Robots go haywire in a mechanical market

Powered by automated translation

Review: Beyond Mechanical Markets: Asset Price Swings, Risk, and the Role of the State, US$22.97 (Dh84.35) Amazon.com, by Roman Frydman and Michael Goldberg

If the central argument of Beyond Mechanical Markets were accepted, we would effectively have to discard decades of economic and market research. This would involve throwing away mathematical models with which central banks, the IMF and many other prestigious bodies make their predictions and decisions.

Roman Frydman and Michael Goldberg, two professors at US universities, offer a fairly compelling case that just such a brash rethinking is necessary. While early modern economists such as John Maynard Keynes and Friedrich Hayek understood properly the limits of economics as a forecasting tool, the authors say time and technology ultimately led to wide acceptance of mathematical models that hinged on the presumption that people behave rationally and uniformly.

This, of course, is not the case - people buy and sell stocks, for example, for reasons that are inherently unpredictable. They also change their strategies and outlooks based on what Frydman and Goldberg term "nonroutine" events that mathematical models do not and indeed cannot account for. When people invest, they first get what relevant information they can, and then, facing an unknowable future, take their best stab at it.

So how should economists confront a chaotic market composed of humans armed with varying degrees of information and insight? Frydman and Goldberg do not yet have bountiful answers, aside from proposing a Contingent Market Hypothesis to replace the old Efficient Market Hyphothesis.

The new approach scraps the simple formulation that undergirded the old one: that market prices, on the whole, accurately reflect the "correct" value of the stock, bond or other security in question. In its place comes a view of markets as an agglomeration of understandings and philosophies that vary over time in response to unpredictable events.

It is hard to argue with eschewing mechanical models of markets that treat people as mere robots and the future as a problem to solve rather than an uncertainty to stare in awe at.

But while Beyond Mechanical Markets is an instructive look at the nature of markets and forecasts, its railing against the fragile underpinnings of conventional economics is facile. Its authors do recommend a few real-world responses, including having regulators announce price ranges they deem acceptable as a means of containing asset price bubbles.

What other major adjustments in policy spring from the Contingent Market Hypothesis, though, remain to be seen.