Forecasting and Pricing in the Carbon Credits Market

Research by: Yiyang Chen, Rogemar Mamon, Fabio Spagnolo, Nicola Spagnolo, & Heng Xiong

Executive Summary

This article develops flexible forecasting and pricing models for carbon credits. It focuses on the European Union Emissions Trading System (EU ETS). The EU ETS is central to global climate policy. As carbon markets expand, understanding price dynamics becomes essential. This matters for firms, financial institutions, and regulators.

This study examines the European Union Allowances (EUAs). It addresses a key challenge: carbon prices are complex and non-linear. Prices respond to policy shifts, economic shocks, and structural change. Traditional models often fail to capture these dynamics. This limits their usefulness for forecasting and risk management. This paper proposes a modelling framework that addresses this gap. It combines continuous-time models with regime-switching hidden Markov models.

Four stochastic models are extended to allow regime-dependent parameters. These include Geometric Brownian Motion, Ornstein–Uhlenbeck, Cox–Ingersoll–Ross, and jump-diffusion models. Regimes reflect volatility shifts, policy interventions, and market disruptions. Models are estimated using rolling windows. This enables continuous recalibration as new data arrive. It suits fast-changing regulatory environments.

The empirical analysis uses daily EUA data from 2009 to 2024. Results show that regime switching improves overall model performance. The regime-switching jump-diffusion model performs best overall. It provides the strongest fit and most reliable density forecasts. It captures jumps and changing volatility more effectively. This improves derivative pricing and financial risk assessment.

No single model dominates across all forecasting tasks. Simpler models perform well for short-term point forecasts. Regime-switching models better capture full price distributions. This distinction matters for risk managers and regulators. They focus on probability distribution’s tail risks and stress scenarios.

The results also include an evaluation of the cost-of-carry relationship. Standard theory explains only a small portion of price movements. Carbon prices reflect risk premia, liquidity, and policy uncertainty. This highlights the need for richer modelling approaches.

For hedging, the findings provide practical guidance. Models with jumps improve hedge effectiveness and reduce downside risk. However, regime-switching models do not always outperform simpler models. This is especially true for short-horizon hedging. Model choice should match the specific risk objective.

This work provides a practical and adaptable modelling toolkit. It is relevant for practitioners, policymakers, and instructors. It offers applied insight into modern stochastic modelling in carbon markets. This study supports better decision-making in the low-carbon transition.

To cite this article:

Chen, Y., Mamon, R., Spagnolo, F., Spagnolo, N., & Xiong, H. (2026). Forecasting and pricing in the carbon credits market. Journal of Forecasting, 1–27. https://doi.org/10.1002/for.70168

To access the article: https://doi.org/10.1002/for.70168

About the Journal