Examining the causes of the financial crisis
Speech by Adair Turner, Chairman, FSA
The Economic Club of America and National Journal Group
29 October 2009
The financial crisis which began in Summer 2007 and intensified so dramatically in September last year, was the worst for at least 70 years; indeed, by some measures it was the worst financial crisis in the history of market capitalism. Economic catastrophe was only prevented by extreme government policy responses: even with these responses, the world has faced huge economic cost, in unemployment and in lost output and wealth. It is therefore vital that we identify the root causes of this crisis as well as the symptoms: and that our policy response is adequately radical.
To ensure that, we have to recognise that what occurred was not just a crisis of specific institutions and a failure of specific regulations, but a crisis for the entire intellectual theory of efficient, rational and self-equilibrating markets.
In the aftermath of the crisis we must therefore challenge previous assumptions – for instance that increased liquidity in all markets is always beneficial, or that financial innovation will always deliver beneficial results.
We need to ask questions about the role which the financial system plays within the economy, recognising the danger that in some circumstances and in some activities it can swell beyond an economically useful size.
And we need to consider a wide range of policy options to ensure greater alignment between private action and beneficial social effect.
Finally we must recognise that there are some key questions to which we do not yet have clear answers, but on which we need to focus our attention because they are so crucial.
This morning I will therefore do three things:
- First, tell a story of the causes of the crisis which is now fairly familiar but I believe also largely correct.
- Second, highlight specific features which illustrate that this was not just a crisis of specific institutions or regulations, but a crisis of economic theory.
- Third, explore two issues where we have more thinking to do and/or where we need to be open to a wide range of different policy options.
There is now considerable consensus – and I believe well-based consensus – on many of the features of what went wrong. I described those features in my Review of regulation published by the FSA in March1.: reports by Jacques de Larosière for the European Commission, by Paul Volcker for the G30, and several others, have reached similar conclusions. I will not go through the arguments of those reports in detail, but briefly remind us of five points.
- First, there was a macroeconomic context, very large current account imbalances (Exhibit 1 – see attached slides in ‘related information’ for all exhibits) which, combined with fixed exchange rate policies, drove huge accumulation of official holdings of low-risk government securities, (Exhibit 2) driving real risk-free interest rates down to historically low levels.
- Second, these low real and also nominal interest rates, in turn drove a frantic search for yield uplift among investors seeking low risk, or at least apparently low risk return. A demand which was met by financial innovation – the explosion of securitised credit, structured credits and credit derivatives, the alphabet soup of CDS and CDOs and CDO squareds – their apparent ‘value added’ predicated on the belief that by securitising and structuring and hedging, slicing and dicing, one could somehow deliver to investors combinations of liquidity, risk and return, more favorable than the underlying pool of credits.
- Third, financial innovation combined with low nominal and real interest rates helped drive rapid credit growth in several countries, with lowering credit standards in many markets, particularly in residential and commercial real estate.
- Fourth, there was a significant increase in leverage across the system, (Exhibit 3) both at the level of institutions – banks and investment banks – and embedded within products, that leverage apparently justified by confidence in the fact that the system had become less risky, and that leverage in turn further fueling credit growth.
- Fifth, and quite as important as the rise in leverage, there were profound changes in the scale and nature of maturity transformation in the system – with increasingly reliance on the idea that contractually long assets could be considered liquid because saleable in liquid markets; increasing reliance by many banks on potentially volatile wholesale funding: and the growth of maturity transformation arising not on the balance sheets of banks, but in off balance sheets (SIVs and conduits), on investment bank balance sheets, and in mutual funds.
These factors created a system of greatly increased systemic risk and fragility: a system highly susceptible to a surge of irrational exuberance and highly vulnerable when that exuberance turned to doubt and then despair.
