Search Results
Title: Identifying Sheepskin Effects in the Returns to Education
Resulting in 1 citation.
| 1. |
Flores-Lagunes, Alfonso Light, Audrey L. |
|
Identifying Sheepskin Effects in the Returns to Education Working Paper, Department of Economics, University of Arizona, Tucson AZ, April 2004. Also: http://gsbwww.uchicago.edu/labor/flores2004.pdf Cohort(s): NLSY79 Publisher: Graduate School of Business, University of Chicago Keyword(s): Education; Educational Attainment; Educational Returns; Schooling Permission to reprint the abstract has not been received from the publisher. [Exerpt from Introduction] In this study, we ask how to interpret the range of estimated sheepskin effects that arise from different model specifications. We begin with the observation that sheepskin effects are identified because individuals with a given amount of schooling differ in their degree attainment or, stated differently, because S varies among individuals within a given degree category. It is important to recognize that the variation needed for identification can represent 'signal,' 'noise,' or a combination of both, and that the source of variation determines which parameters are identified and how we should interpret the estimates. To illustrate the identification issues, suppose we observe variation in S among individuals who hold a bachelor's degree. This variation might accurately reflect the underlying behavioral process if, for example, some college students choose full-time (or even part- time) employment at the expense of more rapid progress toward a degree. The variation might also reflect measurement error: data that cross-classify individuals as college graduates with only 12, 13 or 14 years of school should certainly be met with suspicion, and even reports that appear more logical can be error-ridden. Variation in S within degree category is essential if we wish to identify S-D interactions in a wage model, but it is equally important to consider the source of variation. Under the first scenario, we might predict that the sheepskin effect increases with S because S is positively correlated with omitted in-school work experience, however, the estimate should be interpreted as a reward for work experience rather than evidence that employers use degrees to screen for unobserved traits. If the variation arises primarily from measurement error, it might pay to restrict the functional form of our wage model rather than rely on 'noise' for identification. |
|
|
Bibliography Citation
Flores-Lagunes, Alfonso and Audrey L. Light. "Identifying Sheepskin Effects in the Returns to Education." Working Paper, Department of Economics, University of Arizona, Tucson AZ, April 2004. |