Predicting carbon allowance prices with energy prices: A new approach

The global economy is consistently concerned with what can accurately predict the price of carbon allowance, a trading concept that resulted from the need to mitigate carbon emissions. This motivates this paper to examine the predictive role of energy prices in the forecast of European Union carbon allowance prices. Using the novel Feasible Quasi Generalized Least Squares estimator which is able to account for the effects of conditional heteroskedasticity, serial correlation, persistence and endogeneity in the predictors, the symmetric predictive model is compared with the asymmetric predictive model to determine if the latter outperforms the former in the forecast performance. The results show that the carbon allowance prices are significantly predicted by all the energy prices considered. However, the asymmetric predictive models for oil and coal prices offer a better forecast performance than their symmetric versions. For natural gas, the results appear mixed, as asymmetries only matter when lower sample size, i.e. 50% of the full sample size, is used. Comparison of the proposed models with traditional models reveals that the former outperform the latter. Robustness is provided for these findings through different sample sizes, various forecast horizons and alternative forecast test. Policy inferences for carbon emissions mitigation and investment purposes are drawn accordingly. (C) 2020 Elsevier Ltd. All rights reserved.

Automated summary

This study examines the predictive role of energy prices in forecasting EU carbon allowance prices, using a novel estimation method. The results show that energy prices significantly predict carbon allowance prices, with asymmetric predictive models outperforming symmetric models for oil and coal prices, and mixed results for natural gas prices. The proposed models also outperform traditional models, providing robust policy inferences for carbon emissions mitigation and investment purposes.