Use the feeds below to stay connected with updates to this site's content and discussions.

What's this?
How to view RSS Feeds with Outlook?

Comment on a post

We need your comments to help our community flourish. Provide your Professional thoughts and opinions by replying to a post that interests you.

Become a Guest Blogger

Are you a expert in the topics being discussed on this site? Connect with our site moderators to request guest blogger privileges.

Commenting Policy

Be sure to check our Comment Policy before participating!

Analysis of More Than 200 Employment and Training Evaluations

A new National Bureau of Economic Research (NBER) working paper has reviewed the findings of 207 worldwide employment and training evaluations, finding generally long-term positive effects, especially for training and subsidized private sector employment.  The review pays careful attention to the type of intervention (job search assistance, training, private or public sector employment subsidy, or sanction/threat for failure to look for work); type of participant (age, gender, long-term unemployed, economically disadvantaged); length of impact studied (less than a year, 1-2 years, or longer than 2 years); and other factors.

The studies began as long ago as 1980, but most were conducted between the early 1990s and the early 2000s.  Only about one-eighth of the evaluations came from English-speaking countries (the number of U.S.-specific studies isn’t provided), and nearly half from European countries other than the UK (the rest came from primarily non-OECD nations).  Only a fifth of the studies used an experimental, random-assignment design, although the studies’ conclusions didn’t vary by whether an experimental or non- experimental approach was used.

Taking the 207 studies altogether, the average impact (generally speaking, most studied the impact on employment, unemployment, and/or earnings) was relatively small and not statistically significant within the first year after the conclusion of the program, but progressively grew to a highly statistically significant effect thereafter.  However, this overall average comprised contrasting trends for two different program types:  short-term interventions like job search assistance or sanctions had better short-term and less positive long-term effects, while training had a less positive short-term impact but a greater effect in the long term.  Despite the relatively large number of studies examined, the user should be cautioned that for some types of studies and participants, the sample size greatly diminished, especially among those studies that followed participants for a longer period (rendering some findings provisional rather than definitive). 

Job search assistance programs generally have a positive effect through the first two years after the program‘s end, with results falling off sharply thereafter.  By contrast, training (including on-the-job training ) programs had increasingly better impacts the longer that individuals were tracked.  Private sector subsidized employment showed the same pattern as training, but the impact after 2 years was even better and particularly large.  By contrast, publicly-subsidized jobs had little if any effect in the short- or long-term. 

 The best effects were found with programs for women and the disadvantaged, with lesser long-term impacts on youth and adults older than 24 (although the average impact for youth was relatively better than for other age groups in the short run).  The authors also found that employment programs had larger impacts in periods of slow economic growth and higher unemployment.

See What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations.  For the detailed findings, see especially tables 1 (an overview of the studies); 3a and 3b (which examine the impacts of various types of studies); and 4 (which examines impact trends over time for individual studies, in contrast to 3a, and 3b, which includes all the studies — some of which did not track participants for longer periods).  The remaining tables attempt to statistically control for various elements.