Student Abilities during the Expansion of U.S. Education, 1950-2000

Abstract

Since 1950, U.S. educational attainment has increased substantially. While the median student in 1950 dropped out of high school, the median student today attends some college. In an environment with ability heterogeneity and positive sorting between ability and school tenure, the expansion of education implies a decrease in the average ability of students conditional on school attainment. Using a calibrated model of school choice under ability heterogeneity, we investigate the quantitive impact of rising attainment on ability and measured wages. Our findings suggest that the decline in average ability depressed wages conditional on schooling by 31-58 percentage points. We also find that the entire rise in the college wage premium since 1950 can be attributed to the rising mean ability of college graduates relative to high school graduates. JEL: I2, J24.

Accounting for Changing Returns to Experience

“…Cohort effects (q s,c ) that are a function of average years of schooling, similar in spirit to Hendricks and Schoellman (2014).…”

Decomposing the Education Wage Gap: Everything but the Kitchen Sink

“…Other supply (percent of CZ 0.0572*** 0.4465*** -0.0409*** -0.4448*** 0.0179 population of worker's skill group, (0 of college and high school graduates. Hendricks and Schoellman (2011) present evidence that a fair amount of the growth in the college wage premium can be attributed to the growth in the relative ability (or "quality") of college graduates compared with high school graduates. Such a change in quality is unmeasured and will, thus, be captured only by the constant term.…”

The Return to College: Selection and Dropout Risk

“…Without preference shocks, school sorting would be perfect in the sense that all agents of a given type j would make the same college entry decision. This would cause us to overstate the role of ability selection (see Hendricks and Schoellman, ). The preference shocks affecting the college dropout decision mainly improve the model's ability to account for the timing of dropout decisions and for the dropout rates of high‐ability students.…”

“…Hendricks and Schoellman () bound the noise in test scores using the correlation of multiple tests taken by the same individuals. Their approach only yields a lower bound for test score noise.…”