Latest Issue

Explaining the financial crisis

The credit crunch was an accident waiting to happen, thanks to a long period of benign market conditions which encouraged riskier behaviour by financiers. But how did problems in the US mortgage market spread to become a crisis of bank capital?

By Charles Goodhart   February 2008

One of the striking things about the current financial crisis is the extent to which it was foreseen. Almost every major central bank, as well as the international financial institutions like the IMF, had been warning for some time about a serious underpricing of risk in the system. They pointed to low differentials between the financial returns paid on risky assets and safe assets and to increased levels of debt relative to underlying capital. In the face of very low interest rates, financial institutions were buying riskier assets to improve returns, often “leveraging” themselves several times over (by buying assets with borrowed money). Just before the crisis broke, in August 2007, the authorities were actually cautiously welcoming evidence that the underpricing of risk was being reversed.

How had this underpricing come about? In part it had resulted from the long period of low interest rates that had continued from the ending of the high-tech bubble in 2001, until central banks began to raise rates again in 2005. After the bubble burst, there was a fear of deflation in the US. Moreover, there appeared to be a world glut of savings. These two factors prompted expansionary monetary policies, with nominal interest rates at low levels and accelerating monetary growth in several countries.

This period of low interest rates did not lead to higher inflation. Indeed, these years saw a continuation of what is known as the “great moderation.” Ever since the early 1990s, the developed countries, with the partial exception of Japan, have enjoyed a golden age. Inflation has been low and growth steady with few, if any, cycles. This benign macro-economic picture led many to believe that financial conditions were less risk-prone than in the past.

Moreover, whenever financial markets in the US had weakened sharply over the previous 20 years—black Monday in October 1987; the housing crisis in 1992; the Asian crisis in 1997-98; or the bursting of the high-tech bubble at the end of 2001—the Federal Reserve always moved in swiftly to prevent the downturn spreading more widely into the economy. A view had developed that the Fed would support financial markets from any serious collapse. Whatever the truth of this assumption, there was a general belief among bankers that if a particular bank got into trouble, its difficulties would be seen as part of a systemic problem and hence the central bank would intervene, protecting them from losses.

The coming together of all these factors helps to explain the widespread underpricing of risk.

Originate and distribute

Financial markets are highly innovative and the last decade has seen many new forms of “securitisation,” whereby loans to individuals and companies which used to be held on banks’ own books are sold off, in bundles, on to capital markets. (For a bank or financial institution, a loan is an asset comparable to an interest-paying tradeable bond.)

The rise of securitisation has been combined with a new banking strategy that began in the US, known as “originate and distribute.” Under this strategy banks originate loans, in the form of residential mortgages, say, and then put baskets of them together in various ways to create securities which are sold to non-bank financial institutions—pension funds, insurance companies and so on—so that the loans leave the originating banks’ balance sheets.

In many cases, this transfer of assets off balance sheets is more apparent than real. Banks have, for example, established various special purpose investment companies which they control but which, under current accounting rules, do not show up on the banks’ balance sheets. These vehicles have been used to hold the securities created by the banks. This was done in part to avoid the banking regulatory regime (“Basel 1”), which governs the amount of capital that banks have to put behind the assets (loans, for example) that they have on their balance sheets. By originating loans and then placing them, in the form of securities, into these vehicles, banks have been able to lend more without increasing the amount of capital required. Indeed, it has been very profitable for banks to play “pass the parcel” in this way, earning fees from loan origination, but then securitising and selling on those loans into the capital markets. In countries where capital markets are most developed, such as the US, the relative size of banks has fallen sharply relative to other financial institutions.

