Abstract: Colombia is a highly unequal country in terms of income and education. The Covid-19 pandemic exacerbated that inequality. Can motivated and studious recent college graduates help reduce this inequality by serving as teachers for short periods of time? The program Teach for Colombia does precisely that: select college graduates from top universities and passionate about teaching to serve as teachers in schools that provide instruction to low-income children. Using data from Teach for Colombia and from the Colombian Institute of Education Evaluation (ICFES) we evaluate the academic impact amongst low-income children of having a Teach for Colombia teacher. Based on a triple difference design (including school, cohort and subject fixed effects as well as all the second order interactions), we find that having a Teach for Colombia teacher in a given subject for at least one school year is associated with scoring 0.017 standard deviations higher on that subject in the high school exit exam. This effect is not statistically significant. Anecdotal evidence suggests that Teach for Colombia teachers usually replace temporary teachers that are not on the formal teaching career. Hence we interpret our main result as suggesting that Teach for Colombia teachers likely perform on average at least as well as those temporary teachers. Using simpler difference-in-differences estimators we find evidence of negative selection at the school-by-cohort level that justifies using the triple difference design. Exploring heterogeneous effects, we find that the program is most effective in schools that are rural, that have less access to internet, or that are located in the Pacific Coast region, where the program has an effect of 0.05 standard deviations that is statistically significant at the 10% level.

About the exhibitor: Federico González es estudiante del PhD en Economía de la Universidad de Harvard. Es economista de la Universidad de los Andes.

Exposure time: 1 hora y 15 minutos