Analysis on Logistics Efficiency Measurement of New Western Land-Sea Corridor under the Background of “Double Carbon” and Ecological Environment Protection

Under the research background of ecological environment protection and “double carbon” goal, this paper applies panel data on the logistics industry from 2010 to 2019 in 12 provinces of China's new western land-sea corridor to statically measure the logistics industry's technical efficiency after taking into account the impact of different environmental elements and to analyze the dynamics trends of total factor productivity in the logistics sector. It is measured by using the three-stage SBM model and the Malmquist–Luenberger productivity index, which considers undesirable output. The findings indicate the following: (1) In the context of “double carbon,” the overall technical efficiency of the logistics sector in the new western land-sea corridor seems to be relatively low; however, the average technical efficiency of the logistics sector in the southern portion of the new western land-sea corridor does seem to be higher than that of the northern part. (2) The logistics industry's technical efficiency varies greatly by region, with locations near central China having much higher technical efficiency than remote inland areas. (3) The fundamental reason for the improvement of technical efficiency in the logistics industry is pure technical efficiency, and the driving force behind the increase in total factor productivity is technological advancement. (4) Economic development, informatization development, industrial market scale, and import and export all have a substantial influence on the logistics industry's technical efficiency. Finally, depending on the findings, policy recommendations are offered.


Introduction
With the rapid development of China's economy, the pollution and destruction of the ecological environment have been unable to make the economy sustainably developed, therefore, China has formulated the corresponding implementation of ecological protection. A major strategic objective for high-quality economic development and environmental sustainability in China is to accomplish the aim of "peak carbon dioxide emissions and carbon neutrality" [1]. During the fourteenth Five-Year Plan phase (2021-2025), the logistics industry needs to shoulder the responsibility of reducing emissions while developing at a high quality. Te logistics industry's efciency is a comprehensive indicator to measure the whole logistics operation and resource allocation in regional logistics development. In the regional coordinated development pattern based on the external environment, it is especially crucial to explore the logistics industry efciency, enhance the allocation efciency of input factors in the logistics industry, and enable the growth of "high quality" with "good efciency." "Increase the efciency of the logistics sector by strengthening the construction of transport infrastructure in western China, synergizing with the expansion of the Yangtze River Economic Belt," according to the comprehensive plan for the new western land-sea corridor [2]. Te new western land-sea corridor connects the Maritime Silk Road with the Overland Silk Road, ofering a crucial commercial and logistical corridor for the opening up of western China's inland areas. Te new western land-sea corridor's competitive development index of 113.2 in 2020 demonstrates a strong development trend as well as a considerable increase in the logistics industry's service and operational efciency [3]. Te container throughput of Beibu Gulf seaport in Guangxi has increased by 64% yearon-year, from 3.08 million TEUs in 2018 to 5.05 million TEUs in 2020. Rail-sea trains in the western area have climbed by 299 percent year-on-year, from 1,154 trains in 2018 to 4,607 trains in 2020 [4]. At the same time, with the fast expansion of the logistics sector, the energy consumption of logistics industry will increase at a high speed. Statistics on the overall energy consumption of the logistics sector in 12 provinces of the new western land-sea corridor increased from 73,272,801 tons of standard coal in 2010 to 101,375,500 tons of standard coal in 2019, with 3.67 percent average annual growth rate.
In the context of "double carbon," can logistics sector in the new western land-sea corridor expand at an efcient and sustainable speed? Are all kinds of input resources adequately distributed and utilized? Is there a signifcant difference in the growth of the logistics sector among provinces in the new western land-sea corridor? Tis paper investigates the growth of the logistics sector along the new western landsea corridor to show the extent to which the external environmental variables infuence logistics industry's efciency, and it serves as a reference for the improvement of logistics sector's efciency from the perspective of rational resource allocation. Furthermore, it serves as a decisionmaking foundation for the construction of China's international commerce corridor with the shortest transportation time to ASEAN countries.

