Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

We test the effects of two childhood vaccination trails, one for Cholera and one for Measles, in Matlab, Bangladesh on the children's eventual educational outcomes. The Cholera vaccine was given as a double blind randomized controled trial to 92,838 participants in 1974. The vaccination provided significant protection against cholera for the 5-14 year age group over the first 90 days. Measles vaccination for children between 9 months and 5 years of age was introduced in two areas in Matlab in March 82 and extended to four areas in December 1985 and had a substantial effect on the prevalence of measles and childhood mortality. This trial involved about 8000 children equally split between the early and late phase-in areas. While the short run health effects of these vaccinations have been studied, we investigate the long run effects of these health interventions on children's educational outcomes in the form of years of schooling completed, as measured in the 1982, 1996 and 2005 socio economic surveys of the area. With the randomized trial of the cholera vaccine our approach to estimating the effect is very simple; we simply compare the outcomes of the treated and control groups, though we do examine the effect of vaccination at different ages. For the measles vaccine we use an approach similar to 'differences in differences' to exploit the phasing in of the program with the inclusion of household, village and pre-vaccination child characterizes as controls. There is high out migration of young adults from the area, to Dhaka and abroad, which may create a sample selection bias when measuring education outcomes since these out migrants are lost from the sample. We avoid this problem by using educational measurements at young ages, before out migration becomes a serious concern; we use data from 1982 for the 1974 cholera trial and data from 1996 for the 1982-85 measles trial. These early measures are right censored since some children are still in school at the time of the measurement, but this issue can be addressed in the estimation, and we avoid having to model the migration process. We find that, in addition to the avoidance of acute health conditions in the short run, these vaccination have a significant long term effect on education attainment