Study on the Change in the Total Factor Carbon Emission Efficiency of China’s Transportation Industry and Its Influencing Factors

Abstract

The transportation industry is a high carbon emission industry, and China has also put forward strict requirements for the transportation industry to achieve carbon emission reduction. By measuring the total factor carbon emission efficiency of the transportation industry, we can understand the change trend and the influencing factors of the total factor carbon emissions. To fully consider the problem of multiple inputs and outputs in the transportation industry and obtain a more accurate efficiency evaluation value, this paper adopted the slack-based model-data envelopment analysis method and global Malmquist—Luenberger index to study the change in the total factor carbon emission performance of the transportation industry. The combination of static analysis and dynamic analysis was used to calculate the TFP of the transportation industry and increase the content of output indicators. The results indicate that the average TFP and GML index values exhibited significant heterogeneity nationwide. The values in Anhui and Hebei Provinces were greater than 1, and the average GML index values in Shanxi, Guangxi, and Yunnan were greater than 1. The eastern region performed well in terms of technical efficiency and scale efficiency. The technical efficiency in the central, western, and northeastern regions was optimal. In terms of influencing factors, the influencing factors causing the different carbon emission efficiencies in the four regions varied. Finally, corresponding policy suggestions were proposed.

1. Introduction

The transportation industry, acknowledged as a pillar industry of the national economy, plays an important role in regional development. Moreover, this industry drives economic development. However, the rapid expansion of the transportation industry has increased the demand for and consumption of fossil fuels such as gasoline and kerosene [1]. As one of the industries with a high energy consumption level, the transportation industry could cause an increase in regional carbon dioxide emissions and even serious environmental pollution problems [2].

In 2020, China formally proposed at the 75th United Nations General Assembly to achieve a carbon peak by 2030 and carbon neutrality by 2060, which is known as the “double carbon target”. The carbon peak refers to the peak of carbon emissions before 2030 and then gradually falls back. Carbon neutrality refers to enterprises, organizations, or individuals offsetting the total amount of greenhouse gas emissions directly or indirectly produced within a certain period via plant afforestation, energy conservation, and emission reduction efforts to achieve zero CO₂ emissions [3]. This objective has forced transportation managers to more closely consider carbon reduction. In recent years, the Chinese government has committed to implementing various measures to alleviate the problem of excessive carbon emissions of the transportation industry.

The carbon emissions of the transportation industry are affected by many factors, such as the level of economic development, the level of human input, and the level of energy input. Therefore, many factors need to be considered when studying the carbon emission efficiency of the transportation industry. The TFP is the carbon emission efficiency value calculated under comprehensive consideration of multiple input factors, that is, the ratio of actual carbon emission efficiency to better carbon emission efficiency, which can measure whether carbon emissions reach the optimal state to understand the current carbon emissions status of the industry. Moreover, carbon emission efficiency of the transportation industry varies greatly among different regions due to the large economic differences among the various regions in China and the unbalanced development of transportation infrastructure. The improvement in TFP of the transportation industry plays an important role in the realization of the dual carbon goal, so it is necessary to study the total factor carbon emission efficiency of the transportation industry in all provinces of China. By elucidating the current emission efficiency of the transportation industry, clarifying regional differences, and identifying influencing factors in China, local governments can introduce targeted policies to help achieve the goal of carbon neutrality.

Recently, the carbon emission efficiency of the transportation industry has attracted extensive attention among scholars, but few scholars have comprehensively analyzed the total factor carbon emission efficiency. For example, Wang and Guo [4] used the superefficiency SBM model to study the carbon emission efficiency of the public transport sector in Beijing, thereby selecting the number of employees, capital investment, and operating vehicles as input indicators; the passenger volume as the expected output; and transportation carbon emissions as the unexpected output. Lu et al. [5] adopted cargo turnover, passenger turnover, and carbon dioxide emissions as outputs to analyze the environmental efficiency of the transportation industry in eastern China. Yu et al. [6] also used cargo turnover, passenger turnover, and carbon dioxide emissions as outputs to study the carbon emission efficiency of the railway transportation industry in China.

China’s transportation industry is large in size, resource consumption is growing rapidly, and the industry is significantly affected by policies. For example, China’s low-carbon policies and subsidies for new energy vehicles have promoted the transportation industry to reduce fossil energy consumption and transform and upgrade toward the low-carbon green goal. It is precisely because of the comprehensive effect of internal and external factors that the energy structure of the transportation industry is constantly changing and adjusting, which makes carbon emission efficiency research more complex.

At present, scholars’ analysis of carbon emission efficiency is mainly divided into two categories. One is static analysis, that is, using the SBM-DEA model. This method can obtain the annual TFP value of the transportation industry, obtain the change trend, and analyze from the time dimension. The other is dynamic analysis, that is, measuring the GML index. This method can obtain the change rate of carbon emission efficiency and further obtain the factors affecting carbon emission efficiency through index decomposition to further study the internal causes of carbon emission efficiency change. Thus, by combining the two methods, we can obtain a more accurate and comprehensive analysis of carbon emission efficiency.

To accurately calculate the carbon emission efficiency of the transportation industry in China, this paper used the superefficiency slack-based model-data envelopment analysis (SBM-DEA) method and global Malmquist—Luenberger (GML) index to statically and dynamically analyze, respectively, the TFP of the Chinese transportation industry at the provincial level. As such, we divided China into four regions for comparative analysis to investigate the efficiency heterogeneity and underlying causes in different regions of China. At the same time, the technical efficiency change index (GECH) could be further decomposed into pure technical efficiency (GPECH) and scale efficiency change index (GSECH) for specific analysis to help policy makers formulate suitable management policies targeting the transportation industry.

The paper is organized as follows: Section 2 provides a literature review. Section 3 introduces the model and data sources used. Section 4 analyzes the total factor carbon emission efficiency of the transportation industry in China from both static and dynamic perspectives. Section 5 studies the influencing factors of the TFP of the transportation sector. Section 6 provides conclusions and findings, in addition to suggestions.

2. Literature Review

The DEA model, which considers economic, technical, and other factors, has been widely used to analyze TFP and green energy efficiency [7,8] in specific areas, such as the manufacturing industry [9], agricultural industry [10,11], and tourism industry [12,13] Based on the DEA model, the SBM-DEA model has been widely used in efficiency research with slack variables, and it has been verified that this model can achieve a higher calculation accuracy and smaller errors than those obtained with the traditional DEA model [14].

The traditional DEA model has some deficiencies in measuring carbon emission efficiency. One is that it is impossible to solve the problem of ranking when multiple DMUs are valid at the same time. Second, the relaxation of multiple inputs and outputs has not been fully considered. The superefficient SBM-DEA model can solve these problems and accurately measure the total factor carbon emission efficiency. Therefore, the SBM-DEA model with undesirable outputs has been used by many scholars to study carbon emission efficiency. Yuan et al. [15] added environmental variables and employed the SBM-DEA model to analyze the carbon emission efficiency of the transportation industry in China.

Considering that the highest efficiency value of different decision-making units (DMUs) reaches 1, the traditional DEA model cannot compare different DMUs with the same efficiency values, so the superefficiency DEA model has been applied because the efficiency value of each DMU in this model differs. Additionally, the superefficiency DEA model has been combined with the SBM model, referred to as the superefficiency SBM-DEA model. Sun et al. [16] utilized the superefficiency SBM-DEA model to evaluate the carbon efficiency in Shandong Province. Considering that the actual input index of the transportation industry changes according to different proportions, the nonradial DEA model prevents bias due to radial selection [17]. Therefore, through the nonradial SBM-DEA model, we could obtain more accurate TFP values, reflecting the static changes in the transportation industry.

Compared with the ML index, the GML index not only contains the observations in all periods but also solves the problems of the ML index that cannot be accumulated by bad cycles and is not transitive, and there is no feasible solution for linear programming, thus showing a long-term growth trend [18]. At the same time, the index decomposition of the GML index can further study the influencing factors of the change in TFP. Therefore, scholars adopt the GML index to reflect the dynamic changes in TFP. Oh [19] constructed the GML index based on the traditional Malmquist—Luenberger index. Adopting the total set of production technologies during all research periods as the reference set, he constructed a single production frontier that could be used to measure the gap between the technical efficiency and the frontier during each period for each DMU.

Moreover, in dynamic analysis, the GML index is usually decomposed into several factors. Zhang and Wei [20] decomposed the meta frontier nonradial Malmquist—Luenberger carbon emission performance index into efficiency, best-practice gap, and meta technology gap changes to analyze the dynamic changes in the TFP of the transportation industry in China. Zhang et al. [21] combined a nonradial distance DEA model with a bootstrapped nonradial Malmquist carbon emission performance index and further decomposed the Malmquist index into efficiency and technology change components. Li et al. [22] calculated Malmquist and GML index values to study the dynamic change in the carbon emission efficiency of the Chinese transportation industry. This paper referred to the research of Zhao and Sun [23]. The Malmquist index was decomposed into the GECH, technological change, GPECH change, and GSECH change for further analysis.

Considering that the minimum value of the TFP is 0, the data were truncated. If ordinary linear regression were used, the regression results would exhibit bias [24]. Therefore, scholars have usually selected the Tobit model to explore the influencing factors of TFP. Guo and Boa [25] combined the unexpected output SBM model with the Tobit model to measure the impact of human capital and industrial structure on carbon emission efficiency in 30 mainland China provinces. This paper also used the Tobit model for empirical analysis.

In previous research on the transportation industry, the SBM-DEA model has been directly employed to calculate the dynamic total factor carbon emission efficiency, but few scholars have considered the static and dynamic changes in carbon emission efficiency. Therefore, the key indicators affecting the TFP of the transportation industry remain unclear, and there are only general policies and very few targeted policies in the different regions of China. Therefore, this paper combined static and dynamic models to illustrate the changes in carbon emission efficiency and investigated the carbon emission efficiency of the transportation industry in the different regions.