Literature Review
Charnes et al. [5], used data envelope analysis (DEA) to evaluate the relative efciency of decision-making units (DMUs) with multiple inputs and multiple outputs. DEA is a nonparametric linear programming approach with a fexible calculation mechanism. It is not required to perform a certain type of production function and allows the existence of inefciencies. Most study adopts DEA to assess the efciency of the regional logistics sector. Tian and Li [6] utilized the DEA-Malmquist model to assess the logistics sector's total factor productivity in 30 provinces of China between 1999 and 2006 and found that there were disparities in total factor productivity and scale inefciency among the provinces. Markovits-Somogyi and Bokor [7] employed DEA and DEA-PC to assess the logistics industry efciency in 29 European nations, both methods are considered more appropriate after comparison. Yu and Qian [8] analyzed the logistics industry's technological efciency in 11 provinces of the Yangtze River Economic Belt between 2006 and 2015 through the DEA-Malmquist method, concluding that it was generally not high, with the eastern area being better than the western area, and that the logistics industry's efciency increase was primarily infuenced by economic growth, informatization development, and degree of openness. Lei et al. [9] studied the technical progress index of China's 49 listed logistics businesses from 2008 to 2017 and concluded that the logistics sector's technical progress had a major benefcial infuence on the skill structure of employment. Although DEA has unique advantages in measuring the input-output efciency of DMUs, it does not account for the efect of random errors on output, which easily leads to errors in calculation results.
Stochastic frontiers analysis (SFA) is a standard representation of the parametric method with the beneft of utilizing a production function to evaluate the input-output efciency while accounting for stochastic error. Fan and Wang [10] measured the service efciency of 11 logistics corridors in China from 2000 to 2013 and found that the logistics corridors difered signifcantly in time and space, with lower service efciency when running through western China or across eastern, central, as well as western China. Zhang et al. [11] examined the technological efciency of low-carbon development in China's logistics sector from 2007 to 2016, and believed that the overall situation was low and the regional diferences were obvious. Te proportion of secondary sector in the provincial GDP and the average size of logistics enterprises have a favorable infuence on the improvement of logistics industry's technical efciency, while fnancial support and energy consumption have a negative infuence on technical efciency, and the impacts of environmental regulation are not obvious. Han and Liu [12] investigated the efciency of 80 Chinese logistics listed enterprises from 2013 to 2017, concluding that total logistics company efciency is increasing, and the average technical efciency of enterprises in the eastern area is higher than enterprises throughout the western and central areas. Although the infuence of random error is considered in the SFA model, there is no precise theoretical support when choosing the production function, and there are strict assumptions about the distribution of inefciency terms [13].
A three-stage DEA model was proposed, which integrates the benefts of parametric and nonparametric analysis methods, excludes the efects of external environmental factors and statistical noise on efciency evaluation, and allows measurement results to more accurately describe the internal managerial level of the decision-making unit [14]. According to Zhong [15], when the infuence of external environmental factors and statistical noise are removed, the technological efciency, pure technological efciency, and scale efciency of China's logistics sector change signifcantly. J. Zhang and J. Zhang [16] examined the logistics industry's efciency in 31 provinces in China from 2010 to 2014 and discovered that the scale efciency of logistics industry was increasing as a whole, but the degree of logistics operation and management restricted the logistics expansion. According to Zhang et al. [17], from 2009 to 2014, the overall logistics sector's efciency of the New Silk Road Economic Belt at home and abroad