The most important reason researchers erred so badly on risk measurement is the manner in which EMH and most other economic investigators conduct their research. Since the Second World War the social sciences have attempted to become as rigorous as the physical sciences. No discipline has put more effort into this goal than economics. Starting about fifty years ago, economists held out high hopes that through mathematics they could make the dismal science as predictable as Einstein's theory of relativity or Kepler's laws of planetary motion. Nobel laureate Paul Samuelson, then a young professor of economics at MIT, was the first to integrate the techniques of differential equations, which had met with such success in physics, into a structured approach which could be used to study virtually any economic problem.
The key assumption was rationality: for a firm it meant maximizing profits, for an individual maximizing his or her economic desires. Rational behavior is the bedrock of Samuelson's work. This dubious platform allowed the economist to merrily build the most complex mathematical models. Economics could now be converted into a precise physical science.
The great majority of economic research gravitated in this direction, despite the warnings of some of the important economic thinkers of the past. John Maynard Keynes, for example, was trained as a mathematician but refused to build his classic theory on unrealistic assumptions. Like his teacher, the great Victorian economist Alfred Marshall, Keynes believed economics was a branch of logic, not a pseudo-natural science.
Marshall himself wrote that most economic phenomena do not lend themselves to mathematical equations, and warned against the danger of falling into the trap of overemphasizing the economic elements that could be most easily quantified.
The Samuelson revolution, however, with its emphasis on complex quantification parroting the physical sciences, came to totally dominate economics in the postwar period. Mathematics, which pre-Samuelson was a valuable but subordinate aid to reality-based assumptions, now rules economics. Good ideas are often ignored by economists simply because they are not written down in pages of highly complex statistical formulas, or don't employ equations using most of the letters of the Greek alphabet. The vast amount of research published in the academic journals contains minuscule additions to economic thinking, but is dressed in sophisticated mathematical models. Bad ideas planted in deep math tend to endure, even when the assumptions are questionable and evidence strongly contradicts the conclusions.
Economic ideas and principles once understood by educated readers are now unfathomable to all but the most highly trained mathematical researchers. This would be well and good if economics had achieved the predictability of a physical science.
But without realistic assumptions the dismal science has been broken down rather than rejuvenated by mathematics. As John Cassidy points out in an excellent article in The New Yorker, complex new mathematical theories such as those of Robert Lucas, a Nobel Prize winner from the University of Chicago, while causing a generation of novice economists to build ever more complex models, are discredited in the end, with no agreement on what should replace them.
Lucas's work concluded that the Federal Reserve should not actively guide the economy, but only increase the money supply at a constant rate.5 The research came under sharp theoretical attack, again because at the core of Lucas's complex mathematical formulas were untenable simple assumptions such as supply always equals demand in all markets. (If this were true we could not have unemployment—the supply of workers would never exceed the demand for them.) Once the supply-demand assumption is dropped, few of Lucas's conclusions hold up. Commenting on the impracticality of Lucas's work, Joseph Stiglitz, then chairman of the President's Council of Economic Advisers, said, "You can't begin with the assumption of full employment when the President is worried about jobs—not only this President, but any President."6
Economics, traditionally one of the most important of the social sciences, has suffered a self-inflicted decline. Not all are unaware of this. In 1996 the Nobel Prize for economics was awarded to two men,
William Vickrey, an emeritus professor at Columbia (for a research paper in 1961) and James Mirless, a professor at Cambridge. Although the popular press extolled Vickrey's contribution as breaking fresh intellectual ground in fields as diverse as tax policy and government bond auctions, the professor denied the hyperbole. He said, "[it's] one of my digressions into abstract economics At best it's of minor significance in terms of human welfare."7 When interviewed, he talked instead about other unrelated work he had done, which he considered far more important.
The failure of the complex statistical models to provide much insight into current economic problems has resulted in a cutback in hiring of economists by Wall Street and major corporations. Lawrence Myers, who before becoming a governor of the Federal Reserve, ran one of the nation's most successful economic forecasting firms, St. Louis-based Macroeconomic Advisers, said, "In our firm we always thanked Robert Lucas for giving us a virtual monopoly. Because of Lucas and others, for two decades no graduate students were trained who were capable of competing with us by building econometric models that had a hope of explaining short-term output and price dynamics. We educated a lot of macroeconomists who were trained to do only two things—teach macroeconomics to graduate students and publish in the journals. . . . [These economists] don't care what happens out there. [They] don't try to build models which are consistent with the real world."8
Complicated statistical analysis is no different in the investment arena, nor should it be, since it's another branch of economics. Simple assumptions are usually necessary as a platform for abstruse statistical methods. More complex assumptions, though far more descriptive of the real world, do not allow the development of the statistical analysis the researchers desire, or the academic journals will publish.
The assumption of total rationality is the mother lode of complex statistical analysis. It eliminates the need for any other psychological assumptions, which, though likely to provide better guidelines to investor behavior in the real world, would vastly complicate the analysis, and probably send it in directions completely away from the researchers' paradigm.
Given the simple assumption of rationality, researchers in the best tradition of the Samuelson Revolution can merrily take off to examine how the totally rational investor will approach markets. They can then use the most complex differential equations or other statistical methodology to discover new results. Whether these assumptions have the remotest connection to reality is irrelevant. Who cares?
Was this article helpful?