Estimating Marginal Returns to Education

Abstract

This paper estimates marginal returns to college for individuals induced to enroll in college by different marginal policy changes. The recent instrumental variables literature seeks to estimate this parameter, but in general it does so only under strong assumptions that are tested and found wanting. We show how to utilize economic theory and local instrumental variables estimators to estimate the effect of marginal policy changes. Our empirical analysis shows that returns are higher for individuals with values of unobservables that make them more likely to attend college. We contrast our estimates with IV estimates of the return to schooling. KeywordsReturns to Schooling; Marginal Return; Average Return; Marginal Treatment Effect Estimating the marginal return to policies is a central task of economic cost-benefit analysis. A comparison between marginal benefits and marginal costs determines the optimal size of a social program. For example, to evaluate the optimality of a policy that promotes expansion in college attendance, analysts need to estimate the return to college for the marginal student, and compare it to the marginal cost of the policy. This is a relatively simple task (a) if the effect of the policy is the same for everyone (conditional on observed variables) or (b) if the effect of the policy varies across individuals given observed variables but agents either do not know their idiosyncratic returns to the policy, or if they know them, they do not act on them. In these cases, individuals do not choose their schooling based on their realized idiosyncratic individual returns, and thus the marginal and average ex post returns to schooling are the same. 1 * edward.vytlacil@yale.edu. . The Web Appendix for this paper is http://jenni.uchicago.edu/estimating_returns_ed/. 1 See James J. Heckman and Edward J. Vytlacil (2007b). NIH Public Access Author ManuscriptAm Econ Rev. Author manuscript; available in PMC 2014 August 08. NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author ManuscriptUnder these conditions, the mean marginal return to college can be estimated using conventional methods applied to the following Mincer equation: (1) where Y is the log wage, S is a dummy variable indicating college attendance, β is the return to schooling (which may vary among persons) and ε is a residual. The standard problem of selection bias (S correlated with ε) may be present, but this problem can be solved by a variety of conventional methods (instrumental variables, regression discontinuity, and selection models).The recent literature shows how to empirically test the conditions that justify conventional methods (James J. Sergio Urzua 2010 andHeckman and. Applying these methods on data from the National Longitudinal Survey of Youth of 1979 (NLSY), we find that returns vary (i.e. β is random) and furthermore agents act as if they possess some knowledge of their idiosyncratic return (i.e., β is correlated with S). Selection on gains complicates the estimation of marginal returns.Under assumption...

Estimating a nonparametric triangular model with binary endogenous regressors

“…Unlike existing work, notably CL09 and Carneiro et al. (), hereafter CHV11, we find that returns on a college education are consistently positive. Moreover, the returns curves we estimate are inconsistent with the shape restrictions imposed in those papers.…”

How does education change the relationship between fertility and age-dependency under environmental constraints? A long-term simulation exercise

“…While there had previously been some ambiguity around the specific effect of education on economic growth (i.e., Krueger and Lindahl 2000;Pritchett 2001), more recent literature which also considers the age-structure of human capital yields clear evidence that in virtually every society the more educated are more productive in economic terms and contribute more to macro-economic growth (Lutz, Cuaresma and Sanderson 2008;Carneiro, Heckman and Vytlacil 2010;Patrinos and Psacharopoulos 2011). This effect is captured here by giving the better educated higher weights when calculating the denominator of the EWDR.…”

Marginal Treatment Effects from a Propensity Score Perspective

“…In practice, however, it is difficult to condition on X nonparametrically, especially when X is highdimensional. Therefore, in most empirical work using LIV, it is assumed that (X, Z) is jointly independent of ( , η, V) (e.g., Carneiro and Lee 2009;Carneiro, Heckman, and Vytlacil 2011;Maestas, Mullen, and Strand 2013). Under this assumption, the MTE is additively separable in x and u:…”

“…To see how the two components combine to shape MPRTE(p), let us revisit the classic example on the economic returns to college. In the labor economics literature, a negative association has often been found between η and U, suggesting a pattern of "positive selection," i.e., individuals who benefit more from college are more motivated than their peers to attend college (e.g., Willis and Rosen 1979;Blundell, Dearden, and Sianesi 2005;Moffitt 2008;Carneiro, Heckman, and Vytlacil 2011;Heckman, Humphries, and Veramendi 2016). In this case, the second component of equation (9) would be a decreasing function of p. On the other hand, the literature has not paid much attention to the first component, concerning whether individuals who by observed characteristics are more likely to attend college also benefit more from college.…”