PROFITABILITY, CONCENTRATION AND THE INTERINDUSTRY VARIATION IN WAGES

THIS paper explores empirically the hypothesis that product market imperfections affect the earnings of labor in U.S. manufacturing industries. To the extent that labor shares in the excess return due to product market power, any policies designed to reduce this power may restrain or reduce wages in the affected industry. Thus workers may oppose antitrust action aimed at their own industry. Workers may oppose increased import competition that restrains market power not only because of potential unemployment but also because this increased competition indirectly reduces future wages. In addition, measures of the social loss due to product market power are understated if some portion of costs are actually return to market power. The transfer from consumers to producers, including labor as a factor of production, is also understated. Past results attempting to isolate empirically the relation between product market power, usually represented by product market concentration, and labor earnings have been mixed. This paper argues and demonstrates that economic profitability is a superior measure to concentration in summarizing the relative extent of product market power across industries. After developing a model of the division of the total excess return available to an industry between labor and capital, the paper presents empirical results that support the hypothesis that excess return or economic profitability is superior to product market concentration in explaining the interindustry variation in wages. The results indicate that labor receives 7% to 14% of the total excess return. For empirical analysis other determinants of interindustry wage variation must be controlled. The first section of the paper briefly discusses certain relevant labor force characteristics. The second section examines the relationship between wages and product market power. The third section discusses the data and presents empirical results. If possible, variables are measured as four-year averages over 1967-1970, to avoid single-year disturbances in the data and to approach long-run equilibrium observations. The final section presents concluding comments.