3. Materials and Methods

3.1. Superefficiency SBM-DEA Model

This paper adopted a nonradial superefficiency SBM-DEA model with undesirable outputs to calculate the TFP of the Chinese transportation industry. It was assumed that there were N DMUs in the production process of transportation carbon emissions, and each DMU contained m production inputs, $s_1$ desirable outputs, and $s_2$ undesirable outputs. Referring to Tone [26], the model can be expressed as follows:

$$min\rho =\frac{\frac{1}{m}\left({{\displaystyle \sum }}_1^m\overline{x}/x_{ij}\right)}{\frac{1}{s_1+s_2}\left({{\displaystyle \sum }}_{r=1}^{s_1}\frac{{\overline{y}}^d}{y_{rj}^d}+{{\displaystyle \sum }}_{l=1}^{s_2}\frac{{\overline{y}}^u}{y_{lj}^u}\right)}$$

(1)

Subject to:

$$\begin{array}{c}\overline{x}\ge {{\displaystyle \sum }}_{j=1,j\ne 0}^n{x}_{ij}{\lambda }_j; i=1, \dots , m\hfill \\ {\overline{y}}^d\ge {{\displaystyle \sum }}_{j=1,j\ne 0}^n{y}_{rj}^d{\lambda }_j; r=1, \dots , s_1\hfill \\ {\overline{y}}^u\ge {{\displaystyle \sum }}_{j=1,j\ne 0}^n{y}_{lj}^u{\lambda }_j; l=1, \dots , s_2\hfill \\ \overline{x}\ge x_{ij};\hfill \\ {\overline{y}}^d\le y_{rj}^d;\hfill \\ {\overline{y}}^u\ge y_{lj}^u;\hfill \\ {\lambda }_j>0;\hfill \end{array}$$

where j denotes the different provinces; $x_{ij}$, $y_{rj}^d$, and $y_{lj}^u$ denote the i input factors, r desirable outputs, and l undesirable output factors, respectively; $\overline{x}$, ${\overline{y}}^d$, and ${\overline{y}}^u$ denote the slack variables of the inputs, desirable outputs, and undesirable outputs, respectively; and ${\lambda }_j$ is the weight vector.

3.2. Global Malmquist—Luenberger Index

Referring to the research of Li et al. [22], this paper decomposed the GML index into the GECH and technological change index (GTCH). GECH could be further decomposed into GPECH and GSECH. The GML index from periods t to t + 1 can be expressed as follows:

$$GM{L}_t^{t+1}=\frac{1+\overrightarrow{D_c}\left(x^t,y_d^t,y_u^t;y_d^t,-y_u^t\right)}{1+\overrightarrow{D_c}\left(x^{t+1},y_d^{t+1},y_u^{t+1};y_d^{t+1},-y_u^{t+1}\right)}$$

(2)

$$\begin{gathered} GM{L}_t^{t+1}=GEC{H}_t^{t+1}\times GTC{H}_t^{t+1} \\ =GPEC{H}_t^{t+1}\times GSEC{H}_t^{t+1}\times GTC{H}_t^{t+1} \\ \begin{array}{c}=\left[\frac{1+{\overrightarrow{D}}^c\left(x^t,y_d^t,y_u^t;y^t,-y_u^t\right)}{1+{\overrightarrow{D}}^t\left(x^t,y_d^t,y_u^t;y^t,-y_u^t\right)}\times \frac{1+{\overrightarrow{D}}^{t+1}\left(x^{t+1},y_d^{t+1},y_u^{t+1};y^{t+1},-y_u^{t+1}\right)}{1+{\overrightarrow{D}}^c\left(x^{t+1},y_d^{t+1},y_u^{t+1};y^{t+1},-y_u^{t+1}\right)}\right]\\ \times \left[\frac{1+{\overrightarrow{D}}^t\left(x^t,y_d^t,y_u^t;y_d^t,-y_u^t\right)}{1+{\overrightarrow{D}}^{t+1}\left(x^{t+1},y_d^{t+1},y_u^{t+1};y_d^{t+1},-y_u^{t+1}\right)}\right]|\end{array} \end{gathered}$$

(3)

where $x, y_d, \mathrm{and} y_u$ denote the input factors, desirable outputs, and undesirable outputs, respectively.

3.3. Data Source

This paper studied the TFP of the transportation sector in China from 2003 to 2019. Considering that the quality of the statistical data for Taiwan, Hong Kong, and Macao differed, we omitted these data and only considered the data for the 30 administrative regions. In the industrial classification of national economic activities of China, the transportation industry includes the storage and postal industries, which account for a small proportion and can be neglected.

The increase in the gross domestic product (GDP) of the transportation industry is not the only evaluation metric for the transportation industry. The most important functions of the transportation industry include goods transportation and passenger transfer. Therefore, the cargo and passenger turnover levels are also output indicators of the transportation industry [27]. Zhou et al. [28] chose three outputs, including two desirable outputs, i.e., the cargo and passenger turnover levels, and one undesirable output, i.e., transportation carbon emissions, to analyze the carbon dioxide emission performance of the transportation sector in China. We selected three inputs and four outputs, including three desirable outputs and one undesirable output.

The inputs included capital (K), labor (L), and energy consumption (E). Referring to Li and Zhang [29], the depreciation rate of transportation infrastructure was set to 8.76%. The price index of fixed asset investment was also converted to the base period of 2003. The labor input is the number of employees in the transportation industry. We selected eight fuels closely related to the transportation industry as energy inputs, namely, raw coal, crude oil, gasoline, kerosene, diesel, fuel oil, natural gas, and electricity, and we unified their consumption into standard coal.

The desirable outputs included the added value of the transportation industry (G), passenger turnover (T1), and cargo turnover (T2). All price indicators were calculated at constant 2003 prices. This paper considered three modes of transportation, namely, railway, highway, and waterway transportation, and calculated their respective freight and passenger volumes. Then, we converted the passenger turnover under each transportation mode into freight turnover according to a certain proportion to obtain the converted total turnover [30], and the conversion factor is provided in Appendix A 

Table A1. Conversion factor of the passenger and freight turnover levels under the different transportation modes.

Modes	Railway	Highway	Waterway
Conversion factor	1	1/10	1/3

Data source: Chinese Ministry of Transport.

.

The capital stock of the transportation industry was considered the capital input, and the frequently used method for capital stock calculation is the perpetual inventory method.

$$K_{it}=K_{it-1}\left(1-\delta \right)+I_{it}$$

(4)

where i denotes the various regions, and t denotes the year. K is the capital stock, $I_{it}$ denotes the fixed asset investment, and $\delta$ denotes the depreciation rate. The initial capital stock in each province in the base year can be calculated as follows [31]:

$$K_{i,2003}=\frac{I_{i,2003}\times K_{2003}}{\sum I_{i,2003}}$$

(5)

The carbon emissions in the transportation industry (C) were considered the undesirable output, which can be calculated as follows:

$$C={{\displaystyle \sum }}_{n=1}^{8}A{D}_n\times E{F}_n$$

(6)

where n denotes the energy type, AD denotes the energy consumption, and EF denotes the carbon dioxide emission factor. The correlation coefficients of the various fuels are provided in Appendix A 

Table A2. Carbon emission correlation coefficients calculated for the different fuels.

Fuels	Conversion Coefficient of Standard Coal (kg CE/kg, m3)	Average Low Calorific Value (KJ/kg)	Carbon Content per Unit Calorific Value (Ton Carbon/TJ)	Carbon Oxidation Rate	Carbon Dioxide Conversion Coefficient (kg CO2/kg, m3)
Raw coal	0.7143	20,908	26.37	0.94	1.9003
Crude oil	1.4286	41,816	20.1	0.98	3.0202
Gasoline	1.4714	43,070	18.9	0.98	2.9251
Kerosene	1.4714	43,070	19.5	0.98	3.0179
Diesel	1.4571	42,652	20.2	0.98	3.0959
Fuel oil	1.4286	41,816	21.1	0.98	3.1705
Natural gas	1.33	38,931	15.3	0.99	2.1622

Data source: Guidelines for the preparation of provincial greenhouse gas inventories ([2011] No. 1041); China carbon emission trading network.

. The carbon emission coefficient for electricity in the different regions varies, as summarized in Appendix A,

Table A3. Average carbon dioxide emission factor of the regional power grid in China in 2005.

Region	Provinces, Autonomous Regions, and Cities (30 Provinces, Excluding Tibet)	Carbon Dioxide Emission Coefficient of Electric Power (kg CO2/kw·h)
North China	Beijing, Tianjin, Hebei, Shanxi, Shandong, Western Inner Mongolia	1.246
Northeast	Liaoning, Jilin, Heilongjiang, Eastern Inner Mongolia	1.096
East China	Shanghai, Zhejiang, Jiangsu, Anhui, Fujian	0.928
Central China	Henan, Hubei, Hunan, Jiangxi, Sichuan, Chongqing	0.801
Northeast	Shanxi, Gansu, Qinghai, Ningxia, Xinjiang	0.977
South	Guangdong, Guangxi, Yunnan, Guizhou	0.714
Hainan	Hainan	0.917

Data source: Guidelines for the preparation of provincial greenhouse gas inventories ([2011] No. 1041).

.

In regard to the transportation industry, the added value of the transportation industry, freight turnover, and passenger turnover are all output forms. However, most scholars either only considered the added value of the transportation industry or only considered the freight and passenger turnover levels in the selection of expected output indicators, which could lead to inaccurate measurements of carbon emission efficiency. Therefore, this paper selected four outputs, including three desirable outputs, including the added value of the transportation industry, freight turnover, and passenger turnover, and one undesirable output, namely, the carbon emissions in the transportation industry, to more accurately calculate the TFP of the transportation industry using a nonradial superefficiency SBM-DEA model with undesirable outputs.

Referring to the research of the above scholars, the selected input—output indicators are summarized in

Table 1. Input and output indicators.

Variables	Indicators	Meaning	Unit
Inputs	Capital input	Capital stock of the transportation industry (K)	102 million yuan
	Labor input	Number of employees in the transportation industry (L)	10 thousand persons
	Energy input	Fuel consumption of the transportation industry (E)	Ton standard coal
Outputs	Desirable outputs	Added value of the transportation industry (G)	102 million yuan
		Passenger turnover (T1)	102 million person-km
		Cargo turnover (T2)	102 million ton-km
	Undesirable output	Carbon emissions of the transportation industry (C)	Ton

Source: Provided by the author.

