Priority to unemployed immigrants? A causal machine learning evaluation of training in Belgium

TL;DR

This study uses causal machine learning to evaluate the effectiveness of training programs for unemployed individuals in Belgium, revealing heterogeneity in outcomes and potential for policy improvements through targeted reassignments.

Contribution

It introduces a novel application of Modified Causal Forests to assess training program effects and proposes simple rules to optimize employment outcomes.

Findings

All programs have positive effects after the lock-in period.

Reassigning unemployed to programs based on individual gains can increase employment duration by up to 20%.

A simple policy rule captures about 70% of the potential improvement.

Abstract

Based on administrative data of unemployed in Belgium, we estimate the labour market effects of three training programmes at various aggregation levels using Modified Causal Forests, a causal machine learning estimator. While all programmes have positive effects after the lock-in period, we find substantial heterogeneity across programmes and unemployed. Simulations show that 'black-box' rules that reassign unemployed to programmes that maximise estimated individual gains can considerably improve effectiveness: up to 20 percent more (less) time spent in (un)employment within a 30 months window. A shallow policy tree delivers a simple rule that realizes about 70 percent of this gain.