Keen – Friedman and assumptions in economics

Steve Keen – Friedman and assumptions in economics

Milton Friedman argued that a theory should be judged on its ability to predict, not the realism of its assumptions. Keen is critical of this ‘instrumentalist view, as shown in this extract, taken from pp.163-164 of Keen, S. (2011). Debunking economics : the naked emperor dethroned? (Rev. and expanded ed.). London; New York: Zed Books Ltd.

Economists often imply, when they fob off some critical student, that the unrealistic assumptions in introductory economics courses are dropped in more advanced theory — which portrays these assumptions as heuristic tools. In fact…the assumptions used in more advanced theory are often more unrealistic than those presented in introductory lectures.

Scientific realism versus instrumentalism Musgrave also points out that most scientists reject an instrumental view of science in favor of ‘scientific realism’ — the belief that scientific theories should not merely predict reality but should, in some sense, represent it.

Ironically, this is actually the belief that most economists have about economic theory. Friedman’s instrumentalism is little more than a smokescreen behind which to hide when one wishes to quell a budding class rebellion. It is often evident to the student objector that, though professing that the assumptions don’t matter, his teachers continue to use the same small class of assumptions over and over again: rational utility-maximizing individuals, profit-maximizing firms, and a plethora of ancillary assumptions built on these foundations.

These assumptions are used because economists believe that these assumptions do capture essential elements of reality, and regard any theory which does not use these building blocks as ‘unrealistic.’ This belief is most clearly seen in the manner in which the ‘bibles’ of economics, its academic journals, filter out papers that do not make this core set of assumptions.

Assumptions do matter — to economists The proposition that assumptions don’t matter implies that economists would be quite willing to accept a theory which assumed irrational behavior if the model generated results which accorded with observation. It also implies that the development of economic theory would be driven primarily by the desire to produce theories that provide a closer fit to observed data.

Both these implications are strongly at variance with reality.

As any non-orthodox economist knows, it is almost impossible to have an article accepted into one of the mainstream academic economic journals unless it has the full panoply of economic assumptions: rational behaviour (according to the economic definition of rational!), markets that are always in equilibrium, risk as an acceptable proxy for uncertainty, and so on. When it comes to safeguarding the channels of academic advancement, little else matters apart from preserving the set of assumptions that defines economic orthodoxy.

Similarly, the development of economic theory over time has been propelled by the desire to make every aspect of it conform to the preferred economic model.

Uploaded 15 October 2014