How to tackle long-term unemployment

As 2014 draws to an end, the US labour market finds itself in an unusual situation. The unemployment rate has fallen from its peak of 10% in October 2009 to under 6%, but an unusually large share of the unemployed has been out of work for a long time. In October 2014, the share of long-term unemployed (unemployed for more than 26 weeks) among the unemployed remains above 30%, down from its peak of about 50% in 2010, but still well above 17% observed in 2007 prior to the Great Recession. Western European economies have either experienced similar run-ups in long-term unemployment rates in recent years (the UK) or have experienced high long-term unemployment rates for many decades already (Germany, Italy, France). The presence of large numbers of long-term unemployed represents a fundamental challenge to policymakers in designing labour market institutions that help reintegrate the long-term unemployment into the labour market.

Long-term unemployment, the long slump and the Great Recession

In a recent paper, Krueger et al. (2014) note that the relationships between unemployment and inflation and vacancies summarised by the Philipps and Beveridge curves, respectively, have broken down since the end of the Great Recession. By contrast, the relation between short-unemployment, vacancies (Ghayad and Dickens 2012), and inflation (Gordon 2013, Watson 2014) remains stable. It thus seems as if the breakdown of the Philipps curve and the Beveridge curve is linked to the rise in the long-term unemployment.

In a recent paper, we show how central the experience of the long-term unemployed is for explaining the long malaise in the labour market starting in 2007 and – according to some measures – persisting until today (Kroft et al. 2014). Crucial elements here and in Krueger et al. (2014) are the low job finding rates among the long-term unemployed and the high rate at which they leave the labour force. Interestingly, as Krueger et al. (2014) demonstrate, even when the long-term unemployed find jobs, these jobs tend to be unstable, low-paid, and often part-time.

It follows that labour market policies targeting long-term unemployment might have substantial beneficial consequences for the performance of the aggregate labour market, if they are effective at reconnecting the long-term unemployed with the labour market. The differential experience of the long-term unemployed compared to the short-term unemployed also suggests that labour market policies effective among the short-term unemployed might not be effective among the long-term ones. So, what is known about how effective active labour market policies are at reconnecting the long-term unemployed to the labour market?

Active labour market policies and the long-term unemployed

Developed economies expend significant resources on active labour market policies, such as training or wage subsidies targeting the unemployed. Kluve (2010) reports that in 2002 the US and the UK expended relatively modest amounts (0.13 and 0.4% of GDP), whereas the large continental European and the Scandinavian countries spent between 0.5 and 2% of GDP on labour market policies. What do we know about the overall effectiveness of such policies and particularly about policies that target the long-term unemployed?

Two meta-analyses summarise findings from a great many studies. Card et al. (2010) analysed 199 estimates from 97 studies of active labour market policies covering the years 1997-2007. Kluve (2010) examined 137 evaluations from 19 European Countries.1

These two studies emphasise the heterogeneity in results associated with different interventions and study populations. The evidence does not support a simple conclusion regarding the effectiveness of active labour market policies. Kluve (2010) reports that of the 137 studies reviewed, 75 find a significantly positive effect, 29 a significantly negative effect, and 33 report no statistically significant effect. On balance, active labour market policies tend to have positive effect, but there is a lot of variation across programmes. Card et al. (2010) report similar results of significant positive and negative short-run effects in about 1/3 of studies each. They also find that job search assistance policies tend to have positive short-run impacts in helping reemployment, but that training programmes show more positive medium-term impacts. However, many active labour market policies do not generate positive outcomes or even seem to harm the treated.

Card et al. (2010) and Kluve (2010) also examine whether programme effectiveness varies with the design of the policy. They compare programmes with training components, programmes that focus primarily on job search assistance, and programmes that emphasise employment subsidies. Kluve (2010) reports positive effects for 38 out of 70 training programmes. Interestingly, subsidies to private sector employment are often found to have positive effects (17 out of 23), while public employment subsidies tend to be ineffective (7 positive out of 26 evaluations). Card et al. (2010) report that training or job search assistance tend to be more effective than private or public employment subsidies.

Our reading of the evidence in Card et al. (2010) and Kluve (2010) is that programme effectiveness varies a lot across active labour market policies and that relatively little is known about what generates this heterogeneity. In particular, little is known whether the heterogeneity is linked to the design of the policy or to the treated populations. This reading of the evidence is reinforced by the experimental evidence of the UK’s Employment Retention and Advancement demonstration (ERA).

The Employment Retention experiment covered three different populations, including a group of long-term unemployed.2 Over the medium term, the treatment was only effective for the long-term unemployed. For them, the treatment consisted primarily of financial rewards for searching for jobs and for maintaining employment once a job was found. This treatment was found to generate fairly large effects on employment and earnings over the five-year study period and was found to be cost-effective.

The evidence of the Employment Retention and Advancement demonstration thus suggests that financial incentives can help reintegrate the long-term unemployed into the labour market. The Retention, however, also provides evidence that programme effectiveness varies substantially with the treated population.

