Concluding remarks
This chapter argues that labor market inequalities are shaped by technological change through a variety of economic mechanisms. Within the technology-labor market nexus, however, which of the specific mechanisms we evaluate are most likely to survive the test of time?
Before answering this question, it is useful to put things in perspective and recall that most of the statements we have made in this chapter are not all meant to represent general insights; rather, they allude to a particular historical episode.
Specifically, technology has not always been skill-biased in the past: the transformation from artisanal workplaces to the factory in the 19th century had much the opposite effect [Goldin and Katz (1998)]. Moreover, not all the drastic productivity advancements in the past were embodied in equipment: electricity was to a large extent embodied in new structures, as the electrification of production required a whole new blueprint for the plant [Atkeson and Kehoe (2002)]. Even in reference to this particular historical episode, there are serious dissenting views on the overall impact of IT on the macroeconomy [e.g., Gordon (2000)] and on the role of technology in explaining the observed changes in the U.S. wage structure [e.g., Card and DiNardo (2002)].In returning to the original question, we identify three rather general categories that we find particularly interesting and plausible.
The first idea is factor-specificity of the recent technological advancements. In particular, the embodiment of productivity improvements in equipment capital goods, and the skill-bias of such productivity improvements. Whether in the Nelson-Phelps version of skills as a vehicle of adoption and innovation, or in the version of skills and capital as complementary in production, the skill-bias of the IT revolution is one of the most robust and pervasive in the literature.
Skill-biased technical change and capital-skill complementarity are crucial to explain the climb of the skill premium, notwithstanding the continuous growth in the relative supply of skilled labor. A growing and promising avenue of research is on the endogenous determinants of the factor-bias in technological advancements [Acemoglu (2002b, 2003b)].The second idea is vintage human capital. The technological specificity of knowledge appears to be an important idea to explain some of the most puzzling aspects of the data such as the rise in within-group or “residual” inequality, the fall of the real wages at the bottom of the skill distribution, the growth in the returns to experience, and the slowdown of output growth in the aftermath of a technological revolution.
The third idea is the interaction between technology and the organization of labor markets. Radical technological developments, like those we have witnessed in the past three decades, are bound to interact deeply with the various aspects of the structure of labor markets, like the organization of production within the firm, labor unions, and labor market policies. Through this interaction, the literature has successfully interpreted the move from the Tayloristic to the flatter multi-tasking organizational design of firms, the decline of unionization, and the upward trend in unemployment rate in Europe. In particular, the comparison of the U.S. and European experiences seems a fruitful way of studying this channel.
These ideas are the building blocks of the most successful and influential papers in the first generation of models that we have surveyed in this chapter. Where will the literature go next? We argued in various parts of the chapter that one major weakness of this literature is the scarcity of rigorous quantitative evaluations of the theories proposed. Most of the papers reviewed are qualitative in nature. This is not too surprising, given the young vintage of the literature (which developed only starting from the mid-1990s), and given that, in any field, it naturally takes a long time before a handful of theoretical frameworks emerge as successful and begin to be used for a systematic quantitative accounting of the facts (e.g., the search and matching model in the theory of unemployment, and the neoclassical and the endogenous growth model in the theory of cross-country income differences). Inthis chapter we have highlighted some features that seem important for a successful theory of the link between technological change and labor market outcomes. Quantitative theory should be a priority within this field of research over the years to come.