VoxEU Column Labour Markets

Youth employment and summer employment programmes

Among different youth employment programmes, summer employment programmes have received relatively little attention. This column looks at the effects of the New York City’s summer youth employment programme – the largest programme of this kind in the US – on a number of outcomes. Whereas the programme did not raise later earning or college enrolment, it decreased the probability of incarceration and mortality. This is a new, often neglected, large benefit from a youth employment programme. 

Many federal, state, and local policies attempt to support individuals’ labour market prospects, including public employment and subsidised employment programmes. In many cases, these programmes focus on encouraging youth employment — often summer youth employment in particular. City programmes across the US provide youth with summer jobs — the fifty most populous cities in the country have all had summer youth employment programmes in the last five years — and the Work Opportunity Tax Credit (WOTC) subsidises employment of summer youth. As youth unemployment remains elevated following the Great Recession, youth employment programmes have received increased scrutiny from policymakers. While literature typically finds that other (i.e., non-summer-employment) active labour market programmes for youth have costs that outweigh their benefits (Stanley et al. 1998 Heckman et al. 1999, Lalonde 2003), summer youth employment has “received relatively little attention from program evaluators” (Lalonde 2003, p. 532).

Programmes to support summer employment are justified with various rationales.

  • One rationale is that summer employment could provide income support to youth (and their families) through wages earned in the programme.
  • A second rationale is that summer work experience could improve future employment outcomes, perhaps particularly for disadvantaged youth who would otherwise have low summer employment rates.
  • A third rationale for such programmes is that they could help to keep youth involved in socially productive activities or ‘out of trouble’. Keeping youth out of trouble during the summer could help them avoid dangerous activities and could have benefits such as decreasing incarceration or mortality rates.

The Summer Youth Employment Programme

We investigate the empirical support for these three rationales by analysing the New York City (NYC) Summer Youth Employment Programme (SYEP) in the years 2005 to 2008, inclusive. During these years, SYEP provided summer jobs to NYC youth aged 14 to 21, paid by the NYC government at a total cost of $236 million. Each year, SYEP received more applications than the number of jobs available. In the face of this excess demand, SYEP randomly allocated spots in the programme to applicants by lottery. We compare the outcomes of individuals who participate in SYEP because they were randomly selected to receive a job through SYEP, to the outcomes of individuals not selected. We examine the effect of SYEP on observable outcomes suited to assess each of the rationales for summer youth employment programmes. We link SYEP administrative data on these lottery winners and losers to Internal Revenue Service (IRS) administrative data on the universe of US federal tax data; to New York State (NYS) Department of Corrections and Community Services (DOCCS) administrative data on individuals incarcerated in NYS; and to NYC Department of Health and Mental Hygiene (DOH) administrative data on causes of death in New York City. In the four years of lotteries we study, there were 294,580 SYEP applications subject to the lottery, of which 164,977 won the lottery and 129,603 lost the lottery. 

This context provides a particularly promising setting for studying a youth employment programme.

  • First, the large scale of the programme, the random assignment, and the accurate data allow us to estimate precise causal effects, not only on earnings, the employment rate, and college enrolment, but also on rare outcomes including mortality and incarceration.

Our sample sizes are at least an order of magnitude (and in many cases two orders of magnitude) larger than the sample sizes of other randomised studies. The ability to look precisely at mortality, which other studies have not been able to observe, will prove particularly interesting since it will have important implications for the benefits of the programme and offer a new take on previous results. The data allow us to estimate effects up to almost a decade after programme participation.

  • Second, the IRS and DOCCS administrative data allow us to examine a wide variety of outcomes, including contemporaneous and subsequent earnings, earnings by industry, college enrolment, incarceration, and mortality.
  • Third, NYC SYEP is the largest summer youth employment programme in the US and, therefore, represents a central recent case study of US summer youth employment programmes and an important example of youth employment programmes more generally.

Findings of the effect of the programme

We find that SYEP participation increases earnings and employment in the year individuals are employed through the programme.

In a baseline specification, we estimate that in the year of SYEP participation, SYEP raises average earnings through the programme by $1,085.34, lowers other earnings by a modest $208.87, and, therefore, raises net earnings in the year of participation by $876.26. Thus, crowding out of other earnings was 19.24% of the SYEP transfer in this year. We also estimate that, on net, SYEP raises the probability of having any job by 71 percentage points in the year of the participation, with only a five percentage point decrease in the probability of having a non-SYEP job.

