Negative Congestion Externalities in Job Search
RCTReview
Evaluating active labor market policies using individual-level randomized controlled trials assumes the absence of spillovers between treated and untreated job seekers, violating the Stable Unit Treatment Value Assumption. When treated job seekers increase search intensity, non-participants competing for the same limited pool of available vacancies in local labor markets experience crowding-out effects.
Picture this
Imagine a musical chairs game where 100 players compete for 80 chairs. If a coaching program teaches 30 players how to run faster and react quicker, those 30 players will claim chairs much faster, leaving fewer chairs for the remaining 70 players. The non-trained players end up losing chairs not because they became worse at running, but because the heightened effort of trained players created a bottleneck that crowded them out.
What the evidence says
Participation in activation reduced non-participants' exit rate from unemployment by ~5% (p < 0.10 in proportional hazard; -2.7 percentage points probability of employment at 3 months, p < 0.05). Comparing participants directly to non-participants within the same region overstates program efficacy by ignoring this 25%+ crowding-out spillover.
- Who was studied
- N = 40,403 newly unemployed benefit applicants across Denmark (3,751 in experiment counties Storstrøm and South Jutland: 1,814 treatment, 1,937 control; 36,652 comparison county workers) in 2005-2006.
- How
- Stratified partial likelihood proportional hazard duration model and linear probability model comparing experiment counties with comparison counties via difference-in-differences.
What to do
Incorporate local labor market spillover measurements and regional control comparison groups when evaluating the net impact of mandatory job search assistance programs.
From the source
"The effect of the presence of the activation program on the nonparticipants in the program is a reduction in the exit rate of about five percent."
816b5187-9076-4043-aba7-184043b2e524-Estimating Equilibrium Effects of Job Search Assistance.pdf