Aggregation Trees

41 Pages Posted: 19 Dec 2022 Last revised: 26 May 2024

See all articles by Riccardo Di Francesco

Riccardo Di Francesco

University of Rome Tor Vergata - Department of Economics and Finance

Date Written: December 15, 2022


Uncovering the heterogeneous effects of particular policies or "treatments" is a key concern for researchers and policymakers. A common approach is to report average treatment effects across different subgroups based on observable covariates. However, there is likely to be considerable uncertainty about the appropriate grouping. This paper proposes a nonparametric approach to discovering heterogeneous subgroups in a selection-on-observables framework. The approach constructs a sequence of groupings, one for each level of granularity. Groupings are nested and feature an optimality property. An "honesty" condition allows us to construct valid confidence intervals for the average treatment effect of each group. The utility of the proposed methodology is
illustrated through an empirical exercise that revisits the impact of maternal smoking on birth weight.

Keywords: Causality, conditional average treatment effects, recursive partitioning, subgroup discovery, subgroup analysis

JEL Classification: C29, C45, C55

Suggested Citation

Di Francesco, Riccardo, Aggregation Trees (December 15, 2022). CEIS Working Paper No. 546, Available at SSRN: or

Riccardo Di Francesco (Contact Author)

University of Rome Tor Vergata - Department of Economics and Finance ( email )

Via columbia 2
Rome, Rome 00123

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics