Working Paper: NBER ID: w22999
Authors: Rong Hai; James J. Heckman
Abstract: This paper investigates the determinants of inequality in human capital with an emphasis on the role of the credit constraints. We develop and estimate a model in which individuals face uninsured human capital risks and invest in education, acquire work experience, accumulate assets and smooth consumption. Agents can borrow from the private lending market and from government student loan programs. The private market credit limit is explicitly derived by extending the natural borrowing limit of Aiyagari (1994) to incorporate endogenous labor supply, human capital accumulation, psychic costs of working, and age. We quantify the effects of cognitive ability, noncognitive ability, parental education, and parental wealth on educational attainment, wages, and consumption. We conduct counterfactual experiments with respect to tuition subsidies and enhanced student loan limits and evaluate their effects on educational attainment and inequality. We compare the performance of our model with an influential ad hoc model in the literature with education-specific fixed loan limits. We find evidence of substantial life cycle credit constraints that affect human capital accumulation and inequality. The constrained fall into two groups: those who are permanently poor over their lifetimes and a group of well-endowed individuals with rising high levels of acquired skills who are constrained early in their life cycles. Equalizing cognitive and noncognitive ability has dramatic effects on inequality. Equalizing parental backgrounds has much weaker effects. Tuition costs have weak effects on inequality.
Keywords: human capital; credit constraints; natural borrowing limit; education; wealth
JEL Codes: I2; J2
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Credit constraints (E51) | human capital accumulation (J24) |
| Cognitive ability (G53) | educational attainment (I21) |
| Noncognitive ability (D91) | educational attainment (I21) |
| Parental education (I24) | educational attainment (I21) |
| Wealth (D31) | educational attainment (I21) |
| Cognitive ability (G53) | wages (J31) |
| Noncognitive ability (D91) | wages (J31) |
| Parental education (I24) | wages (J31) |
| Wealth (D31) | wages (J31) |
| Cognitive ability (G53) | consumption (E21) |
| Noncognitive ability (D91) | consumption (E21) |
| Parental education (I24) | consumption (E21) |
| Wealth (D31) | consumption (E21) |
| Equalizing cognitive and noncognitive abilities (I24) | reducing inequality (I14) |
| Equalizing parental backgrounds (I24) | reducing inequality (I14) |
| Tuition costs (I22) | inequality (D63) |