The Impact of Technological Change on LowWage Workers: A Review.

November 2005

David Card, University of California Berkeley and the National Bureau of Economic Research; John DiNardo, University of Michigan and NBER.

Download paper here.


The relationship between technological change and the earnings of less-skilled workers is one of the oldest issues in economics (Berg 1984). Renewed interest in the link was spawned by labor market trends in the 1980s, including the decline in real wages for younger and less-educated workers, and the sharp increase in the wage gap between college and high school educated workers. At the same time, the introduction of the micro-computer was hailed as a revolutionary event that promised to change the nature of work. Two prominent studies written at the close of the decade Bound and Johnson (1992) and Katz and Murphy (1992) argued that the falling fortunes of less skilled workers were caused by adverse demand shocks, specifically "skillbiased" technological changes induced by the new computer technology. A vast subsequent literature has tended to confirm this basic view.1 By now it is widely accepted that technological changes have hurt, and will continue to hurt, the labor market prospects for less skilled workers in the United States and other advanced countries. The technological change hypothesis, in turn, has provided a powerful intellectual foundation for a laissez faire approach to policies for aiding less skilled workers. In this paper we present a critical review of the literature linking technological change to the structure of wages in the U.S. economy. We argue that the evidence for the technological change hypothesis is weaker than many observers have recognized. From a research design perspective we identify two key concerns. First, many studies reason backward from an effect (recent changes in the time series behavior of wage inequality) to a cause. A typical study does not ask "what is the evidence for an effect of this technological invention?" Rather, most have adopted a forensic approach, asking "why has wage inequality increased? " In a world where there are many potential causes, some of which are unknown (or ignored), a forensic approach can only eliminate candidate explanations. Even when such an analysis has ruled out all but one of the enumerated hypotheses, the analysis provides at best only limited support for the remaining explanation, since others could be constructed to explain the same set of facts.

Employment, Unemployment, and the Labor Market