was not high, with large diferences in logistics industry's efciency among regions being more infuenced by scale efciency. Mei et al. [18] evaluated the logistics sector efciency in East China between 2012 and 2016 and concluded that the logistics sector was in a state of increasing scale efciency on the whole. Increasing the retail volume of social consumer goods might enhance the logistics sector's efciency. Yang [19] found that Jiangsu province's logistics industry performed well overall under low-carbon constraints from 2007 to 2016 and scale efciency was the primary problem restricting the logistics industry's efciency. Zhang et al. [20] analyzed the logistics sector's efciency in 19 provinces of the Yangtze River Great Protection Region between 2013 and 2017 and found that the growth of the logistics sector in this area was unbalanced, with scale efciency leading to the largest increase in technical efciency. Te three-stage DEA only evaluates the infuence of desirable output on efciency but does not consider the impact of undesirable output on efciency. It does not objectively refect the true level of industry management and thus measure the efciency of the logistics industry.
Tone [21] proposed a slacks-based measure (SBM) including undesirable outputs relying on DEA improvement, thereby enhancing the scientifc reliability of the efciency measurement. Liu and Guan [22] argued that the logistics industry in China's 30 provinces was generally inefcient under low-carbon constraints from 2003 to 2014, with the inefciencies concentrated in western China. Deng and Shen [23] evaluated the logistics industry's efciency in 30 Chinese provinces subject to carbon emission constraints in 2016 and reported that there were signifcant local diferences in China. Te fundamental restriction to logistics development was inefcient scale, and energy structure was negatively connected to logistics sector efciency. Zheng et al. [24] used the SBM in conjunction with hierarchical regression to measure the logistics industry's efciency in 18 Chinese provinces bound by carbon emission constraints from 2007 to 2017, concluding that the efciency gap across eastern and western regions had narrowed, with external variables positively infuencing efciency having shifted from the degree of openness prior to 2013 to regional economic development. Although a nonradial, nonoriented SBM with undesirable output can avoid the problems of slack and single output variable, it cannot accurately and objectively reveal the efciency of the logistics industry without distinguishing internal managerial inefciencies from external environmental infuences and statistical noise.
In conclusion, although the three-stage DEA may remove the efects of external environment variables and statistical noise from efciency evaluation, it cannot fully account for the slack in input-output variables and undesirable output, which leads to efciency measurement error. Te slack problem can be avoided by considering the undesirable output of the SBM model but not the interference of external environmental variables and statistical noise. Most studies have been undertaken on China as a whole, or on a specifc province, or on a regional economic belt. Results vary, and systematic studies of the new western land-sea corridor are scarce. Based on this, this paper expands as follows: (1) In terms of the research methodology, the technical efciency of the logistics industry in 7 provinces, 4 autonomous regions, and 1 municipality (henceforth referred to as 12 provinces) along the new western land-sea corridor in China from 2010 to 2019 was analyzed using a three-stage SBM model considering nonradial, nonoriented, and undesirable output. In order to complete the multidimensional extension of the study, the Malmquist-Luenberger index is used to dynamically analyze the fuctuation of total factor productivity (TFP) in the logistics sector. (2) In terms of research content, this study uses the panel data to make an empirical analysis of the logistics industry in 12 provinces of the new western land-sea corridor from 2010 to 2019, to assess the overall efciency of the logistics sector, identify the main afecting factors, and analyze the diferent regional situations to expand the study of logistics industry's efciency in less developed regions. Figure 1 shows the research frameworks of this paper.

