The effects of employment change on payroll per employee : implications from size-wage premia and labor composition

Abstract

Employer size has been linked to higher wages across industries and occupations by a host of studies, but a paradox emerges in the relationship between size and payroll per employee. From multivariate analysis on establishment-level, longitudinal data compiled by the United States Census Bureau, an unexpected, negative relationship exists not only between payroll per employee and size but also with growth. The relationship exists whether employment change is positive or negative, over short and long periods, or measured in fixed or relative terms. The negative effect on average payroll is strongest for the most dynamic change rates and weakest for expanding larger establishments hinting at a diminishing effect across size. The presence of workforce compositional changes within the establishment cannot be directly observed but is nonetheless the most logical explanation; lower wage employees are the primary means by which establishments expand and contract. As to the observed shape of the trends—convergence toward zero and compression of predicted changes in average payroll across size—the available data provides no clear indication of the components at work. Plausible factors stem from size-wage differentials, saturation of lower-wage workers, and/or influences of capital on worker bargaining power. All or none of these may be present but their presence and magnitude are little more than conjecture. Nonetheless, there is certainty in that the there is an unquestionable presence of a negative trend across establishment size categories in payroll per employee during growth and a positive trend during downsizing.