. The required data were retrieved from the China Statistical Yearbook, China Energy Statistical Yearbook, and the National Bureau of Statistics of China. Based on the economic development level in the different regions and the practice of the National Bureau of Statistics in 2011 (National Bureau of Statistics. http://www.stats.gov.cn/ztjc/zthd/sjtjr/dejtjkfr/tjkp/201106/t20110613_71947.htm), the 30 provinces were divided into four regions: eastern, central, western, and northeastern regions (summarized in Appendix A,

Table A4. Division of the four regions.

Region	Provinces
East area	Beijing, Tianjin, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, Hainan
Central area	Shanxi, Anhui, Jiangxi, Henan, Hubei, Hunan
West area	Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Xizang, Shanxi, Gansu, Qinghai, Ningxia, Xinjiang
Northeast area	Liaoning, Jilin, Heilongjiang

Source: National Bureau of Statistics.

). Therefore, this paper analyzed China as a whole, and these four regions.

4. Carbon Emission Performance Results and Analysis

4.1. Static Carbon Emission Performance Analysis

We used a nonradial superefficiency SBM model with undesirable outputs to calculate the TFP of the transportation industry in the 30 provinces (cities and autonomous regions) of China. The results of the static carbon emission efficiency of the transportation industry are provided in Appendix B,

Table A5. TFP of the transportation industry from 2003 to 2019.

	Region	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019	Mean
E	Beijing	0.1604	0.1056	0.0889	0.0858	0.0742	0.1258	0.1087	0.1257	0.1258	0.1304	0.1272	0.1269	0.1514	0.1486	0.1573	0.1538	0.1564	0.1266
	Tianjin	1.1829	1.2298	1.2460	1.2417	1.2557	0.4240	1.1266	1.1127	1.1243	1.0678	0.4750	0.4444	0.3925	0.4142	0.3809	0.4221	0.4561	0.8233
	Hebei	1.2309	1.0769	1.0620	1.0490	1.0545	1.0637	1.0859	1.0984	1.1270	1.1600	1.1550	1.1952	1.2711	1.1894	1.3063	1.2897	1.2471	1.1566
	Shanghai	0.2699	0.2425	0.2497	0.2486	0.2276	1.0426	0.1385	0.2294	1.0067	1.0380	0.2144	1.0294	1.1209	1.1598	1.2233	1.2639	1.2601	0.7038
	Jiangsu	0.4011	0.3903	0.4218	0.4724	0.5198	0.4676	0.4753	0.5071	0.5740	0.6247	0.5786	0.5510	0.5286	0.5253	0.5415	0.5481	0.5753	0.5119
	Zhejiang	1.0143	1.0106	0.5270	0.5284	0.5370	0.4475	0.4341	0.4705	0.4801	0.4870	0.5032	0.4999	0.5721	0.5876	0.5351	0.5663	0.6254	0.5780
	Fujian	1.0057	1.0007	0.4833	0.5067	0.4512	0.4176	0.3809	0.3560	0.3507	0.3636	0.3582	0.3664	0.3969	0.4316	0.3874	0.4066	0.4434	0.4769
	Shandong	1.0367	0.6127	0.5201	0.6198	0.6086	1.0683	1.0437	1.0085	0.5616	0.5107	0.4717	0.4796	0.5058	0.5321	0.4967	0.5089	0.5353	0.6542
	Guangdong	0.4450	0.3953	0.3139	0.3164	0.3256	0.3375	0.3091	0.3097	0.3277	0.3939	0.3651	0.4531	0.4726	0.5848	0.6883	0.6702	0.6942	0.4354
	Hainan	0.3874	0.2548	0.3662	0.4049	0.3986	0.2838	0.2747	0.2874	0.3373	0.3642	0.2586	0.3411	0.3416	0.3340	0.3065	0.3147	0.4359	0.3348
C	Shanxi	0.3922	0.3520	0.3255	0.3260	0.3297	0.3757	0.2913	0.2864	0.2899	0.3025	0.3273	0.3249	0.3578	0.3846	0.4007	0.4893	0.6523	0.3652
	Anhui	1.1189	1.1351	1.2552	1.2006	1.1177	1.3345	1.2123	1.2434	1.2711	1.3089	1.3426	1.3404	1.2207	1.1648	1.0959	1.0838	1.0769	1.2072
	Jiangxi	0.3881	0.3888	0.3772	0.4142	0.4710	0.5428	0.5130	0.5031	0.5237	0.6215	1.0087	1.0170	1.0280	1.0378	1.0596	1.0760	1.0849	0.7091
	Henan	1.0163	0.4484	0.4438	0.4791	0.4995	0.5753	0.5962	0.5940	0.6146	0.6891	0.5955	0.6396	1.0010	1.0328	1.0166	1.0093	1.0223	0.7220
	Hubei	0.2824	0.2270	0.2057	0.2114	0.2041	0.2877	0.2956	0.3383	0.3524	0.3898	0.4061	0.4228	0.5063	0.4812	0.4815	0.4807	0.4809	0.3561
	Hunan	0.5528	1.0065	1.0462	1.0537	1.0577	0.5220	0.4279	0.4220	0.4183	0.4995	1.0216	1.0022	1.0305	1.0401	1.0494	1.0409	0.7407	0.8195
W	Inner Mongolia	0.4566	0.3284	0.2297	0.2587	0.2456	0.2842	0.2733	0.2617	0.2673	0.2601	0.2763	0.2741	0.3108	0.3687	0.3705	0.4042	0.3905	0.3094
	Guangxi	0.4277	0.3797	0.3807	0.3927	0.3879	0.3794	0.3530	0.3516	0.3706	0.3917	0.4151	0.3796	0.4899	0.5166	0.4667	0.4835	0.4689	0.4138
	Chongqing	0.3297	0.1890	0.1856	0.2240	0.2412	0.3193	0.3301	0.3194	0.3404	0.3148	0.2610	0.2750	0.3129	0.3150	0.2869	0.3193	0.3315	0.2879
	Sichuan	0.3148	0.2586	0.2374	0.2469	0.2338	0.2995	0.2457	0.2379	0.2587	0.2692	0.2873	0.2418	0.3013	0.2645	0.2363	0.2497	0.2592	0.2613
	Guizhou	0.3498	0.3606	0.3399	0.3443	0.3085	0.2903	0.2814	0.2678	0.2709	0.2812	0.3204	0.3229	0.4239	0.4559	0.5110	1.0174	1.0314	0.4222
	Yunnan	0.2395	0.3551	0.1515	0.1526	0.1549	0.1883	0.1872	0.1824	0.1864	0.2035	0.2240	0.2161	0.2607	0.2650	0.2440	0.2371	0.2242	0.2160
	Shanxi	0.3322	0.2941	0.2526	0.2480	0.2290	0.3072	0.3010	0.2973	0.3111	0.3431	0.3806	0.3833	0.4565	0.5126	0.4776	0.4789	0.4794	0.3579
	Gansu	0.3727	0.3644	0.3639	0.4261	0.5128	0.5246	0.5115	0.4990	0.5360	0.5578	0.5364	0.4973	0.5778	0.5850	0.5913	1.0084	1.0239	0.5582
	Qinghai	0.2781	0.2884	0.2316	0.2399	0.1782	0.2363	0.2454	0.2721	0.2791	0.2873	0.2638	0.2775	0.3137	0.3157	0.2647	0.2629	0.2259	0.2624
	Ningxia	0.3493	0.3986	0.2473	0.2748	0.2756	0.4666	0.4901	0.5225	0.5547	1.0005	0.5436	0.5130	0.5246	0.5086	0.4234	0.3813	0.3981	0.4631
	Xinjiang	0.3575	0.3101	0.2341	0.2568	0.2612	0.3005	0.2791	0.2659	0.2711	0.2981	0.3162	0.2983	0.3312	0.3362	0.3194	0.3437	0.3351	0.3009
NE	Liaoning	0.4972	0.4368	0.3653	0.3608	0.3848	0.3745	0.3735	0.3735	0.3793	0.3955	0.4335	0.4199	0.5363	0.6144	0.6823	0.7095	1.0208	0.4917
	Jilin	0.3417	0.2945	0.2099	0.2019	0.1758	0.3251	0.3131	0.3101	0.3373	0.3543	0.3644	0.3471	0.3555	0.3748	0.3576	0.3865	0.4026	0.3207
	Heilongjiang	0.3033	0.2958	0.2724	0.2626	0.2580	0.2739	0.2354	0.2443	0.2076	0.2212	0.2187	0.2223	0.2471	0.2522	0.2648	0.2782	0.2834	0.2554
National average		0.5278	0.4810	0.4211	0.4350	0.4327	0.4629	0.4378	0.4433	0.4685	0.5043	0.4683	0.4967	0.5447	0.5578	0.5541	0.5962	0.6121	0.4967
Eastern Mean		0.7134	0.6319	0.5279	0.5474	0.5453	0.5678	0.5377	0.5505	0.6015	0.6140	0.4507	0.5487	0.5753	0.5907	0.6023	0.6144	0.6429	0.5802
Central Mean		0.6251	0.5930	0.6089	0.6142	0.6133	0.6063	0.5561	0.5645	0.5783	0.6352	0.7836	0.7912	0.8574	0.8569	0.8506	0.8633	0.8430	0.6965
Western Mean		0.3462	0.3206	0.2595	0.2786	0.2753	0.3269	0.3180	0.3161	0.3315	0.3825	0.3477	0.3344	0.3912	0.4040	0.3811	0.4715	0.4698	0.3503
Northeast Mean		0.3807	0.3424	0.2826	0.2751	0.2729	0.3245	0.3073	0.3093	0.3080	0.3237	0.3389	0.3298	0.3797	0.4138	0.4349	0.4581	0.5699	0.3559

Note: E denotes the eastern region, C represents the central region, W represents the western region, and NE denotes the northeastern region.

.

