Transportation Sustainability, Macroeconomics, and Endogeneity in China: A Hybrid Neural-Markowitz-Variable Reduction Approach

Abstract

This paper analyzes the sustainability of the transportation sector in China by investigating feedbacks or endogenous effects among energy consumption, transportation, and major macroeconomic variables and their consequent CO2 emission levels related to fiscal policy, monetary policy, inflationary pressure, and economic activity. A novel combination of Alternative Variable Reduction Techniques, a Markowitz Portfolio Selection Model, and Neural Networks is applied on monthly data from January 1999 to December 2017. The main idea is to develop a hybrid approach to unveil endogeneity in China's transportation footprint, which is mostly impacted by energy consumption and CO2 emission levels of different transportation modes and major macro-economic factors while simultaneously exploring the epistemic uncertainty that surrounds this issue as captured by Information Entropy, Variance Inflation Factor, and Covariance matrix. It can be concluded that road transportation has played a dominant role in decreasing the entire sustainability level, which is mainly driven by trade and fixed-asset investment as well as monetary and fiscal policy, while consumer expectations and cornerstone economic prices have shown limited impact on the transportation footprint. Railways and waterways help mitigate the harmful impacts of increased economic activity and expansionist monetary and fiscal policy on the transportation footprint.