Posted Content•
The causal effect of education on earnings
TL;DR: This paper surveys the recent literature on the causal relationship between education and earnings and concludes that the average (or average marginal) return to education is not much below the estimate that emerges from a standard human capital earnings function fit by OLS.
Abstract: This paper surveys the recent literature on the causal relationship between education and earnings. I focus on four areas of work: theoretical and econometric advances in modelling the causal effect of education in the presence of heterogeneous returns to schooling; recent studies that use institutional aspects of the education system to form instrumental variables estimates of the return to schooling; recent studies of the earnings and schooling of twins; and recent attempts to explicitly model sources of heterogeneity in the returns to education. Consistent with earlier surveys of the literature, I conclude that the average (or average marginal) return to education is not much below the estimate that emerges from a standard human capital earnings function fit by OLS. Evidence from the latest studies of identical twins suggests a small upward "ability" bias -- on the order of 10%. A consistent finding among studies using instrumental variables based on institutional changes in the education system is that the estimated returns to schooling are 20-40% above the corresponding OLS estimates. Part of the explanation for this finding may be that marginal returns to schooling for certain subgroups -- particularly relatively disadvantaged groups with low education outcomes -- are higher than the average marginal returns to education in the population as a whole.
Citations
More filters
Posted Content•
TL;DR: In this paper, the authors proposed an alternative and less demanding technique based on a recentered influence function (RIF) regression, which is more likely to yield the parameters of interest and is therefore preferred over current best practice.
Abstract: Socioeconomic related health inequality as measured by a bivariate rank dependent index, of which the concentration index is a leading example, is well documented across a wide number of countries and measures. To decompose an inequality index is to ascertain the potential causes of this measured inequality. Current available regression based decomposition methods applicable to bivariate rank dependent indices impose stringent conditions in order for them to recover the parameters of interest. In this paper we suggest an alternative and less demanding technique based on a recentered influence function (RIF) regression. Because of the less stringent conditions this method imposes, it is more likely to yield the parameters of interest and is therefore preferred over current best practice. The RIF regression approach is also simple to estimate and interpret: the regression yields average marginal effects of covariates on the rank dependent index and interpretation resembles that of standard conditional mean analysis and has strong links to the treatment effects literature. Interpretation of the RIF regression method applied to a bivariate rank dependent index is illustrated by way of an empirical example of income related health utility inequality. (Less)
Posted Content•
TL;DR: In this article, the feasibility of using differentially private (DP) methods to implement a validation server for releasing tabular statistics, means, quantiles, and regression estimates is studied.
Abstract: Federal administrative tax data are invaluable for research, but because of privacy concerns, access to these data is typically limited to select agencies and a few individuals. An alternative to sharing microlevel data are validation servers, which allow individuals to query statistics without accessing the confidential data. This paper studies the feasibility of using differentially private (DP) methods to implement such a server. We provide an extensive study on existing DP methods for releasing tabular statistics, means, quantiles, and regression estimates. We also include new methodological adaptations to existing DP regression algorithms for using new data types and returning standard error estimates. We evaluate the selected methods based on the accuracy of the output for statistical analyses, using real administrative tax data obtained from the Internal Revenue Service Statistics of Income (SOI) Division. Our findings show that a validation server would be feasible for simple statistics but would struggle to produce accurate regression estimates and confidence intervals. We outline challenges and offer recommendations for future work on validation servers. This is the first comprehensive statistical study of DP methodology on a real, complex dataset, that has significant implications for the direction of a growing research field.
25 Sep 2013
TL;DR: In this paper, the authors examined the participation effects of the implementation of the subsidized loan system by the Dutch government and found that parental education is strongly associated with the maximum willingness to pay for higher education.
Abstract: This paper examines the participation effects of the implementation of the subsidized loan system by the Dutch government. Our survey among 144 secondary school students reveals that parental education is strongly associated with maximum willingness to pay for higher education. This paper shows that between 0,7 percent and 3,6 percent of the secondary school students will not follow higher education due to increased costs caused by the subsidized loan system.