Modeling the long-run drivers of total renewable energy consumption: evidence from top five heavily polluted countries

ABSTRACT A boom in renewable energy consumption as a percentage of final energy consumption (REC) has a significant impact on cleaner production and environmental sustainability. However, studies on REC as for an individual country-case and panel case approach for top emitters are rare. Against this backdrop, this study investigates the long-run drivers of REC by using a time series data from 1990 to 2015. The FMOLS-grouped results indicate that economic growth and trade openness increase REC whereas population growth has a negative but significant impact on REC. To check for causal links, the innovating accounting approach using variance decomposition analysis is applied. The results show that there is a unidirectional causality running from economic growth, trade openness, and population growth to REC. Further, the predictive accuracy of the FMOLS based econometric output and the bi-directional long short-term memory (Bi-LSTM) is analyzed. The Bi-LSTM formulated algorithm outperformed the econometric output. The Bi-LSTM is utilized to forecast REC to the year 2030. The output from the Bi-LSTM shows that China, the US, India, the Russian Federation’s, and Japan’s REC will hit ∼11.3395, ∼11.1245, ∼34.6969, ∼2.9097, and ∼7.4859, respectively. The US and Japan’s REC levels will increase while that of China, India, and Russia Federation will decrease. As a policy implication, new policy directives for China, India, and the Russian Federation are required to boost REC levels.