The general direction of travel will be towards higher levels of bank capital than have been required in the past, and in particular capital which moves more appropriately with the economic cycle and more capital required against trading books and the taking of market risk.

• On the economic cycle, there has been considerable commentary on the procyclical nature of Basel 2 risk-sensitive capital measures, and it is inevitable that any system which is risk sensitive, unless deliberate countervailing adjustments are made, will be to a degree procyclical i.e. capital requirements will rise as we head into a recession and credit quality declines. But it is important to note that the degree to which Basel 2 is procyclical in relation to the banking books – credit extension on balance sheet – depends crucially on how it is implemented and can be significantly reduced if banks use appropriate through-the-cycle approaches to estimation of probability of default and loss given default. It is therefore important to ensure that the detailed implementation of Basel 2 does not introduce unnecessary and unintended procyclicality, and the FSA on Monday issued a clarification of our approach designed to ensure this.

Looking forward, however, we need to go beyond the avoidance of unnecessary procyclicality and to create a system which introduces significant counter cyclicality, requiring banks to build up substantial capital buffers in good economic times – ratios well above absolute minimum levels – so that they can run them down in bad. Such an approach makes sense from a micro-prudential point of view, reducing the risk of bank failure. But it is also desirable from a macro-prudential and macro-economic perspective: it will tend to place at least some restraint on over-rapid expansion in the boom, and it will reduce the danger that impaired capital makes it more difficult for banks to lend in recessionary times, thus making the recession worse. Ideally, such a regime must be agreed at international level, and the FSA is working closely within the Basel Committee on Banking Supervision and the Financial Stability Forum to design the details. There are many of those details to be worked out; whether the buffer requirements will be defined in formulaic terms, as in the Spanish dynamic provisioning system, or by regulator discretion; and how to deal with the complexity of different economic cycles in the different countries in which a cross-border bank may operate. But the principle is clear.

• And equally it is clear that in respect to the trading books of banks, we need to remove procyclicality and to increase capital requirements not just marginally but by several times. The present system of capital regulation of trading books is from a prudential point of view seriously deficient. Its reliance on value at risk (VAR) measures derived (usually) from the observation of the last year’s movements in market prices is clearly procyclical: if volatility goes down in a year, the measure tells banks that risks looking forward have reduced, and thus fails to allow for the fact that historically low volatility may actually be an indication of irrationally low risk aversion and therefore increased systemic risk. It fails to allow effectively for the low probability tail events which are crucial to extreme idiosyncratic and even more so to overall systemic risk. And, overall, the level of capital required against trading books has been simply too low relative to the risks being taken, given what we now understand about the systemic dangers of relying on liquidity through marketability, and about the susceptibility of securitised credit markets to irrational exuberance, sudden liquidity disappearance and rapid price collapse. Major banks with a large percentage of their balance sheets devoted to trading activity, have been required to hold only very thin slices of capital against it [Exhibit 16]. That will change radically given the proposals already issued by the Basel Committee, and these changes in themselves are likely to result in a significant contraction in the scale of future trading books.

And, looking forward, the FSA believes that a fundamental review is required of how trading books are defined and how risks in trading books are estimated. VAR-based approaches were originally developed to model the risks in trading in markets likely to be continually and deeply liquid (e.g. government securities, major currency FX swaps and interest rates derivatives), and were adopted by regulators in the mid-1990s when a high proportion of bank trading was concentrated in such highly liquid instruments. Over the last decade and a half, however, trading books have expanded rapidly to encompass the holding of many debt securities whose markets are imperfectly liquid even in normal times and which became suddenly illiquid when the downturn occurred; this booking of potentially illiquid assets in trading books was indeed in part driven by the lower capital charge there incurred. The FSA will be proposing to the Basel Committee that a fundamental review of the division between the trading and banking books and of the appropriate use of VAR to measure risk is now required.

New approaches to liquidity

New approaches to the management and regulation of liquidity are equally important. Indeed, we need to ensure that the regulation of liquidity is recognised as being at least as important as capital adequacy, a sense which was to a degree lost over the last several decades, with intense regulatory focus and international debates on capital adequacy, but less focus on liquidity – no Basel 1 or Basel 2 for liquidity to match the equivalents for capital.

