Giovagnoli, Paula Ines
(2013)
From preschool provision to college performances: empirical evidences from a developing country.
PhD thesis, London School of Economics and Political Science.
Abstract
This thesis comprises of three stand-alone papers. The first paper exploits a natural experiment in Argentina to analyse what happens to maternal labour outcomes when there
is an abrupt change in free public preschool provision. Using household survey data, the instrumental variable estimate shows that mothers work on average 9 hours more per week because her child is attending preschool. On the other hand, mothers for whom the access to public childcare was reduced did not change their labour supply, suggesting a shift in the mode of childcare from formal to informal.
The main goal of the second paper is to measure the effect of observable individual characteristics on the whole conditional distribution of performances. Quantile regression methods are shown to provide a flexible framework to model the interactions between observed and unobserved factors, which are the source of non-homogeneous effects on performance that alter its conditional distribution in subtle ways improperly summarised by mean OLS based methods. Using a database of students at public universities in Argentina, the empirical results strongly suggest the presence of heterogeneous effects, which leaves
room to question whether relevant factors like parental education or secondary school type are stronger or weaker for certain individuals. The third paper examines the role of labour market conditions on youth schooling behaviour using discrete time proportional hazards models. The findings show that, as predicted by human capital theory, labour demand has a significant effect on the hazard rate for dropping out of school. However, the results only hold for young males but not for females. The paper also tests whether each student’s hazard rate for leaving school without completion changes autonomously over time. Using a non-parametric specification, the results indicate that the risk of dropout is increasing over time for both males and females.
Actions (login required)
 |
Record administration - authorised staff only |
Download statistics
Download statistics