Public spending in education and student’s performance in Colombia

Abstract

This paper investigates if higher public spending in education and better teacher qualifications are related to student’s performance, using data from Saber 11, a national standardized test conducted by Instituto Colombiano para la Evaluacion de la Educacion. The estimation exploits differences in both policy variables across regions and employs interactions to study if more investment in public education and higher teacher qualifications can help increase average performance and reduce the impact that socioeconomic factors, such as family income, have on student performance. The analysis proposes a model where student performance in Mathematics and Language are dependent not only on the variables of interest of this paper, but also on economic, social and cultural status, sex and age of students, and school characteristics. The results show that students’ characteristics and their environment, school features and departmental differences in the policy variables explain roughly 20% of the variation in education performance in Colombia, a relatively high percentage when compared to those found by other studies focusing on OECD countries and based on PISA. After controlling for students’ and school characteristics, the results show that in Colombia, public spending per student and teacher qualifications are positively related to better learning outcomes. For the first one, the results suggest that if all regions reach the level of spending per student of Bogota – the region with the highest spending – average math scores can increase by 3.8 to 4.3 points (around 8%), depending on the regions, with the highest improvement for low income students.