The Great Recession, accompanied by soaring unemployment rates, has brought unemployment insurance back to the centre of public attention. The debate rages on whether generous unemployment insurance programmes provide remedy or exacerbate the problem. Proponents of such programmes point out the value of insurance against income loss from being laid off and long-term unemployment. Opponents cite a robust empirical finding in labour economics, namely that unemployment insurance reduces the incentives for job search and increases the duration of joblessness.
A key factor missing from the current debate has been the effect of unemployment insurance on job match quality. According to the theory of job search, unemployment insurance prolongs jobless durations for two reasons:
(i) Agents spend less time and effort looking for a job; and
(ii) Agents become more selective in the type of job they seek.
The latter channel implies that access to unemployment benefits may help job-seekers find better jobs. In the past, however, unemployment insurance proponents were cautious about making the case for a positive effect on job quality, since “the evidence here is very thin” (Layard et al. 2005).1 In fact, until very recently, most of the available estimates of the unemployment benefit effect on job quality – mainly measured by wages – were not significantly different from zero (Card et al. 2007, Lalive 2007, Van Ours and Vodopivec 2008).
Evidence for positive effects of unemployment insurance on job quality
In recent work, we demonstrate empirically that access to more generous unemployment insurance does indeed lead to agents finding better jobs (Nekoei and Weber 2015). Moreover, we explain why previous studies had a hard time identifying this effect. In particular, we exploit a quasi-experimental design to study the effects of an extension of the potential unemployment benefit duration on job search outcomes, using 19 million job separations recorded in Austrian administrative data.
Consistent with prior research, we estimate that an increase in potential benefit duration causes workers to stay jobless longer. But in contrast to previous studies, we find that the benefit extension also causes workers to find jobs that, on average, pay 0.5% higher wages. Moreover, the positive wage effect persists over time and does not substitute any other desirable job characteristics.
Unemployment benefits help workers avoid large wage drops
Investigating the mechanisms driving the positive effect of unemployment insurance on wages, we find that it results from an attenuation of wage declines between pre- and post-unemployment jobs. In particular, access to more generous benefits reduces the likelihood of experiencing a wage loss that is larger than 40%, and increases the likelihood of achieving a wage increase between 0 and 10%.
Unemployment benefits help workers join better firms
Exploiting the matched employer-employee component of the Austrian data, we also investigate if unemployed workers benefitting from unemployment insurance find jobs in higher-paying firms, or if they find better-paid jobs in an average firm. Our results show that unemployment insurance helps job-seekers move toward ’better’ firms – they find jobs in larger firms with higher proportions of male and older employees, which typically proxy for a higher bargaining power of workers. Most importantly, we document that these unemployed are moving to firms that, on average, pay their other employees higher wages.
Consider two popular explanations for unemployment. According to the first explanation, unemployment is due to randomness in job search.2 The second one argues that randomness is not an issue, but the mismatch between workers’ skills and firms’ demand for skills creates unemployment. In this world, unemployed workers have to wait for a suitable vacancy to open or acquire necessary skills for available vacancies.
Unemployment insurance subsidises job search in either scenario, because it gives unemployed individuals more time to sample job offers or prepare for the right vacancy. Following this argument, unemployment insurance should raise jobless durations and, at the same time, increase job quality. Additionally, it also subsidises leisure so that agents may delay starting an accepted offer, which would lead to agents lowering their effort with no effect on job quality. Folk wisdom can thus explain positive or zero effects of unemployment insurance on job quality.
Folk wisdom is wrong
Contrary to folk wisdom, the effect of unemployment insurance on subsequent job match quality is not necessarily positive, as it is determined by two offsetting forces. On the one hand, unemployment insurance increases agents’ selectivity which, in turn, has a positive effect on job quality. On the other hand, unemployment insurance raises jobless durations and may thus reduce job quality, as job opportunities decline over the jobless spell. This decline in job opportunity over time can be caused by multiple factors, such as loss of human capital, stigma, screening by employers, or diminished savings.3
If we assume that job seekers are rational and forward-looking in their decisions, how can a more generous unemployment benefit lead to a lower subsequent job quality? This is theoretically possible if agents care about well-being or consumption, rather than just earnings when they are employed. It can be shown that unemployment insurance creates a wedge between the two and can thus reduce wages, albeit that it always increases consumption.
