Elsevier

Journal of Public Economics

Volume 154, October 2017, Pages 67-94
Journal of Public Economics

The effect of Medicaid expansion on crime reduction: Evidence from HIFA-waiver expansions

https://doi.org/10.1016/j.jpubeco.2017.09.001 Get rights and content

Highlights

  • We study the crime-reduction effect of Medicaid expansions to adult population between 2001 and 2008.

  • We find that Medicaid expansions led to an economically meaningful reduction in the rates of robbery, aggravated assault and larceny theft.

  • Much of the crime-reduction effect of Medicaid expansions likely occurred through increasing substance use disorder treatment rate and reducing substance use prevalence.

Abstract

Substance use figures prominently in criminal behavior. As such expanding public insurance and improving access to substance use disorder (SUD) treatment can potentially reduce substance use and reduce crime. We examine the crime-reduction effect of Medicaid expansions through the Health Insurance Flexibility and Accountability (HIFA) waivers. We find that HIFA-waiver expansion led to a sizeable reduction in the rates of robbery, aggravated assault and larceny theft. We also show that much of the crime-reduction effect likely occurred through increasing SUD treatment rate and reducing substance use prevalence. The implied benefit-cost ratio estimate of increased treatment on reducing crime ranges from 1.8 to 3.2.

Introduction

Substance use and crime are two of the most intractable social ills facing the United States, and they are inextricably linked. A positive correlation between substance use and crime has been observed in arrestee drug test results and inmate drug reports. Among arrestees who were booked on violent or property crimes, one in every four tested positive for illicit drug use at the time of arrest (ONDCP, 2012). Moreover, among prison inmates charged with violent crimes, 52% reported being under the influence of alcohol or drugs when committing the crime, or committing the crime to acquire money to purchase drugs; among those charged with property crimes, this number is 39% (Miller et al., 2006).

To the extent that this observed correlation involves causality running from substance use to crime, interventions to reduce substance use should also reduce crime. Nonetheless, empirical evidence suggests that punitive approaches to substance control such as prohibition and the “war on drugs” have not led to significant crime reduction (Miron, 1999, Kuziemko and Levitt, 2004, Markowitz, 2005).1

In this paper we explore an area that has garnered relatively little attention in the economic literature on crime reduction, namely public health insurance policy. Using county-level panels of crime data between 2001 and 2008 across the United States, we examine the crime-reduction effect of state Medicaid expansions through Health Insurance Flexibility and Accountability (HIFA) waivers (CMS, 2001). The HIFA initiative provides states with federal matching funds to expand Medicaid to all low-income adults with family incomes up to 200% FPL in states. We also explore the extent that state HIFA-waiver expansions provide plausibly exogenous shocks for local SUD treatment rate, which serves as one of the potential pathways to substance use reduction and eventually leads to crime reduction. Our estimates reveal that state HIFA-waiver expansions are associated with an economically meaningful reduction in the rates of specific types of crimes for which theory suggests an increase in the SUD treatment rate should have an effect (i.e., robbery, aggravated assault and larceny theft). Our estimates also suggest that the effect of the HIFA-waiver expansions on increasing SUD treatment rate and reducing substance use prevalence is likely to be one of the driving forces behind the estimated crime-reduction effect.

This study has implications for both public health insurance policy and public safety policy. It provides previously undocumented evidence of significant reductions in crime rates arising from state Medicaid expansions. This has direct relevance to the current health care reform discussions surrounding insurance expansion and “mainstreaming” of SUD treatment. While the political sea change may lead to repeal of the Affordable Care Act (ACA), the effect of insurance expansion on social outcomes, such as crime reduction, may still be of interest for policy and research. We show that a set of state Medicaid expansions preceding the ACA benefitted people with SUDs by providing a cost-effective public health approach to crime reduction, partially through increasing their treatment use and reducing their substance use.

Previous studies of the economic benefits of SUD treatment have often emphasized the direct health returns on treatment through recovery from addiction and the related productivity gains (Belenko et al., 2005). We instead focus on the public finance aspects of SUD treatment and take a more comprehensive view of the cost of crime to the public sector, including direct, indirect and opportunity costs. Our instrumental variable (IV) estimates demonstrate a benefit-cost ratio of 1.8 to 3.2, that is, a 10 percent relative increase in the SUD treatment rate at an average cost of $1.6 billion yields a crime reduction benefit of $2.9 billion to $5.1 billion. This downstream benefit to public safety represents a sizable fraction of returns on SUD treatment. Specifically, as the U.S. criminal justice system scales back mandatory minimum sentences for low-level drug and other minor offenders who may also be substance users, replacing incarceration with better access to SUD treatment can be a cost-effective investment in public safety.

