The Effects of Summer Employment on Disadvantaged Youth: Experimental Evidence

The unemployment rates facing American youth are bleak. Youth employment over the summer, when teenagers are most likely to work, is at a 60 year low. The last decade has shown a dramatic drop: 44.1 percent of teens were employed in July 2000, but only 25.6 percent were working in July 2010. The situation for minority and low-income youth is even worse: the 2010 employment rate for low-income black teens in Illinois was less than one-fourth the rate for higher-income white teens (9 vs. 39 percent).

There are good reasons to think that this level of teen unemployment will create significant social costs as well as life-long consequences for youth. Teen employment has been shown to significantly increase employment outcomes later in life; one study from the late 1990s finds that working 20 hours a week as a high school senior increases earnings by 22 percent and wages by 10 percent 6-9 years later. Increasing wages for teens - both immediately through the provision of a job itself and later through increased earning potential - may also have a substantial impact on their crime rates. There is evidence that the higher the wage rate available to an individual, the less likely he is to commit a violent or property crime. Increasing income more generally (through income transfers, housing vouchers, tax credits, etc.) has also been convincingly shown to reduce crime, as has investing in individuals' skill development. Since the provision of a summer job is likely to perform all three functions - increase the available wage, provide additional income, and improve individuals' skills (not to mention keep youth busy during the summer months when crime usually spikes) - it is reasonable to expect that it would also decrease criminal behavior.

Surprisingly, there is almost no direct evidence on the effects of providing teens with summer jobs. Some early evaluations of programs in the 70s and 80s showed promising but mixed results, yet the research designs were too weak to draw strong conclusions. In addition, none of those evaluations looked at criminal behavior or violence involvement as outcomes, which seem likely to be one of the key effects of such programs. Less direct evidence shows that intensive residential job training programs have created substantial decreases in arrests, convictions, and incarceration for participants, and that job placement programs can reduce crime among parolees and increase incomes among welfare recipients. Taken together, the evidence suggests that using employment as a crime reduction strategy is quite promising, if not yet proven. But for summer jobs programs in particular, there is a startling lack of evidence.

Despite the dearth of research, policymakers already seem convinced that summer employment support is a good idea. The federal government dedicated $1.2 billion of the 2009 stimulus to employment for disadvantaged youth, with an emphasis on summer jobs programs. These spending levels are not new. Summer jobs for disadvantaged youth have been federally funded since 1964; from 1998 to the present, they have been part of the annual appropriation for Youth Activities of about $1 billion. The President proposed another $1.5 billion specifically for youth summer jobs last fall; when Congress did not pass his proposal, he committed the Department of Labor to arranging for 250,000 youth summer opportunities regardless.

Given the amount of resources spent on youth summer jobs programs over the past half century, the lack of evidence on such programs' effects is startling. Because of the way Chicago's jobs program for disadvantaged youth - One Summer +PLUS - is structured, the program will produce some of the only rigorous evidence to date on the effects of summer jobs programs. It will also measure the incremental effectiveness of adding a social-cognitive skill development component. Similar social-cognitive programming has been shown to reduce violent crime and increase school engagement in a recent randomized control trial. The evaluation will also assess the cost effectiveness of both treatment arms.