Engineering the Financial Crisis: Systemic Risk and the Failure of Regulation
by Jeffrey Friedman & Wladimir Kraus
(University of Pennsylvania Press, 2011)
Within the mainstream media a standard narrative has developed about the causes of the financial crisis and the subsequent global economic slowdown. This narrative suggests that: bankers were encouraged by the structure of their bonus-schemes to take irresponsible risks; that banks held far too little capital to support the scale of the risks they were taking; and that the failure of policy makers to curtail these behaviours, contenting themselves instead with the new orthodoxy of "light-touch regulatory regimes", was due to the insidious political influence of bank-lobbyists.
As a result, so the story continues, the manufacture and sale by investment banks of vast quantities of sub-prime "toxic" assets, exploiting a bubble in the US property market, led to the collapse of the financial system and the socialisation of huge banking losses. While public sector workers and taxpayers now face a decade or more of austerity, bankers continue grossly to overpay themselves, unabated in their greed.
That this narrative has become standard does not mean that it is well supported by evidence. Quite the contrary. It is one of the strengths of Jeffrey Friedman and Wladimir Kraus's new book, Engineering the Financial Crisis, that it undertakes a careful examination of the evidence for several elements of the standard narrative. Their conclusion - that responsibility for the financial crisis lies more with the regulators than with the regulated - has important implications for current policy debates about the nature of regulatory reform.
The authors' principal focus is a specific set of financial regulations - the Basel-II capital adequacy rules and the US equivalent, the Recourse Rule - which, unintentionally, created an incentive for banks to manage their balance-sheet risks in a homogeneous way, using a combination of sovereign debt and mortgaged-backed debt to lower their regulatory-capital burden. Small falls in the value of these assets combined with credit-downgrades, led to systemic de-leveraging of bank balance sheets, asset "fire-sales" and a loss of confidence in the credit-worthiness of the financial system. This, in turn, precipitated a global recession.
Freidman and Kraus argue that the transmission of falls in US property valuations onto bank balance-sheets, where the financial impact was significantly amplified; and, subsequently, the transmission of the capital costs of impaired bank balance-sheets into the wider economy causing a sharp recession, were both the avoidable consequences of poorly thought out regulatory policy.
Thus, causal responsibility for the crisis should be attributed not so much to capitalism - which is to say, the errors made by capitalist investors, including bankers - as to capital adequacy regulations - which is to say, the errors made by regulators of capitalism. (p.78).
The book provides a cautional lesson for those policy makers who place their faith in more regulation and higher capital requirements, heedless of evidence suggesting that these behaviours created the current mess. Walter Bagehot observed that a well run bank does not need capital, whereas there is no amount of capital that will be sufficient to protect a badly run bank. The current policy response to the financial crisis - the introduction of the Basel III rules - requires banks to hold significantly more capital, which may simply mean that when the next financial crisis occurs more capital will be lost; and the consequent economic downturn will be wider, deeper and longer.1
Friedman and Krauss start their account by asking whether there is good evidence to support "the conventional wisdom about the causes of the financial crisis" (p.5). The standard narrative can be broken down into an argument containing a number of steps, each of which could be subjected to tests of its empirical validity. One example is the claim that "bankers deliberately took excessive risks because their compensation packages encouraged excessive risk taking" (p.36). Reviewing the (rather slight) academic literature on the relationship between compensation and risk-taking (pp. 36-42, and Appendix I) the authors find no serious evidence that corporate compensation structures played any role in the crisis.
Not only was it the case that many of the most highly paid bank executives, in the most senior roles, lost the largest amounts of money when their firms collapsed (James Cayne, CEO at Bear Stearns, and Richard Fuld, CEO at Lehman Brothers, are estimated to have lost around $1bn each; Sanford Weill, Chairman at Citigroup, is estimated to have lost around $500mn), but it was also the case that many of the most highly paid bank executives were unaware of the size and scale of the risks being taken by their staff. One could certainly argue that senior bank executives who did not fully understand the risks being taken by their employees were negligent and did not deserve to keep their jobs, let alone receive high levels of compensation. This, however, is very different from an argument which says that the structure and levels of executive bonus schemes encouraged irresponsible risk-taking, for which there appears to be little evidence.