That system collapsed last year and we need to build a more stable system for the future. But to do that, I think we have to recognise that what collapsed last year was not just specific institutions, what failed were not just specific regulations, but that last year saw the collapse of an intellectual theory, an ideology indeed, of automatically self-correcting and equilibrating markets, based on an efficient market hypothesis in which the rational behaviour of agents would lead necessarily to beneficial and adequately stable results.
Three points which help illustrate the scale of that failure:
First the path of CDS spreads and equity prices for major banks from 2002 to 2008 (Exhibit 4). In efficient markets, these market prices are meant to provide us with information and to ensure market discipline. They are supposed to capture thoughtful rational expectations of future events, and as a result to provide information which leads to better decision-making. And even today, you can find not only market commentators, but central banks and regulators quoting CDS spreads for banks as providing external and useful information about the degree of risk in the financial system. – ‘CDS spreads have reduced therefore we infer that risk has reduced’.
But as shown here, overall CDS spreads and equity prices for major banks provided us with no forewarning whatsoever of the crisis: indeed, those who used CDS spreads to infer from the wisdom of markets the level of risk and the appropriate price of risks, would have concluded from these figures that the financial system had reached a point of historically low risk in spring 2007, the point which we now recognise as that of maximum unrevealed fragility.
The idea that market prices are always in some efficient market sense ‘correct’ should have died and been buried in this crisis. And instead we need to recognise that market prices in liquid traded markets, while they may provide useful information as to the relative value of different assets within a specific asset class, at the level of the total asset class can be subject to herd and momentum affects, to self-reinforcing cycles of exuberance and then despair, well described by John Maynard Keynes, by Charles Kindlebeger, Hyman Minsky, and Robert Shiller, but largely assumed away in much recent economic theory, and even more ignored in the deliberations of ‘practical men’, market analysts, traders, and many policymakers, who were not so much, as per Keynes’s phrase, relying on the wisdom of some long-dead economist, but on the dominant conventional wisdom of economists today.
A second illustration is a quote from the IMF Global Financial Stability Report in April 2006 (Exhibit 5), which mirrored the confidence which those low CDS spreads both reflected and reinforced.
‘There is growing recognition that the dispersion of credit risk by banks to a broader and more diverse group of investors, rather than warehousing such risk on their balance sheets, has helped make the banking and overall financial system more resilient.
The improved resilience may be seen in fewer bank failures and more consistent credit provision. Consequently the commercial banks may be less vulnerable today to credit or economic shocks’.
And the IMF was certainly not alone at that time. The assumption that financial innovation had made the world safer was a dominant conventional wisdom, and one of the explanations advanced for what looked to many economists like the ‘Great Moderation’, ‘Great Stability’. It reflected a confidence that, even if few people understood the intricacies of the mathematics of structured credit and credit derivatives trading, it must in some way be adding value and dispersing risk, since any innovation which did not do so would not survive in a competitive marketplace. Alan Greenspan has subsequently stated that he assumed that we could rely on privately motivated decision-makers to make rational decisions in their own self interest, but that that assumption has turned out to be wrong.
And third (Exhibit 6) are figures which chart rapidly increasing income leverage within the US and UK economies, increasing debt to GDP. But an also intriguing feature of that increasingly leverage is that while there was some increase in the leverage of household and commercial sectors – in their debt as a % of GDP, which must have been matched by increasing assets and liabilities of the financial sector – the increase in the size of the financial systems balance sheets is far bigger than can be explained by this effect. Instead, the most striking feature of the figures is the dramatic increase in intra-financial system claims, an explosion which was matched by huge increases in trading volumes relative to underlying real activities, and by significant increases in the % of measured GDP accounted for by financial services.
If you accept the Greenspan doctrine, then by definition all of this increased activity was economically useful, truly valuable, a true contribution to useful welfare enhancing income. But if we do not accept that, by definition, it becomes possible that at least some of this increased financial system activity reflected economic rent extraction rather than value added.
In the aftermath of the crisis, we must therefore be more willing than before to challenge two assumptions:
- First that all markets are by definition self-correcting and in some sense rational.2.
