Econ 210c, Spring 2002, Reading Notes for March 6

Brad DeLong

Claudia Goldin and Lawrence F. Katz (1998), "The Origins of Technology-Skill Complementarity," QJE (August).

Richard Freeman (1998), "Spurts in Union Growth."

Claudia Goldin and Lawrence F. Katz (1999), "The Returns to Skill in the United States Across the Twentieth Century."

Claudia Goldin and Lawrence Katz (1999), "Education and Income in the Twentieth Century: Evidence from the Prairies."

Claudia Goldin and Robert A. Margo (1992), "The Great Compression: The Wage Structure in the United States at Mid-Century," QJE 107 (February).

This is Claudia Goldin week...

In 1940 in America the man earning more than 90% of other men made five times as much as the man earning more than only 10% of other men. By 1950 this 90-10 measure of wage inequality had shrunk from five to three. This "Great Compression," as Goldin and Margo call it, was the single most significant change in the relative distribution of male (or, rather, white male) earnings in the United States (at least until the tremendous widening of the income distribution that has taken place in the last two decades). The Great Compression persisted for a generation: in 1975 white male wage differentials were very similar to what they had been thirty years before at the end of World War II. Almost half of the Great Compression is due to falling returns to schooling; the other half is due to a complex mix of other factors--falling within-group variation, reductions in the spread of educational attainment, and so forth.

Goldin and Margo's main finding in this "Great Compression" paper is that a large chunk of it took place during World War II--when the National War Labor Board [NWLB] enforced wage controls--and persisted after the war, long after the NWLB and its controls had vanished. The timing thus poses a big problem for supply-and-demand-based explanations of the wage structure: how can one understand supply and demand forces that seem to be 'created' by bureaucratic fiat. Goldin and Margo try some strategies--references to union effects and the G.I. bill's impact on labor supply. How successful are they?

Goldin and Katz's look at the returns to education at the start of the century finds that even in rural Iowa the returns to formal education--in the sense of higher wages paid to the highly-educated--were remarkably high: 12% for an extra year of education. These high returns existed even in places, like within agriculture, where one would not have expected to necessarily find large benefits from formal education. Is formal education much more important even in non-white-collar industries than our prejudices allow? Is formal education a sorting and screening device, so that cross-sectional estimates of schooling premia are very bad estimates of the importance of education for economic growth? The high returns to education at the start of this century appear to have triggered an enormous social investment in education in America, a social investment not matched elsewhere.

This investment in education seems to have, over the century, greatly lowered the returns to formal schooling, which Goldin and Katz at times call "skill." By the late 1970s economists like Richard Freeman were marvelling at the near-disappearance of the college-high school wage premium, and writing about the "overeducated American." But then--almost the moment Freeman published his book, in fact--the formal education premium (and other sources of wage inequality among white males) took off like a rocket. Why? It's not as if the baby-boomers and their children stopped going to school.

The natural place for a neoclassical labor economist to look for a sudden shift in inequality not driven by changes on the supply side is on the demand-for-labor side, in "capital-skill complementarity": a large positive cross-derivative in the production function for physical capital and the skills built by formal education. Goldin and Katz's "Origins" paper looks at the micro level for such a cross-derivative, and finds it in electricity-using and in capital-intensive assembly-line-using industries back before World War II.

The last paper, Richard Freeman's "Spurts in Union Growth," looks at another very important determinant of the wage structure, but one that cannot be easily fit into neoclassical economic models. Freeman provides a model of both the upward pressure on unionization rates in the late 1930s once an initial wave of Depression-era unionization had boosted unionization above some critical density, and his model can be turned around to account for the large and aggressive efforts of employers to reduce union density in the 1980s and 1990s once the initial wave has started to ebb.

In the end, how we understand the wage distribution (for white males, at least), depends on whether we find ourselves "neoclassicals" or "institutionalists" in our analyses of the labor market. Do we focus on production function cross-derivatives and on relative supplies of different kinds of workers? Or do we focus on the NWLB, on the high school, and the union?