Friedman and Krauss point out that there is an important distinction between the effect of our actions and the intention that lies behind our actions. Bankers might have acted recklessly without acting in an intentionally reckless manner: their recklessness might have been inadvertent. Further, while there were losses on so-called "risky assets", these losses were not as high as they would have been if banks had owned the maximum amount of these assets that their capital allowed. In early 2007, bank capital levels were well above their regulatory minimum. The argument that bankers' compensation schemes encouraged them to take excessive risk needs to show that the risk-taking was both intentional as well as excessive, and there is simply no evidence to support either claim.
The second major theme of the book is the role played by bank regulation, which created incentives for banks to manage their capital position by the use of specific asset classes, notably AAA-rated mortgage backed securities, which attracted regulatory relief at a discount of 80% by comparison with unsecured loans to individuals and businesses. It is understandable that regulators would judge a loan, in the form of ownership of a tradable security, secured against a diversified portfolio of mortgages, to be much less risky than an unsecured loan. However, by institutionalising this judgement in the form of regulatory relief for certain types of asset the regulators provided an significant incentive for all banks to follow a similar pattern of investment to manage their regulatory capital position. While the securitisation process created the possibility for US property sector risks to be disaggregated and widely dispersed, the Basel-II rules (and, in the US, the Recourse Rule) encouraged the aggregation and concentration of these risks onto bank balance sheets.
If bankers had believed that the US property bubble was about to burst then they would have divested themselves of some or all of their exposure to this sector by replacing the mortgage-backed securities with other assets. Likewise if the financial regulators had foreseen the falls in US property prices and their consequent impact on bank balance sheets then they would have changed the risk weightings attached to mortgage-backed securities, making them less attractive for banks to hold. In the absence of such foresight, the regulatory framework encouraged banks to increase their exposure to the property market via securitised debt. As Friedman and Krauss observe,
Regulatory arbitrage need not increase leverage levels and thus risk; it may instead increase usable capital levels at the expense of regulatory capital levels, reducing risk (other things equal). (p.84)
The problem, as we now know, is that the bankers and the regulators were both wrong about the safety of mortgage-backed assets, and consequently the regulatory rules aimed at lowering levels of risk ended up increasing them.
The problems that the banking sector faced in 2007-08 were compounded by the fact that when mortgage-backed assets were down-graded the cost of holding them, in regulatory capital terms, rose quickly. While AAA-rated mortgage-backed securities attracted a risk weighting of only 20%, A-rated mortgage-backed securities attracted a risk weighting of 50%. At the point of downgrade from AAA to A, the capital cost of owning the asset rose from 20% to 50% of the nominal value, a 150% increase in the real cost. Rather than raise significant extra capital - which would be difficult to achieve in the midst of a financial crisis - banks decided to sell assets instead. And because nearly all banks had adopted the same strategy for managing their regulatory capital, nearly all banks wanted to sell the same assets at the same time; for which reason the prices of mortgage-backed securities fell far more sharply than was warranted by the more modest falls in value in the US housing market.
Despite the frequently-repeated description of mortgage-backed securities as "toxic assets", there was nothing, and still is nothing "poisonous" about the process of bundling loans secured against property into tradable securities, which are structured into tranches offering investors a range of risk and return profiles. It is now evident that some of these securities were badly designed and some others were collateralised by mortgages granted through poor underwriting processes. Not all mortgage-backed securities suffered these defects: default rates from US issued structured finance during the period H2 2007 to end 2010 were 7.7%, only marginally higher than the 6.3% figure for US corporate bonds (i.e. non-structured corporate debt). In Europe the default rates for structured finance issues over the same period was lower than 1%. (Source: Standard & Poor's, quoted in OECD Journal, 2011).
Yet, ill-considered regulation, based on the credit ratings assigned to mortgage-backed securities, created herd behaviour among banks with respect to owning these assets; and created cliff-risk with respect to the trigger for banks wanting to be rid of them. When a market is full of sellers prices fall precipitously. The fault lies not with the product being sold, but with the rules that create industry-wide incentives for banks to invest heavily in a particular asset class and then, suddenly and collectively, to divest from it.
The third important theme presented in the book follows from the second, and concerns the ignorance of the different parties - bank executives and policy makers alike - with regard to the likely consequences of their actions. Friedman and Kraus distinguish between acting irrationally and acting in ignorance of the relevant information. It was not irrational for regulators to encourage banks to own lower-risk assets, by granting regulatory capital relief to mortgage-backed securities; nor was it irrational for bank executives to buy such assets once regulatory capital relief was offered. Thus, without anyone so intending, the conditions germane to triggering a major financial crisis were created, owing to the abundance of such assets on nearly all bank balance sheets at the moment when prices in the US property market began to fall.