One particularly fashionable type of investment vehicle is the “structured investment vehicle” (SIV), often referred to as a type of “conduit.” Such vehicles fund their long-term loans to customers by issuing “asset-backed commercial paper” (ABCP)—very short-term bills (generally repaid in less than three months) sold to all kinds of investing institutions. As the assets (loans) held by the SIVs, such as securitised mortgages, typically don’t get repaid for several years, these vehicles ran a clear funding risk: if ever the money markets became illiquid they would be unable to repay their ABCP as it fell due. To protect themselves, the SIVs typically had arrangements with their sponsoring bank whereby it would lend them money if that ever happened.

Indeed, in general, commercial and investment banks have remained lenders of last resort to capital markets and recently have had to bail out the conduits they sponsored, or their own hedge funds and private equity funds. Examples include Banque Nationale de Paris in France, Barclays Capital in Britain and IKB and Sachsen LB in Germany. In most of these cases, the parent bank perceived that it had moral obligations (or at least reputational risk) that went beyond its legal obligations. In other cases, however, such as two hedge funds run by Bear Stearns, investors have been left with almost nothing.

One continuing accusation is that the process of “originate and distribute” made originating banks less concerned about the quality of credit assessment, since the loans were not on their books. But the credit quality of loans bundled into conduits was not only assessed by the originating bank; it was also supposed to be assessed by credit rating agencies.

As their name suggests, these agencies usually only assess the credit default risk of the assets—whether the borrower is likely to repay the loan when it falls due. But many investors who bought ABCP issued by conduits misinterpreted the ratings as covering market and liquidity risk as well—that is the risk that, should the holder need to sell the asset to raise funds, the price might fall sharply. So government debt with a rating of AAA has a superior overall quality than the AAA of securitised mortgages, because even in adverse conditions, the government bond market remains liquid.

It is also possible that the agencies made a faulty assessment of the credit default risk in the US mortgage sector. Since the second world war there had been pockets of weakness but no example of house prices falling across the whole of the US. So long as the price of a house is greater than the value of the mortgage, mortgage borrowers will attempt, as far as they can, not to default, because they would then lose the valuable equity in the house. Under those circumstances, defaults are usually restricted to cases of personal misfortune, such as chronic illness or unemployment, which leave borrowers unable to meet the payment. But this condition changes sharply whenever house prices fall below the value of the mortgage. Under those circumstances, there is a clear benefit to the mortgager in handing the keys to the house back to the lender. From late 2006, house prices began to fall in most areas of the US. This was against a background of expanding mortgage loans, particularly in the sub-prime area (loans to the most risky customers). The extension of such loans, together with the increase in interest rates, meant that many borrowers were starting to default. Since this had not happened before, the credit ratings agencies may have failed to assess the risk properly.

The liquidity problem

Until the end of the 1960s, commercial banks in Britain, and across the developed world, held 25 per cent or more of their assets in real liquid assets, assets which could be sold in open markets easily and with little price impact. These were largely very short-term government bonds. Since then there has been a continuous decline in the holding of liquid assets, and a shift from government bonds to private sector-assets. While these assets, notably in the form of residential mortgages, have had relatively high credit ratings, they are not as liquid in the sense that there is a broad, resilient secondary market on which they can be sold without moving the piece much. So the scale and quality of bank liquid assets has been declining.

Furthermore, in the 1960s and 1970s, commercial banks funded their assets largely with deposits by retail customers. Although many of these deposits were nominally on demand at very short notice, in practice the assets remained extremely stable. In recent decades there has been an increasing reliance on wholesale funding and on the short-term credit market (usually called the money market, where lending and borrowing can be anything from overnight to six months). The practice of using short-term borrowing to fund longer term loans has been increasing rapidly; Northern Rock is an extreme example of this. Northern Rock obtained some 80 per cent of its funding from wholesale markets. When these markets dried up in August, the bank was left exposed. In view of the worldwide constipation in money markets, it is difficult to believe that the monetary authorities could by then have taken any preventative action which would have avoided the very public exposure of its funding difficulties.