That lack of a defined international standard has reflected the extreme complexity of the liquidity risk, which makes it difficult to achieve effective regulation through generally applicable quantitative ratios equivalent to capital adequacy ratios. Many developed economies did in the past require limits to defined ratios, such as loans to deposits; but it is less clear today that deposits are inherently more sticky than other categories of funding in a world of internet retail deposits and wholesale depositors (corporates, local authorities, charities, etc) with multiple options to redeploy spare funds. [3] And the emergence of the originate and distribute securitised credit model has been accompanied by increasing options to manage liquidity through secured funding (e.g. repos), and an increased reliance on liquidity through marketability, making bank liquidity risk assessment crucially dependent on assumptions about the liquidity of specific asset and secured funding markets, which it is difficult to reduce to simple quantitative rules.

Measuring and limiting liquidity risk is, however, crucial and reforms to regulation need to include both far more effective ways for assessing and limiting the liquidity risks which individual institutions face and a better understanding of market-wide liquidity risks. The FSA’s Consultation Paper on Liquidity published in December, therefore proposes a major reform of our liquidity supervision. It will put in place:

• Significantly enhanced reporting requirements focused on a detailed mismatch ladder analysis and applicable to all banks and building societies.

• Regulations which will require all banks to focus on the combined liquidity effect of their holdings of liquid assets, the maturity (on both a contractual and behavioural basis) of their assets and liabilities, and the term, diversity and reliability of funding sources.

  • For simpler mortgage banks and building societies this will be underpinned by a quantitative “buffer ratio” rule, which will restrain reliance on wholesale funds.
  • And for larger banks it will be achieved by a regime which requires the development by each bank and review by supervisors of a detailed Individual Liquidity Assessment, on the basis of which we will give Individual Liquidity Guidance, including the required level of a liquid assets buffer, which will be defined on a standardised basis but whose required level will be tailored to individual circumstance. At the core of the assessment and guidance will be stress-test scenarios, rather than models which seek to infer the probability distribution of risks from the observation of past fluctuations. This reflects the fact that liquidity risk assessment is inherently concerned with low probability but extreme events. And crucially the stress tests will need to cover potential market-wide stresses as well as idiosyncratic stresses, reflecting the lesson of the financial crisis that market-wide collapses in the liquidity of specific asset or funding markets can have huge impacts which analysis of individual specific risks will not capture.

This new regime and the related reporting requirements will greatly enhance our ability to understand emerging liquidity risks in individual institutions and across the whole economy, and to conduct sectoral analysis which can identify outlier business models and management practices. On the basis of this increased understanding, we will keep under review the appropriate balance of quantitative rules, stress test based analyses, and discretionary guidance. We anticipate that the new regime will result in major changes in the extent and nature of maturity transformation in the overall banking system, with banks holding more liquid assets and a greater proportion of those assets held in government securities, an incentive for banks to encourage more retail time deposits and less instant access, less reliance on short-term wholesale funding, and, as a result, a check on rapid and unsustainable expansion of bank lending during favourable economic times.

Regulation by economics substance: shadow banks and near banks

The third key priority is to ensure that in future financial activities are always regulated according to their economic substance not their legal form. One of the striking features of the years running up to the crisis, as I stressed earlier, was that a core banking function – maturity transformation – was increasingly being performed by institutions which were not legally banks, but the off balance sheet vehicles of banks, (SIVs and conduits), investment banks and mutual funds. To different degrees in different countries these ‘near banks’ or shadow banks escaped the capital, leverage and liquidity regulation which would apply to banks. In the case of SIVs they also escaped the degree of disclosure and accounting treatment which would have applied if the economic activities were performed on balance sheet. In future it is essential that if an economic activity is bank-like and poses a significant risk to consumer or financial stability, regulators can extend banking-style regulation. And essential that accounting treatment reflects the economic reality of risks being taken.

Applying these principles will have more implications for legal powers and regulatory structures in some other countries – and in particular in the US – than in the UK. European approaches to the bank capital adequacy have always applied to investment banks: the requirements set down for trading book capital were in retrospect inadequate, but in Europe investment banks did not lie outside a regulatory boundary. In the US, with a fragmented regulatory structure, there is a greater need to look at structures and powers. But, across the world, regulators need to continually assess how evolving industry structures and institutional roles are changing the nature of risk, both for individual institutions and for the whole system, and if necessary to adapt the coverage of prudential regulation over time.

