A scenario-based integrated approach for modeling carbon price risk

Carbon prices are highly dependent on government emission policies and local industrial compositions. When historical data does not exist or limited price data can only be sourced from another country, scenario analysis becomes the only tool for the modelling of future carbon prices. However, various plausible but equally possible scenarios can produce large variations in forecast carbon prices. In a traditional approach of scenario analysis, investment decisions or risk management strategies are proposed and analysed for each given scenario, optimal solutions are determined. However, when the number of scenarios becomes large, it often becomes too complex and intractable to have a clear view on the selection of investment decisions or risk-management strategies because these decisions and strategies are closely linked with each of the many scenarios. In this paper, it is proposed to use a stochastic mean-reversion model to represent future carbon price movements, but this model is calibrated to the forecast carbon prices of all the scenarios. In this approach, a single model is used to capture the underlying uncertainty and expectation of the stochastic carbon prices as projected by all the scenarios, carbon price risk can thus be modeled and analysed without the need for direct references to any specific scenarios. The modelling and management of long-term carbon-price risk are therefore purely dependent on future carbon price levels and volatilities of these scenarios, instead of on the scenarios themselves. Through such an approach, the optimization of investment decisions and risk management solutions can be much simpler because the forecasted carbon prices are the only input data.