Western banking systems have become much less liquid. In effect, the banks are dependent on the support of central banks to help them through any liquidity crisis. This, of course, raises the question of how far it is appropriate for central banks to absorb all the downside risk on liquidity, and for commercial banks to take the potential upside in the form of the normal interest rate differential between borrowing short at lower rates of interest and lending long at higher rates (which is one of the key ways that banks make money). Should the central bank always bail out commercial banks from liquidity difficulties, which were in part brought upon themselves?

So the 2007 financial crisis was an accident waiting to happen. And, as noted earlier, central banks could, and did, see it coming, warning about the mis-pricing of risk and dangers in the over-expansion of the credit markets. But while the central banks foresaw the difficulty, they did nothing about it. This may have been because they did not have sufficient instruments to tackle the worsening risk profile in financial markets; or it may be that they did not have the will to do anything about it.

Why did the crisis spread?

The trigger for the crisis was, as everyone knows, the rising defaults in the US sub-prime mortgage market, but the trigger could have been almost anywhere else. Many commentators expected a failure in a hedge fund to be the trigger. So far, hedge funds have been notable for their comparative stability.

One of the most interesting features of this crisis is that the US sub-prime mortgage market is a small part of the overall US mortgage market, and, as its name indicates, was confined to the US. How then did this financial crisis spread so widely across many countries? Here we need to take a slight digression to explain how the mortgage pools worked, and how the original mortgages from the sub-prime market in the US ended up in bank conduits, and SIVs, in the rest of the developed world, notably in Europe.

Under normal circumstances, the probability of default of a prime mortgage borrower is extraordinarily low; frequently less than .003 per cent per year. But in the sub-prime market, even under normal circumstances, the probability of default is substantially higher, from 5 to 7 per cent, say over the course of the first five years in which the mortgage is held. But because they are caused by individual accidents of ill luck, the probability of a lot of people defaulting at the same time is very much less. Mortgages are pooled together to be securitised and sold on the capital market or placed in conduits, and the pool is divided into several tranches—from the riskiest (with the highest returns) to least risky (with the lowest). Each of these tranches can then be sold to different investors. The most risky “equity” tranche, for example, may be held by hedge funds, who can afford to take such risks, particularly since the risk can be partially hedged by investing in assets whose value goes up when house prices fall. The middle tranches may be held by longer term holders, such as pension funds; and the (safest) senior tranches were very largely held by these bank conduits, or SIVs. However, when house prices actually fall, particularly at a time when effective interest rates are rising, the probability of default will rise in a non-linear, indeed possibly an exponential, way. This means that there is a real credit risk, even on the senior tranches.

Throughout 2005 and 2006, US interest rates began to rise back to more normal levels. As a result, in early 2007 housing prices began to falter. This led to increasing losses in the lower tranches of the mortgage pools. Since these were often held by hedge funds, a number of hedge funds were hit, notably Bear Stearns in late June. There were other cases, some rumoured, some probably real, of big losses to hedge funds. Nothing much happened to the financial system more widely, but there was growing uncertainty about how further defaults might eat into the higher tranches. As already noted, the vehicles which hold these higher tranches are largely financed by very short-term bills known as asset-backed commercial paper (ABCP). And in turn these ABCP were largely held by money market fund managers. These money managers in general have a “convertibility commitment” whereby they guarantee to their investors that they will always be able to transform the funds held with them back into cash without any loss. In addition they generally have very little capital, and there is no “lender of last resort” for them. Accordingly they are very risk-averse. So when there began to be a suspicion of doubt about the credit risk of the higher tranches of the mortgage pools, they fled en masse. They refused to roll over their holdings of ABCP, and ABCP holdings began to decline dramatically. But the securitised mortgages, which were held by the conduits and SIVs, were impossible to sell in the open market without driving the price down sharply.