Section snippets

Theories of substance use, SUD treatment and crime

Contemporary criminological theories suggest that substance use is one of the root causes of crime. The most cited criminological theory on this causal relationship is Goldstein's (2003) tripartite model, in which three hypotheses are provided to explain how substance use causes violent and property crimes. First, the pharmacological hypothesis states that violence may occur as a direct result of the intoxication. Intoxication of certain substances may trigger aggression and lead to violent

Data

Our data consists of a panel of annual, county-level observations between 2001 and 2008. Data sources include the Uniform Crime Reports (UCR), the National Survey of Substance Abuse Treatment Services (N-SSATS), and other nationally representative datasets that provide supplementary information on important local-level socioeconomic and policy contextual measures (Table 1).

Estimating the main effect of HIFA-waiver expansions on crime rates

We used a two-way (i.e., county and year) fixed effects model to estimate the effect of state HIFA-waiver expansions on county crime rates to isolate the within-local variations over time: Crime Rate c , s , t = β 1 + β 2 HIFA s , t + β 3 X 1 c , s , t + β 4 X 2 s , t + ρ c + τ t + ε c , s , t where c denotes county, s denotes state, t denotes year. ρc represents county fixed effects and τt represents year fixed effects. The two-way (i.e., county and year) fixed effects account for the time-invariant county heterogeneity and the national

Estimating the pathway effects of HIFA-waiver expansions on SUD treatment rate and of SUD treatment rate on crime rates

We first used a two-way fixed effects model to estimate the pathway effect of state HIFA-waiver expansions on county SUD treatment rate similar to the one used to estimate crime rates: SUD Treatment Rate c , s , t = β 1 + β 2 HIFA s , t + β 3 X 1 c , s , t + β 4 X 2 s , t + ρ c + τ t + ε c , s , t

As in Eq. (11), we included in Eq. (12) ρc and τt to adjust for the time-invariant county heterogeneity and the national secular trend. We also included the full set of covariate vectors Xc,s,t and Xs,t to account for the time-varying county-level

Estimating the pathway effect of SUD treatment rate on substance use prevalence

We took one step further to explore the pathway effect of SUD treatment rate on substance use prevalence using a TSLS model similar to the one used to estimate the pathway effect of treatment rate on crime rates. The previous Eqs. (13), (14) capture the TSLS Stage I, and the following Eq. (16) captures the TSLS Stage II. Note that the level of aggregation here is substate region denoted by r. Our final sample includes a balanced panel of 347 substate regions over 8 years. Substance Use Prevalence r , s

Discussion

Expanding public insurance coverage holds the potential to increase individual SUD treatment use, reduce substance use, and promote public safety through crime reduction. The main contribution of our study is that we uncovered a heretofore unrecognized relationship between the implementation of Medicaid expansions and the reduction in crime rates: we found that Medicaid expansions under HIFA-waiver reduced the rates of robbery, aggravated assault and larceny theft. Our study also contributes to

References (61)

  • Katherine Baicker et al.

    The effects of Medicaid coverage—learning from the Oregon experiment

    N. Engl. J. Med.

    (2011)
  • Katherine Baicker et al.

    The Oregon experiment—effects of Medicaid on clinical outcomes

    N. Engl. J. Med.

    (2013)
  • R. Richard Banks

    Beyond profiling: race, policing, and the drug war

    Stanford Law Rev.

    (2003)
  • Gary S. Becker

    Crime and punishment: an economic approach

  • Gary S. Becker et al.

    A theory of rational addiction

    J. Polit. Econ.

    (1988)
  • Gary S. Becker et al.

    Rational addiction and the effect of price on consumption

    Am. Econ. Rev.

    (1991)
  • Gary S. Becker et al.

    The market for illegal goods: the case of drugs

    J. Polit. Econ.

    (2006)
  • Steven R. Belenko et al.

