Beyond the average: a synthetic control approach to estimate heterogeneous treatment effects with panel data

This paper introduces a new method for analyzing panel data to estimate heterogeneous treatment effects based on the synthetic control approach. It targets binary treatments applied to multiple units. First, a synthetic control is created for each treated unit in order to estimate the individual treatment effect (ITE). These ITEs are aligned on a common timescale, with the treatment time set to zero. Next, kernel density estimation is used to derive the distribution of effects over time. Then, the highest density region (HDR) is identified, and its mean values serve as estimates of the heterogeneous average treatment effects on the treated (HATT). Simulations confirm that the method is reliable and more precise than standard models. Finally, we demonstrate its real-world relevance by applying it to assess the euro's impact on economic growth.