Downstream Effects: Understanding the Consequences of Policy Decisions for Marginalized Groups

Abstract

This dissertation provides three chapters focused on intersections between race, class, ability, and geography within the economics and sociology of education.

In Chapter One, I evaluate the impacts of the New York City Department of Education District and Citywide GT programs through regression discontinuity and lottery-based designs. I show that acceptance into a District and Citywide GT program induces non-NYC, wealthier, White, and Black students to enroll in public school compared to ineligible and non-offer-receiving applicants. I also show that GT positively affects early-grade absences for students in poverty, but the study’s methods produce imprecise results for students’ ELA and math achievement.

In Chapter Two, I evaluate Massachusetts Chapter 222, a state legislative reform designed to reduce student discipline incidents and suspensions. I leverage a difference-in-differences and event study design to compare the outcomes between high and low incident rate school-grades before and after the implementation of Chapter 222. Chapter 222 caused significant reductions in student incidents and suspensions, particularly for students at high risk of committing incidents or being suspended (i.e., students with disabilities and Black, Hispanic, and low-income students). Chapter 222 also improved ELA achievement, absences, and dropout rates.

In Chapter Three, we explore the association between 1935-40 Home Owners' Loan Corporation A-D security ratings and present-day public districts’ and schools’ outcomes. We evaluate whether those districts and schools located in historically “D” neighborhoods have worse modern-day outcomes. Leveraging fixed-effects models, we find that districts located today in historically redlined "D" neighborhoods have less district per-pupil total revenues. We find that schools in D neighborhoods have higher per-pupil revenues, larger shares of Black and low-income student bodies, and worse average test scores than those in higher-rated HOLC neighborhoods. Finally, we document a persistence in these patterns across time, with overall positive time trends regardless of HOLC security rating but widening gaps between redlined "D" and "A," "B," and "C" outcomes.