The national average TFP value of the transportation industry varied between 0.4 and 0.5, exhibiting an increasing trend in recent years. The average TFP value of the transportation industry in the eastern, central, western, and northeastern regions in each year is shown in Figure 1. It could be concluded that the TFP of the transportation industry in China maintained a stable trend as a whole and exhibited a gradual upward trend in recent years. The TFP value in the eastern region significantly decreased in 2003 and rebounded in 2005, which basically coincides with the trend of the carbon emission efficiency in the central region. Since 2012, the TFP value in the eastern region first exhibited a downward trend and then remained stable. The TFP value in the western region basically coincided with that in the northeastern region, and the fluctuation trend was the same. However, the TFP in the western region maintained an upward trend in 2017 after a sudden decline in 2016. Moreover, the TFP in the central and western regions decreased in 2019. The average value in the central region was higher than the national average, reaching a maximum of 0.7134 in 2003 and a minimum of 0.4507 in 2013. The average value in the central region ranked first among the four regions, and the TFP maintained an upward trend. The carbon emission efficiency values of the transportation industry in the western and northeastern regions were similar, and the upward and downward trends were also consistent.

Based on the efficiency value in the various provinces and cities, only the average carbon emission efficiency of the transportation industry in Hebei and Anhui was greater than 1 from 2003 to 2019, indicating that the TFP index in Hebei and Anhui occurred along the effective frontier. The TFP in Tianjin, Shanghai, Jiangxi, Henan, and Hunan varied between 0.7 and 0.9, while the value in the other provinces was below 0.7, indicating that there remains much room for improvement in the TFP of the transportation sector in these areas. The TFP of the urban transportation sector has not reached the optimal state.

4.2. Dynamic Global Malmquist—Luenberger Index Analysis

4.2.1. Dynamic Change Analysis of the National Transportation Industry

We decomposed the GML (GML index takes 1 as the critical value. If it is greater than 1, the efficiency will increase year on year; if it is less than 1, the efficiency will decrease year on year) index of the transportation industry in China from 2003 to 2019 into four indicators for further analysis, as summarized in

Table 2. GMCPI and decomposition index values of the national transportation industry.

Year	GECH	GECH		GTCH	GML
		GPECH	GSECH
2003–2004	0.9231	0.8911	1.1075	0.9714	0.8821
2004–2005	0.8817	1.0282	0.9122	1.0621	0.9095
2005–2006	1.0442	1.0254	1.0200	0.8983	0.9379
2006–2007	0.9857	0.9600	1.0518	1.0187	1.0014
2007–2008	1.2547	1.1140	1.1692	1.0299	1.1498
2008–2009	0.9816	0.9942	1.0069	1.1608	0.9559
2009–2010	1.0321	0.9988	1.0410	1.0078	1.0328
2010–2011	1.1313	1.0201	1.1108	0.9889	1.0200
2011–2012	1.0806	1.0448	1.0358	0.9462	1.0122
2012–2013	0.9925	1.0417	0.9718	0.9125	0.7714
2013–2014	1.1332	0.9859	1.1535	0.9625	0.9967
2014–2015	1.1228	1.1607	1.0001	0.8189	0.9111
2015–2016	1.0325	1.0123	1.0287	0.9383	0.9677
2016–2017	0.9809	1.0273	0.9801	1.0449	1.0213
2017–2018	1.0856	1.0737	1.0349	0.9592	1.0191
2018–2019	1.0401	1.0230	1.0177	0.9214	0.9535

Source: Provided by the author.

. From 2007 to 2008, the GML index was the highest, and the GML index was the lowest from 2012 to 2013. From 2003 to 2008, the GML index gradually increased to the maximum, indicating that the TFP of the transportation industry in these years gradually improved and then showed a fluctuating trend. Fluctuations and changes may be due to the combined effect of the economic development level, technical level, and other factors, resulting in annual differences in carbon emission efficiency.

The GTCH value in the periods 2004–2005, 2006–2010, and 2016–2017 was greater than 1, indicating that technological progress in the transportation industry played a positive role in promoting a change in the efficiency of the overall transportation industry in China in recent years, which shows that the CO₂ emission efficiency has been improved by improving the technical level.

In terms of technical efficiency index decomposition, GSECH was higher than GPECH, indicating that the development scale of the overall transportation industry in China was suitable, driving improvement in the carbon emission efficiency of the transportation industry. This shows that GSECH has a higher effect on the carbon emission efficiency of the transportation industry than GPECH. At the same time, the GSECH on carbon emission efficiency is also different each year, so the index value shows fluctuations.

Source: Provided by the author.

We calculated the average values of the GML and decomposition indices in China overall and in the four regions from 2003 to 2019, as summarized in

Table 3. Average GML and decomposition index values of the regional transportation industry.

Regions	GECH	GECH		GTCH	GML
		GPECH	GSECH
Eastern	1.0691	1.0081	1.0700	1.0150	0.9642
Central	1.0423	1.0438	1.0134	0.9606	0.9807
Western	1.0263	1.0284	1.0385	0.9615	0.9749
Northeast	1.0278	1.0322	1.0000	0.9459	0.9638
Whole country	1.0439	1.0251	1.0401	0.9776	0.9714

Source: Provided by the author.

. Table 3 reveals that the average GML value in the northeastern region was the lowest, while that in the central region was the highest. Moreover, the GPECH and GSECH values in the four regions were greater than 1, indicating that both GPECH and GSECH can improve the carbon emission efficiency of transportation. The GTCH value in the eastern region was greater than 1. However, GTCH is less than 1 in the other three regions, indicating that technological progress has improved in the eastern region but not in other regions. In regard to the GECH decomposition index, the GPECH value in the central region was the highest, while the GSECH value in the eastern region was the highest. The economic development of the eastern region was inseparable from transportation infrastructure optimization, so the development scale of the transportation industry was larger than that in the other regions. Due to the geographical location, the proportion of railway transportation was very large in the central region.

4.2.2. Dynamic Change Analysis of the Subregional Transportation Industry

In this section, we analyze the changes in each decomposed indicator in the four regions from 2003 to 2019. Figure 2 shows the GML of the transportation industry in the four regions from 2003 to 2019. The value in the eastern area fluctuated near approximately 1 but suddenly fell to its lowest point from 2012 to 2013 and then recovered to reach a stable level. The GML of the transportation industry in the central region greatly fluctuated, reaching a peak from 2007 to 2008 and its lowest point from 2012 to 2013. The change trends in the western and northeastern regions were basically the same, reaching a peak from 2007 to 2008, but the decline range in the northeastern region from 2012 to 2013 was narrower than that in the western region, and the value fluctuated near approximately 1. The index in the eastern region gradually changed from indicating an invalid state to reflecting an effective state. However, the GML index in the central, western, and northeastern regions demonstrated that the carbon emission efficiency of the transportation industry in these three regions fluctuated.

Figure 3 and Figure 4 show the GECH and GTCH values of the subregional transportation industry from 2003 to 2019. The GECH value in the eastern region revealed the largest fluctuation range. It exhibited significant declines in the periods 2008–2009, 2011–2013, and 2014–2015 and reached its lowest point from 2012 to 2013 and its highest point from 2013 to 2014. Due to the high economic level, high population density, and large scale of the transportation industry in the eastern region, the GECH value greatly changed. The value in the central region reached its highest point from 2012 to 2013. In the western and northeastern regions, the fluctuation trends of the GECH values were basically the same, and the GTCH value fluctuated near approximately 1. This shows that the low TFP may be related to the development level of the region. The western region is relatively backward in technology due to the restrictions of unbalanced development. Compared with the eastern and central regions, the western and northeastern regions are relatively rich in energy and are less likely to be affected, so the change trend is basically the same.

The GTCH value of the transportation industry in the eastern area from 2003 to 2014 was greater than 1, but the GTCH change value in recent years was less than 1. This may be because the GTCH in the eastern region was in a period of rapid progress from 2003 to 2014 and gradually reached a relatively stable stage, which slowed the pace of improvement and showed a retrogressive trend to the TFP. The value in the central region exhibited a trend of first rising and then falling, and the GTCH value was less than 1. GTCH refers to technological change, which is mainly related to the input/output ratio within the transportation industry. This suggests that the GTCH of the transportation industry in the four regions still exhibits much room for improvement.

The values of the decomposition indices GPECH and GSECH of GECH are shown in Figure 5 and Figure 6, respectively.

The GPECH value in the eastern area remained basically constant at approximately 1, exhibiting a significant decline from 2012 to 2013. Additionally, the GPECH value in the central area fluctuated at approximately 1, with the lowest point from 2008 to 2009 and the highest value from 2012 to 2013. The development of the central region was relatively slow, which may be driven by the “spillover effect” of the eastern region, so it is shown as the “follower” of the eastern region. The overall performance of GPECH in the western region indicated an upward trend, and the maximum value was achieved from 2014 to 2015. From 2003 to 2005, the value in the central and western regions first decreased, then increased, and fluctuated at approximately 1. The GPECH in the northeastern area was less than 1 before 2007, decreased after reaching its highest point in 2008, and then maintained an upward trend, which indicates that the GPECH of the transportation industry was constantly improving. This may have driven the improvement of GPECH by increasing capital investment in Northeast China, introducing foreign capital, talent, and other strategies.

The GSECH value in the eastern area continued to rise from 2003 to 2008, reached a maximum value from 2007 to 2008, greatly fluctuated between 2008 and 2015, and reached a minimum value from 2008 to 2009, after which it stabilized. The rapid development of the economic level in the eastern region has led to the improvement of the GSECH of the transportation industry, so it rose rapidly from 2003 to 2008. During 2011–2015, China implemented a low-carbon policy, and the eastern region was the first to respond. However, due to the imperfect technology level at the initial stage of the low-carbon policy, the pure technical efficiency of the transportation industry shows a downward trend. The GSECH value in the northeastern area reached approximately 1, with a narrow fluctuation range, which may have occurred because the scale of the transportation industry in the northeastern area remained relatively constant. The GSECH of the transportation industry is related to the level of local economic development. The GSECH value in the eastern region was the highest, and its change was the most obvious.

4.2.3. Dynamic Change Analysis of the Provincial Transportation Industry

By calculating the GML index, we could analyze the dynamic change in the transportation industry in 30 provinces (cities and autonomous regions) from 2003 to 2019, as summarized in Appendix B,

Table A6. GML index in the four regions of the 30 provinces.

