Should Workers Care About Firm Size?

The question of wage differentials by firm size has been studied for several decades with no commonly accepted explanations for why large firms pay more. In this paper, we reexamine the relationship between firm size and wage outcomes by estimating the returns to unmeasured ability between large and small firms. Our empirical methodology, based on non- linear instrumental variable estimations, allows us to directly estimate the returns to unmeasured ability by firm size and, therefore, to test the two main theories of wage determination proposed to explain the relationship between firm size and wages, namely ability sorting and job screening. We use data from the Survey of Labour and Income Dynamics (SLID), which provides longitudinal information on workers' and firms' characteristics, including establishment and firm size. We find significant differences in the returns to unmeasured ability across firm size. In particular, we find that the returns to unmeasured ability seem to follow a non-linear pattern. The returns to unmeasured ability are significantly higher in medium size (above 500, but below 1000 workers) firms relative to small firms. However, the returns to unmeasured ability are not significantly greater in large firms relative to medium or small firms. Overall, it seems that ability sorting dominates for moves from small to medium size firms in that ability is more productive and, therefore, more rewarded in the latter than the former. On the other hand, when firms become "too large," the monitoring costs hypothesis seems to dominate in that ability is not more rewarded than in smaller firms.