I consider a problem of decision under uncertainty. A student wishing to study Social Sciences chooses a degree among Economics, Education, Political Science or Psychology but does not know how well she will do in each and the payoff in the labour market. For example, evidence about students at the University of Torino in North-Western Italy shows that the mean time to degree is 7 years versus 5 de iure and Economics offers the greatest mean earnings. Then, I assess whether information on predicted earnings by degree can inform students’ choice. I divide the econometric strategy in two steps to deal with selection by students into degrees first and then jobs. Firstly, I model degree choice by exploiting a selection on observables strategy and rich covariates to deal with open admission policy by universities. I also deal with non-random variation in achievement and time to degree by exploiting exclusion restrictions that the institutional setting offers. Secondly, I obtain predicted earnings for students in each degree and I allocate them to the degree paying the greatest ones. I find that students with different observed ability at high school would be recommended different sets of degrees although with some overlapping. Completing degrees that pay different predicted earnings also shows variation in predicted time to degree. Moreover, using different rules to allocate students I find that constraints to overall enrolment capacity by degree decrease mean predicted earnings with respect to unconstrained allocations. Informing students with different characteristics that choosing the same degree may lead to different outcomes such as achievement, time to degree and earnings should persuade policy makers to disclose such information rather than mean population values to inform degree choices.
Presented by:
Marcello Sartarelli (IOE, London)
Date & time:
November 29, 2010 4:00 pm
External seminars home