Methodology
subject to where ρ is the efciency and 0 ≤ ρ ≤ 1. When ρ � 1, s − , s g , s b are all zero, and DMU is valid, s − ∈ R m is the input slack variables, s g ∈ R s 1 is the desirable output slack variable, s b ∈ R s 2 is the undesirable output slack variable, and λ ∈ R n is the weight [21].

Te Second Stage: Stochastic Frontiers Analysis (SFA).
Te primary purpose of the second stage is to divide the input slack variable from the frst stage into three variables: internal managerial inefciency, external environmental infuences and statistical noise, and to reduce the efects of the environmental and statistical noise on efciency evaluation. Firstly, the dependent variables are the input slack elements obtained in the frst phase, and the independent variables are the external environmental variables, which are regressed by using a Stochastic frontier analysis (SFA) to generate new input variables.

Journal of Environmental and Public Health
where s ni is the slack variable of the ith DMU's nth input, f n (z i ; β n ) indicate the efect of external environmental variable on input slack variables, z i � (z 1i , z 2i , · · · , z ki ) are k external environmental variables, β n are the parameter vectors, v ni represent statistical noise, and μ ni represent inefcient management.
Secondly, the SFA results are used to adjust input variables for all DMUs to a more favorable external environment and to eliminate the impact of external environmental variables and statistical noise on efciency measurements.
where x A ni is the modifed input variable and x ni is the obser ved input variable, indicates modifcation to external environmental variables, and [max (v ni ) − v ni ] represents the adjustment of statistical noise.  (3), the recalculated efciency can more properly represent the effciency of the logistics sector.

Te Tird
When measuring the efciency of the logistics industry in the 12 provinces of the new western land-sea corridor, the advantages of the three-stage SBM model are as follows: First, by reducing the efects of external environmental variables and random errors and taking into account undesirable output, the third-stage results can more precisely refect the actual efciency of DMUs. Second, it is an empirical study on the impact of external environmental variables on efciency, which can quantify the degree and mechanism of environmental variables.

Malmquist-Luenberger Productivity Index.
By replacing the distance function in the Malmquist productivity index with the directional distance function, Chung et al. [25] introduced a Malmquist-Luenberger productivity index that relies on the directional distance function to calculate the total factor productivity of undesirable output, including carbon dioxide emissions. where the ML productivity index represents the change in DMU's total factor productivity between period t and period t + 1. If ML > 1, the total factor productivity rises, if ML < 1, the total factor productivity falls, and if ML � 1, the total factor productivity remains unchanged.
Te ML productivity index can also be separated into efciency change index MLEC and technology change index MLTC between period t and period t + 1.
Te efciency change index MLEC may also be subdivided into two components: the pure technological efciency change index MLPTEC and the scale efciency change index MLSEC.
Te Malmquist-Luenberger productivity index is actually a modifed Malmquist index. Te traditional Malmquist index is based on the output distance function, but the Malmquist-Luenberger index is based on the directional distance function, which can improve the good output while reducing the bad output, and the output distance function cannot be realized.

Indicators and Data
Te 2006 China Tird Industrial Statistics Yearbook shows that the value added of transport, storage, and postal services account for more than 80 percent of the logistics sector's value added. Terefore, the existing research essentially uses the indicators related to transportation, warehousing, and postal services to represent the logistics industry, and this method has also been adopted in this paper. Te input indicators are defned to include capital, labor, and energy consumption, while the output indicators include industry value added and carbon dioxide emissions. Indicators and data processing are described below.

Input Indicators.
Te fxed asset investment of the whole society in the logistics industry is chosen as the capital indicator, and the capital stock of the logistics industry is calculated by a perpetual inventory system. Te formula for calculation is as follows: where K it and K it− 1 are, respectively, the logistics sector's capital stock in region i between t and t − 1 periods, K i0 is the logistics sector's capital stock in region i during the base period, calculated by dividing the amount of fxed assets investment in 2010 by 10% (at 2010 constant price), δ is the depreciation rate of capital at 9.6 percent [26], and I it and P it denote fxed assets investment and fxed assets investment price index in region i in period t, respectively. Te total number of the logistics sector's urban employees, as well as urban private frms and individuals, is used to calculate the labor force input indicator.
Using the logistics sector's energy consumption as the energy input indicator, converting the various kinds of logistics energy consumption in diferent provinces into standard coal, and measuring total energy consumption, the following is the formula for processing converted standard coal: where E means the total quantity of energy consumption after conversion of all forms of energy consumption to standard coal, M i represents the various types of energy consumed by the logistics sector, and P i represents the conversion factor for energy i into standard coal.

