Elsevier

European Economic Review

Volume 62, August 2013, Pages 1-16
European Economic Review

Does unemployment insurance crowd out home production?

https://doi.org/10.1016/j.euroecorev.2013.04.006Get rights and content

Highlights

  • We study the interaction between self-insurance and public insurance.

  • In particular, we focus on home production (HP) and unemployment insurance (UI).

  • Doubling UI benefits is associated with a 22% decline in HP in the U.S. data.

  • A heterogeneous-agent model matches the empirical facts regarding HP and UI.

  • We interpret the results as a substitution between UI and HP.

Abstract

In this paper, we study the interaction between self insurance and public insurance. In particular, we provide evidence on a negative correlation between unemployment insurance benefits and home production using the American Time Use Survey (ATUS) and the state-level unemployment insurance data of the U.S. The empirical results suggest that moving to a two times more generous state would decrease time spent on home production about 22% for the unemployed. Then, we pursue a quantitative assessment of this empirical finding using a dynamic competitive equilibrium model in which households do home production as well as market production. The model is able to generate the empirical facts regarding the unemployment benefits and home production. The fact that unemployment insurance benefits crowd out home production is interpreted as a substitution between the two insurance mechanisms against loss of earnings during unemployment spells.

Introduction

This paper investigates whether households substitute public insurance (unemployment benefits) for self insurance (home production) against loss of earnings during unemployment spells. We do so in two steps. First, we provide empirical evidence on the relationship between the level of unemployment insurance and home production using the American Time Use Survey (ATUS) and the state-level unemployment insurance data. Second, we pursue a quantitative assessment of the documented empirical facts using a dynamic heterogenous-agents model of unemployment with incomplete asset markets.

Although there is a vast literature on the effects of unemployment insurance policies on market production, surprisingly there is a lack of theory and evidence on the effects of unemployment benefits on non-market production (in particular, home production).1 Recent studies provided evidence on the use of home production as a self insurance mechanism against lost/reduced earnings.2 Since unemployment benefit programs provide another channel of insurance against lost earnings, we would like to investigate whether people tend to substitute these two insurance mechanisms.

We use the American Time Use Survey (ATUS) and state-level unemployment insurance data to provide empirical evidence on the relationship between home production and the level of unemployment insurance. The empirical results suggest a negative correlation between the former and the latter. Quantitatively, moving to a two times more generous state would – on an average – decrease time spent in home production about 22% for the unemployed households. Moreover, we find that the results are sensitive to family composition. In particular, the negative relationship between the level of unemployment benefits and home production is stronger for single households compared to the married ones; and stronger for those who are married with a non-employed spouse compared to those married with an employed spouse. We interpret this result as a higher degree of substitutability between self insurance and public insurance (home production and unemployment insurance) for the households who have less additional insurance opportunities.

In order to pursue a quantitative assessment on the substitution between home production and unemployment benefits, we present a dynamic model featuring a heterogeneous-agents framework, where households receive idiosyncratic (un)employment shocks. The asset markets are incomplete so that households can partially insure themselves through a non-interest bearing asset. An additional channel of partial insurance is home production through which households can increase their home production to insure against lost earnings during unemployment spells. The model implies a reduction in average hours of home production in response to an increase in the level of the unemployment benefit. In particular, the average fraction of time spent in home production for unemployed is decreasing from 0.19 to 0.15 as we increase replacement rate from 0.20 to 0.90 gradually. We also replicate the benchmark empirical regression with the model data and find a coefficient of −0.25, that is doubling benefits decrease time spent in home production by 25%. This result is consistent with the empirical one and in line with the interpretation of substitutability between self-insurance (home production) and public insurance (unemployment benefits) against loss of earnings during unemployment spells. The negative correlation between unemployment benefits and time spent in home production is stronger when agents’ willingness to substitute home goods with market goods is greater and time is more intensely used in home production process. Similarly, it is weaker when risk aversion and time discount factor is greater. The fact that the time spent for home production and unemployment benefits are negatively correlated is sensitive to the elasticity of substitution between market goods and home goods. However, it is robust to various values of other parameters. The model also implies a negative relationship between the level of unemployment benefits and mean savings of households for a reasonable range of unemployment benefits which is consistent with the empirical counterpart documented in Engen and Gruber (2001).

Substitutability between various insurance channels has been studied in several papers. For instance, Engen and Gruber (2001) consider precautionary savings as a self insurance mechanism and examine the relationship between unemployment benefits and precautionary savings. They find that households increase their precautionary savings in response to a decrease in unemployment benefits. They interpret this as a substitution between self insurance and public insurance. Cullen and Gruber (2000) consider spousal labor supply as a self insurance mechanism against loss of earnings during unemployment spells. They find that unemployment insurance crowds out this kind of family insurance mechanism. Cutler and Gruber, 1996a, Cutler and Gruber, 1996b provide empirical evidence on the fact that households substitute public health insurance for the private one using the policy changes in the U.S. in 1980's and 1990's. Chetty and Saez (2010) emphasize the role of informal self insurance mechanisms such as loans from relatives and spousal labor supply – which does not generate moral hazard problem – in determining the optimal level of social insurance. Among others, Golosov and Tsyvinski (2007), Attanasio and Rios-Rull (2000), and Ortigueira and Siassi (2011) study the interaction between self insurance and public insurance and determine the optimal level of the latter under availability of various sources of self insurance. Taskin (2011) studies the optimal rate of unemployment insurance in an economy where agents do home production as well as market production and finds that the optimal rate of unemployment insurance is significantly smaller in the economy where agents are able to do home production. A number of papers including Moffitt and Nicholson (1982), Meyer (1990), Card and Levine (2000), and Nakajima (2011) study the effect of unemployment insurance policies on labor supply using the U.S. data. We contribute to the literature on the interaction between self and public insurance by studying home production and unemployment benefits in a dynamic competitive equilibrium framework, and providing empirical evidence on the interaction between these two insurance mechanisms.