	Region	2003–2004	2004–2005	2005–2006	2006–2007	2007–2008	2008–2009	2009–2010	2010–2011	2011–2012	2012–2013	2013–2014	2014–2015	2015–2016	2016–2017	2017–2018	2018–2019	Mean
E	Beijing	0.8169	0.8536	0.8393	0.9102	0.9175	0.8933	1.1043	1.0004	0.9847	0.9663	0.9388	0.8573	0.8926	1.0803	0.9950	1.0372	0.9430
	Tianjin	1.0340	0.9942	0.8163	1.2175	0.3392	1.6275	0.9540	1.0178	0.7437	0.6751	0.9569	0.8244	0.9682	0.9997	1.0731	0.9690	0.9507
	Hebei	0.8447	0.6961	0.9970	1.4409	0.9957	0.8997	1.0373	1.0857	0.9903	0.7194	0.9566	0.9057	0.8803	1.2177	0.9372	0.8964	0.9688
	Shanghai	0.9986	0.9667	0.9171	0.8590	0.8594	0.8902	1.1498	0.9123	1.0689	0.8769	1.0806	0.9800	0.9340	1.0354	1.0061	1.0314	0.9729
	Jiangsu	0.9707	1.0961	1.0637	1.0493	0.8993	1.0203	1.1031	1.1388	1.0510	0.6784	0.9502	0.8532	0.9195	1.0468	0.9964	0.9752	0.9883
	Zhejiang	0.8925	0.9569	1.0190	1.0065	0.9387	0.9540	1.0617	0.9722	0.9364	0.8282	0.9772	0.9770	0.9729	0.9791	1.0612	1.0176	0.9719
	Fujian	0.3857	0.9664	1.0341	1.0524	0.8522	0.8850	0.9655	0.9637	1.0020	0.8017	1.0283	0.9516	1.0172	1.0009	1.0113	0.9926	0.9319
	Shandong	0.5102	0.8294	1.0179	0.9766	1.5933	0.7403	0.9148	0.9510	0.8877	0.6761	1.0127	0.9804	0.9853	1.0488	1.0076	1.0308	0.9477
	Guangdong	0.8807	0.7906	0.9164	0.9924	0.9105	0.9334	1.0612	1.0821	1.2414	0.6932	1.2144	0.8527	1.1005	1.0807	0.9740	0.9952	0.9825
	Hainan	0.7428	1.3463	0.9990	0.9900	0.8030	0.9787	1.0248	1.0812	1.0387	0.5546	1.2973	0.7997	0.9273	0.9468	0.9680	1.2580	0.9848
C	Shanxi	0.8643	0.8895	0.9228	1.0046	0.9952	0.8263	1.0415	0.9924	0.9869	0.8574	0.9610	1.0031	0.9912	1.0747	1.1573	1.0563	0.9765
	Anhui	0.7903	0.9021	0.9015	0.8985	1.6306	0.9850	1.0110	0.9592	1.0819	0.5881	0.9874	0.7688	0.9316	0.9251	0.9926	0.9496	0.9565
	Jiangxi	1.0700	0.9476	0.9688	1.0087	1.7090	0.9262	0.9761	1.0330	1.0927	0.7230	0.9885	0.8944	0.9938	1.0004	0.9845	0.9335	1.0156
	Henan	0.3849	0.9781	0.9595	1.0027	1.4898	1.0414	1.0089	0.9968	0.9940	0.6367	1.0360	0.8514	0.9875	0.9961	1.0586	0.9581	0.9613
	Hubei	0.9340	0.8659	0.9212	0.9475	1.2976	0.9879	1.2459	1.0809	1.1194	0.8003	1.0373	0.9591	0.8990	1.0131	0.9887	0.9870	1.0053
	Hunan	1.0408	1.8160	0.4841	0.9057	1.0407	0.7617	1.0598	0.9848	1.1221	0.8130	0.9864	0.8754	0.9762	1.0149	0.9962	0.6294	0.9692
W	Inner Mongolia	0.6590	0.7428	1.0330	0.9907	1.1285	0.9425	0.9723	1.0058	0.9664	0.7875	0.9858	0.9675	1.0539	1.0898	1.0430	0.9236	0.9558
	Guangxi	0.8964	0.8470	0.9066	0.9504	1.2629	0.9457	1.0560	1.0582	1.0241	0.7238	0.9223	1.0320	0.9904	1.0002	1.0136	0.9261	0.9722
	Chongqing	0.6349	1.0072	1.1051	1.1119	1.1481	1.0387	1.0141	1.0783	0.9180	0.6244	1.0237	0.9228	0.9386	0.9584	1.1348	0.9978	0.9785
	Sichuan	0.9175	0.8891	0.8699	0.8941	1.3368	0.8061	1.0484	1.1228	0.9997	0.8134	0.8375	0.9789	0.8484	1.0267	1.0853	0.9930	0.9667
	Guizhou	1.0024	0.9104	0.9223	0.9200	0.9375	1.0049	0.9764	1.0114	1.0804	0.8468	1.0433	0.8917	0.9720	1.1045	1.0773	0.7292	0.9644
	Yunnan	1.6835	0.4610	0.9281	1.0735	0.9193	0.9795	1.0190	1.0493	1.0666	0.9267	0.9724	1.0214	0.9995	1.0445	1.0204	0.8546	1.0012
	Shanxi	0.9331	0.8699	0.9141	0.9407	1.3402	0.9616	1.0068	1.0495	1.0804	0.8534	0.9689	0.9215	1.0310	0.9954	0.9763	0.9650	0.9880
	Gansu	0.9781	0.9925	1.0327	1.0941	1.2664	0.9434	0.9969	1.0666	0.9849	0.7215	0.9186	0.9022	0.9197	1.0053	0.9884	1.0259	0.9898
	Qinghai	0.9353	0.8405	0.9400	0.8265	1.4314	0.9968	1.0436	1.0155	0.9670	0.7138	1.0523	0.9375	0.9500	0.9555	0.9768	0.7538	0.9585
	Ningxia	1.1639	0.6728	0.9985	1.0140	1.6027	0.9850	1.0552	1.0105	0.9505	0.7763	0.9546	0.9262	0.9434	0.8989	0.8629	0.9732	0.9868
	Xinjiang	0.9431	0.7518	1.0108	1.0336	1.1853	0.9306	0.9737	0.9778	1.0356	0.8092	0.9648	0.8710	0.9616	0.9721	1.0896	0.8816	0.9620
NE	Liaoning	0.7608	0.8093	0.9439	1.0422	1.0370	0.9810	0.9994	0.9943	0.9991	0.8981	0.9638	0.9579	1.0659	1.0089	0.9775	0.9394	0.9612
	Jilin	0.8835	0.7497	0.8844	0.9016	1.4711	0.9450	1.0325	1.0842	0.9735	0.9041	0.9585	0.8272	1.0238	1.0488	1.0760	0.9653	0.9831
	Heilongjiang	0.9116	0.8443	0.8706	0.9867	1.1559	0.8638	1.0705	0.8229	0.9792	0.8541	0.9259	0.8406	0.9566	1.0693	1.0429	0.9607	0.9472
National average		0.8821	0.9095	0.9379	1.0014	1.1498	0.9559	1.0328	1.0200	1.0122	0.7714	0.9967	0.9111	0.9677	1.0213	1.0191	0.9535	0.9714
Eastern Mean		0.8077	0.9496	0.9620	1.0495	0.9109	0.9822	1.0376	1.0205	0.9945	0.7470	1.0413	0.8982	0.9598	1.0436	1.0030	1.0203	0.9642
Central Mean		0.8474	1.0665	0.8597	0.9613	1.3605	0.9214	1.0572	1.0079	1.0662	0.7364	0.9994	0.8920	0.9632	1.0040	1.0296	0.9190	0.9807
Western Mean		0.9770	0.8168	0.9692	0.9863	1.2326	0.9577	1.0148	1.0405	1.0067	0.7815	0.9677	0.9430	0.9644	1.0047	1.0244	0.9112	0.9749
Northeast Mean		0.8519	0.8011	0.8996	0.9768	1.2213	0.9299	1.0342	0.9671	0.9839	0.8854	0.9494	0.8753	1.0154	1.0423	1.0321	0.9552	0.9638

Note: E denotes the eastern region, C represents the central region, W represents the western region, and NE denotes the northeastern region.

.

From the perspective of the temporal trend, the GML index values in China and the four regions first increased, then decreased, and again increased. Among them, the average values from 2007 to 2008 and 2009 to 2012 were greater than 1, indicating that during these periods, the GML index of the transportation industry in all regions exhibited a growth trend, followed by a decline. After 2016, the index began to show a growth trend. Among the four regions, there was a slight difference in the average GML value between the eastern and northeastern regions. The GML value in the central region ranked first, which indicates that the increase rate of the TFP of the transportation industry in the central region was the highest.

From the perspective of the regional trend, the GML values in Jiangxi, Hubei, and Shanxi were greater than 1, indicating that the GML index in this region increased from 2003 to 2019, while the TFP of the transportation sector in the other provinces declined.

We further decomposed the GML index into GECH and GTCH, as provided in Appendix B,

Table A7. GECH index of the transportation industry in the four regions of the 30 provinces.

