The credit crisis continues unabated. In order to understand the nature of the crisis, it is useful to think of it as a failure to price liquidity risk correctly; a failure that has an interesting dynamics. During the last five years prior to August 2007, the market systematically underestimated liquidity risk and therefore systematically put the liquidity risk premium too low. This led to excessive liquidity creation. Now the reverse has been going on: the market is overestimating liquidity risk and puts the liquidity risk premium too high, leading to a destruction of liquidity. The central banks try to increase liquidity but this is more than offset by liquidity destruction by the financial sector.
How could this happen? In order to answer this question we should go back to the basics of liquidity and credit creation. Liquidity and credit are created whenever financial institutions (banks) borrow short and lend long. In so doing, they become less liquid themselves, making it possible for the non-bank sector to become more liquid. In this sense banks create liquidity (and credit) for the non-banking sector. They take a risk in performing this service, though. The risk arises from the fact that their assets (loans) have a longer maturity than their liabilities (deposits). As a result, they are vulnerable to a” run on the bank”, i.e. a simultaneous withdrawal by many depositors. If this liquidity risk materializes banks find out that they cannot pay out the depositors.
The liquidity risk banks take in this whole process is a very special one. It is special because it is a “tail risk”, i.e. a risk that materializes very rarely, but when it does, its consequences are catastrophic. Such a risk is difficult to price. There are several reasons for this. First and foremost, liquidity risk has a very special structure. Underlying it are two possible equilibria. A good one, that exists most of the time, when the deposit-holders trust the banks’ ability to honor their commitments. As a result, nothing happens and the banks have the impression that the risk is very low. There is also a bad equilibrium that arises when for some reason deposit-holders loose the confidence they have in the banks. They then have an incentive to all run simultaneously to the bank, forcing the bank to sell its assets. Many banks default Thus, a liquidity crisis quickly becomes a default crisis. The main characteristic of this crisis is that it occurs when agents loose confidence because they think that others have lost confidence. And these others loose confidence because they think that others have lost confidence, etc.
It is this collective panic that makes tail risk so difficult to quantify. As a result, financial modeling has routinely not given this kind of risk much attention. Instead it has focused on normally distributed risks which can be easily quantified. Thus almost all internal risk models produced by large banks assume that returns are normally distributed. These models typically vastly underestimate the probability of liquidity crises. This was made very clear when after August 8 a spokesman of a major bank declared that according to the internal risk model of his bank a credit meltdown of the size observed in August could only occur once every ten thousand years. It occurred several times in the last one hundred years.
A second reason why liquidity risk is difficult to price correctly is the result of the lender of last resort function of central banks. When a liquidity crisis erupts, central banks are forced to intervene and to provide liquidity in the market. They are the only ones capable of doing so, because they are the ultimate providers of liquidity. The knowledge that central banks will inevitably intervene in times of crisis creates a perception during normal times that liquidity risk is not something banks have to worry about in their pricing of the risk they take. This is a kind of moral hazard but not in the traditional sense. Traditionally, moral hazard has been seen to occur when the central bank bails out individual banks, thereby giving banks an incentive to take on excessive risk later on. It is not in this traditional sense that moral hazard occurs when the central bank provides liquidity in times of crisis. When the central bank injects liquidity through open market operations it aims at averting the generalized liquidity crisis (without focusing on the liquidity needs of a specific bank). If central banks are successful, this success will create a belief among bankers that there is no reason why they should price liquidity risk when taking decisions about the maturity structure of their assets and liabilities.
This is exactly what has happened during credit bubble before August 2008. Convinced that liquidity risks are nothing to worry about, banks have failed to price it. They have been supported in this belief by the conviction that central banks would take care of it.
Why have central banks not reacted in time? They must have seen that banks were increasingly taking on more liquidity risk without pricing it correctly. The fact is that they did not see this. There are a few exceptions. The BIS has been crying wolf for a long time but was politely disbelieved by the others (with the possible exception of the ECB which emitted timid warnings).
In a way it was not that easy to see what was going on. Using sophisticated new financial instruments, banks systematically pushed many of their lending activities outside their balance sheets. As a result, they were funding illiquid investments made by special conduits with short-term loans, thereby circumventing the normal supervisory and regulatory framework. The smokescreen created by the banks was very successful in hiding the true nature of the risks to those who should have taken action to prevent this from happening.
Thus, the credit crisis was made possible because banks failed to price liquidity risk correctly and central banks were fooled by the same banks. Today a movie in reverse is unfolding. Banks cannot get out of their torpor, because nobody dares to trust anybody. We are now in the phase that banks systematically overestimate the liquidity risk. As a result, they give it too high a price. This leads them to drastically change the maturity composition of their assets and liabilities. They all want to shorten the maturity of their assets and to lengthen the maturity of their liabilities. Since they all want to do this at the same time they cannot succeed. The assets of one bank are the liabilities of another. The former wants to shorten its maturity that the latter wants to lengthen it. This leads to a massive destruction of credit and liquidity, exactly the opposite of what happened during the bubble phase.
How to get out of this conundrum? Central banks are trying hard to create liquidity but until now they are unsuccessful. The extra liquidity they bring in the market is used by the banks to shorten the maturity of their assets. As a result, the impact of the central banks’ liquidity injection on credit and liquidity creation in the system as a whole is zero. And banks continue to destroy liquidity.
The way out is to make banks price liquidity correctly. This means in the present circumstances to convince banks that the liquidity risk premium they are now applying is too high. This is easier said than done. The reason why banks now use an excessively high liquidity risk premium arises from the same collective action problem we identified earlier. The lack of trust each bank has vis-à-vis the others prevents them from doing their normal business of borrowing short and lending long. Thus the financial system is stuck in the bad equilibrium.
Banks will not easily get out of the bad equilibrium. Collective action must be triggered by the authorities. The ingredients of this collective action are the following. First, liquidity injections by the central banks will continue to be necessary. These liquidity injections, however, are insufficient. Other measures will have to be taken, including lowering interest rates. A recession would exacerbate the credit crisis and lead to paralysis in the banking system. If the banks remain stuck in their torpor, the authorities may have to bypass the banking system and become the main providers of credit to the non-banking sector. The most obvious way would be to finance government budget deficits directly. All this sounds terrible for most finance students who have been brainwashed to believe that governments cannot do anything useful and that only agents in markets are capable of taking rational decisions. .
In the longer run, when the present credit crisis is overcome, the authorities will have to bring their act together by making sure that supervision on the creation of credit and liquidity is restored. Otherwise banks will start repeating the same mistakes as soon as they return to normal credit conditions. And there is one thing supervisors now know for sure. This is that they should never again trust the banks’ internal risk models.
There is an important lesson to be learned by those who believe that financial markets are always and everywhere efficient. Financial markets are very good at pricing risks that arise from normally distributed shocks. They fail completely to price tail risks. The latter often arise as a result of collective movements of panic and madness. That’s when the government has to step in. There is no place for these scenarios in our finance textbooks. It’s time we start teaching business students the whole story. This will also teach them that governments can be useful, an idea that is still considered to be devilish in many business schools.