Principal investigator: Bruno Benevit
Original title: University Choice: The Role of Expected Earnings, Non-pecuniary Outcomes and Financial Constraints
Authors: Adeline Delavande and Basit Zafar
Location of the Intervention: Paquistão
Sample Size: 2149 students
Sector: Education
Primary Variable of Interest: Choosing an educational institution
Type of Intervention: Perceptions about universities
Methodology: GMM
Summary
Upon completing secondary education, young people face numerous higher education options. This choice is related to the course and the institution, and can be affected by both market expectations and constraints, as well as non-pecuniary aspects. This article investigated which determinants affect students' choice of higher education institution in Pakistan, observing factors such as monetary returns, non-pecuniary returns associated with universities, and financial constraints. Through a school choice model, the estimates obtained using the Generalized Method of Moments (GMM) indicated that non-pecuniary factors, such as the ideology of the educational institution, are the main determinants of these choices. Job market prospects played a significant, but less relevant, role.
- Policy Problem
Choosing a higher education institution is a crucial step in an individual's career, potentially influenced by a variety of factors. Each student faces a unique and highly personalized decision when selecting the higher education institution that best suits their educational needs and goals.
Beyond expectations of future earnings, which are often considered based on salary projections and job opportunities associated with different programs and universities, there are significant subjective factors that play a crucial role (Eisenhauer, Heckman, and Mosso, 2015). Factors such as alignment with the institution's culture and values, geographic location, faculty quality, and the availability of specific academic resources can influence this decision (Delavande and Zafar, 2019). In this sense, each student considers a unique combination of objective and subjective factors to find the institution that best aligns with their academic and personal aspirations.
- Policy Implementation Context
Pakistan's higher education system encompasses a wide range of institutions, similar to the rest of South Asia. The types of institutions vary from public and private universities to madrassas, Islamic religious schools commonly found in the region. Both public and private universities have their own entrance exams, similar to those in the United States and Brazil. Additionally, universities also conduct admission processes based on secondary school exams and/or interviews.
The quality of higher education in the country varies widely, with some institutions internationally renowned, while others face infrastructure and governance challenges. In this sense, private universities stand out, adopting a Western teaching style and offering the best job market returns, but implying high costs for students. Madrassas are at the other extreme, generally offering their services free of charge. Islamic universities, which are usually public and accessible to poorer segments of the population, perform in an intermediate way. Despite the variety of options, the enrollment rate of Pakistani students aged 17 to 23 was 5,1% in 2011 (Delavande and Zafar, 2019).
- Evaluation Details
To gather data regarding students' expectations of educational institutions, this study conducted a questionnaire among male university-age students enrolled in different colleges in two urban centers in Pakistan. The five universities presented in the questionnaire represented the diversity of higher education institutions in Pakistan, divided into: (i) Very Restricted University (UR), (ii) Restricted University (UR), (iii) Islamic University (UI), (iv) Madrasa in city 1 (M1), and Madrasa in city 2 (M2).
This questionnaire was designed to understand students' preferences regarding university choices. Students were presented with a hypothetical scenario of selecting higher education institutions and asked to rank five existing universities in terms of enrollment preference. These choices were made under two distinct conditions: the first considering the student's current financial situation (declared as financially constrained), and the second disregarding the costs of higher education institutions (declared as having no financial constraints). This approach allowed for isolating students' preferences at the time of the survey, free from external influences such as financial constraints or other factors.
The research also collected data on students' beliefs regarding various academic and job market outcomes, providing information on students' perceptions of the returns associated with different educational institutions (in relation to the student themselves and the general population). Additionally, the questionnaire also examined students' perceptions related to parental acceptance, likelihood of dropping out, ideological alignment, quality of teaching, and monthly costs of a given educational institution.
- Method
This study was based on a school choice model developed by the authors, considering parameters of students' utility function, such as consumption, ideological alignment, costs associated with courses, costs of (possible) relocation of the higher education institution, school-specific factors, and the probabilities of graduating and dropping out. Information shocks after entering the course also affect the probability of dropping out. Additionally, the model also considers the probability of obtaining employment in both scenarios of course completion or non-completion and its effects on students' subjective expectations regarding their future income (at age 30).
To estimate the structural parameters that affect Pakistani students' choice of higher education institution, the Generalized Method of Moments (GMM) procedure was adopted. Specifically, the GMM considered the joint estimation of two regressions: (i) a multinomial logit regression to identify the probability of choosing a particular institution and (ii) a fractional logit regression to identify the probability of dropping out of a particular institution. The authors verified the validity by observing the difference in results between subsamples of students in their first year and in subsequent years.
The subjective approval of students' parents by higher education institutions was verified considering the costs, family income, type of institution (Madrassa or not), distance from home, degree of religiosity of the students, and expectations regarding the return in the job market provided by the course.
The authors conducted linear regressions to estimate expectations regarding employability (conditional on being a graduate and/or not being a graduate), pecuniary returns (conditional on being a graduate and/or not being a graduate), level of education, and ideological alignment. The models considered fixed effects of the current educational institution, fixed effects of the educational institutions assessed in the questionnaire, and the student's age.
Finally, the study presented several estimates on the impact of three policy simulations: (i) reduction of financial constraints, (ii) provision of information regarding the returns of each type of higher education institution, and (iii) ideological homogenization of educational institutions.
- Main results
The estimates found indicated a strong variation in students' beliefs about the outcomes considered among various educational institutions, along with a significant diversity in beliefs among individuals within each educational institution. The subjective belief data offered a coherent representation of this situation, insofar as expectations regarding their earnings at age 30 conditioned by the institutions were consistent compared to the patterns observed in the actual data.
Furthermore, the results indicated that students tend to choose higher education institutions based on non-pecuniary outcomes, such as parental approval and graduation ranking, implying a higher valuation for the institutions in which they were enrolled at the time. Similarly, average beliefs related to dropping out tended to be lower for the institution in which students were enrolled at that time.
Regarding the policy simulations, the results revealed that reducing financial constraints through the provision of loans or free education could have benefited approximately 60% of the students in the sample, and that almost 20% would have chosen a different higher education institution. These results suggest that financial constraints play a significant role in university choice in Pakistan, where efficient credit markets are scarce (Delavande and Zafar, 2019). Concerning the other two simulated policies, the results highlighted the relevance of non-financial factors and the diversity of student preferences regarding the specific ideology of each institution: any policy that had led to the standardization of institutions in terms of teaching would have had a relatively insignificant impact on enrollment choices; however, it would have resulted in welfare losses for one-third of the students.
- Lessons in Public Policy
This article investigated the role of financial constraints, expected monetary returns, and non-pecuniary factors associated with higher education institutions in the higher education choice process of students in Pakistan. To this end, the authors explored the context of higher education in the country, which exhibits significant ideological heterogeneity among its conventional universities and madrassas.
The study's evidence demonstrated that financial constraints are important in determining students' choice of university, being strongly influenced by parental approval and subsidies. In this sense, subsidy policies focused on mitigating these constraints have the potential to increase student enrollment rates in expensive universities, raising incomes and substantially increasing the lifetime utility of the students who benefit.
References
DELAVANDE, A.; ZAFAR, B. University Choice: The Role of Expected Earnings, Nonpecuniary Outcomes, and Financial Constraints. Journal of Political Economy, v. 127, no. 5, p. 2343–2393, Oct. 2019.
EISENHAUER, P.; HECKMAN, JJ; MOSSO, S. Estimation of Dynamic Discrete Choice Models by Maximum Likelihood and the Simulated Method of Moments. International Economic Review, v. 56, n. 2, p. 331–357, May 2015.