    The Economic Benefits of Drug Treatment: A Critical Review of the Evidence for Policy Makers

    (2005)
  • Marianne Bertrand et al.

    How much should we trust differences-in-differences estimates?

    Q. J. Econ.

    (2004)
  • Lawrence D. Bobo et al.

    Unfair by design: the war on drugs, race, and the legitimacy of the criminal justice system

    Soc. Res. Int. Q.

    (2006)
  • Robert Brame et al.

    Selective attrition and the age-crime relationship

    J. Quant. Criminol.

    (2003)
  • Bureau of Labor Statistics (BLS)

    National Compensation Survey: Health Plan Provisions in Private Industry in the United States, 2008

    (2009)
  • Jonathan Caulkins et al.

    What price data tell us about drug markets

    J. Drug Issues

    (1998)
  • Centers for Medicare & Medicaid Services (CMS)

    Health Insurance Flexibility and Accountability (HIFA) Demonstration Initiative

    (2001)
  • Redonna K. Chandler et al.

    Treating drug abuse and addiction in the criminal justice system: improving public health and safety

    JAMA

    (2009)
  • Christopher Chen et al.

    Massachusetts' health care reform and emergency department utilization

    N. Engl. J. Med.

    (2011)
  • Teresa A. Coughlin et al.

    An early look at ten state HIFA Medicaid waivers

    Health Aff.

    (2006)
  • Charles Dackis et al.

    Neurobiology of addiction: treatment and public policy ramifications

    Nat. Neurosci.

    (2005)
  • Dhaval Dave et al.

    Mental health parity legislation, cost-sharing and substance-abuse treatment admissions

    Health Econ.

    (2011)
  • Stephen G. Donald et al.

    Inference with difference-in-differences and other panel data

    Rev. Econ. Stat.

    (2007)
  • Cited by (70)

    • The effect of opioids on crime: Evidence from the introduction of OxyContin

      2023, International Review of Law and Economics
    • Harm reduction or amplification? The adverse impact of a supervised injection room on housing prices

      2023, Regional Science and Urban Economics
      Citation Excerpt :

      Facilities such as MSIRs might even improve the profile of the local area and, thus, the price. For example, substance use disorder treatment centres (SUDTCs) have been shown to reduce crime and traffic fatalities (Bondurant et al., 2018; Wen et al., 2017; Swensen, 2015; Freeborn and McManus, 2010). There are noticeable limitations to using these findings to learn about the effect of MSIRs.

    • Losing insurance and psychiatric hospitalizations

      2023, Journal of Economic Behavior and Organization
      Citation Excerpt :

      Demand theory predicts that insurance, by reducing the out-of-pocket price faced by consumers, should increase the quantity of healthcare demanded (Grossman, 1972). To date, the literature has explored the effect of gaining insurance on psychiatric healthcare use due to available sources of experimental and quasi-experimental variation (Golberstein and Gonzales, 2015; Wen et al., 2015; Baicker et al., 2017; Wen et al., 2017; Maclean et al., 2018a; Meinhofer and Witman, 2018; Grooms and Ortega, 2019; Maclean and Saloner, 2019). While there are exceptions, public insurance expansions appear to increase psychiatric healthcare use and shift financing away from patients and towards insurance.3

    • Health, crime, and the labor market: Theory and policy analysis

      2022, Journal of Economic Dynamics and Control
    • Local access to mental healthcare and crime

      2022, Journal of Urban Economics
      Citation Excerpt :

      Our study adds to a small but growing literature which suggests that supportive, rather than punitive, policies are effective in reducing crime related to behavioral health conditions. For example, Wen et al. (2017) exploit changes in state Medicaid Health Insurance Flexibility and Accountability waivers to non-traditional populations (low-income, childless adults) that cover SUD treatment to study the effect of SUDs on crime. The authors document that improved access to SUD treatment through waivers reduces crime, in particular robbery, aggravated assault, and larceny.

    View all citing articles on Scopus

    We appreciate helpful comments on earlier drafts of this work from Joseph Doyle, Harold Pollack, Chad Meyerhoefer, Sara Markowitz, Alison Cuellar, as well as anonymous journal reviewers, participants at the 2014 ASHEcon Fifth Biennial Conference and the 2013 AcademyHealth's Annual Research Meeting, and seminar participants at Emory University. All errors are our own.

    View full text