Numerous books have been written on the causes of the global financial crises of 2007-2009, some having more value than others.
The Number that Killed Us is perhaps unique in its single focus on the role that Value at Risk models played in the crisis. Unfortunately, this single focus is taken to such an extreme that that book removes the crisis from its broader historical context and recasts the underlying causes of the crisis with a single lens. It also extrapolates its points into an all out attack on the use of all quantitative methods in finance.
The second chapter explains how the big investment banks got their regulators to adopt capital requirement rules based on VaR figures which the regulated entities could develop using their own models. While the narrative is informative, the book removes it from its historical context and characterises this event as “one of the most puzzling happenstances in financial history.” It even goes so far as declaring it as the true beginning of the crisis. However, the reality is that there has been such regulatory capture as long as there has been regulation. Furthermore, identifying the beginning of the crisis with the beginning of VaR-based regulation obscures the true roots of the crisis which is overleveraging of risky assets throughout the financial system.
In later chapters, the book does a reasonable job of explaining the fundamental flaw in the strategy of being highly levered in normal times and planning on liquidating positions should levered bets fail; namely, because everyone else is trying to do the same thing, the formerly highly liquid assets become illiquid. However, while it may be true that such faulty strategies were justified during this period using faulty VaR models, that does nothing to explain why history is replete with such events before the advent of VaR models. (Think the Panic of 1907, the Great Crash of 1929, the UK crash of 1973-74). The crisis of 2007-2009 was merely the latest in a long history of such crises. VaR was simply the instrument of choice for rationalising overleveraged positions this time.
Towards the end of the book, the author cites several well-respected academics who support the argument that banks should be far less leveraged. However, he then goes on to chastise those same academics for not taking his position that all modelling needs to be banned from risk management.
His justification for this position is however self-contradictory, arguing on the one hand that risk is too complicated to be modelled mathematically in some places while in other places arguing that managing risk is so simple that mathematics is not required. He argues that risk-taking and regulation be done using only “common-sensical” approaches, as if such approaches themselves do not assume the validity of certain models.
My suggestion is that anyone reading this book first become familiar with financial crises in general so that they can separate the factual information it contains from its bluster.
Paul Kaplan is quantitative research director with Morningstar Europe. This Morningstar article first appeared in Financial Adviser, part of the Financial Times Group.