The conduits and SIVs therefore needed alternative funding to replace the ABCP. Remember that the commercial banks had links with these conduits and SIVs, and such banks were supposed to provide the alternative funding in just such a contingency. This initially was the case with the German banks, IKB and Sachsen. But as the value of the mortgage pool assets in their conduits declined, these banks had to absorb the losses, reducing their own capital severely, and had to be bailed out by their respective regional governments. Many other banks had conduits and SIVs to which they were connected, with some sponsoring relationship. People did not know what the liability of many banks to such institutions might be. There came to be a growing concern as to whether banks wanting to borrow in the money markets were always fully solvent, irrespective of what might be shown in their recent balance sheets.

So on 9th August, the inter-bank market—the money market in which banks lend to and borrow from each other—and in particular the three-month market, dried up. It became almost impossible for those who were regular borrowers in the wholesale markets to fund themselves, except at high, or indeed any rates at all. This of course, applied particularly to Northern Rock, around 80 per cent of whose total funding was from the wholesale markets.

What should the Central Banks do?

Interest rates usually rise when inflation is high and the real economy booming. Long-term rates are usually slightly higher than short rates, and there are also numerous differences between interest rates on differing instruments depending on the default risk of the underlying asset, the liquidity of its market and so on. But under normal circumstances these risk premia remain relatively stable. So normally, interest rates for London inter-bank lending of different maturities, known as Libor, are just a fraction higher than those for riskless government debt.

But under the influence of the turmoil of late 2007, risk premia rose and widened sharply. So for a given level of official interest rates the effective interest rates that were being faced by private-sector borrowers were rising; in addition, credit quality requirements were tightened, so the availability of credit to the private sector was cut back quite sharply. So even if the authorities did not worry about the financial markets for their own sake, they would be disturbed that the crisis was starting to have a deflationary effect in reducing demand.

This highlights a dilemma for central banks. They have two main objectives, maintaining the stability of the financial system and conducting monetary policy so as to maintain stable prices, in many cases to achieve an inflation target. But whereas the first objective might require lower interest rates, the second can pull in the other direction. Indeed, the crisis occurred at a time when the world economy was growing strongly, and when there was upward pressure on inflation (and so interest rates) from a rise in food and energy prices and the declining ability of China to bring ever more low-paid workers into manufacturing production. So it was dubious how far official interest rates could be lowered, were central banks to hold to their mandates to maintain price stability as a primary objective.

If central banks were constrained in their ability to lower interest rates, could they intervene in other ways? They could obviously inject additional liquidity by lending to banks against the value of assets on their balance sheets. However, as noted earlier, the commercial banks were no longer holding large volumes of clearly liquid and high-quality assets. This meant that the central banks would have to widen the range of collateral that they would accept. Given that commercial banks had to some extent brought this problem on themselves by failing to maintain sufficient liquidity, was it appropriate to accept as collateral anything that the commercial banks wanted to offer? Would this introduce moral hazard, allowing reckless behaviour to go unpunished?

Moreover, the problem was not about the availability of cash as such. Commercial banks balance their funding through the inter-bank market, borrowing and lending among themselves, either overnight or for fixed periods such as three months, at interest rates determined by supply and demand: the central banks act as a backstop lender of liquidity if this system falters, lending at an officially declared rate. Almost all the time in recent months the central banks have provided as much, or more, immediate cash than the commercial banks wanted, it was longer term money that they had stopped lending to each other. Could the central bank undertake a kind of “operation twist,” whereby they could make available to the commercial banks three-month funds at the official rate, and mop up funds from the overnight market at the lower market rate?

In the particular case of Northern Rock, the Bank of England was forced to undertake massive lender of last resort (LOLR) actions for several tens of billions of pounds after the inter-bank market dried up and Northern Rock could no longer fund its rapidly growing mortgage business. The governor of the Bank of England, Mervyn King, had hoped to keep this LOLR action secret, in order to avoid the potential bad effect on retail depositor perceptions of the safety of their money. As we now know, the announcement of this support led the depositors to believe that their money might not be safe and to run to withdraw it. There is, however, a question as to whether an LOLR action on this scale could ever be kept secret even if its disclosure had not been forced by Britain’s own laws (the Market Abuses directive, whose purpose is to ensure that all commercially sensitive information is made available in a timely fashion to all investors to prevent insider trading).