- Second that financial innovation which results from market competition is by definition useful.
And we must be more willing to ask searching questions about whether the financial system is delivering its vital economic functions as efficiently as possible, or whether parts of it can and before the crisis did swell beyond their economically efficient size.
It is against that background that we must design our regulatory response. Many of the required features of that response are clear:
- Capital and liquidity regulation reforms to make the banking system a shock absorber rather than shock amplifier in the economy: more capital and more liquidity, and countercyclical capital built up in the good times and able to be drawn down in bad.
- Reforms to deal with large systemically important, potentially too-big-to-fail banks, with possible capital surcharges for the largest and most interconnected, and/or resolution procedures which would enable controlled wind-down.
- And action to reduce interconnectedness in OTC derivative markets, migrating as many contracts as possible to central counterparty clearing systems, and ensuring adequate capital and collateral is held against the remaining bilateral contracts.
Those elements of the agenda are already agreed and being pursued at both national and international levels. But I would like to concentrate today on two issues where either appropriate policy is not yet clear, or where regulation alone might be insufficient but where other instruments such as tax reform might conceivably have a role to play. These two key issues are:
- the optimal level of capital in the banking system and of leverage across the economy; and
- the optimal size of the wholesale financial services industry and in particular of trading activities.
In addressing each of these issues we need to go back to fundamental questions about the functions which the financial system performs, and the different ways in which it has performed those functions in different countries and periods.
Optimal capital requirements and leverage
There is very strong consensus that the global banking system should in future run with more capital, lower leverage than in the past. But how much more? Do we know what the optimal level of capital in the banking system should be? Is the optimal capital ratio roughly 5%, or roughly 10%, or roughly 20%?
Two things are striking when we pose that question.
The first is how little focus there has been on that question over the last several decades of regulatory debate:
- When the Basel I capital regime was introduced in the 1980s, there was a clear objective to level up capital ratios to a best international practice level, but the latter was simply interpreted as being the level which existed in the countries with somewhat higher capital ratios, and there was little debate about whether this level was in any sense optimal.
- And while the introduction of Basel II involved immensely complex consideration of what relative weight should be attached to different assets, when it came to the aggregate level of capital across the system, the overt approach followed was to leave the overall level broadly unchanged.
Second, comparisons across countries and periods tell us that banking systems can perform their economic functions with very different levels of capital. In both the UK and the US, there were past periods where banks had not just slightly, but much higher capital ratios than today, much lower leverage (Exhibit 7), but still managed to perform credit intermediation functions in growing market economies. And there are some developing countries with current capital ratios significantly higher than in the developed economies but still achieving rapid economic growth.
So it is certainly possible to run banking systems with significantly higher levels of capital than today. And it is clear that, with all other factors equal, higher capital ratios should mean a more stable financial system.
Against that benefit of increased stability, however, there must be a potential disadvantage that higher capital requirements could impose a cost on the economy as a result of the increased cost of credit intermediation. Bank capital is more expensive than senior debt or deposits: bank equity capital more expensive than debt capital. Therefore the more capital and the higher quality capital regulators demand, the higher the margin that banks need to charge between deposit rates and lending rates. And that would seem to impose a cost on the economy, restraining lending through increasing its price.
But while that must be true to a degree, two caveats must be made:
- The first reflects Modigliani and Miller’s famous insights on the impact of capital structure on a firm’s cost of capital. Their theory argues that while lower leverage might seem to increase capital/funding costs by raising the proportion of more expensive equity capital, this effect is offset because as leverage falls, debts becomes safer and therefore less expensive and equity returns become less volatile and the cost of equity falls. 3.The tax deductibility of interest payments, of course, means that these offsets are not complete, so that a private incentive to seek higher leverage still remains. But that raises the issue of whether our tax systems are sensibly designed.