There was an additional regulatory requirement to mark assets to market prices, even at times when market prices did not reflect the real long-term value of assets, but instead reflected the preponderance of sellers compared with buyers. All over the developed world prices of mortgage-backed securities fell, bank balance sheets weakened, the amount of capital available to banks to lend reduced, and economies moved into recession. This was, the authors note, an "ignorance cascade" with devastating effects:
The financial regulators and the accounting regulators had accidentally engineered not only a financial crisis but a calamitous recession. (p.103).
Nobody - neither bankers, nor academics, nor journalists nor policy makers - was able to foresee the potential consequences of the accumulation of regulatory and policy initiatives which, notwithstanding the original intentions of their authors, combined to create the worst global recession in living memory.
One problem is that there is too much information for any individual to come to know, to understand fully in its multiple ramifications and then to give appropriate weight to its importance. Another problem is the complexity of the interactions between different parts of the financial and economic systems, whereby a policy which has been adopted because it appears to reduce risk in one part of the system might turn out to increase risk in another part. Combined, these two problems make it difficult to predict the effects of emerging trends in the financial markets, such as the growth in issuance of mortgaged-backed securities and their retention on bank balance-sheets.
The authors argue that in a capitalist economic system none of the participants in the market place need be "well informed, prescient, or even particularly intelligent. They need only to be heterogeneous" (p.137). The diversity of business strategies employed by the participants will, over time, lead to a weeding-out of poor practice. Good practice will survive and will be replicated by other participants, in turn, until a new and superior form of business practice comes along, which is itself supersessive. This model, parallel to biological models of evolution through successful adaptation, explains why the capitalist system has been so successful.
"regulators are required to be synoptic perceivers, codifying in law their predictions about which practices will lead to market failure and should therefore be curbed" (p.138).
The regulatory system therefore imposes homogeneity, and does so through a process of "intelligent design". Policy makers insist that certain practices be followed and that others be prohibited, and all market participants are forced to comply. Fallible designers, however, make for imperfect systems. The financial regulators aimed to make the banking system safe, but ended up making it uniformly vulnerable.
Friedman and Kraus cover a great deal of ground in a modest number of pages. There are two ways, I think, in which their argument could be developed and strengthened. First, by including some testimony of participants in banking system; second by examining the logic of the causal relationships between the various elements of the financial crisis.
In the standard narrative much is made of the apparent willingness of bankers to take excessive risk. The claim is that bankers knew that they would benefit if their risk-taking activities were successful: higher profits would mean higher pay. If these risk-taking activities were unsuccessful, the bankers also knew they would be able to pass their losses onto society, as governments would be forced to bailout the banks to prevent them from collapsing. (Let us leave aside whether there is any evidence that bankers did think this way; or evidence that all governments lost money by rescuing their banking systems). The standard narrative assumes that the purchase of mortgage-backed securities by banks reflected a desire to "take risk"; yet there seems good reason to believe the contrary.
One of the characteristics of the international financial system in recent years has been the growing demand for safe, high-quality assets. Across the world investors had holdings of excess cash, well above their immediate investment and consumption needs, and wanted to invest this cash in safe assets in order to preserve its value. Banks and investment management firms, which act as agents on behalf of these principal investors, were therefore looking for asset classes that provided, first and foremost, security rather than high returns.2 Since government-insurance of bank deposits is capped and, given an insufficiently large supply of government debt, there was a strong market demand for AAA-rated paper for investors to hold.
At the same time, banks and investment management firms increasingly wanted access to AAA-rated securities for their own use, to act as collateral for a variety of transactions including margin for swaps, futures, options and other synthetic instruments. The growing use of derivatives in the financial markets had created the need for high quality assets that could be pledged by counterparties to the trade, either as the initial margin or, subsequently, as variation margin. In addition, the "repo" market, which allows banks to borrow cash by pledging securities to the lender, had become a huge collateralised inter-bank deposit market, that allowed participants to lend or borrow without the need for time-consuming and expensive credit checks.3 The quality of the collateral allowed the market to function efficiently, in the same way that deposit insurance allows retail investors to place money in a bank, assured that their cash is guaranteed by government.
