Macroeconomic Implications of Variation in the Workweek of Capital

IN A TYPICAL workweek, one in every four manufacturing production workers in the United States is employed at night. This fraction fluctuates sharply over the business cycle, accounting disproportionately for business cycle changes in employment. The variation in work at night amounts to over 40 percent of the cyclical changes in the employment of manufacturing production workers. While cyclical movements in shiftwork are pervasive in many manufacturing industries, there are also some nonmanufacturing industries in which such variation is important. Changes in the number, as well as the length, of shifts affect the workweek of capital. Demonstrating the importance of the workweek of capital for studying business cycles is the main goal of this paper. In most business cycle models, the capital stock is taken as quasi-fixed. Cyclical variation in output arises from applying more or less labor to a fixed capital stock. This property of capital should lead to diminishing marginal product of labor, to countercyclical real wage and average product of labor, and potentially to capacity constraints. Keynesian students of cyclical productivity, however, have long realized the importance of variable utilization of capacity. Their models, which admit the persistent unemployment or underemployment of both capital and labor, can readily dispense with the implications of diminishing marginal product of labor. Work on equilibrium business cycles also finds