Where do we go from here? At the time of writing (January 2008), the answer is not clear. The ABCP market seems to have stopped shrinking. But sales of securitised mortgages on the open market remain difficult at other than distress prices, which would cause further difficulties to banks’ balance sheets. In the light of this, three or four of the largest US banks got together in October to try to establish a “super-conduit” (known as MLEC), but this was challenged on the grounds that it would lead to an artificial price for such assets. In the face of opposition, and lack of enthusiasm from other banks, this plan was quietly dropped in December.

Following the failure of the MLEC initiative, financial markets became even more fragile in November and early December, with almost all banks wanting to borrow and few to lend. In response to this, central banks took an unprecedented joint initiative, in the late autumn, to try to meet commercial banks’ demand for funding. Each central bank took some new step that it had not taken before: the European Central Bank (ECB) to supply dollar funding, the Fed to auction longer maturity funds and the Bank of England to accept a wider range of collateral. The joint nature of this initiative lessened public claims that each individual central bank was, in effect, confessing failure. However, this initiative did not appear at the time to lower the premia on one and three-month Libor markets significantly, and was dismissed by some as ineffective. Nevertheless, the holiday period passed quietly enough, and when credit markets reopened in the new year these Libor rates fell back sharply towards the official rates. It is now quite likely that the worst of the liquidity crisis is behind us.

But another cause of the late autumn relapse was that the losses arising from defaults and write-downs were being taken on board, and were proving far worse than had initially been estimated. Princely heads rolled. Chuck Prince, CEO of Citigroup, Stan O’Neal, CEO of Merrill Lynch, and other senior figures lost their jobs as the impact of the crisis on large banks became clear. In mid-January, the Fed estimated that writedowns were already above $100bn and would rise to nearer $250bn. What had once been seen as an almost pure liquidity crisis was now perceived as a crisis of bank capital—which in truth it always had been. There was a sharp switch from the belief, widespread in early 2007, that bank capital was abundant after a run of years of strong bank profits, to an appreciation that it was eroding fast.

It was mitigated in a few cases by the injection of outside capital into big financial companies, such as Citigroup, by “sovereign wealth funds”—owned by foreign governments mainly in the developing world. Each announcement led to a relief recovery on stock markets (though also to political concerns). But it is dubious whether these funds can meet the capital needs of the whole of the west’s banking systems. And without extra capital, western banks will be more reluctant to extend new, or increased, loans. There will be tougher credit conditions and higher interest rate spreads above the official rate. Mortgages will become harder to get, and will cost more. All this has only just started to happen. Whereas the financial crisis started in August, effects on the wider economy in Britain were minimal before December. This will now change. The worst of the liquidity crisis may now be over; we cannot yet tell how serious a shortfall of bank capital may develop, the effect on the real economy could be just starting.

Now that the financial horse has bolted, there will be a rush by regulatory agencies in all countries to slam shut the stable door, especially in Britain, where the Northern Rock saga revealed the weaknesses of both our deposit insurance scheme and the lack of a proper special insolvency regime for banks. The FSA, Britain’s bank regulator, faces awkward questions about why it did not take action sooner, since Northern Rock’s business plan clearly left it vulnerable once the inter-bank market became illiquid. Likewise, the Bank of England’s response to the liquidity problems in the money market looks to have been rather slow—and less effective than that of the ECB. Meanwhile, the treasury gave the impression of dithering. But such issues require a second, more prescriptive, article that I hope to write soon.

Based on a paper prepared for the Journal of International Economics and Economic Policy, Vol 4 No 4

We want to hear what you think about this article. Submit a letter to

More From Prospect