The Rise and Fall of the College Graduate Wage Premium

Featured in print Digest

Simple supply and demand specifications do a remarkable job explaining the long-run evolution of the college wage premium.

The wage premium for workers in occupations requiring high levels of education was exceptionally high in 2005. But this is not the first time that the gap has been so wide. In 1915, for example, the premium for a college education was also large. In the decades in between, the United States saw the earnings gap between the more educated and the less educated narrow dramatically, up until the early 1950s, and then begin to widen rapidly again after 1980.

What caused these changes? In The Race Between Education and Technology: The Evolution of U.S. Educational Wage Differentials, 1890 to 2005 (NBER Working Paper No. 12984), co-authors Claudia Goldin and Lawrence Katz conclude that "strong secular growth in the relative demand for more educated workers combined with fluctuations in the growth of relative skill supplies go far to explain the long-run evolution of U.S. educational wage differentials."

Using this supply-demand framework, the authors find that from 1915 to 1940, the relative demand for college graduates (those with 16 or more years of schooling) grew at an average rate of 2.16 to 2.41 percent per year. But the supply of college-educated workers grew at an average 3.19 percent annually during the same period. Not surprisingly, the wage premium for college graduates over high school graduates narrowed dramatically during the period.

Starting in 1980, the supply-demand picture flipped, the study shows. The rise in the supply of college-educated workers slowed to 2.00 percent per year while demand increased to somewhere between 3.27 to 3.66 percent per year. That's a major reason behind the rise in the premium back to the levels of 1915. "Overall, simple supply and demand specifications do a remarkable job explaining the long-run evolution of the college wage premium," the authors write, with the exception of two periods.

The first is the 1940s, when the actual premium declined more sharply than predicted. In all likelihood, the authors argue, the lingering effects of wartime wage policies, the bargaining power of industrial unions, strong wartime demand for production workers, and the postwar consumer boom all acted to narrow the premium below its long-term equilibrium level. (In the 1950s, the premium rebounded more strongly than predicted, apparently bringing the wages of skilled workers back into balance.)

In the mid- to late-1970s, the college wage premium narrowed again. Slowing productivity, high inflation, and oil-price shocks complicated the picture. Unions whose members had wages fully indexed to inflation and whose contracts provided real wage growth helped boost the income of non-college-educated workers. These "institutional" factors may have led, again, to the college premium narrowing more than supply-and-demand fundamentals would have warranted. In any case, the college premium rebounded dramatically in the 1980s after a deep recession and employers' tougher line with unions led to concession bargaining in the early part of the decade.

The study finds similar supply-demand forces influencing the wage premium to high school graduation during the first part of the twentieth century. A jump in the number of high school graduates starting around 1910 marked the beginning of the decline in that premium that lasted until the 1940s.

The authors also examine what role immigration has played in the fall and rise of the college wage premium since 1890. Since immigrants have tended to swell the ranks of less-educated workers, especially during the initial and latter parts of that period, their influx has had an impact. But "immigration had a far smaller effect on relative skill supplies in all periods we examine than is generally presumed and thus it had a smaller impact on changes in the premium to education than is often asserted," the authors argue.

After 1980, for example, when there was a slowdown in the growth of the relative supply of college graduates, the decline in growth of the relative supply of native-born graduates accounted for 86 percent of the change. Thus, only 14 percent was due to immigrants. For high school graduates, the immigration effect was far more pronounced. But again, the slowdown in the growth of the supply of native-born high school graduates accounted for 57 percent of the effect, they calculate, while 43 percent of the effect was caused by immigration.

"Technological change is the engine of economic growth." Yet, the authors conclude, it also "creates winners and losers and can sometimes have adverse distributional consequences that may foment social tension". [But] [i]f workers have flexible skills and if the educational infrastructure expands sufficiently, then the supply of skills will increase as demand increases for them. Growth and the premium to skill will be balanced and the race between technology and education will not be won by either side and prosperity will be widely shared."

-- Laurent Belsie