A new methodology for carbon price forecasting in EU ETS

Abstract

This paper proposes a new methodology for carbon price forecasting. It posits a finite distributed lag (FDL) model and then applies a GA‐ridge algorithm to determine a set of proper predictors with coefficient estimates. An empirical study was conducted in the European Union Greenhouse Gas Emissions Trading market, revealing that our methodology not only yields good forecasting results but also provides some interesting analysis on the carbon price market. It turns out that the combination of the FDL model and GA‐ridge algorithm is desirable for forecasting and analyzing the complicated carbon price market because of its capability of selecting proper predictors from a class of predictors of itself.

Interval Forecasting of Carbon Futures Prices Using a Novel Hybrid Approach with Exogenous Variables

“…First, the forecast accuracy of all estimated models is examined in terms of I , which is defined in (7). Second, we use the analysis of variance (ANOVA) test to investigate if the means of the accuracy measure (i.e., I ) are significantly different among the six prediction models.…”

“…After an extensive review of the extant literature, we found that in recent years great research efforts have been expended in two areas: (1) understanding the underlying mechanisms that determine carbon futures prices [1][2][3] and (2) the development of various models suitable for forecasting carbon futures prices [4][5][6][7][8][9][10][11][12][13][14][15]. A slight significant progress in forecasting carbon futures prices is notable.…”

Forecasting carbon price using a multi‐objective least squares support vector machine with mixture kernels

“…Due to the complicated nonlinear and nonstationary features of the carbon market, machine learning models have achieved excellent performances in carbon price forecasting as well (Wei et al, 2015). Han et al (2015) combined a finite distributed lag model and genetic algorithm ridge model for carbon price forecasting. They found that the proposed model could select proper predictors by itself and achieved a higher forecasting accuracy than ANN and partial linear models.…”