How does k-12 school economic segregation impact lower-income students’ college enrollment and employment prospects? Using data from Texas, I find that the typical (median) lower-income student goes through grades 5 to 12 having 1 in 20 upper-income classmates compared to 1 in 2 for the typical upper-income student. The low exposure to upper-income classmates is mainly driven by residential and school socioeconomic segregation, classroom assignment contributes little. I use within-school residual variation from school trends in the proportion of upper-income students to capture the impact of having a higher share of upper-income peers. I distinguish between two possible explanations for the relationship between exposure to upper-income classmates and long-term outcomes: spillover from peers with higher family income and spillover from higher-achieving peers. The slope of the relationship between the proportion of upper-income peers and college enrollment and wages implies that, in the absence of income segregation, lower-income students' 4-year college enrollment would increase by 0.44 percentage points and their young adult quarterly wages would increase by 1.9%, on average. The impact of exposure to upper-income peers appears to be independent of peer academic achievement.

Recent work has highlighted the link between cross-income friendships and economic mobility. Schools can facilitate cross-group friendship formation. Although prior work has examined the impact of school choice policies on segregation, less is known about the impact of school factors such as school course offering on cross-income exposure. I use data from Texas to capture the impact of the addition of advanced courses on lower-income students' exposure to upper-income students. To capture the impact of advanced courses, I exploit variation in when an advanced course is first added to a school subject area. I find evidence that adding an Advanced Placement (AP) course in a subject area increases lower-income students' share of upper-income classmates by 1.4 percentage points. This effect is driven by three subject areas: science, foreign language, and fine arts. The increase in the proportion of upper-income students in the subject area counters any increase in sorting by income after the addition of an AP course.

Although US public schools have made substantive progress in reducing the number of students who dropout of high school, there remains a substantial portion of students nationwide every year who do not obtain a high school degree. One strategy to address persistent high school dropout is to provide students with choice and flexibility in their schooling schedules and environments, often done through alternative schools. The expansion of alternative schools presents an interesting case of increased school heterogeneity, increasing the number of schools targeting lower performing students. In this paper, we use the variation in exposure to alternative schools to identify the impact of exposure to alternative schools on persistence, graduation, college enrollment, and arrests.