Learning From Our Market Mistakes

We recently conducted our own “after action review” of the events that roiled the world’s credit markets in August and September. Our objective was to identify any critical pieces of research that we had overlooked that would have better helped our readers prepare for what transpired. By far the most important was a report published in the Banque de France Financial Stability Review in May 2006. In “Market Liquidity Risk and Its Incorporation into Risk Management”, Arnaud Bervas presciently concluded that:

The excessively optimistic assessment of market liquidity (i.e., the belief that transactions can be settled at current prices without any notable delays or transaction costs) may be a serious threat to financial stability…

Admittedly [in comparison with the near failure of the Long Term Capital Management hedge fund in 1998], the financial community today appears to have a better grasp of the risks arising from the liquidity illusion. The fact nonetheless remains that current risk management tools, particularly the most common Value at Risk (VaR) measures, do not capture this complex component of market risk satisfactorily. In fact, standard VaR calculations do not take specific account of the risk to which a portfolio is exposed at the time it is liquidated.

Bervas continues:

Liquidity crisis is illiquidity risk that has reached its paroxysm. It may be defined as the market’s inability to absorb order flows without provoking violent price adjustments that are unrelated to fundamental value. It is characterized by the sudden widening of the bid-ask spread, or even the total disappearance of buy (or sell) flows and the inability to trade. It often leads to an increase in short-term volatility as well as the slump of the primary markets. It therefore contains the seeks of serious systemic upheaval…Extreme liquidity risk is not a sum of minor independent risks, but rather systemic risk that leads to a major break in the usual statistical relationships between risk factors. It is, admittedly, a rare risk, but one that is inherent to liberalized financial systems where phases of excessive optimism alternate with sharp market decline.”

Read the whole paper; it’s worth it, even after the fact.

Equally interesting is the second chapter of the IMF’s just published Global Financial Stability Review. It asks the question, “Do Market Risk Management Techniques Amplify Systemic Risks?” and finds that they do. "Having institutions that employ the same risk model [in this case, the VaR approach] is destabilizing both in terms of the covariance structure and volatility of returns relative to the historical baseline…Overall, VaR based systems provide the scope for self-reinforcing mechanisms to arise” as the institutions that use them all simultaneously sell assets to reduce their leverage and risk, triggering further asset price declines, more margin calls and reductions in liquidity, and even more asset sales. Finally, one last point about VaR models that has always bothered us was the source of their underlying assumptions about asset return volatility.

As we have noted before, these assumptions usually come from one of three sources: (a) historical data (with the risk that the future may not resemble the past); (b) a model based forecast (with the risk that the model may be wrong); or (c) the volatility implied by market prices (with the risk that everyone may be making an incorrect assumption). As you can see, all of the alternatives have a good chance of being wrong; yet without volatility assumptions, most models used in modern finance would cease to work. More than anything else, the events of August and September made it painfully clear (as we have tried to do in our writing) that for all its pretensions of precision, there is much unavoidable uncertainty at the heart of modern finance, which will always remain as much an art as a science.

Comments

[ikkyu:]

Have you not read "the black swan" (2007) by Nicholas Taleb? your conclusions are nearly the same.

Cheers,
john

2007 Oct 29 08:34 AM