A reconciliation of empirical findings
A large body of existing empirical work has not found any effect of unemployment insurance on job quality. For instance, three prominent papers that use quasi-experimental designs and administrative data provide estimates of unemployment insurance effects on wages that are not significantly different from zero (Card et al. 2007, Lalive 2007, Van Ours and Vodopivec 2008). Moreover, there is recent evidence of a statistically significant negative wage effect for Germany (Schmieder et al. 2014), while the evidence discussed above finds a positive effect for Austria.
These different empirical findings do not contradict theory once we take into account duration dependence, meaning that the job-seeker’s opportunities and skills deteriorate the longer she remains out of a job, while at the same time unemployment benefits decrease. In Nekoei and Weber (2015), we show that a job search model incorporating duration dependence can reconcile the contrasting empirical estimates of the effect of unemployment insurance on job quality.
Moreover, the model highlights a potential source of heterogeneity that drives empirical estimates, namely the relative importance of search effort versus selectiveness in determining the job-finding rate. This heterogeneity is reflected in a negative correlation between the effect of unemployment insurance on jobless durations and its effect on job quality. We show that the negative correlation holds in a meta-analysis of existing estimates – studies that estimate stronger effects of unemployment benefits on jobless durations also tend to find smaller effects of benefits on job quality, and vice versa.
As the unemployment insurance system is designed to balance the value of insurance against job loss with the cost of extra taxes, an important policy question is: What is the optimal level of generosity of unemployment insurance? The conventional answer to this question has focused on the effect of unemployment benefits on unemployment durations. As insurance reduces the incentives of jobless agents to find a job, it raises aggregate benefit payments and creates higher taxes for the rest of the population. In the conventional model, this negative fiscal externality is balanced against the insurance or consumption-smoothing value of unemployment insurance to determine the optimal benefit generosity.
However, if unemployment insurance also affects job quality, it might change future tax revenues. The total fiscal externality of the programme should thus be calculated as the sum of the traditional negative externality from increased unemployment durations and the externality from job quality, the sign of which depends on the sign of the effect of unemployment insurance on job quality (and is theoretically undetermined).4
In the Austrian case, the externality from job quality is positive and has the same order of magnitude as traditional duration externality, but with the opposite sign. Based on our theoretical insights and this empirical estimate, we conclude that the optimal generosity of unemployment insurance varies depending on the relative importance of the effort versus selectivity margins in job search. These results suggest that taking gains in job quality into account could significantly change the policy recommendations regarding the optimal generosity of unemployment insurance.
Card, D, R Chetty and A Weber (2007) “Cash-on-hand and competing models of intertemporal behavior: New evidence from the labor market”, The Quarterly Journal of Economics 122(4): 1511–1560.
Chetty, R (2010) “Should unemployment benefits be extended? An economic framework and empirical evidence”, Presentation to the Economic Policy Institute, May 26.
Lalive, R (2007) “Unemployment benefits, unemployment duration, and postunemployment jobs: A regression discontinuity approach”, The American Economic Review, 97(2): 108–112.
Layard P R G, S J Nickell and R Jackman (2005) Unemployment: Macroeconomic performance and the labour market, Oxford University Press.
NYTimes (2013) “Benefits ending for one million unemployed”, 27 December.
Nekoei, A and A Weber (2015) “Does extending unemployment benefits improve job quality?”, CEPR Discussion Paper 10568.
Schmieder, J F, T von Wachter, and S Bender (October 2014) “The causal effect of unemployment duration on wages: Evidence from unemployment insurance extensions”, Technical report.
CEA (2013) “The economic benefits of extending unemployment insurance”, Council of Economic Advisors, US Department of Labor, Washington DC, December.
Van Ours, J C and M Vodopivec (2008) “Does reducing unemployment insurance generosity reduce job match quality?”, Journal of Public Economics, 92(3): 684–695.
1 For examples of such policy recommendations, see CEA (2013) and Chetty (2010). For anecdotal evidence of such a channel, see the New York Times (2013).
2 Randomness either stems from sampling vacancies due to the lack of information about them (random search models), or from the lack of coordination among applicants (directed search models).
3 Many papers offer theoretical reasons for this duration dependence and several papers offer supportive empirical evidence; see Nekoei and Weber (2015) for references.
4 Following the literature on optimal unemployment benefits design, we neglect potential general-equilibrium effects of benefits on non-recipient workers. We provide supportive empirical evidence for this assumption in our setting (see Nekoei and Weber 2015).