The efficient markets hypothesis is the catwalk supermodel of economics.
Strutting down the runway in haute couture, catwalk supermodels present an elegant stylized vision of fashion that bears scant resemblance to the reality of buying clothes on the high street.
Through its failure in successive crises – the 1987 stock market crash, the collapse of Long-Term Capital Management and the current financial crisis – the EMH and the models it has spawned has been shown to have a similar lack of relevance to how financial markets actually work.
The recent series of articles in the Financial Times highlighting the shortcomings of the EMH are important because, despite the flaws, the ideas underpinning the theory remain the well-entrenched orthodoxy.
Trading rooms are not filled with bankers pontificating on the finer points of economic theory, but the tools they employ on a daily basis – from derivatives pricing models to risk management metrics such as Value at Risk – are all ultimately based on a vision of the world that assumes constant liquidity and the near impossibility of extreme events.
Several of these critiques view the failure of EMH as an opportunity to replace it with something else. Writing on behavioural finance, Jonathan Davis showcases the work of Professor Andrew Lo of Massachusetts Institute of Technology, who admits financial markets simply do not lend themselves to deductive theory as well as the physical world. Similarly, Paul De Grauwe, Professor of Economics at the University of Leuven, suggests the crisis creates the opportunity to develop better models, but notes “the interaction between… imperfectly informed individuals regularly creates collective movements of euphoria and panics. These phenomena are hard to model. Yet this is what macroeconomists will have to do if they want to regain respectability as scientists.”
This last statement contains the essence of the problem. Over the past 50 years economics has attempted to turn itself into a “hard” science through mathematical rigour. But as Professors Lo and De Grauwe both admit, the real world does not lend itself readily to this form of analysis. Rejecting the simplicity of the EMH and focusing on how economic agents actually behave may well produce models that are an improvement on their predecessors, but this is still asking the wrong question. Using the catwalk analogy, we shouldn’t ask how supermodels could be made to look more like normal people and dressed accordingly. Instead, the important question is “do we really need supermodels and haute couture fashion shows to shop on the high street?”
The first step is to accept the futility of attempting to describe human interactions with mathematical models, particularly ones that return automatically to an equilibrium state. Economics used to be a descriptive discipline – Keynes’s original work was not a mathematical treatise – and would benefit from becoming descriptive once more.
Interestingly, a search of the Financial Times database for the economist Hyman Minsky, yields several references at the height of the crisis, primarily focusing on the “Minsky moment” when a speculative bubble bursts. His descriptive model, to my knowledge, is not referenced in the more recent discussions on EMH, perhaps because it considers instability to be the norm: “the internal workings of a capitalist economy generate financial relations that are conducive to instability and that the price and asset value relations that will trigger a financial crisis in a fragile financial structure are normal functioning events”.
The second step is to accept that extreme events are not that rare. Models based on the EMH are often used as an excuse when things go wrong.
The one-in-a-million year or 25 sigma event that no model could predict is regularly trotted out as an excuse – every few years or so.
The important point about extreme events, Black Swans to quote prominent critic Nassim Taleb, is that they cannot be predicted or modelled. The conclusion, put forward in a new book, Lecturing Birds on Flying by Pablo Triana, is that “[having] no model is better than a dangerous model”.
If we discard the flawed models, what do we replace them with? Taleb and Triana’s answer is experience honed by common sense.
If this sounds anathema, it is a reflection of how powerfully entrenched the false sense of precision offered by models has become.
But, in fact, this approach has already been successfully used during the credit crisis.
At the end of 2006, when their peers were still accumulating ever-larger credit exposures, senior management at Goldman Sachs took a negative bet on the US sub-prime mortgage market. Goldman’s subsequent financial performance relative to other investment banks is testimony to the power of experience and common sense over slavish adherence to mathematical models.
Get alerts on Fund management when a new story is published