GECH	Region	2003–2004	2004–2005	2005–2006	2006–2007	2007–2008	2008–2009	2009–2010	2010–2011	2011–2012	2012–2013	2013–2014	2014–2015	2015–2016	2016–2017	2017–2018	2018–2019	Mean
E	Beijing	0.6588	0.8416	0.9650	0.8649	1.6947	0.8644	1.1559	1.0010	1.0368	0.9754	0.9977	1.1933	0.9812	1.0586	0.9779	1.0165	1.0177
	Tianjin	1.0396	1.0132	0.9966	1.0112	0.3377	2.6570	0.9877	1.0104	0.9498	0.4448	0.9356	0.8833	1.0553	0.9195	1.1082	1.0805	1.0269
	Hebei	0.8749	0.9862	0.9878	1.0052	1.0087	1.0208	1.0116	1.0261	1.0292	0.9957	1.0348	1.0635	0.9358	1.0983	0.9873	0.9670	1.0021
	Shanghai	0.8985	1.0296	0.9956	0.9155	4.5819	0.1329	1.6560	4.3882	1.0311	0.2065	4.8024	1.0889	1.0347	1.0547	1.0332	0.9970	1.6154
	Jiangsu	0.9731	1.0806	1.1199	1.1004	0.8997	1.0164	1.0670	1.1318	1.0884	0.9261	0.9524	0.9593	0.9937	1.0309	1.0122	1.0496	1.0251
	Zhejiang	0.9963	0.5215	1.0028	1.0162	0.8334	0.9700	1.0838	1.0203	1.0143	1.0333	0.9934	1.1444	1.0271	0.9107	1.0584	1.1043	0.9831
	Fujian	0.9950	0.4830	1.0483	0.8905	0.9256	0.9120	0.9348	0.9850	1.0367	0.9854	1.0228	1.0832	1.0875	0.8976	1.0495	1.0905	0.9642
	Shandong	0.5910	0.8488	1.1917	0.9820	1.7552	0.9770	0.9663	0.5569	0.9093	0.9236	1.0169	1.0546	1.0518	0.9335	1.0246	1.0519	0.9897
	Guangdong	0.8883	0.7942	1.0078	1.0292	1.0364	0.9158	1.0019	1.0582	1.2022	0.9269	1.2411	1.0430	1.2374	1.1770	0.9737	1.0358	1.0355
	Hainan	0.6578	1.4370	1.1056	0.9844	0.7122	0.9677	1.0464	1.1738	1.0796	0.7102	1.3188	1.0014	0.9778	0.9176	1.0269	1.3849	1.0314
C	Shanxi	0.8975	0.9249	1.0015	1.0112	1.1396	0.7754	0.9829	1.0123	1.0435	1.0820	0.9928	1.1012	1.0749	1.0418	1.2210	1.3332	1.0397
	Anhui	1.0144	1.1058	0.9565	0.9309	1.1940	0.9084	1.0256	1.0223	1.0297	1.0258	0.9984	0.9107	0.9542	0.9408	0.9889	0.9937	1.0000
	Jiangxi	1.0020	0.9701	1.0979	1.1373	1.1524	0.9451	0.9807	1.0410	1.1867	1.6230	1.0083	1.0108	1.0095	1.0210	1.0155	1.0083	1.0756
	Henan	0.4412	0.9897	1.0797	1.0426	1.1516	1.0364	0.9962	1.0348	1.1210	0.8642	1.0741	1.5649	1.0318	0.9843	0.9929	1.0129	1.0261
	Hubei	0.8038	0.9064	1.0277	0.9654	1.4094	1.0276	1.1444	1.0418	1.1059	1.0420	1.0410	1.1975	0.9505	1.0006	0.9985	1.0004	1.0414
	Hunan	1.8208	1.0394	1.0072	1.0038	0.4936	0.8196	0.9862	0.9912	1.1943	2.0452	0.9810	1.0282	1.0093	1.0089	0.9919	0.7116	1.0708
W	Inner Mongolia	0.7193	0.6994	1.1262	0.9494	1.1572	0.9617	0.9575	1.0212	0.9731	1.0623	0.9920	1.1342	1.1860	1.0050	1.0909	0.9659	1.0001
	Guangxi	0.8876	1.0027	1.0315	0.9879	0.9779	0.9305	0.9960	1.0542	1.0569	1.0597	0.9146	1.2905	1.0545	0.9034	1.0360	0.9698	1.0096
	Chongqing	0.5733	0.9823	1.2070	1.0764	1.3242	1.0337	0.9675	1.0659	0.9247	0.8292	1.0534	1.1378	1.0067	0.9111	1.1128	1.0381	1.0153
	Sichuan	0.8214	0.9181	1.0398	0.9471	1.2806	0.8206	0.9679	1.0876	1.0407	1.0673	0.8415	1.2461	0.8777	0.8935	1.0568	1.0381	0.9966
	Guizhou	1.0308	0.9427	1.0128	0.8960	0.9409	0.9693	0.9518	1.0118	1.0379	1.1393	1.0079	1.3128	1.0754	1.1209	1.9908	1.0138	1.0909
	Yunnan	1.4827	0.4265	1.0074	1.0151	1.2159	0.9940	0.9743	1.0222	1.0918	1.1004	0.9647	1.2064	1.0165	0.9207	0.9719	0.9456	1.0223
	Shanxi	0.8853	0.8588	0.9817	0.9234	1.3418	0.9797	0.9879	1.0464	1.1027	1.1092	1.0071	1.1910	1.1230	0.9318	1.0026	1.0009	1.0296
	Gansu	0.9777	0.9986	1.1710	1.2035	1.0230	0.9750	0.9756	1.0741	1.0407	0.9617	0.9271	1.1619	1.0124	1.0107	1.7055	1.0154	1.0771
	Qinghai	1.0369	0.8031	1.0361	0.7427	1.3263	1.0385	1.1087	1.0258	1.0291	0.9185	1.0517	1.1307	1.0063	0.8386	0.9931	0.8591	0.9966
	Ningxia	1.1409	0.6204	1.1113	1.0031	1.6928	1.0503	1.0661	1.0616	1.8037	0.5434	0.9436	1.0227	0.9694	0.8325	0.9006	1.0442	1.0504
	Xinjiang	0.8673	0.7548	1.0972	1.0172	1.1505	0.9288	0.9528	1.0195	1.0996	1.0606	0.9435	1.1103	1.0151	0.9500	1.0760	0.9749	1.0011
NE	Liaoning	0.8784	0.8365	0.9876	1.0665	0.9734	0.9973	0.9999	1.0155	1.0429	1.0960	0.9686	1.2773	1.1456	1.1105	1.0398	1.4388	1.0547
	Jilin	0.8620	0.7129	0.9617	0.8706	1.8494	0.9631	0.9906	1.0875	1.0506	1.0285	0.9525	1.0242	1.0543	0.9540	1.0809	1.0416	1.0303
	Heilongjiang	0.9754	0.9209	0.9639	0.9826	1.0614	0.8595	1.0377	0.8497	1.0655	0.9887	1.0166	1.1115	1.0208	1.0496	1.0507	1.0186	0.9983
National average		0.9231	0.8817	1.0442	0.9857	1.2547	0.9816	1.0321	1.1313	1.0806	0.9925	1.1332	1.1228	1.0325	0.9809	1.0856	1.0401	1.0439
Eastern Mean		0.8573	0.9036	1.0421	0.9799	1.3786	1.0434	1.0911	1.3352	1.0377	0.8128	1.4316	1.0515	1.0382	0.9998	1.0252	1.0778	1.0691
Central Mean		0.9966	0.9894	1.0284	1.0152	1.0901	0.9188	1.0194	1.0239	1.1135	1.2804	1.0159	1.1355	1.0050	0.9996	1.0348	1.0100	1.0423
Western Mean		0.9476	0.8189	1.0747	0.9783	1.2210	0.9711	0.9915	1.0446	1.1092	0.9865	0.9679	1.1767	1.0312	0.9380	1.1761	0.9878	1.0263
Northeast Mean		0.9052	0.8234	0.9711	0.9732	1.2947	0.9400	1.0094	0.9843	1.0530	1.0377	0.9793	1.1377	1.0736	1.0380	1.0572	1.1664	1.0278

Note: E denotes the eastern region, C represents the central region, W represents the western region, and NE denotes the northeastern region.

and

Table A8. GTCH index of the transportation industry in the four regions of the 30 provinces.

