No abstract

No abstract

No abstract

The key instrument for regulating the reduction of greenhouse gas (CO2) emissions is the emissions trading system (ETS), which Russia chose to create. One of the key issues for the ETS participants is the mechanism of CO2 pricing. Energy sources are the most important factors in the CO2 price. The study of the relationship between them and the CO2 price at foreign ETSs gives contradictory results. The paper investigates the relationship between energy variables and CO2 price using exploratory analysis tools as the first stage of machine learning. The univariate analysis showed that the laws of price distribution of futures contracts for CO2 and coal are further from normal in comparison with the prices of futures contracts for gas and Brent oil. Logarithmization improved the statistics of the data. Bivariate analysis showed a close relationship between the prices of CO2 and coal futures contracts. The price data for the other energy variables showed a weak to moderate relationship. Correlation analysis, taking into account the different time lags between the energy variables and CO2, indicates that it is appropriate to include past energy price information in the model. The close linear relationships of the energy variables suggest exploring opportunities to reduce the dimensionality of the data. Exploratory analysis revealed groups of data that would be appropriate to describe with different machine learning models. The presence of data groups in the resulting variable indicates the presence of other CO2 pricing factors that should be identified and taken into account in modeling.