中国交通能耗核心影响因素提取及预测

Exploring the Core Factors and Forecasting the Energy Consumption in China's Transport Sector

随着我国工业化城镇化进程的加快,交通能耗已连续多年位居各行业能源消耗第二,仅次于工业用能,占总能耗的10%-20%,因此分析提取影响我国交通能耗的主要因素,并预测交通运输能耗甚是必要。本文首先利用贝叶斯结构方程模型(BSEM)从经济活动总量、技术进步和交通运输结构三方面刻画交通能耗各影响因素的作用机理。然后运用通径分析(Path-analysis)法提取出交通能耗的主要影响因素包括:交通运输总周转量、城镇居民家庭人均可支配收入、单位周转量能耗、交通运输行业固定资产投资占比和公路民航周转量占比。进而构建VAR模型研究交通能耗与各主要因素间的动态时滞关系,脉冲响应分析发现运输周转量受某一冲击后对交通能耗具有显著的促进作用和较长的持续效应,单位周转量能耗这一指标对交通能耗的影响具有阶段性,短期产生的"回弹效应"会促使交通能耗的增加,长期的技术进步效应会抑制能耗增加,公路民航周转量占比对交通能耗同样具有正向冲击效应。最后进行Gibbs抽样构建基于贝叶斯估计的BVAR模型,预测出2015-2020年我国交通能耗分别为43 393.28、47 076.06、51 085.70、55 452.38、60 209.16、65 392.30万吨标准煤,且拟合效果表明相对预测误差为2.87%,模型预测结果良好。最后根据本文研究结论,提出关于中国交通节能的政策建议。

With the acceleration of industrialization and urbanization,transportation has become the second largest（ around 10%-20% of total energy demand） energy intensive sector in China,next only to industry. In this context,it is of great necessity to figure out the main factors that affect energy consumption in China＇s transport sector,and make reasonable prediction. Firstly,this paper builds a Bayesian structural equation model（ BSEM） to describe the influencing mechanism from three aspects： economic aggregate,technical progress and transportation structure. Next,path-analysis technology is applied to extract the main factors,including transportation converted turnover,per capita disposable income of urban households,energy consumption per unit of turnover,the ratio of transport fixed asset investment and the ratio of road and civil aviation turnover. Based on this analysis,VAR model is built to explore the dynamic time-delay relationship between energy consumption in transport sector and the main factors. The impulse response analysis indicates that when imposing a positive impulse on transport turnover,the significant motivating effect for transportation energy consumption lasts for a long period; the response of transport energy consumption to energy consumption per unit turnover are distinct periodically — in the short term,it will cause the increase of traffic energy consumption due to ＂ rebound effect＂,but in the long-term,it inhibits the increase of energy consumption due to technological progress effect; the ratio of highway and civil aviation turnover also has a positive impact on transportation energy consumption. Finally BVAR model is constructed on the basis of Gibbs sampling to predict traffic energy consumption from 2015 to 2020,which respectively is 433.93,470.76,510.85,554.52,602.09,653.92 million tons of standard coal. The fitting results show that the relative prediction error is 2.87%,implying the predicted results are pretty reliable. Lastly,this paper puts forward related suggestions about energy saving in China＇s transport sector.