Estimating heterogeneous impacts using causal forest

Abstract

The labor shortage within small and medium-sized enterprises (SMEs) represents a critical concern in South Korea. In response, the government implemented the Youth Tomorrow Chaeum Deduction policy to promote the influx of SMEs and support the long-term employment and asset-building of young workers. However, the policy concluded in 2022, and as of 2023, its scope was restricted to the manufacturing and construction industries. To assess the effectiveness of the policy, the present study introduces two analytical approaches: a difference-in-differences and Causal Forest. Causal Forest estimates the conditional average treatment effect (CATE) to understand the heterogeneity in the effect of Youth Tomorrow Chaeum Deduction for different demographic subgroups. This study suggests that predicting heterogeneous treatment effects through CATE can be a suitable method for policy planning to target policy beneficiaries within a limited budget.