on 5 August 2010
Standard finance theory imagines that prices move independently of agents' behaviour. But if enough agents buy an asset, its price will be bid up. Shin shows that in a number of very standard frameworks, the behaviour of optimising financial agents generates financial instability, bidding prices up further in response to initial price increases, and vice versa. Consider a bank marking-to-market, and managing its risk according to a Basel II mandated VaR target: when the value of a risky asset that it holds increases, so does the equity on its balance sheet, allowing it to expand its holdings of the risky asset. Equivalently, a dynamic delta hedge requires fund managers to sell the underlying in response to falls in its price, further driving down its price in a way that makes the initial calculation of delta inaccurate even as a linear approximation.
By implication, the debate over fat-tails versus Gaussian distributions may only have a second order relevance to the recent credit crunch: failing to understand that these distributions are determined by financial agents' behaviour seems a far greater failing.
A decade ago, Shin and colleagues presented these arguments about the weaknesses of VaR to the committee drafting the Basel II regulations. This book, and the Clarendon lectures on which it is based, give cause for hope that these arguments are finally being listened to.