This article analyzes the consequences of integration in public education. I show that the flight from the integrated multicultural public schools to private education increases private educational expenditures and, as a result, decreases fertility among more affluent parents whose children flee. In contrast, among less prosperous parents, integration in public education decreases their children’s human capital levels. I demonstrate that the poor, who cannot opt out, incur greater costs than the rich, who can resort to private education. I also analyze the overall society-wide effect of the integration policy and derive a condition that determines precisely whether this policy increases or decreases the average level of human capital in society
This paper uses representative student data from St. Petersburg, Russia to analyze school segregation by parental socioeconomic status and student academic performance. The proposed systematic segregation indices account for ordinal variables and take expected segregation into account. We decompose segregation by school type, school, and classes and compare the results to results obtained from PISA for urban areas in Russia and six European countries. Segregation by socioeconomic status is moderate in St. Petersburg and Russia, as compared to other European countries. Segregation between schools and school types reflects parental choice, whereas within-school segregation along the lines of student performance reflects school policies.
In this study, we test how the level of relative teacher wages affects educational outcomes. Russia provides a unique setting for testing this relationship given its high regional heterogeneity. We use two measures of educational outcomes at different levels of the school system. Our results show that the level of relative teacher wages has a significant positive effect on both test scores. Institutional reforms in teacher wage setting in Russia further allow us to estimate an instrumental variable model and difference-in-difference model, which confirm the robustness of our main result. We also provide some evidence on the possible channels of this effect.