Prediction Market Theory: How a book published in 1921 can help explain the 2008 Wall Street Crash!

The Betfair Prof / Leighton Vaughan Williams / 29 June 2009 / Leave a Comment

The Betfair Prof, Leighton Vaughan Williams, explains it all...

Born and brought up on a farm in McLean County, Illinois, he claimed that the only reason he turned to economics was that plowing was too hard on the feet. Well, farming's loss was economics' gain, and Frank Hyneman Knight's seminal volume, Risk, Uncertainty and Profit, published in 1921, was to show just how great that gain was.

In this major contribution to economic thought, Knight makes an important distinction between insurable and uninsurable risks. According to Knight, profit--earned by the entrepreneur who makes decisions in an uncertain environment--is the entrepreneur's reward for bearing uninsurable risk. Essentially, he was highlighting the distinction between measurable risk and what he terms 'uncertainty'. The world of business, he argued, falls into the realm of this uncertainty, since it deals with "... situations which are far too unique for any sort of statistical tabulation to have any value for guidance. The conception of an objectively measurable probability or chance is simply inapplicable."

What makes Knight's insight so topical is that these ideas are particularly relevant to financial markets, where surprises occur regularly. For years Knight was ignored by the theorists of mainstream financial economics, who either didn't know or didn't care for Knights' belief in "the sheer brute fact that the results of human activity cannot be anticipated". Knight was perhaps over-stating his case, but the theorists who followed him, with their mathematical models of risk management, went to the diametrically opposite position, constructing a mathematically tractable universe in which rational agents can measure and value risk and return, and trade them with confidence.

In this alternative universe, risk can be reduced to statistical variance and forecasts of the future can be encapsulated in measurable probabilities. The implication is that economic agents are indeed rational, or failing that, act 'as if" they are (the famous 'as if" theorem first proposed by Milton Friedman in his 1953 'Essays in Positive Economics'). The other implication is that the financial world is 'ergodic', in the sense of possessing a stable, recurrent underlying structure.

In this alternative universe, all risk is insurable and the collapse of the sub-prime mortgage market shouldn't have been such a big deal. After all, the risk of that happening was insurable, wasn't it? Of course it was, using so-called 'credit default swaps' (CDSs). In this context, these are basically bets (otherwise termed 'private insurance contracts') based on whether people will default on their mortgages. Unlike bookmakers and betting exchanges, those laying the bets were, however, unregulated and investment houses that sold them weren't required to set aside funds sufficient to cover their potential liabilities. At their height, the nominal value of the CDSs was in excess of $63 trillion (63,000 billion dollars!). And when the 'insurance' was called in, the layers just didn't have the money to pay.

Goodbye Bear Stearns, Goodbye Lehman Brothers, Goodbye conventional wisdom.

Could a well-constructed prediction market have predicted all this? I guess it depends who was placing the bets!