- The second caveat is that an increased cost and decreased volume of lending, is only harmful if we are confident that the lending which would otherwise occur, in a more lightly capitalised banking system, would be beneficial. Constrained lending could be at the expense of useful corporate investment projects: or at the expense of households’ ability to shift consumption across the life cycle, with some households saving while others borrow. But decreased lending might also constrain wasteful products and unaffordable borrowing commitments, particularly in an irrationally exuberant upswing. It therefore turns out that we cannot answer the question – what is the optimal leverage of the financial sector? – without considering the optimal leverage of the household and corporate sectors, and whether and to what extent we can rely on those sectors to choose optimal debt levels. But it is not clear that either economic theory or empirical evidence provides clear answers to those questions.
The reason why the regulatory authorities of the world have in the past evaded the fundamental question of the optimal level of capital may therefore simply be that it is a very difficult question to answer. And our decisions on how much to increase capital may therefore be inevitably judgemental. But we will I think make better judgments, if we are aware of the complexity of the arguments and the tradeoffs that we are implicitly making. And the FSA believes that we can use modeling work to at least illustrate the tradeoffs potentially involved, and last Thursday published a Discussion Paper which describes potential modeling approaches.4.
It may be, however, that one of the key insights from this consideration of optimal capital structures, relates not to regulation but to tax. In Modiglaini and Miller’s analysis the one factor which clearly makes it rational for either banks or corporates to increase their leverage, is the fact that in almost all tax regimes across the world, returns to debt providers are tax deductible at corporate level, but returns to equity providers not. Even, therefore, if we had sufficient confidence in free market rationality to assume that both corporate and banks would naturally gravitate in an unbiased world to capital structures which optimally balance equity and debt, it is clear that tax policy has introduced a huge bias into the system, a bias towards sub-optimally high leverage. If we could start with a clean slate, it would almost certainly be better to have a less biased system; but we do not have a clean slate, and reform would create large windfall gains and losses. But if change to the tax bias is too difficult, we do at least need to understand that the bias exists and that our regulatory approach needs to lean against it.
Trading and market liquidity
Let me turn to my second open issue, the optimal size of the financial system and in particular the optimal scale of trading activities. Exhibit 6 highlighted the dramatic growth of the scale of the financial sector as measured by financial sector assets and liabilities as percent of GDP, with the most rapid growth being of claims between the financial institutions, rather than between the financial sector and the rest of the economy.
This growth of financial activity relative to real economic activity can also be observed in measures of trading volume:
- In the 1970s, world trade and long-term investment flows were supported by foreign exchange transactions which were roughly double the value of those real flows. Today, foreign exchange transactions are well over 20 times the value of underlying trade and long term investment flows, with short-term capital flows the dominant driver.
- The value of interest rate derivatives has increased from $18 trillion in 1995 to $400 trillion today – a 20 times increase against a doubling in nominal global GDP.
- In 1990 credit derivatives did not exist: in 2007, about $60 trillion of gross nominal value was outstanding, a huge multiple of the value of the underlying credit instruments hedged through the derivative markets.
- The value of daily oil trading on major exchanges in 1990 amounted to roughly the same as the value of underlying oil produced and consumed in the world now ten times.
A crucial question is whether this increased volume of trading has made the economy more efficient or less. The dominant conventional wisdom of the last 20 years has denied that this is even a legitimate question. If you trust the rationality of the efficient market, the beneficial effects of free competition between financial institutions, and the wisdom of corporate treasurers only to buy financial products that make sense, then this expansion in trading volume and the related increase in intra-financial system claims must be axiomatically beneficial. But after this crisis, and the clear evidence of markets which can sometimes be collectively irrational, that argument by axiom is no longer adequate. We have to ask fundamental questions about whether the financial system is performing its economic functions cost efficiently, and whether and to what extent and in what markets increased trading activity adds value.
