ABSTRACT
Few subjects in labor economics have received more attention than the esti mation of union-nonunion wage differentials (Lewis 1986; Jarrell and Stanley 1990; Andrews et al. 1998; Blanchflower 1999). Although there is a consen sus that union density and organizing strength in the United States have de clined over time, there is no agreement as to how union wage premiums have changed. The theoretical relationship between union density and the union wage gap is ambiguous. For any given level of union labor demand, there exists an inverse relationship between union density and the wage premium, while declining union premiums are more likely if demand curves for union labor are shifting inward.1
