Global minimax rates for heterogeneous treatment effect estimation
Abstract
Recent work established minimax rates for pointwise estimation of the conditional average treatment effect (CATE), but the fundamental limits under global losses remained unresolved. We derive matching minimax lower and upper bounds under the L2 and L∞ risks on certain covariate-density and support classes. For the lower bounds, we develop a many-fuzzy-hypotheses method that replaces each fixed alternative in a multiple-hypothesis construction with a mixture over nuisance functions that leaves its CATE unchanged. A localized product-mixture χ2 bound is used to control the divergence of each resulting mixture from a common baseline, enabling multiple well-separated CATE hypotheses to be handled simultaneously. For the upper bounds, we analyze a modified local polynomial R-learner. The L2 argument extends the pointwise bias--variance analysis, while the L∞ argument combines concentration inequalities for U-statistics with a continuity argument enabled by the use of a continuous kernel and additional regularity of the second-order basis. The resulting rates show that the pointwise elbow phenomenon persists globally: the L2 transition threshold is unchanged, while the L∞ rate has an additional logarithmic factor and a transition threshold adjusted by 1−loglog n/log n.
Related
- Minimax Rates for the Estimation of Eigenpairs of Weighted Laplace–Beltrami Operators on Manifolds
- Minimax rates for pointwise CATE estimation: the role of design smoothness
- Minimax Rates for Robust Gaussian Mean Estimation Near the Breakdown Point under Constraints
- Non-Adaptive 1-Bit Mean Estimation: Minimax Rates and the Sample-Interval Tradeoff
- Global minimax risk and acquisition laws for heterogeneous information fusion
BibTeX
@misc{liu2026global,
title = {{Global minimax rates for heterogeneous treatment effect estimation}},
author = {Shuhan Liu and Matey Neykov},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.37496},
eprint = {2610.37496},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.37496},
}
RIS
TY - UNPB
TI - Global minimax rates for heterogeneous treatment effect estimation
AU - Shuhan Liu
AU - Matey Neykov
PY - 2026
DA - 2026/10/27
PB - arXiv
DO - 10.48550/arxiv.2610.37496
UR - https://arxiv.org/abs/2610.37496
AB - Recent work established minimax rates for pointwise estimation of the conditional average treatment effect (CATE), but the fundamental limits under global losses remained unresolved. We derive matching minimax lower and upper bounds under the L_2 and L_∞ risks on certain covariate-density and support classes. For the lower bounds, we develop a many-fuzzy-hypotheses method that replaces each fixed alternative in a multiple-hypothesis construction with a mixture over nuisance functions that leaves its CATE unchanged. A localized product-mixture χ^2 bound is used to control the divergence of each resulting mixture from a common baseline, enabling multiple well-separated CATE hypotheses to be handled simultaneously. For the upper bounds, we analyze a modified local polynomial R-learner. The L_2 argument extends the pointwise bias--variance analysis, while the L_∞ argument combines concentration inequalities for U-statistics with a continuity argument enabled by the use of a continuous kernel and additional regularity of the second-order basis. The resulting rates show that the pointwise elbow phenomenon persists globally: the L_2 transition threshold is unchanged, while the L_∞ rate has an additional logarithmic factor and a transition threshold adjusted by 1−loglog n/log n.
ER -