Friedman and Machlup on the Significance of Testing Economic Assumptions

This article questions Milton Friedman’s methodological position in a famous essay dating to 1948 where he questions the validity of tests of the assumptions of economic theory. Valid tests, he maintains, by and large, concern the empirical implications of hypotheses derived from the theory. The truth or falsehood of assumptions is “largely irrelevant.” In response, this article argues that Friedman’s position is questionable but for different reasons depending on the nature of the assumptions. “Auxiliary assumptions” concern the environment in which the test is supposed to take place. “Generative assumptions” concern the postulates or theorems from which the hypothesis is derived. If auxiliary assumptions are false, all test results bear less weight, whether are confirmatory or disconfirmatory. If generative assumptions are false, positive results confer less confirmation on the hypothesis. The article further questions Friedman’s famous proposal to treat the assumptions of economic theory as “as if” statements, which are not really supposed true but simply taken for granted. The article goes on to question a related one by Fritz Machlup agreeing with Friedman about the irrelevance of tests of “generative assumptions.” Unlike Friedman, however, Machlup stands on the authority of philosophers of science, who, he claims, maintain that theoretical postulates in science should be regarded as “rules” that must be followed.