  • At the same time, we do not find that youth employment has a positive effect on subsequent earnings or on college enrolment. In each of the three years following SYEP participation, we estimate that SYEP participation causes a modest decrease in earnings of around $100 per year.

This effect is driven by those who had worked prior to SYEP participation and by those in the upper half of the earnings distribution; meanwhile, SYEP participation has an insignificant impact on subsequent earnings among WOTC-eligible individuals. Starting in the fourth year following SYEP participation, we find insignificant impacts of SYEP participation on earnings. We also find that SYEP has no impact on college enrolment, with an extremely precise 95% confidence interval that rules out a positive or negative effect greater than one one-hundredth of a year of college during the period we examine. It is notable that even for this young group with typically little prior job experience, and even during the Great Recession period that we examine separately, an employment programme did not provide a path to greater future earnings.

Over the year of SYEP participation and the subsequent four years, SYEP participation on net raises average earnings by $536.53. Thus, SYEP succeeds on net in transferring money to youth selected for the programme, though with significant crowding out of other earnings. The average decrease in total earnings in the year of SYEP participation and the subsequent four years is 54.02% of average SYEP earnings, representing an aggregate decrease of $104.03 million over these four subsequent years. The degree of crowding out of other earnings is small relative to likely lifetime earnings but is substantial relative to the size of the programme.

  • Strikingly, consistent with the hypothesis that SYEP keeps youth ‘out of trouble’, we find that SYEP participation decreases the probability of incarceration and decreases the probability of mortality.

SYEP causes a decrease in the probability of incarceration of 0.10 percentage points, driven by a decrease among males. While this effect is small in percentage point terms, it represents a substantial 10.36% reduction relative to the baseline incarceration rate of 0.95%. The SYEP-induced decrease in mortality, also driven by males, is 0.08 percentage points, again small in percentage point terms but a substantial 19.92% of the 0.38 percent baseline mortality rate. Evidence from analysing cause of death suggests that SYEP may prevent death by external causes. SYEP appears to put youth on a path that leads away from dangerous outcomes.

The point estimates suggest that by October 2014, around 86 lives were saved by the four years of the SYEP programme from 2005 to 2008. Under standard cost-benefit analysis calculations that assign a value of $9 million to the lives of prime-aged males in real 2013 dollars (Viscusi and Aldy 2003), this would imply benefits of $773 million. Thus, the effects on mortality have the potential to change greatly one’s view of the magnitude of the benefits of a youth employment programme and how such benefits compare to its costs. Past literature has typically found negative net benefits of (non-summer-employment) active labour market programmes for youth, but it has not examined this important outcome.

Concluding remarks

Like most previous work on active labour market programmes for youth, we find that the effects on future earnings cannot justify the programme in a cost-benefit analysis; in fact, we find that SYEP modestly reduces participants’ subsequent earnings. Adding a new twist to previous work, we also find a very large new source of benefits that could well be pivotal to the cost-benefit analysis.

Other youth active labour market programmes may or may not have such mortality benefits, but it is worth noting that like SYEP, other youth programmes have been found to keep youth out of trouble (Schochet et al. 2008, Heller 2014). Future work should examine further whether other programmes show similar results.

Authors' note: this post does not necessarily reflect the views of the US Treasury.


Gelber, A, A Isen and J B Kessler (2014), “The Effects of Youth Employment: Evidence from New York City Summer Youth Employment Program Lotteries”, NBER WP122314

Heckman, J J, R J Lalonde, and J A Smith (1999), “The Economics and Econometrics of Active Labour Market Programs”, Handbook of Labour Economics, Volume 3A. Eds. O Ashenfelter and D Card. Amsterdam and New York: Elsevier, 1865–2095.

Heller, S B (2014), “Summer jobs reduce violence among disadvantaged youth”, Science, Vol 346, 5 December 2014.

LaLonde, R (2003), “Employment and Training Programs”, Means-Tested Transfer Programs in the United States, Ed. R Moffitt, University of Chicago Press.

Schochet, P Z, J Burghardt, and S McConnell (2008), “Does Job Corps Work? Impact Findings from the National Job Corps Study”, The American Economic Review 98.5 (2008): 1864-86.

Stanley, M, L F Katz, and A B Krueger (1998), “Developing Skills: What We Know about the Impact of American Employment and Training Programs on Employment, Earnings and Educational Outcomes”, G8 Economic Summit.

Viscusi, W K, and J E Aldy (2003), “The Value of a Statistical Life: A Critical Review of Market Estimates Throughout the World”, Journal of Risk and Uncertainty 27.1 (2003): 5-76. 

735 Reads