College Attrition and the Dynamics of Information Revelation

Abstract

This paper investigates the role played by informational frictions in college and the workplace. We estimate a dynamic structural model of schooling and work decisions, where individuals have imperfect information about their schooling ability and labor market productivity. We take into account the heterogeneity in schooling investments by distinguishing between two-and four-year colleges, graduate school, as well as science and non-science majors for four-year colleges. Individuals may also choose whether to work full-time, part-time, or not at all. A key feature of our approach is to account for correlated learning through college grades and wages, whereby individuals may leave or re-enter college as a result of the arrival of new information on their ability and productivity. Our findings indicate that the elimination of informational frictions would increase the college graduation rate by 9 percentage points, and would increase the college wage premium by 32.7 percentage points through increased sorting on ability.

Bumpy Rides: School‐to‐work Transitions in South Africa

“…Although other researchers have developed models to estimate the parameters associated with learning about academic ability (Arcidiacono, ; Stange, ) or labor market skill (James, ; Miller, ), few have incorporated both in settings of imperfect information (to my knowledge, Belzil (), Arcidiacono et al . (), and Altonji et al . () are the only other studies to do so, and the latter did not structurally estimate their model).…”

“…Although future states may be probabilistic in each of the aforementioned studies (agents face risk of grade failure or unemployment, for instance), they all assume the agent is perfectly informed about how current endowments map into future states. This article follows a smaller number of sequential schooling choice and job search models (including Jovanovic, , ; Miller, ; Arcidiacono, ; Belzil, ; James, ; Stange, ; Trachter, ; Arcidiacono et al ., ) that do not assume perfect information, but instead model how agents learn about choice‐relevant parameters as they accumulate new information. As in those models, I posit that agents are imperfectly informed about their academic ability and employment opportunities, formulate a model of how they learn about these characteristics, and estimate the parameters underlying the learning process.…”

The Return to College: Selection and Dropout Risk

“…Other studies of risky college completion that use transcript data include Arcidiacono (, ) and Stange (). In these models, college grades affect either earnings or the utility derived from attending college.…”

“…Other recent models with learning about student abilities include Arcidiacono et al. (), Stange (), Stinebrickner and Stinebrickner (), and Trachter ().…”

Updated Expectations and College Application Portfolios

“…Zafar, 2011;Arcidiacono, Hotz, and Kang, 2012;Stange, 2012;Stinebrickner, 2012 andWiswall and Zafar, 2015;and Arcidiacono et al 2015). This paper provides a direct analogue at another crucial time for human capital investment -when students decide where to apply to college.…”