The new empirical finding in this paper is not the well-known correlation between wages and capital intensity but rather the correlation between weekly hours and capital intensity. This correlation may either be explained by, contradict, or complement the efficiency wage findings, depending on other assumptions. First the contradiction. The efficiency wage literature generally assumes that weekly hours are fixed and intensity of effort is variable and costly to observe. But if hours of work are variable and if workers prefer fewer hours, then an “efficiency contract” can stipulate both a higher hourly wage rate and also fewer weekly hours than the prevailing market contract. This would make us expect a negative correlation between hours and wage rates.

If instead employers are indifferent to the number of hours worked by each individual employee per week (monitoring problems but no fixed costs) then the relationship between hours and capital intensity could merely be a secondary labor supply effect caused by the high hourly wages offered in capital-intensive sectors. Workers rationally choose to work more weekly hours because of the higher opportunity cost of leisure. This view of course makes the additional assumption that substitution effects outweigh income effects.

Our findings could also complement the efficiency wage literature. Some monitoring costs are like capital costs in that they are paid per worker rather than per hour. Other may be subject to learning curves and other economies of scale. Then our theory suggests that industries that incur high monitoring costs would also require their workers to exert additional effort. Thus the inter-industry wage differential would be part wage premium, part compensation for higher effort.

Distinguishing between these possibilities is not necessary for our purposes and clearly outside the scope of this paper. Our whole approach of tracing out the wage-effort offer curve at different points in time and connecting its movements to changes in product prices, technology and worker benefits represents a substantial departure from the efficiency wage literature. This comes from the very different conceptual frameworks that underlie wages determined to solve monitoring problems from wages that are instead payment for observable effort.