Clearly, a significant level of trading activity is beneficial to the economy. There is an economic value in considerable market liquidity, in customers (corporate, household or institutional) being able to buy and sell contracts in liquid efficient markets at fine bid-offer spreads. Economically beneficial trade is lubricated by the existence of forward foreign exchange markets: and you cannot have a liquid forward market without some position takers, which means speculators. Speculating traders are not exactly the most favoured citizens today: but they play a useful role in a market economy.
But the fact that market liquidity has an economic value, does not mean that more market liquidity, supported by more speculation, is limitlessly beneficial in all markets. Liquidity provision, like most economic activities, is subject to diminishing marginal value. Professor Benjamin Friedman put it in an article in the Financial Times in August 5. it is difficult to discern that there is great economic value in devoting high intelligence and large computing power to the task of anticipating market movements a few seconds before the rest of the market anticipates them. And as Professor John Eatwell has pointed out global trade was actually growing faster in the 1970’s, when FX volumes were only twice the volume of underlying trade and investment than today, when they are 50 times.6.
It is possible indeed that there is a point beyond which increased trading activities delivers not just diminishing marginal returns, but negative returns for the non-financial sector, due to a combination of instability and rent extraction effects.
- All liquid financial markets, I argued earlier, can be subject to herd and momentum effects, to swings of irrational exuberance and then despair which can produce significant diversion from equilibrium values. And the more capital and human intelligence devoted to position taking activities, the greater volume of trading, the greater may be the potential for those herd and momentum effects. More trading can under some circumstances produce more volatility not less: volatility against which the non-financial sector then has to hedge, paying the financial sector for the service. The financial sector thus has an almost unique capability to create demand for its own services.
- The financial sector has a peculiar capability to charge high but hidden margins for some of its services. In its market-making activities, position-taking profits can be made by exploiting superior knowledge of underlying flows: a hidden form of margin.
- And as Paul Woolley of the London School of Economics has shown, complex principal/agent relationships, combined with opaque products the value of which is inherently difficult to understand, create further large opportunities for rent extraction.7.
Some categories of wholesale financial services, and in particular but not exclusively those linked to the trading of more complex instruments such as structured credit and credit derivates, if left entirely to the free market, may have an inherent tendency not only to create instability but to grow beyond their economically optimal size.
And if they do that, they will tend to produce remuneration levels in financial services which some will see as excessive, and which while acceptable as long as economic times are good for everybody, will become a focus of extreme resentment when times are bad, as they are for many people today.
The question then becomes what if any policy levers can be used to address the potential for trading activity to swell beyond its economically efficient size, for traded related profits to be supernormal.
Three are sometimes proposed:
- One is popular, indeed populist, but ultimately unlikely to be effective: direct regulation of bonuses.
- The second, is clearly appropriate and at the core of the required regulatory response: appropriate capital requirements.
- The third is an option which practical implementation difficulties may argue against, but which should not be excluded from discussion: financial transaction taxes.
1. Direct regulation of pay
The popular, and indeed populist policy, is to limit bankers’ bonuses. If some banking activity is unnecessary, some profits deriving from what even the Chairman of the British Bankers’ Association has described as ‘banks have chased short-term profits by introducing complex products of no real use to humanity’,8. and if traders get paid enormous sums for trading these products, why not just limit the bonuses – for instance by setting a maximum percentage pay out rate of investment bank profits, the idea put forward by the French and German governments before the G20 meeting in Pittsburgh?
One can understand why that policy is politically appealing – particularly at a time when investment banks and the trading arms of major commercial banks are making high profits precisely because of conditions which the crisis has created – higher market shares for the survivors, large government bond issues, near zero short-term interest rates.
But the trouble is that in the long run a sector specific incomes policy is totally unenforceable. It would take investment banks no time at all to work out the many ways round such rules, such as shifting people from employee to self-employed service provider contracts.
