Displacement Externalities in Active Labor Market Programs
Standard microeconometric evaluations assume that treating job seekers does not impact untreated individuals. However, when activated job seekers search more intensively, they overcrowd the labor market and reduce job finding rates for nonparticipants competing for the same positions.
Picture this
Imagine a crowded musical chairs game where some players are suddenly given fast running shoes. While those runners grab available chairs much faster, they bump into and push away the barefoot players, making it significantly harder for the non-runner players to find a seat.
What the evidence says
Mandatory job search assistance increased participant exit rates from unemployment by 17.9% to 20% (p < 0.01), but reduced nonparticipant exit rates by 4.8% to 5.0% overall (p < 0.10) and by 7.9% in South Jutland (p < 0.05). At three months, nonparticipants were 2.7 percentage points less likely to find employment (p < 0.05).
- Who
- N = 89,466 unemployed benefit applicants (40,403 in the 2005–2006 experiment cohort and 49,063 in the 2004–2005 pre-experiment cohort) across 15 Danish counties, including 3,751 trial participants in Storstrøm and South Jutland [5].
- How
- Stratified partial likelihood duration model and difference-in-differences linear probability model comparing nonparticipants in trial counties against unemployed job seekers in non-experiment counties [6-9].
What to do
Evaluate active labor market interventions by comparing non-treated individuals in treatment regions to external control populations rather than relying on internal regional control groups.
From the source
"Using a difference-in-difference model we show that the nonparticipants in the experiment regions find jobs slower after the introduction of the activation program (relative to workers in other regions)."
Estimating Equilibrium Effects of Job Search Assistance