This will at times require judgements about the appropriate treatment of institutions which have some of the risk characteristics of banks but not others. Two examples:

• The first is mutual funds taking consumer investments which are liquid in nature (immediate or very short redemption) and investing in long-term securities. These are not banks, the crucial distinction being that the liquidity of the promise to savers is not matched by certainty of capital value at redemption. But if they have made assurances that they will maintain a stable net asset value, that they will not “break the buck”, they may in a liquidity crisis act in a fashion which exacerbates that crisis, selling rapidly to meet redemptions and fuelling the deflationary cycle. That is why the G30 report recommends that “money market funds which want to continue to offer bank-like services… and assurances” should be reorganised as special purpose banks and regulated as such. This is a pressing issue for the US but not the UK: for a variety of reasons mutual funds of this character have not developed here. But if they ever did develop in future, we would need to keep under review at what point bank-like characteristics justify bank like regulation.

• The second is hedge funds, whose asset managers are present in the UK, and who are regulated as asset managers by the FSA, though the actual legal fund is usually registered offshore and not subject to prudential regulation. Here the issue, now being considered by fora such as the Financial Stability Forum, is whether the funds themselves should be regulated and subject to prudential limits on leverage or liquidity. The argument against is that these institutions are not banks: they do not deal directly with the general population but with professional sophisticated investors; leverage levels are in general (though with some exceptions) far lower than those of banks (typically two or three not twenty); and they do not perform contractual maturity transformation, since they have the power to limit the speed of redemptions via redemption gates. In general (again with some possible exceptions) they neither have the scale nor the characteristics which would require that individual hedge funds be treated as systemically important and thus too big to fail. But in aggregate, they may nevertheless play a role with systemic effects which regulators and central banks need to understand, but which currently we lack the data to analyse effectively. Rapid deleveraging by hedge funds has probably over the last six months played a non-trivial role in exacerbating financial instability. It is for these reasons that the G30 report suggested that regulators should have the power to gather detailed information from hedge funds, so that we and central banks are better able to judge their evolving systemic importance: and recommended that “for funds above the size judged to be potentially systemically significant, the prudential regulator should have the authority to establish appropriate standards for capital, liquidity and risk management”. That halfway house on hedge funds – information and the power to respond to future developments in size, concentration, leverage levels, and practices – would be in line with the principle of focusing on economic substance not legal form.


To conclude, let me return to my opening thought. We are in the middle of an economic downturn which, to a far greater extent than any since the 1930s, is the result of developments which were to a degree internal to the global financial system. Developments in the banking and the near-bank system, which had been lauded as improving allocative efficiency and financial stability, have in fact caused serious harm to the real economy. The changes which we need to make to create a sounder system for the future will be profound. Their guiding principle should be that they should create a banking system focused on the delivery of the value-added functions of banking which are so essential to a market economy.

[1] Even the banks which were largely doing “originate and distribute” would often however have to warehouse significant quantities on balance sheet before packaging and distributing, and could be left with liquidity strains and future potential losses if liquidity suddenly dried up (e.g. Northern Rock)
(go back)

[2] This assumption seems also at time to have been based on a misunderstanding of the inference that could be drawn from a credit rating, with high ratings treated as carrying the inference of liquidity, rather than simply lower credit default.
(go back)

[3] Loan to deposit ratios limits continue however to be used in some emerging economies. A number of emerging economies e.g. India, also continue to use reserve asset ratios as monetary policy instruments, with significantly liquid balance sheets a resulting byproduct.
(go back)

Both comments and trackbacks are currently closed.

One Comment

  1. Peter Morgan
    Posted Friday, February 27, 2009 at 9:36 pm | Permalink

    This sums it up and is to be applauded for its comprehension and conciseness. Only thing I would query is the hedge fund leverage being “typically” lower than banks (2-3 versus 20, respectively). I know of many hedge funds using strategies requiring leverage of around 30 plus to produce worthwhile returns after transaction costs and fees are taken out.