On Parsimonious Explanations of Production Relations

This paper examines three bodies of macroscopic evidence that are relevant to testing the classical theory of production. No one of these bodies of evidence offers any persuasive support for the classical theory. Fits to data of the Cobb-Douglas and ACMS functions appear to be artifactual, the data actually reflecting the accounting identity between values of inputs and outputs. Similarly, there is little or no support for the U-shaped long-run cost curve, the data on firm size distributions being better explained by the Gibrat hypothesis and constant costs. Finally, observed distributions of executive salaries are explainable in simple sociological terms, without recourse to classical theory.