Medicaid Expansion's Spillover Effects on the Criminal Justice System

Evidence regarding the Affordable Care Act’s (ACA) effect on physical, mental, and financial health is extensive. Yet, there is a dearth of evidence on how the ACA’s coverage expansions impact other social sectors, including the criminal justice system. Previous research suggests that connecting justice-involved individuals to health insurance coverage and community-based healthcare and social services may lower rates of re-arrest. Using a unique data source and a cohort difference-in-differences approach, this study seeks to quantify the relationship between Medicaid expansion and rates of re-arrest in three matched county pairs.

Data and Methods

Working with local officials, we obtained booking and release dates from the county jail in six counties. These data spanned a total of 48 months total: 24 months pre-expansion and 24 months post-expansion. Three of these counties were in a state that expanded its Medicaid program, while three of these counties were in a nearby state that did not expand; these county pairs were Hennepin, MN and Dane, WI; Pima, AZ and El Paso, TX; and East Baton Rouge Parish (EBRP), LA and Hinds, MS. Minnesota and Arizona expanded Medicaid on January 1, 2014, while Louisiana expanded its Medicaid program beginning on June 1, 2016.

First, we estimated descriptive statistics for the outcome of interest and relevant covariates in the pre- and post-period across the treatment and comparison groups. Next, we conduct separate difference-in-differences regression analyses for each county pair using a linear probability model. The probability of re-arrest, conditional on having a first arrest, was the outcome of interest. Model covariates included individual-level demographic characteristics, quarter and year fixed effects, the severity of the crime, and whether the arrest was for parole violation. Standard errors were clustered at the individual level.

Results

The pre-period level of the outcome varied within county pairs: it was higher in Hennepin (28.1%) compared to Dane (26.5%) and Pima (25.7%) compared to El Paso (14.4%), and lower in EBRP (14.2%) compared to Hinds (18.2%). Difference-in-differences regression analyses suggested that expansion is associated with a small decrease in the probability of re-arrest after Medicaid expansion in Pima vs. El Paso (-0.5%; 95%CI:-0.9,-0.02; p=0.04) and EBRP vs. Hinds (-0.9%; 95%CI:-1.6,-0.1; p=0.02). However, expansion was associated with a slight increase in the probability of re-arrest in Hennepin compared to Dane (0.5%; 95%CI:-0.2,1.2; p=0.16).

Discussion

We find that changes in the probability of re-arrest in expansion counties relative to non-expansion counties before and after expansion are relatively small. In light of these findings, it is important to remember that Medicaid expansion is not occurring in isolation. Increased coordination across and integration of health and criminal justice systems may mediate the relationship between Medicaid expansion and re-arrest. Using interview data from qualitative work in these six counties, we can and will provide important context regarding potential mediating factors to our quantitative results. Additionally, future work will consider new functional forms of the outcome (number of re-arrests and time-to-first re-arrest), heterogeneous treatment effects by race and gender, and methods to address the challenge posed by the semi-competing risks to re-arrest.