If super normal profits are being made, they will in some way or other flow into the pocket of employees as well as shareholders. If you do not think that there is a problem of super normal profits, then you should accept bankers’ bonuses as earned for useful activity: if you think there is a problem, you have to address it at the level of pre-remuneration profit, not by asking regulators to do the impossible job of regulating pay levels.
2. Appropriate capital standards
A second policy option – and the one which is clearly required and at the core of our policy response to the crisis – is higher capital requirements on trading activity. In the past the capital regime for many categories of trading activity was not just a bit wrong, but radically wrong, based on simplistic faith in value-at-risk models simplistically applied, and on the unjustified assumption that presently liquid markets would always remain so. That capital regime allowed major trading banks to trade complex and risky instruments with clearly inadequate capital support.
Changes to the trading book capital regime already agreed for implementation by January 2011, will increase requirements over three times on some categories of trading activity, and among the Basel Committee’s most important agenda items is a fundamental review of the whole trading book capital regime. Higher capital against trading activity will greatly reduce the risks of future financial instability: if it also reduces the volume of trading in some instruments and the aggregate remuneration of some traders that may be a perfectly acceptable side effect.
3. Financial transaction taxes
But it is possible that even with higher capital requirements, financial trading activity will continue to grow both within the banking system and outside it, and that even if regulation can guard against the consequences of this for the stability of particular banking institutions, those high levels of trading activity will continue to support unnecessary rent extraction by the financial sector, and may continue to generate economic instability. If that is the case, and if society is worried about the consequences, either for financial instability or for the size and remuneration of the financial sector, then it cannot exclude consideration of tax instruments – such as taxes on financial transactions, Tobin Taxes as they are often labelled after James Tobin’s proposal of such a tax specifically for a foreign exchange transaction. There are of course many complex issues to be considered in the design of such taxes, and anyone who thinks that such taxes would either prevent all or even most rent extraction in the financial sector, or that they could be designed to tune the liquidity of markets to precisely its optimal level – neither too liquid nor insufficiently liquid – is fooling themselves. But in the real world of imperfect instruments with which we try to achieve results at least a bit better than those we see today, they should not be excluded from consideration.
But my main point today has not been to propose any specific policy measures – other than to strongly endorse those related to capital and liquidity on which the global regulatory community is already actively engaged – but to stress the need, in the aftermath of this huge financial crisis, to go back to basics, and to ask searching questions about the functions the financial system performs in our economy, about optimal levels of leverage in the whole economy as well as in the financial system itself, about optimal levels of trading activity and the optimal size of the wholesale financial services activity. These are questions we used to avoid, because a dominant ideology said that they were invalid or irrelevant, with the level of leverage and the scale of trading activity axiomatically optimal because chosen by the free markets. In the face of this financial crisis, we can no longer avoid these issues.
1.The Turner Review: A regulatory response to the global banking crisis, FSA, March 2009
2. The use of the word ‘irrational’ to describe herd and momentum effects can be challenged on the grounds that such effects can derive from actions which are rational at the level of individual agents. George Soros, for instance, making this case, prefers to say that markets have no tendency to reach equilibrium, rather than that they are irrational.
3. Miller M and Modigliani F.: The cost of capital, corporation finance and the theory of investment, American Economics Review, 1958, 48:3, pp 261-297; Corporate income taxes and the cost of capital: a correction, American Economics Review, 53:3, 1963 pp 443-453
4. Turner Review Conference Discussion Paper: A regulatory response to the global banking crisis: systemically important banks and assessing the cumulative impact, FSA, 2009
5. Benjamin Friedman’s article Overmighty Finance Levies a tithe of Growth, Financial Times, 26 August 2009
6. John Eatwell, Why Turner is Right, Prospect Magazine, Issue 164, 21 October 2009
7. Bruro Biais, Jean-Chartes Rochet and Paul Wooley, Rents, learning and risk in the financial sector and other innovative industries, September 2009
7.Stephen Green’s speech: Good Value in Banking, Banking in the Process of Change Conference (Frankfurt, 8 September 2009)