Essays on Health Insurance and Labor in Economics

Abstract

This dissertation consists of three chapters. Chapters 1 and 2 explore the relationships between domestic oustourcing and employer health plans. Between 1975 and 2012, the share of low-skill workers in the US that are domestically outsourced increased by approximately 10 percentage points to 14 percent. This development coincided with three trends: rising medical care prices, rising income, and shifting labor demand towards high-skill workers. Using a general equilibrium model, I explore the relationship between these trends. As firms increased the generosity of their employer-provided health plans for high-skill employees, anti-discrimination statutes from the US tax code required that they offer the same health plans to all their employees. This fact raised the relative cost of low-skill employees, increasing low-skill domestic outsourcing. With the calibrated model, I find that the rising price of medical care interacting with the anti-discrimination statutes accounts for 10 percent of the rise in low-skill domestic outsourcing between 1975 and 2012, while rising income and the shift towards high-skill workers had minimal effects. These results suggest that the rising price of medical care is an economically meaningful driver of the increase in low-skill domestic outsourcing in the US. Chapter 1 develops a new novel theory featuring both domestic outsourcing and health plans. Chapter 2 applies these theory to the data. In Chapter 3 I explore the effects of Medicaid on employment status. We measure the effect of Medicaid Expansion on the work arrangements of low-income adults. Before the expansion, many of these individuals only had access to subsidized health insurance through a traditional full-time job. After the expansion, they also had access to Medicaid, potentially allowing them to change their labor supply decisions. Using American Community Survey data, we built a sample of low-income adults who only had access to subsidized health plans through a typical full-time job or the Medicaid Expansion. Adults in the sample are childless, spouseless, non-disabled, and reside in states without confounding state-level policies. To identify the effect of newly-found Medicaid access on this sample, we use a difference-in-difference design from \cite{callaway2021difference} . We find that the expansion had a statistically insignificant effect on the share of our sample in several labor market arrangements: traditional full-time employment, part-time employment, self-employment, unemployment, and not participating in the labor force. These results are robust to including pre-treatment covariates and adjustments to our underlying sample framework. We further decompose treatment effects into short- and long-term effects and find both are statistically insignificant. We conclude that Medicaid Expansion had a negligible impact on both the work arrangements for low-income workers.