Abstract:Understanding treatment effects in extreme regimes is important for characterizing risks associated with different interventions. This is hindered by the unavailability of counterfactual outcomes and the rarity and difficulty of collecting extreme data in practice. To address this issue, we propose a new framework based on extreme value theory for estimating treatment effects in extreme regimes. We quantify these effects using variations in tail decay rates of potential outcomes in the presence and absence of treatments. We establish algorithms for calculating these quantities and develop related theoretical results. We demonstrate the efficacy of our approach on various standard synthetic and semi-synthetic datasets.

What problem does this paper attempt to address?

### Problems Addressed by the Paper The paper aims to address the issue of evaluating treatment effects under extreme conditions. Specifically, the authors focus on how to quantify and understand the risks associated with different interventions under extreme conditions, which are often related to the tail behavior of potential outcomes. Traditional causal inference methods primarily focus on Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE), which are based on the expected value of the outcome variable but fail to adequately consider behavior under extreme conditions. ### Main Challenges 1. **Unavailability of Counterfactual Data**: In practice, it is difficult to collect data under extreme conditions, leading to a lack of counterfactual outcomes. 2. **Data Scarcity**: Extreme data is inherently rare, resulting in limited sample sizes and making accurate statistical inference challenging. 3. **Unobserved Confounders**: In causal inference, it is impossible to observe the outcomes of the same individual under different intervention conditions, which is one of the fundamental problems of causal inference. ### Solution To address the above issues, the authors propose a new framework based on Extreme Value Theory (EVT) to estimate treatment effects under extreme conditions. Specifically: - **Extreme Treatment Effect (ETE)**: Defined as the difference in the tail decay rates of the potential outcome distributions between the treatment and control groups. - **Conditional Extreme Treatment Effect (CETE)**: Defined as the difference in the tail decay rates of the potential outcome distributions at the individual level. ### Methodology 1. **Tail Unconfoundedness Assumption**: A weaker assumption than the traditional unconfoundedness assumption, requiring unconfoundedness only in the tail. 2. **Parameter Estimation**: Estimating the parameters of the GEV distribution by maximizing the conditional log-likelihood function. 3. **Computation Method**: Using the k-means algorithm to identify clusters of individuals with similar characteristics and selecting the maximum outcome within each cluster as an approximate maximum for that cluster. ### Experimental Validation The authors validate the effectiveness of the proposed framework using synthetic data and semi-synthetic high-dimensional datasets. The experimental results show that the proposed estimator performs better in estimating ETE and CETE compared to traditional simple estimators, especially as the sample size increases, with the estimates gradually converging to the true values. ### Conclusion The paper provides a new framework for evaluating treatment effects under extreme conditions, which is particularly important in fields such as healthcare where understanding risks under extreme conditions is crucial. By introducing ETE and CETE, the authors provide new tools and methods for assessing the impact of interventions under extreme conditions.

Treatment Effects in Extreme Regimes

Extreme Continuous Treatment Effects: Measures, Estimation and Inference

Extreme Treatment Effect: Extrapolating Dose-Response Function into Extreme Treatment Domain

Estimation and Inference of Extremal Quantile Treatment Effects for Heavy-Tailed Distributions

Extreme Quantile Treatment Effects under Endogeneity: Evaluating Policy Effects for the Most Vulnerable Individuals

Causality and extremes

Estimating individual treatment effect: generalization bounds and algorithms

Treatment Effect Risk: Bounds and Inference

Time series conditional extremes

Extremal Quantile Regression: An Overview

Improved bounds and inference on optimal regimes

On Restricted Optimal Treatment Regime Estimation for Competing Risks Data.

Treatment Effect Quantiles in Stratified Randomized Experiments and Matched Observational Studies

Some benefits of standardisation for conditional extremes

Direct Estimation of the Mean Outcome on Treatment when Treatment Assignment and Discontinuation Compete

A Robust Method for Estimating Optimal Treatment Regimes

Treatment effect estimation under covariate-adaptive randomization with heavy-tailed outcomes

Inference on Extreme Quantiles of Unobserved Individual Heterogeneity

Using extreme value statistics to reconceptualize psychopathology as extreme deviations from a normative reference model

A nonparametric framework for treatment effect modifier discovery in high dimensions

Some methods for heterogeneous treatment effect estimation in high-dimensions