Cod – Liver Oil and Methodology

Paul Samuelson Free Trade and Comparative

The old-fashioned virtues of administering cod-liver oil to children, as a means of correcting vitamin deficiencies, have been generally forgotten in the affluent West. The medicine might have been unpleasant, but it was good for you! The gradual neglect of methodological and philosophical considerations within econ­omics has, similarly, deprived many students of distasteful but essential knowledge. The deficiencies are, often, readily apparent. What had disturbed Samuelson, in this respect, is the damage caused by the uncritical (or unknowing) acceptance of the plausible methodology of Friedman and the Chicago School. Friedman's well-known view (1953) is that economic science advances through positive economics - the formulation of hypo­theses that are to be judged according to the precision and scope of their conformity with the evidence of the predictions that the theory provides. Insofar as this mirrors the search of Karl Popper (Magee, 1973) for a method of distinguishing science from non-science and echoes his recommendations that theories should be formulated unambiguously in order that they should be sus­ceptible to refutation, Friedman's view is unexceptional. However, Friedman continues by arguing that it is of no consequence nor relevance whether the assumptions behind a theory are unrealistic and that, in general, the greater the fruitfulness of a theory then, necessarily, the more unrealistic the assumptions.

Samuelson's response to this has been quite unequivocal; to regard a theory as all the better for its shortcomings is a 'monstrous perversion of science'. In describing the position adopted by Friedman above as the 'F-twist', Samuelson drew attention to the empirical invalidity of theorems deduced from counter-factual hypotheses. Perhaps the most important example of such a counter-factual hypothesis would be the perfectly competitive laissezfaire model of economics that Friedman has consistently supported over the years. But, in criticising Friedman, is Samuelson inadvertently criticising himself? It has been suggested by Machlup (1964) that the 'F-twist' might also be described as the 'S-twist' since the operationally meaningful theorems that Samuelson derives are, of a necessity, empirically invalid since they draw upon unrealistic assumptions. The role of such theorems might then be said to illuminate the divergence between the predictions of a hypothesis and observed reality, by reference to the departure of the assumptions from real world conditions. For example, does Samuelson's famous theorem (1948, 1949) that the perfect mobility of commodities under free trade will equalise the rewards to factors of production in a two-region (USA and Europe), two-commodity, two-factor model (with no factor movement but partial specialisation according to the law of comparative advantage) fall into the 'F-twist'? Samuelson's response was that although neither the assumptions nor conclusions were empirically valid (when measured against the complexity of the real world), the accurate prediction of the theorem reflected the realism of the model, viz: the movement of goods by trade would improve European living standards by nearly as much as large-scale emigration. If there is a suspicion of hand-waving here, it is undeniable that the logic Samuelson has brought to bear on the discussion of economic methodology has yielded some extremely interesting concepts.