Melissa Kearney (Chair), Judith Hellerstein, Nolan Pope, and Katharine Abraham
In this dissertation, I study policies to bridge disparities in broadband access and job search, school discipline, and local economic development in low-income neighborhoods. In Chapter 2, I present evidence on the relationship between broadband pricing and labor market outcomes for low-income individuals. Specifically, I estimate the effects of a Comcast service providing discounted broadband to qualifying low-income families. I use a triple differences strategy exploiting geographic variation in Comcast coverage, individual variation in eligibility, and temporal variation pre- and post-launch. Local program availability increased employment rates and earnings of eligible individuals, driven by greater labor force participation and decreased probability of unemployment. Internet use increased substantially where the program was available. In Chapter 3, I quantify the tradeoffs of school suspension policies on student performance and teacher turnover. I use administrative data from the Los Angeles Unified School District, where suspension rates fell by over 90 percent since 2003. I instrument for school suspension rates by interacting districtwide suspension rate changes with initial school suspension rate levels. The results indicate that falling suspension rates decreased math and English test scores, decreased GPAs, and increased absences. Teacher turnover also increases, particularly for inexperienced teachers. The overall negative impact of reducing suspension rates is driven by small but diffuse behavior spillovers, which are only partially offset by large and concentrated benefits for the small number of students who are no longer suspended. In Chapter 4, I analyze the job impacts of thousands of spatially-targeted local investments to spur economic development in low-income areas, funded by $3-4 billion in annual federal block grants from the Community Development Block Grant. Using a hybrid approach combining synthetic control methods with differences-in-differences, I find that job counts increase by 7% over ten years in census tracts where CDBG investments occurred, at approximately $18,000 per job. Investments disproportionately benefited nearby low-income workers. Job impacts are greatest when investments are made to support local firms. Impacts appear larger in commuting zones that are less affluent, but within-commuting zone impacts are largest in less-disadvantaged neighborhoods. I estimate that each dollar of block grant generates $3.16 of public spending.