GTCH	Region	2003–2004	2004–2005	2005–2006	2006–2007	2007–2008	2008–2009	2009–2010	2010–2011	2011–2012	2012–2013	2013–2014	2014–2015	2015–2016	2016–2017	2017–2018	2018–2019	Mean
E	Beijing	1.2400	1.0143	0.8697	1.0523	0.5414	1.0334	0.9553	0.9994	0.9498	0.9907	0.9410	0.7184	0.9097	1.0205	1.0174	1.0203	0.9546
	Tianjin	0.9946	0.9812	0.8191	1.2040	1.0045	0.6125	0.9659	1.0074	0.7830	1.5177	1.0227	0.9333	0.9175	1.0872	0.9683	0.8968	0.9822
	Hebei	0.9655	0.7058	1.0093	1.4334	0.9871	0.8814	1.0254	1.0582	0.9622	0.7225	0.9244	0.8517	0.9407	1.1087	0.9493	0.9270	0.9658
	Shanghai	1.1115	0.9389	0.9212	0.9383	0.1876	6.6996	0.6943	0.2079	1.0367	4.2462	0.2250	0.9000	0.9026	0.9817	0.9738	1.0345	1.3750
	Jiangsu	0.9976	1.0144	0.9498	0.9536	0.9996	1.0038	1.0339	1.0062	0.9657	0.7326	0.9977	0.8894	0.9253	1.0154	0.9844	0.9291	0.9624
	Zhejiang	0.8958	1.8350	1.0162	0.9905	1.1262	0.9835	0.9795	0.9528	0.9232	0.8015	0.9837	0.8537	0.9472	1.0751	1.0027	0.9215	1.0180
	Fujian	0.3876	2.0010	0.9864	1.1818	0.9207	0.9704	1.0329	0.9783	0.9666	0.8136	1.0054	0.8785	0.9354	1.1151	0.9636	0.9103	1.0030
	Shandong	0.8633	0.9771	0.8542	0.9944	0.9077	0.7577	0.9467	1.7078	0.9762	0.7320	0.9959	0.9296	0.9367	1.1235	0.9834	0.9799	0.9791
	Guangdong	0.9915	0.9954	0.9093	0.9643	0.8785	1.0193	1.0591	1.0226	1.0326	0.7478	0.9785	0.8176	0.8894	0.9182	1.0004	0.9608	0.9491
	Hainan	1.1293	0.9369	0.9035	1.0057	1.1275	1.0114	0.9794	0.9212	0.9621	0.7809	0.9836	0.7985	0.9484	1.0318	0.9426	0.9084	0.9607
C	Shanxi	0.9630	0.9617	0.9214	0.9935	0.8733	1.0657	1.0596	0.9803	0.9458	0.7924	0.9679	0.9109	0.9221	1.0316	0.9479	0.7923	0.9456
	Anhui	0.7791	0.8158	0.9425	0.9652	1.3657	1.0843	0.9857	0.9383	1.0507	0.5733	0.9890	0.8442	0.9764	0.9833	1.0037	0.9557	0.9533
	Jiangxi	1.0679	0.9768	0.8824	0.8870	1.4830	0.9801	0.9953	0.9923	0.9208	0.4455	0.9803	0.8849	0.9844	0.9798	0.9695	0.9258	0.9597
	Henan	0.8725	0.9883	0.8887	0.9617	1.2937	1.0048	1.0127	0.9632	0.8866	0.7367	0.9645	0.5440	0.9570	1.0120	1.0662	0.9459	0.9437
	Hubei	1.1620	0.9552	0.8963	0.9814	0.9207	0.9614	1.0887	1.0376	1.0122	0.7680	0.9964	0.8009	0.9458	1.0126	0.9902	0.9866	0.9698
	Hunan	0.5716	1.7471	0.4806	0.9023	2.1085	0.9294	1.0746	0.9935	0.9395	0.3975	1.0055	0.8514	0.9672	1.0059	1.0043	0.8844	0.9915
W	Inner Mongolia	0.9162	1.0620	0.9173	1.0435	0.9752	0.9801	1.0154	0.9849	0.9932	0.7414	0.9937	0.8530	0.8886	1.0843	0.9560	0.9562	0.9601
	Guangxi	1.0099	0.8447	0.8790	0.9620	1.2914	1.0164	1.0602	1.0038	0.9690	0.6830	1.0084	0.7997	0.9392	1.1071	0.9784	0.9549	0.9692
	Chongqing	1.1075	1.0254	0.9156	1.0330	0.8670	1.0049	1.0482	1.0116	0.9928	0.7530	0.9718	0.8111	0.9324	1.0520	1.0197	0.9612	0.9692
	Sichuan	1.1169	0.9684	0.8366	0.9440	1.0438	0.9823	1.0831	1.0323	0.9606	0.7621	0.9953	0.7856	0.9667	1.1491	1.0270	0.9565	0.9757
	Guizhou	0.9724	0.9657	0.9107	1.0267	0.9963	1.0367	1.0259	0.9996	1.0409	0.7432	1.0351	0.6792	0.9038	0.9854	0.5411	0.7193	0.9114
	Yunnan	1.1354	1.0809	0.9213	1.0575	0.7560	0.9855	1.0459	1.0265	0.9768	0.8421	1.0079	0.8466	0.9832	1.1344	1.0499	0.9037	0.9846
	Shanxi	1.0540	1.0129	0.9312	1.0187	0.9988	0.9816	1.0191	1.0030	0.9798	0.7694	0.9621	0.7738	0.9181	1.0682	0.9737	0.9641	0.9643
	Gansu	1.0004	0.9939	0.8819	0.9091	1.2379	0.9676	1.0218	0.9930	0.9464	0.7502	0.9908	0.7765	0.9084	0.9947	0.5795	1.0103	0.9352
	Qinghai	0.9021	1.0466	0.9073	1.1128	1.0793	0.9598	0.9413	0.9899	0.9397	0.7771	1.0006	0.8292	0.9441	1.1394	0.9836	0.8774	0.9644
	Ningxia	1.0201	1.0845	0.8984	1.0109	0.9468	0.9379	0.9897	0.9519	0.5270	1.4286	1.0117	0.9057	0.9731	1.0798	0.9582	0.9320	0.9785
	Xinjiang	1.0874	0.9960	0.9213	1.0161	1.0303	1.0020	1.0220	0.9591	0.9418	0.7630	1.0226	0.7845	0.9472	1.0233	1.0127	0.9043	0.9646
NE	Liaoning	0.8661	0.9676	0.9558	0.9772	1.0654	0.9837	0.9996	0.9791	0.9580	0.8195	0.9950	0.7500	0.9304	0.9086	0.9400	0.6529	0.9218
	Jilin	1.0250	1.0516	0.9196	1.0357	0.7955	0.9813	1.0423	0.9970	0.9266	0.8791	1.0062	0.8077	0.9711	1.0994	0.9954	0.9268	0.9663
	Heilongjiang	0.9346	0.9168	0.9032	1.0041	1.0890	1.0050	1.0316	0.9684	0.9190	0.8638	0.9107	0.7562	0.9372	1.0188	0.9926	0.9431	0.9496
National average		0.9714	1.0621	0.8983	1.0187	1.0299	1.1608	1.0078	0.9889	0.9462	0.9125	0.9625	0.8189	0.9383	1.0449	0.9592	0.9214	0.9776
Eastern Mean		0.9577	1.1400	0.9239	1.0718	0.8681	1.4973	0.9673	0.9862	0.9558	1.2085	0.9058	0.8571	0.9253	1.0477	0.9786	0.9488	1.0150
Central Mean		0.9027	1.0742	0.8353	0.9485	1.3408	1.0043	1.0361	0.9842	0.9593	0.6189	0.9839	0.8060	0.9588	1.0042	0.9970	0.9151	0.9606
Western Mean		1.0293	1.0074	0.9019	1.0122	1.0203	0.9868	1.0248	0.9960	0.9335	0.8194	1.0000	0.8041	0.9368	1.0743	0.9163	0.9218	0.9615
Northeast Mean		0.9419	0.9787	0.9262	1.0057	0.9833	0.9900	1.0245	0.9815	0.9345	0.8541	0.9707	0.7713	0.9462	1.0089	0.9760	0.8409	0.9459

Note: E denotes the eastern region, C represents the central region, W represents the western region, and NE denotes the northeastern region.

, respectively. We found that the average change in the technical efficiency of the national total sample and the four regions from 2003 to 2019 was greater than 1, indicating that the technical efficiency of the transportation industry followed an upward trend both in China overall and in the local areas. However, in terms of GTCH, the value in only the eastern region was greater than 1, whereas the value in the other three regions exhibited a downward trend. Combined with the GML index values in Appendix B, Table A6, the results indicated that the driving factor of the improvement in the transportation industry TFP mainly involved a change in technical efficiency.

Appendix B, Table A7 provides the change value of the technical efficiency of the transportation industry in the 30 provinces (cities and autonomous regions). We found that except for Zhejiang, Fujian, Shandong, Sichuan, Qinghai, and Heilongjiang, the average technical efficiency in the other regions was higher than 1, indicating that the technical efficiency of the transportation industry in most regions exhibited an upward trend, while the GECH of the transportation sector in these six provinces exhibited a downward trend. Appendix B, Table A8 provides the GTCH value of the regional transportation industry. Among them, the average GTCH value in Shanghai, Zhejiang, and Fujian Provinces was greater than 1, indicating that the GTCH of the transportation industry in these three regions increased faster than that in the other regions and that there was much room for improvement in the other regions.

5. Empirical Analysis of the Total Factor Carbon Emission Performance of the Transportation Industry

5.1. Model Setting, Variable Definitions, and Descriptive Statistics

Considering that the minimum TFP value is 0, the data were truncated. If an ordinary linear regression were used, the regression results would be biased [24]. At the same time, in regard to panel data, the estimated value obtained via a random effect Tobit model regression could be more accurate than that obtained via a fixed-effect Tobit model regression [32]. Therefore, this paper used the random effect Tobit model to further analyze the influencing factors of the transportation industry carbon emission efficiency in the 30 provinces of China from 2003 to 2019.

We selected several factors closely related to the carbon emission performance of the transportation industry, including the population size, economic level, transportation intensity, transportation structure, energy intensity, energy structure, factor endowment, road infrastructure, and opening-up level, based on the literature [7,33,34,35]. These variables are listed in

Table 4. Overview of the variables.

Variables	Symbol	Unit	Definitions
Population Size	PS	104 persons	Resident population
Economic Level	PG	104 yuan/person	GDP per capita
Transportation Intensity	TI	Ton-kilometer/yuan	Passenger and freight turnover/regional actual GDP
Transportation Structure	TS	-	Road turnover/total turnover
Energy Intensity	EI	Tons of standard coal/104 ton-kilometer	Standard coal volume/turnover
Energy Structure	ES	-	The proportion of gasoline, kerosene, diesel, and fuel oil consumption in total energy consumption
Factor Endowment	FE	104 yuan/person	The ratio of capital to employees
Road Infrastructure	OM	104 kilometers	Highway mileage
Opening Up Level	OP	-	Total import and export value/actual regional GDP

Source: Provided by the author.

. Data were retrieved from the National Bureau of Statistics and China Energy Statistics Yearbook. The TFP can be obtained with Equation (7):

$$T{E}_{it}=\alpha +X_{it}+u_i+{\epsilon }_{it}$$

(7)

where i denotes the province, t denotes the year, TE denotes the TFP of the transportation industry, X denotes the series of independent variables, $u_i$ denotes the individual unobserved effects, and ${\epsilon }_{it}$ is a random interference term.

Table 5. Descriptive statistical results of the variables.

Variables	Mean	Std. Dev.	Min	Max
TE	0.497	0.309	0.070	1.340
LnPS	8.174	0.752	6.280	9.433
PG	1.385	0.756	0.370	4.000
TI	0.833	0.671	0.100	5.960
TS	0.327	0.183	0.010	0.720
EI	0.277	0.181	0.020	1.210
ES	0.867	0.096	0.480	0.980
FE	136.908	115.450	5.220	786.450
OM	12.590	7.568	0.650	33.710
OP	0.032	0.038	0.001	0.171

Source: Provided by the author.

provides the descriptive statistical results for the variables. The minimum value of variable TE was 0.070, and the maximum value was 1.34, indicating that the TFP of the transportation industry in the various regions was inconsistent. The minimum per capita GDP reached 0.37 × 10⁴ yuan per person, and the maximum per capita GDP reached 4 × 10⁴ yuan per person, indicating that there were still differences in the level of economic development among the various regions. The maximum transportation intensity was attained in Tianjin in 2007, which suggests that the transportation industry is closely related to regional economic development. There was a large difference between the maximum and minimum values of the transportation structure, indicating that the transportation structure differed among the various regions. The minimum value was observed in Beijing in 2016, and the maximum value was obtained in Tianjin in 2007. Moreover, the maximum value of the energy structure was 98%, which indicates that gasoline, kerosene, diesel, and fuel oil are still the main sources of energy consumption in the energy structure of the transportation industry. The minimum and maximum factor endowment values were attained in Jilin Province in 2003 and Yunan Province in 2019, respectively. The factor endowment index comprehensively considers the level of regional economic development and the input of human factors. The change in highway mileage indicated that road infrastructure construction between the various regions was constantly improving.