Displacement effects of policies

So far, we have summarised studies that examine the direct effects of active labour market policies on treated individuals. From a macroeconomic perspective, an important question is whether such policies help some job seekers at the expense of others that are not treated? Such displacement effects are plausible at least in the short run, when fixed inputs generate diseconomies of scale so that the marginal product of labour declines in aggregate with the number of workers hired. Crepon et al. (2013) investigate this question for a job search assistance programme using a two-stage clustered randomised experimental design. They find that those individuals that received job search assistance were indeed more likely to find jobs. By varying the fraction treated experimentally across regions, Crepon et al. (2013) could also investigate whether there are displacement effects in more heavily treated regions. Although their reported findings here are imprecise, the point estimates do suggest that such displacement effects exist and that they might be greater in weak labour markets. Overall, such displacement effects will lower the social benefits from active labour market policies and suggest that it might not be possible to stimulate labour markets using policies during recessions.

Overall, we find that the literature does not provide strong conclusions on how the effectiveness of active labour market policies varies across subgroups of the jobless, nor specifically on effectiveness of active labour market policies targeted to the long-term unemployed. There is some evidence that job search assistance as well as financial incentives can effectively help treated groups, but it is possible that displacement effects can considerably reduce the benefit from these programmes from a societal point of view, particularly during weak labour markets. One speculative possibility based on the (weak) existing evidence is to focus on providing unemployment assistance and long-term training to the long-term unemployed in the depths of a downturn but then moving toward more aggressive use of active labour market policies, such as job search assistance and hiring subsidies to try reemploy the long-term unemployed as the labour market tightens in a recovery.

More questions than answers

The current state of the literature is thus unsatisfying from the perspective of designing active labour market policies, especially those that target the long-term unemployed. We do not know with much certitude which groups benefit most from these policies. Similarly, we do not know enough about whether they should emphasise training, employment subsidies, or job search assistance. Crucially for thinking about how to respond to business cycles, we have too little information on how active labour market policies perform over the cycle or whether displacement effects are important. Thus, policymakers need more reliable evidence to help design more effective active labour market policies to reintegrate the long-term unemployed into the labour market during a recession and its aftermath.

The significant resources expended on these policies in various countries, however, suggest large social benefits from answering these questions. We call on policymakers to add an evaluation component to any new or existing policy, preferably based on a randomised design. We encourage those evaluating these randomised experiments to embed their evaluations in behavioural, structural models that incorporate contextual information on the social systems, as well as on the challenges facing particular populations. Combining the experimental evidence with such behavioural models using contextual information will help synthesising results from various settings and designing future active labour market policies to achieve better labour market outcomes for groups at the margin of the labour market, such as the long-term unemployed.

References

Card, D (2013), “Active Labor Market Programs: What have we learned?”

Card D, J Kluve, and A Weber (2010), “Active Labor Market Policy Evaluations: A Meta-Analysis”,The Economic Journal, Vol.120 November, pp. 452-477.

Crepon B, E Duflo, M Gurgand, R Rathelot and P Zamora (2013), “Do Labor Market Policies Have Displacement Effects? Evidence from a Clustered Randomized Experiment”, The Quarterly Journal of Economics, Vol. 128, No. 2, pp.531-580.

Ghayad R and W Dickens (2012), “What can we learn by disaggregating the unemployment-vacancy relationship?” Federal Reserve working paper, Public Policy Briefs.

Gordon R J (2013), “The Phillips curve is alive and well: Inflation and the NAIRU during the slow recovery”, NBER working paper, No. 19390.

Hendry, R, J A Riccio, R Dorsett, D Greenberg, G Knight, J Phillips, P K Robins, S Vegeris, and J Walter (2011), “Breaking the low-pay, no-pay cycle: Final evidence from the UK Employment Retention and Advancement (ERA) demonstration”, Department for Work and Pensions, UK, Experiment on job search assistance and incentive for work programs. Low effectiveness on single parents. Effective on LTU, No. 765.

Kluve J (2010), “The effectiveness of European active labor market programs”, Labor Economics, Vol.17, Issue 6, p 904-917.

Kroft K, F Lange, M Notowidigdo and L Katz (2014), “Long-Term Unemployment and the Great Recession: The Role of Composition, Duration Dependence, and Non-Participation”, The Journal of Labor Economics, forthcoming.

Krueger A, J Cramer, and D Cho (2014), “Are the long-term unemployed on the margins of the labor market?”, Brookings Papers on Economic Activity, Spring.

Watson M W (2014), “Inflation Persistence, the NAIRU, and the Great Recession”, The American Economic Review, Vol.104, No.5, pp. 31-36.

Footnotes

[1] In Card (2013), David Card presents his view of what we know about the effectiveness of active labour market policies and what the most urgent challenges to research in this area are.

[2] The other two groups were two groups of lone parents, one unemployed and the other working part-time. For these two groups treatment included an important job coaching and training components. The ERA found some evidence of treatment heterogeneity within these groups. In particular, the ERA found positive effects on earnings among the unemployed parents who were better educated.

Published in collaboration with VoxEU

Authors: Lawrence F. Katz is the Elisabeth Allison Professor of Economics at Harvard University and a Research Associate of the National Bureau of Economic Research. Kory Kroft is the Assistant Professor in the Department of Economics and the School of Public Policy and Governance, University of Toronto. Fabian Lange is the Assistant Professor of Economics, Yale University. Matthew Notowidigdo is the Associate Professor of Economics, Northwestern University; Fellow, Institute for Policy Research

Image: A man looks at a job board posted at a job fair in Toronto, April 1, 2009. REUTERS/Mark Blinch (CANADA BUSINESS).