Fostering Educational Enrolment Through Subsidies: The Issue of Timing

Abstract

The purpose of this paper is to build a dynamic structural model of educational choices in which cognitive skills shape decisions. The model is estimated by maximum likelihood using cohort data where individuals are observed from birth to the middle of their working life. These data are unique in that they include cognitive skills test scores collected as early as age 7. We then investigate how alternative policies foster educational enrolment. In particular, we simulate the effect of two subsidies different in the timing of disbursement. The first consists of grants assigned directly to individuals aged between 16 and 18. The second is assigned to the parents earlier on, when the cohort is still in its childhood. The latter subsidy affects cognitive skills accumulation and in turn educational choices. Our results suggest that a grant fosters enrolment at the lowest cost but the parental income subsidy generates more welfare as measured by a class of social welfare functions. Nevertheless, these differences in costs and welfare are small. Overall, the results reinforce the view that government investments in cognitive skill accumulation during childhood are worthwhile. However, the results also indicate that such investments should be well structured to ensure a high return. 742FOSTERING EDUCATIONAL ENROLMENT which are the most relevant to the policy question addressed here. Section 3 describes the UK education system. Section 4 presents the cohort data that are used. Section 5 outlines the economics of the model, while Sections 6 and 7 explain, respectively, its identification and estimation. Section 8 begins by presenting the maximum likelihood estimates and the model fit, and then describes the simulations. Section 9 concludes. RELATED LITERATUREThe recent US literature on educational choices looks at the importance of financial incentives and budget constraints in schooling decisions. Keane and Wolpin (1997) investigate the educational and occupational decisions of a US male cohort born in the 1960s (NLSY79) in a dynamic discrete-choice structural model estimating the impact of a tuition fees subsidy on the college participation decision. In the model individuals are not financially constrained and from age 16 onwards decide year by year whether to stay in education, stay home, or work in a white-or blue-collar occupation. The model allows for ex ante (age 16) heterogeneity in endowment. Keane and Wolpin find that a $2000 college fee subsidy would increase high school graduation by 3.5 percentage points and increase college graduation rates by 8.4 percentage points. However, when analysing the lifetime utility effects, those who would benefit most from the subsidy are the individuals with high endowment for school education and white-collar occupations, who would have gone to college even without the subsidy. Those induced to go to college by the subsidy are individuals with low endowment for education and a comparative advantage in blue-collar occupations, with only a minor increase in their lifetime utility...