“Instrumentalism” and Friedman’s methodology: a short objection

Abstract

‘Instrumentalist’ interpretations of economic theories have been held to provide justification for the rightness of being indifferent to the ‘realism’ or otherwise of the assumptions underlying a theory. In this view, what is important for an economist engaged in formulating policy is to have a theory which is capable of generating accurate predictions. This note asserts, with the help of a very simple example, that theories based on unrealistic assumptions are—precisely contrary to its claim—inimical to the interests of an instrumentalist view of theories. Tests of predictive accuracy can be narrow or broad. When narrow, the likelihood is high of there being a large number of theories passing the test—but this could also raise the chances of survival of (unrealistic) theories which may have unintended and undesirable policy outcomes. On the other hand, when the scope of the test of predictive accuracy is widened, the set of contending theories is likely to be small, with the survivors restricted to those based on ‘realistic’ assumptions.