There was then, both among end-investors and among banking and investment intermediaries, a huge and growing appetite for tradable, high-quality assets that would provide a liquid and secure haven for surplus cash holdings. This demand was met by an increase in the supply of AAA-assets, notably by the growth in the securitisation of assets from the balance sheets of banks, including mortgages. While these assets themselves - considered individually - were neither highly rated nor highly liquid, once they were bundled together into an appropriate legal structure, they could be used to create AAA-rated assets that were both liquid and secure. The "drive to securitize" was not, therefore, a sign of excessive or irresponsible risk-taking on the part of banks. It was quite the opposite: a demand driven response to the undersupply of secure assets.
That the standard narrative interprets the growth of securitisation wrongly is not surprising since this narrative is mostly presented by journalists and commentators who have little direct experience of working in structured finance. Whether the regulators fully understood the drivers of the securitisation process is a moot question. While Friedman and Kraus are right to argue that the volume and credit quality of bank holdings of mortgage-backed securities, in the years leading up to 2007-08, suggests that banks were not primarily seeking to take investment risk but to reduce regulatory capital costs, their case could be strengthened by the addition of testimony from industry participants.
The bank treasury teams who were responsible for managing their firm's regulatory risk capital position, the trading teams who were responsible for posting collateral to custodians, brokers and exchanges to cover the margin on their derivative positions, and the investment teams who were responsible for finding investment opportunities for cash-rich clients, could all provide insight into the supply and demand dynamics of the AAA-rated mortgage-backed securities markets. It would also be interesting to know why, if it is true that their main aim was to find liquid and secure assets to hold, they did not undertake better due diligence of the origination processes for mortgage-backed securities. If their need for safe assets had been better understood by the mortgage underwriters and security manufacturers, then the quality of the mortgage-backed securities issued might have been higher, which would have significantly lessened the impact of the bursting of the US property bubble on the banking system.
A second problem with the standard narrative is the routine confusion of cause with correlation, that is, the assumption that when event B follows event A, then A must have caused B. Logicians describe this fallacy by the Latin phrase post hoc ergo propter hoc, which translates literally as "after which therefore because of which" or, more technically, as "the subsequent is the consequent". The error is to confuse the temporal sequence of events with the causal sequence.
Friedman and Kraus draw attention to the fact that key elements of the standard narrative, such as the relationship between compensation and risk-taking, are not supported by empirical evidence. However they could have made more of the fact that the precise causal mechanisms that connect the fall in value of the US property market, the seizing up of wholesale liquidity markets, the fall in value of mortgage-backed securities, the failure of Lehman Brothers and other financial institutions, and finally the sharp drop in economic activity in the US and Western Europe, remain obscure and under-researched. The standard narrative presumes that each of these events was caused by the immediately preceding temporal event, but it is hard to see why that should be the case.
It would have been possible for the fall in value of the US property market to cause an economic recession without there being a financial crisis: falling house prices might have caused consumer confidence to fall, leading to reduced demand and lower investment. The banking sector might have seen its profitability fall, without the widespread collapse of balance sheet valuations that did in fact occur. Likewise, the credit ratings of mortgage-backed securities might have been lowered in response to a deteriorating US housing market, without causing sharp falls in the prices for these securities if, for example, capital charges for mortgage-backed loans were not so closely linked to changes in credit ratings.
There are no logical ground for assuming the inevitability of the sequence of events that did in fact occur; nor are there logical grounds for assuming that the temporal sequence of events provides a accurate map of the causal sequence. To understand what happened and why requires careful empirical research combined with careful conceptual analysis. When, eventually, this research and analysis is properly done I am sure that it will show the standard narrative to be almost wholly wrong. Friedman and Kraus's book does not provide all the answers, but it makes an invaluable contribution to the revisionary task.
William McChesney Martin, Chairman of the Federal Reserve Bank from April 1951 to January 1970, famously joked that his job was "to take away the punch bowl just as the party gets going." The current regulators, by contrast, appear to think that their job is to insist upon a bigger punch bowl.
Zoltan Pozsar, 'Institutional Cash Pools and the Triffin Dilemma of the US Banking System', IMF Working Paper, WP/11/190, August 2011.
Gary Gorton, Slapped by the Invisible Hand, Oxford University Press, 2010.