Efficiency and ex post screening in higher education - a dynamic discrete choice approach.
Many students drop out of secondary or higher education without obtaining a degree. A recent policy report (OECD, 2012) illustrates the problems: 16% of today's young adults in OECD countries do not complete secondary education. Up to 62% enter a university-level program, but only 39% are expected to complete it. Among the students who complete a degree, there is a large fraction that incurs substantial delays.
In this thesis, we analyze study decisions and degree completion in secondary and higher education. We investigate whether educational policies can influence study outcomes. We apply our empirical analysis to the region of Flanders. The thesis consists of four chapters. The first three chapters analyze enrollment and study progression in higher education. The last chapter analyzes the causal impact of the organization of secondary schools on study success.
Chapter 1: Socio-economic status and enrollment in higher education: Do costs matter?
In the first chapter, we analyze the impact of socio-economic status on enrollment in higher education. We observe a large gap between students from disadvantaged backgrounds and students from advantaged backgrounds. We distinguish between three channels through which socio-economic status may affect study decisions. First, students from disadvantaged backgrounds may be more sensitive to the costs of education. Second, they may have lower preferences for education. Third, they may have developed less academic ability during previous schooling.
We quantify the importance of the above three channels and estimate a mixed logit discrete choice model to control for unobserved heterogeneity (Brownstone and Train, 1999). We find that preferences and especially (acquired) ability are more important than cost sensitivity in explaining the lower enrollment of disadvantaged students. As in Kelchtermans and Verboven (2010), we simulate the impact of tuition fee increases. Tuition fees have a fairly small impact on enrollment but especially reduce enrollment of disadvantaged students. We find that discriminatory policies, with lower tuition fees for disadvantaged students financed by higher tuition fees for other students can increase total enrollment and decrease the enrollment gap between disadvantaged and advantaged students.
Chapter 2: Enrollment and degree completion in higher education without admission standards
In the second chapter, we analyze enrollment and degree completion in a higher education system without admission standards. Many students drop out without obtaining a degree or obtain their degree after a substantial delay. To minimize unsuccessful participation, some countries restrict access to higher education through admission standards. Other countries select students on an ex post basis: admission standards are low and students are selected based on their performance during higher education.
We analyze how a system with ex post selection affects initial enrollment and final degree completion. We develop a dynamic discrete choice model of college/university and major choice, where students trade off their current costs and benefits from studying against the future expected benefits on the labor market. The outcome of the enrollment decision is uncertain, as in Arcidiacono (2004) and Bordon and Fu (2015). Upon observing past performance, students may decide to continue, reorient to another major, or drop out. We control for unobserved factors that can be correlated over time as in Arcidiacono and Miller (2011). We simulate the effects of alternative, ex ante admission policies in a system without admission standards. In this way, we are able to compare study outcomes under both systems and assess the effectiveness of both systems. We find that a suitably designed ex ante screening system (with moderate admission standards) decreases unsuccessful enrollment and can even increase degree completion in higher education.
Chapter 3: Study progression in higher education: The impact of ex post selection
In the third chapter, we further study ex post selection in higher education. Many students who do not perform well in their first year switch to other programs or drop out from higher education. However, a substantial fraction of these students still continues the program. We evaluate whether introducing academic dismissal policies and subsidies can affect study outcomes. Dismissal policies force students to switch to another program or drop out if their performance is below a certain minimum threshold. This paper is related to a small literature that analyzes how financial incentives can influence study duration, see for example Garibaldi et al. (2012) and Gunnes, Kirkeboen and Ronning (2013).
We develop a dynamic discrete choice model where the outcome of enrollment is uncertain and use the estimates to simulate the impact of academic dismissal or ex post selection policies. Low dismissal standards have limited effects on participation and degree completion. Moderate or strict standard can lead to a lower number of graduates. To increase degree completion, dismissal policies can be combined with additional subsidies. We find that limiting subsidies to students who perform well in higher education is more effective than additional subsidies for all students. Performance based subsidies lead to a lower increase in enrollment but a higher increase in degree completion.
Chapter 4: Tracking and specialization of high schools: Does school choice matter?
We study the impact of the organizational structure of high schools on success in secondary education. One aspect of the organization of schools concerns the tracking of pupils into specific programs. Pupils first start at a comprehensive program and choose for a more specific program during secondary education. While some schools offer programs in more tracks (academic, technical, artistic or vocational), other schools specialize and offer only programs in one specific track.
We analyze the effect of starting at a school that offers only programs preparing for academic higher education on finishing high school on time. We take the potential self-selection into account by using distance as an instrument for school choice. The conventional 2SLS method assumes a homogeneous treatment effect. This homogenous treatment effect implies that the program has the same effect for all students. We estimate marginal treatment effects and thereby allow for a heterogeneous treatment effect, following Heckman and Vytlacyl (2005). We find that pupils who started at these specialized schools are less likely to graduate from high school without study delay. A possible explanation for this finding is that when schools offer programs in more domains, pupils are more likely to choose for a program that best matches with his or her interests and ability because more study options are available within the same school. We also find that disadvantaged students suffer more from these specialized schools.
Aakvik, A., Heckman, J. and Vytlacil, E. (2005), Estimating treatment effects for discrete outcomes when responses to treatment vary: an application to Norwegian vocational rehabilitation programs, Journal of Econometrics 125 (1-2), 15-51
Arcidiacono, P. (2004), Ability sorting and the returns to college major, Journal of Econometrics 121, 343-375
Arcidiacono, P. and Miller, R. (2011), Conditional choice probability estimation of dynamic discrete choice models with unobserved heterogeneity, Econometrica 79 (6), 1823-1867
Bordon, P. and Fu, C. (2015), College-major choice to college-then-major choice, The Review of Economic Studies, forthcoming
Brownstone, D. and Train, K. (1999), Forecating new product penetration with flexible substitution patterns, Journal of Econometrics 89, 109-129
Garibaldi, P., Giavazzi, F., Ichino, A. and Rettore, E. (2012), College cost and time to complete a degree: evidence from tuition discontinuities, Review of Economics and Statistics 94 (3), 699-711
Gunnes, T., Kirkeboen, L. and Ronning M. (2013), Financial incentives and study duration in higher education, Labour Economics 25, 1-11
Heckman, J. and Vytlacil, E. (2005), Structural equations, treatment effects, and econometric policy evaluation, Econometrica, 73 (3), 669--738
Kelchtermans, S. and Verboven, F. (2010), Participation and study decisions in a public system in higher education, Journal of Applied Econometrics 25, 355-391
OECD (2012), Education at a Glance: Highlights, OECD Publishing