How Income Changes During Unemployment: Evidence from Tax Return Data∗

We use a panel of tax returns spanning 1999 to 2011 to provide new evidence on household experiences during unemployment. Unemployment is associated with roughly a 20% reduction in household wage earnings. Unemployment insurance compensates for half of these wage losses. Households also partially compensate by using a variety of income sources. Distributions from retirement accounts increase in the short run. Selfemployment income and disability insurance payments increase over longer periods. More generous UI benefits crowd out wage income and are associated with increased distributions from retirement accounts. This combination of responses is consistent with UI benefits lengthening unemployment spells. ∗We are grateful for feedback from Jim Cilke, Ann Huff Stevens, Dan Silverman, and Mel Stephens; conference participants at the 2014 Allied Social Science Association Annual Meeting, the 2014 International Conference of the Association for Public Policy Analysis and Management, the 2012 National Tax Association and Michigan Tax Invitational; seminar participants at Williams College and the Office of Tax Analysis; and two anonymous referees. The views expressed in this paper are those of the authors and do not necessarily reflect the policy of the US Department of Treasury. †Postal Address: 1500 Pennsylvania Ave. #4217-A, Washington, D.C., 20220. Work Telephone: 202-622-5186. Email: [email protected] ‡Corresponding author. Postal Address: 24 Hopkins Hall Drive. Williamstown, MA 01267. Work Telephone: 413-597-4886. Email: [email protected]. Designing policies to help workers after job loss requires information about the extent of financial hardship they face. Numerous studies have documented an immediate negative impact on an individual’s earnings, with most estimates suggesting a 10% to 25% decline in wage income, as well as persistently lower wage income for several years (Jacobson, LaLonde and Sullivan 1993, Farber 1997, Stevens 1997, von Wachter, Handwerker and Hildreth 2008, Couch and Placzek 2010). Consumption also declines as a result of job loss, but by much less than wage income (Dynarski and Gruber 1997, Stephens 2001). Unemployment insurance (UI) compensation can facilitate consumption smoothing but, to the extent that this smoothing is incomplete, households may turn to other income sources. Understanding which income sources are utilized, and how the reliance on such funds is related to the UI system, has important public policy implications. This paper uses individual tax return data spanning 1999 to 2011 to provide further evidence on how income evolves over the course of unemployment. We provide new estimates of the wage losses associated with unemployment, and explore the extent to which households use a variety of non-wage income sources to smooth consumption through an unemployment spell. Because of the breadth of income sources reported on a tax return, we are able to analyze a variety of income types, including capital gains realizations, self-employment income, and distributions from retirement accounts. Information returns filed with the IRS allow us to examine spousal labor income, considered in a companion paper (Kawano and LaLumia 2014), and disability insurance, which we examine here. Considering these different dimensions of compensatory behavior broadens our understanding of how families respond to the negative income shocks associated with unemployment spells. Our data also provide the opportunity to examine the extent to which responses may have differed during the Great Recession. We construct a panel of income tax returns for households that have evidence of an unemployment spell. We rely on UI compensation being taxable income and thus reported on a tax return to identify unemployment. We estimate fixed effects regressions comparing