Unemployment insurance generosity and crime

ABSTRACT

Previous studies show that unemployment has a positive effect on crime rates; however, unemployment insurance (UI) benefits could mitigate these effects. Using county-level unemployment and crime data and taking advantage of the UI law changes across states and over time (1990–2016), we provide new evidence that the UI benefits act as a buffer against local labour market shocks. A one-standard-deviation increase in benefits is associated with 2.4% and 1.9% lower property and violent crime rates for a county at average unemployment.

KEYWORDS:

• Crime
• unemployment insurance
• unemployment
• positive externality
• public economics

JEL CLASSIFICATION:

• D62
• J65
• K40

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ FBI’s Uniform Crime Reporting programme (UCR) (Kaplan 2018).

² The UI maximum weekly payments and maximum duration of payments are hand-entered from annual reports of Archived Significant Provisions of State UI Laws provided by Bureau of Labour Statistics.

³ Local Area Unemployment Statistics provided by Bureau of Labour Statistics.

⁴ Bureau of Economic Analysis.

⁵ Surveillance, Epidemiology, and End Results (2019).

⁶ Quarterly Census of Employment and Wages.

⁷ March Current Population Survey, from Flood et al. (2018).

⁸ March Current Population Survey, from Flood et al. (2018).

⁹ Government Finance Database (Pierson, Hand, and Thompson 2017).

¹⁰ The results (not reported here) are robust to clustering at state, year, and also two-way clustering on county and state-year levels.

¹¹ The CPS files are extracted from the IPUMS website (Flood et al. 2018). The IPUMS constructs a county variable for individuals. However, this variable is available starting from the year 1995. Another problem is that less than 10% of counties are identified each year. For example, the average annual number of identified counties over the years 1995–2016 is 246 counties.

¹² Extracted from Local Area Unemployment Statistics from Bureau of Labour Statistics.