The rest of the paper is organized as follows. We provide empirical evidence in Section 2. In Section 3, we present the dynamic model. In Section 4 we discuss both the theoretical and quantitative results of the model together with some robustness checks. We finally conclude in Section 5.

Section snippets

Empirical evidence

In this section we provide empirical evidence on the correlation between the level of unemployment insurance and home production.

Model

In this section, we develop a quantitative dynamic model to study the relationship between unemployment benefits and home production. We do so by extending the model of Hansen and İmrohoroğlu (1992) with a home production technology (Greenwood and Hercowitz, 1991). In general, the model features a partial-equilibrium heterogeneous agents framework with incomplete asset markets. The details are explained in the following subsections.

Results

Before proceeding to the quantitative results we present several theoretical properties of the model.

Conclusion

We study the interaction between public insurance and self insurance in this paper. In particular we investigate the relationship between unemployment insurance benefits and home production. Previous literature provided both theoretical insights and empirical evidence on the effects of unemployment insurance policies on market production. However, to our best knowledge, the effects of unemployment benefits on non-market production (in particular, home production) have not been studied. This

References (25)

  • Michael C. Burda et al.

    Unemployment, market work and household production

    Economics Letters

    (2010)
  • Alan B. Krueger et al.

    Job search and unemployment insurancenew evidence from time use data

    Journal of Public Economics

    (2010)
  • Mark Aguiar et al.

    Life-cycle prices and production

    American Economic Review

    (2007)
  • Aguiar, Mark and Erik Hurst (2005), ``Consumption versus Expenditure," Journal of Political Economy, 113(5):...
  • Gary D. Becker

    A theory of the allocation of time

    The Economic Journal

    (1965)
  • Benhabib, Jess, Richard Rogerson, Randall Wright (1991), ``Homework in macroeconomics: household production and...
  • Card, D., Levine, P.B., 2000. Extended benefits and the duration of UI spells: evidence from the New Jersey Extended...
  • Canova, Fabio and Angel Ubide (1998), "International business cycles, financial markets, and household production,"...
  • Clark, Kim B. and Lawrence H. Summers (1982), ``Unemployment insurance and labor market transitions," in Workers, Jobs,...
  • Raj Chetty et al.

    Optimal taxation and social insurance with endogenous private insurance

    American Economic JournalEconomic Policy

    (2010)
  • Julie Cullen et al.

    Does unemployment insurance crowd out spousal labor supply?

    Journal of Labor Economics

    (2000)
  • David M. Cutler et al.

    The effect of medicaid expansions on public insurance private insurance and redistribution

    American Economic Review

    (1996)
  • Cited by (6)

    • Income changes and consumption: Evidence from the 2013 federal government shutdown

      2017, Review of Economic Dynamics
      Citation Excerpt :

      In contrast, we find that changes in beliefs about permanent income likely played little role in driving the financial behavior we observe. Other work by Guler and Taskin (2013) suggests similar effects hold for unemployment spells. Burda and Hamermesh (2010), demonstrate cyclical variation in unemployment leads to variation in home production but little impact on long-term unemployment.

    • Families and social security

      2017, European Economic Review
      Citation Excerpt :

      They are also in line with other papers that highlight the insurance role of families such as Brown and Poterba (2000) or Ortigueira and Siassi (2013). Our paper also sheds more light on the role of home production, an issue that has just recently deserved increasing attention, see Guler and Taskin (2013), Achou (2015) or Dotsey et al. (2015). However, our simulation approach indicates a novel way to isolate the importance of all these effects for different household types.

    • Home production and Social Security reform

      2015, European Economic Review
      Citation Excerpt :

      For example, Rogerson and Wallenius (2009) demonstrate that home production alone is qualitatively capable of generating realistic retirement patterns. Guler and Taskin (2013) find a negative correlation between unemployment insurance and home production indicating a substitution between the two insurance mechanisms against loss of earnings during unemployment spells. Dotsey et al. (2014) document the close relationship between time use over the life cycle with the consumption patterns of the related goods and show that a life-cycle home production model can help explain this behavior.

    We are grateful to Mark Aguiar and Yavuz Arslan for their insightful feedback. We would also like to thank the editor, two anonymous reviewers, Arpad Abraham, Mark Bils, Gregorio Caetano, Gaetano Gaballo, Matthew Hoelle, Erik Hurst, Tarιk Kara, Masanori Kashiwagi, Phillip Lederer, Curtis Signorino and the session participants at the European University Institute, Midwest Macroeconomics Meeting at the University of Notre Dame, Royal Economic Society Annual Meeting in London, and Stockman Conference in Rochester for helpful discussions. Temel Taskin would like to thank European University Institute for financial support. The views expressed here are those of the authors and do not reflect the views of the Central Bank of the Republic of Turkey.

    View full text