Implications of the Capital-Embodiment Revolution for Directed R&D and Wage Inequality

W age inequality has increased dramatically in the United States since the late 1970s. In particular, we have witnessed growing wage differences between groups defined by observed skills such as education or experience. For example, the college premium—that is, the percentage difference between the average wages of college-educated and noncollege-educated workers—increased by a factor of four. Since at the same time the relative supply of college-educated workers increased, we would have expected to see a fall of the college premium. The fact that a decrease did not occur suggests that something else changed too. A natural candidate is technical change that has been “biased” toward skilled labor over this time. If the nature of technical change makes skilled workers relatively more productive than unskilled workers, then the wage gap will widen, assuming that market wages reflect marginal productivities. But why should technical change be biased more toward skilled labor? In fact, technical change sometimes has been biased the other way. From a perspective of understanding the evolution of wage inequality, then, it is important to determine the possible bias of technical change. In this article we investigate the long-term determinants of the bias of technical change using a dynamic model where R&D is endogenous and can be directed to specific inputs. One of the key determinants of the form of technical