Forecasting carbon price in the Western Climate Initiative market using Bayesian networks

Abstract

Many studies have been conducted to forecast and analyze the price of carbon in an emission trading scheme, also known as a cap-and-trade market. Exhaustive forecasting studies have been seldom (if ever) performed in the Western Climate Initiative (WCI) market. Because of the distinctive market dynamics in the WCI, the results of research on other markets may not be applicable to forecast the carbon price in this market. Moreover, there is a continuous debate about modelling the causality between driving forces (including energy prices, economic growth, weather, etc.) and carbon price. The objective of this paper is to forecast carbon price in the WCI market, by modelling the uncertainty of the driving forces and their causal relationship with carbon price. A probabilistic model is developed using Bayesian networks to infer the possible ranges of each driving force that could have an escalation/depreciation effect on price as well as the magnitude of this impact. The model is developed based on retrospective and prospective information on the selected driving forces in all the jurisdictions of the WCI market, providing the most probable price(s) over the period 2018–2030.

Keywords

• emission trading scheme (ets)
• the western climate initiative
• bayesian networks (bns)