A hybrid machine learning approach for carbon price forecasting
Type
conference speech
Date Issued
2025-05-26
Author(s)
;
Chen, Zezhun
;
Christopoulos Dr Dimitrios
;
Joe MEAGHER
;
Tsanakas, Andreas
;
Tzougas, George
;
Rui ZHU
Abstract
We investigate the impact of Brexit on the EU and UK Emissions Trading Systems (ETS), highlighting the risk of potential carbon leakage arising from differing carbon pricing dynamics. To analyze post-Brexit carbon market differences, we develop a novel hybrid ARIMA-LSTM machine learning model which captures both linear and nonlinear patterns, providing more accurate predictions and insights into carbon pricing trends than benchmark models. Our results reveal divergence between the two carbon markets post-Brexit underscoring the need for coordinated policies to address these disparities and emphasizing the importance of effective forecasting models to manage carbon pricing risks and promote fair competition.
Event Location
Tartu, Estonia
Event Date
July 1-4, 2025
Additional Information
28th International Congress on Insurance: Mathematics and Economics