5.2. Regression Results

Before basic regression, multicollinearity test results were obtained, as summarized in Appendix C,

Table A9. VIF values of the variables.

Variables	VIF	1/VIF
PG	4.19	0.239
LnPS	3.54	0.283
TI	2.02	0.496
TS	2.33	0.428
EI	1.85	0.540
ES	1.37	0.728
FE	1.54	0.651
OM	4.75	0.211
OP	4.70	0.213

Source: Provided by the author.

. The variance inflation factor (VIF) values of the variables were all less than 5, indicating that there was no multicollinearity among the variables. The Tobit model was assessed via the likelihood ratio test method, and finally, we obtained a random regression model. The regression results obtained with the random Tobit model are provided in

Table 6. Regional Tobit model regression results.

	(1)	(2)	(3)	(4)	(5)
Variables	Full Sample	Eastern Area	Central Area	Western Area	Northeast Area
PG	−0.00562	0.0144	−0.450 *	−0.0108	0.768 ***
	(0.0411)	(0.0307)	(0.2443)	(0.0921)	(0.1969)
LnPS	0.205 ***	0.303 ***	0.474 **	0.0499	0.238
	(0.0438)	(0.0461)	(0.211)	(0.0357)	(0.1792)
TI	0.218 ***	0.280 ***	0.0681	0.0133	−0.0976
	(0.0179)	(0.021)	(0.0797)	(0.0369)	(0.0677)
TS	0.248 ***	0.489 ***	0.501 ***	0.16	−0.615 ***
	(0.0747)	(0.1655)	(0.1723)	(0.1069)	(0.2209)
EI	−0.191 **	−0.216 **	−0.448	−0.389 ***	−0.432 **
	(0.0922)	(0.0982)	(0.2964)	(0.1066)	(0.2084)
ES	−0.277 **	−1.104 ***	−0.219	−0.165	−0.514 ***
	(0.13)	(0.2998)	(0.348)	(0.1245)	(0.1944)
FE	1.92 × 10−5	−0.000873 ***	0.00145 ***	0.000377 ***	0.000505
	(0.0001)	(0.0002)	(0.0004)	(0.0001)	(0.0004)
OM	−0.00993 ***	−0.0222 ***	−0.0140 **	−0.00525 *	−0.00729
	(0.0026)	(0.0062)	(0.0061)	(0.0028)	(0.007)
OP	−0.945 *	−3.291 ***	25.68 ***	−3.700 **	−11.68 ***
	(0.543)	(0.5534)	(7.8189)	(1.6809)	(3.4494)
_cons	−0.987 ***	−0.685	−3.105 *	0.209	−1.466
	(0.3686)	(0.4834)	(1.8879)	(0.2933)	(1.3239)
N	510	170	102	187	51
LR test	257.29	1.86	31.84	43.99	0.00
Wald chi2	250.45	414.86	85.26	97.95	108.27

Note: Robust standard errors in parentheses; *, **, *** represent significant differences at significance levels of 10%, 5%, and 1%, respectively.

, including the national sample and the eastern, central, western, and northeastern areas.

Considering the national total sample, the population size attained a significant positive relationship with the TFP of the transportation industry, which suggests that an increase in the population size could improve the TFP of the transportation industry. Transportation intensity could increase the TFP of the transportation sector by 0.218 units. As a value output form of the transportation industry, the turnover of passengers and goods drives regional economic development. Additionally, the higher the utilization efficiency of highways is, the more conducive this could be to improving the carbon emission efficiency of the transportation industry. The energy intensity and energy structure both had a negative effect on TFP. The larger the scale of the transportation industry is, the more energy it consumes. The road infrastructure indicated that the scale and quantity of roads did not affect the TFP but rather the road utilization level. The opening-up level could increase the carbon emissions of the transportation industry of the host country, thus reducing the carbon emission performance.

The four regional samples of the variables varied. The per capita GDP index was significantly positive in Northeast China, significantly negative in Central China, but not significant in eastern and western China. The economic development level in Northeast China was insufficient. Expansion of the economic scale in the central region could increase energy consumption and reduce the TFP of the transportation industry. The population size was significantly positive in the eastern and central regions, and the population size in the central region was greater than that in the eastern region. The western and northeastern areas were not significant. This indicates that the dividend attributed to the increase in population size in the eastern and central regions could enhance TFP. The transportation intensity was significantly positive only in the eastern region but not in the other regions, indicating that the transportation scale in the eastern region was larger, and that the driving effect on the economy was more obvious.

The transportation structure was significantly positive in the eastern and central regions and significantly negative in the northeastern region. There are many transportation hubs and well-developed transportation networks in the eastern and central regions. The energy intensity index in the four regions was negative, which was only significant in the eastern, western, and northeastern regions. This suggests that the higher the energy consumption, the more unfavorable this was to the TFP. In the eastern and northeastern regions, the energy structure was significantly negative, but it was not significant in the central and western regions. The energy structure in the eastern and northeastern regions was significantly negative, but it was not significant in the central and western regions. The economic development level in the eastern and northeastern regions depends on the transportation industry, which consumes much fossil energy.

The factor endowment demonstrates that the capital and labor structures in the eastern region did not match. The proportions of capital and manpower investments in the central and western regions were matched. The road infrastructure showed that an increase in the highway operation mileage could not improve the TFP. The level of openness to the outside world was significantly positive in the eastern, western, and northeastern regions and significantly negative in the central region, which indicates that the higher the opening-up level in the central region, the more TFP improvement could be facilitated.

6. Conclusions and Suggestions

6.1. Research Conclusions

This paper used the superefficiency SBM model with undesirable outputs to calculate the TFP of the transportation industry in 30 provinces of China from 2003 to 2019. The SBM-DEA model with undesirable outputs and the GML index were used to statically and dynamically analyze the TFP of the transportation industry, respectively. The random Tobit model was used to explore the influencing factors of the TFP of the transportation sector. By analyzing the influencing factors of the TFP of the transportation industry in China as a whole and in these four regions, this paper carefully formulated policy suggestions to improve the TFP of the transportation industry.

First, it was found that the average TFP value in only Hebei and Anhui was greater than 1, and that in the other provinces and regions it was less than 1. The TFP in the central region was higher than that in the eastern region, while the TFP in the western region was the lowest. The average TFP of the entire transportation industry in China and the four regions was less than 1, indicating that the TFP of the overall transportation industry in China was relatively low. However, from a temporal trend perspective, the TFP exhibited an upward trend in recent years, indicating that the TFP of the transportation industry in China was gradually improving.

Second, the average GML values for the national and regional samples were less than 1. Only the average GML value in Shanxi, Guangxi, and Yunnan was greater than 1, indicating that the transportation industry in these areas occurred in the optimal state. The GECH value in the central, western, and northeastern regions was greater than 1, while the GTCH value was less than 1. The GECH and GTCH values in the eastern region were greater than 1. This indicates that the change in GECH was the main factor impacting the GML index of the transportation industry. Because of the high economic level and well-developed transportation network in the eastern region, GECH and GTCH reached the optimal state. In addition, the GSECH value of the transportation industry in the eastern region was greater than the GPECH value, while the GPECH value in the central, western, and northeastern regions was higher. Compared to the other three regions, the transportation industry in the eastern region was larger in scale.

Finally, the Tobit model revealed that the influencing factors causing the different carbon emission efficiencies in the four regions varied. The population size, transportation intensity, and transportation structure played a significant role in promoting the TFP for China as a whole and the eastern region. The population size, transportation structure, factor endowment, and level of openness promoted TFP in the central region. In the western region, the factor endowment played a significant role in promoting TFP. In the northeastern region, the economic level played a significant role. In addition, the economic level inhibited TFP in the central region but promoted TFP in the northeastern region. The transportation structure generated an inhibitory effect in the northeastern region and a promotion effect in the eastern region. Factor endowment inhibited the performance in the eastern region and enhanced the performance in the central and western regions. The opening-up level promoted TFP in the central region but inhibited TFP in the other regions.

6.2. Suggestions

Based on the results of this study, this paper formulated the following suggestions:

(1)

Technical progress in the transportation sector in the various regions should be encouraged to promote TFP improvement. From a transportation industry perspective, the technological level in the transportation sector should be improved. The subsidy policy was introduced to encourage enterprises to increase investment in technological innovation and R&D to form synergy between industrial chains. We will vigorously support the research, development, and use of smart cars and energy-saving cars and introduce low-carbon emission reduction technologies; establish an intelligent transportation system, optimize the transportation layout by taking advantage of the big data Internet of Things, take multiple measures to improve the technical level of the transportation industry, and use technological progress to promote the TFP.

(2)

The energy structure of the transportation industry should be improved, green transportation options should be developed, and green and low-carbon transformation of the transportation industry should be realized. China should continue to formulate green transportation and green travel policies to optimize the TFP of the transportation industry. Therefore, the government should increase the proportion of low-carbon and energy-saving vehicles in the urban transportation structure and ensure the sustainable development of clean energy vehicles. By improving infrastructure construction, such as charging piles and charging stations, and improving battery life technology, the TFP of the transportation industry can be improved, and the consumption structure of fossil energy in the industry can be optimized. Moreover, schools and families should strengthen the cultivation of environmental protection and energy conservation. There should be increases in the publicity of green and low-carbon travel in the community, and improved citizens’ awareness of low-carbon and environmental protection to gradually realize the green transformation of the transportation structure.

(3)

Development of the transportation industry in the different regions varies. The differences between regions should be considered, and corresponding measures should be taken according to the level of economic development among regions and the shortcomings of our own traffic structure when formulating relevant policies. The four regions should identify the constraints and formulate measures to overcome these constraints restricting carbon emission performance improvement. The eastern and central regions should adjust the energy structure and reduce the consumption of fossil energy. For example, the use of new energy vehicles in public transport and private travel should be promoted. The western region should promote transportation infrastructure and increase the mileage of operational roads. For example, capital investment should be increased, and road infrastructure construction should be improved to optimize the traffic structure. The northeastern region should improve the opening-up level, expand the scale to improve the transportation structure, adjust the energy structure, lower the energy intensity, and improve the TFP. The opening-up level should be increased and favorable policies formulated to provide financial and technical support for transportation enterprises.