A Causal Analysis of CO2 Reduction Strategies in Electricity Markets Through Machine Learning-Driven Metalearners

TL;DR

This work tackles the problem of whether incentive-based electricity pricing reduces household $CO_2$ intensity, using CausalML uplift meta-learners to estimate heterogeneous treatment effects under the Rubin Causal Model. It systematically contrasts random and non-random assignment mechanisms, derives efficiency bounds for causal estimates, and applies S-, T-, X-, and R-learners to estimate conditional average treatment effects $\tau(x)$ with cross-fitting. Empirical results on a large regional dataset show that some pricing incentives can paradoxically increase $CO_2$ intensity, with varying performance across meta-learners (e.g., $\text{RMSE}$ values of 0.15–0.25 and $\text{MAE}$ as low as 0.12). The study highlights the importance of accounting for heterogeneity and potential unintended environmental consequences in policy design, advocating more comprehensive causal analyses for sustainable electricity pricing.

Abstract

This study employs the Causal Machine Learning (CausalML) statistical method to analyze the influence of electricity pricing policies on carbon dioxide (CO2) levels in the household sector. Investigating the causality between potential outcomes and treatment effects, where changes in pricing policies are the treatment, our analysis challenges the conventional wisdom surrounding incentive-based electricity pricing. The study's findings suggest that adopting such policies may inadvertently increase CO2 intensity. Additionally, we integrate a machine learning-based meta-algorithm, reflecting a contemporary statistical approach, to enhance the depth of our causal analysis. The study conducts a comparative analysis of learners X, T, S, and R to ascertain the optimal methods based on the defined question's specified goals and contextual nuances. This research contributes valuable insights to the ongoing dialogue on sustainable development practices, emphasizing the importance of considering unintended consequences in policy formulation.

Figures (4)

• Figure 1: Comprehensive Decision Guide Flowchart for Algorithm Selection.(Source; SIGKDD international conference, 2021, Introduction to CausalML, Page 41, 42)
• Figure 2: CATE estimates using learners S, T, R, and X.
• Figure 3: Overlapped histogram of distributions of CATE according to applied algorithms.
• Figure 4: RMSE, MAE, Variance, and Bias estimates for the applied meta-learners.