Working Paper: NBER ID: w21587
Authors: Kerwin Kofi Charles; Erik Hurst; Matthew J. Notowidigdo
Abstract: We study how the recent national housing boom and bust affected college enrollment and attainment during the 2000s. We exploit cross-city variation in local housing booms, and use a variety of data sources and empirical methods, including models that use plausibly exogenous variation in housing demand identified by sharp structural breaks in local housing prices. We show that the housing boom improved labor market opportunities for young men and women, thereby raising their opportunity cost of college-going. According to standard human capital theories, this effect should have reduced college-going overall, but especially for persons at the margin of attendance. We find that the boom substantially lowered college enrollment and attainment for both young men and women, with the effects concentrated at two-year colleges. We find that the positive employment and wage effects of the boom were generally undone during the bust. However, attainment for the particular cohorts of college-going age during the housing boom remain persistently low after the end of the bust, suggesting that reduced educational attainment may be an enduring effect of the housing cycle. We estimate that the housing boom explains roughly 30 percent of the recent slowdown in college attainment.
Keywords: housing booms; college attendance; labor market opportunities
JEL Codes: E24; I21; J24
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Housing boom (R31) | Improved labor market opportunities for young men and women (J49) |
| Improved labor market opportunities for young men and women (J49) | Raised opportunity cost of attending college (D29) |
| Raised opportunity cost of attending college (D29) | Reduction in college enrollment and attainment (I21) |
| Housing boom (R31) | Reduction in college enrollment and attainment (I21) |
| Housing boom (R31) | Decreased likelihood of attending college (D29) |