External Environmental Variables.
External environmental variables primarily comprise aspects that have a considerable infuence on the efciency of the logistics sector but are beyond the control of the logistics industry itself. In this paper, the elements of the external environment are selected from four categories: economic condition, informatization development, industrial market scale, and import and export. First, the provincial GDP is selected as a measure of economic condition and processed through the GDP defator. Te improvement of provincial economic conditions has a favorable infuence on the development of the logistics sector as well as the efciency of the logistics industry. Second, the number of mobile phone users at year-end in each province has been chosen as an essential indicator to assess the level of development of information technology. Te popularity of smartphones makes information more accessible, and the terminal development of logistics information systems shifts to mobile, driving the growth of the logistics sector and infuencing logistics sector efciency. Tird, the industry market scale, the continuous optimization of the industrial market size has led to the fast expansion of the logistics sector, which has had a positive infuence on the efciency of the logistics industry, and the total number of registered legal entities in the logistics business in each province is a key measure for assessing the industry's growth. Fourth, the total import and export in each province is chosen as the import and export measurement, which assists its logistics industry in opening up international markets and promoting the integration of domestic and foreign commodities' markets.

Data Sources.
Tis paper examines the logistics sector in 12 provinces along the new western land-sea corridor between 2010 and 2019, covering inner Mongolia, Guangxi, Hainan, Chongqing, Sichuan, Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang (Tibet is not included due to missing data on energy consumption). Data from transportation, storage, and postal services are used to replace the statistics of the logistics industry. Te collected data is from the websites of the China National Bureau of Statistics, the China Energy Statistics Yearbook, and the China Statistics Yearbook. Table 1 displays the descriptive data for critical factors.
In order to make the measurement result more reasonable, the Pearson's correlation test of input-output indicators is carried out by using Stata 16. Te correlation coefcient of input and output indicators is high, all of which passed the 1% signifcance test and satisfed isotropy requirements; Table 2 displays the results.

Te First Stage.
Te initial input and output data from the logistics business from 2010 to 2019 were substituted into the input-oriented SBM model utilizing 12 provinces as DMUs, and technical efciency, pure technical efciency, and scale efciency in the logistics sector were derived using the MaxDEA Pro8 software, with the results shown in Tables 3-5 Although the pure technical efciency was at the production frontier, scale efciency fuctuated around 0.5 percent, which led to a decline of technical efciency. In the Hainan's logistics business, the trend of technical efciency was more consistent with pure technical efciency, and the increase of scale efciency results in the fuctuation of technical efciency. Te logistics sector in Guangxi, Chongqing, Guizhou, Yunnan, Gansu, and Xinjiang was less technologically efcient overall, but the scale efciency was high and the pure technical efciency was low, which led to low technical efciency. Journal of Environmental and Public Health

Te Second Stage.
In the Stochastic frontier analysis (SFA), explanatory variables are the input slack variables obtained in the frst phase, economic development, informatization development, industry market scale, and import and export are the independent variables, while the frontier 4.1 software can be used to study the impact of external environmental factors on the input slack variables, as detailed in Table 6. Table 6 demonstrates that the one-sided likelihood ratio tests are 34.03, 31.77, and 57.28, respectively, all of which are signifcantly tested at a level of 1%. It suggests that external environmental variables should be excluded when studying the technical efciency of the logistics sector along the new western land-sea corridor, and the estimation result of the SFA model is acceptable. All three gamma values passed the signifcance test of 1%, which shows that the selected external environmental variables are more plausible. When the external environmental variables are positively correlated with the input slack variables, it means that increasing the external environmental variables will increase input redundancy and decrease technical efciency. When the external environmental variables are negatively correlated with the input slack variables, reducing the external environmental variables helps to minimize input redundancy and increase the logistics sector's technical efciency. With the provincial gross domestic product (GDP), the number of mobile phone users at year-end, the number of registered legal entities in   (1) Economic Condition. Te provincial gross domestic product (GDP) is a measure of economic growth, which is negatively related to three slack elements. It implies that the improvement of economic conditions can reduce the input redundancy of capital stock and employees' energy consumption, reasonably allocate the resources of the logistics industry, and improve technical efciency in the logistics industry.
(2) Informatization Development. Te informatization level represented by the number of mobile phone users at year-end shows a substantial positive connection with the slack variables of capital stock and energy consumption and a signifcant negative correlation with the slack variables of employees, all of which pass the 1 percentage point signifcance test. Tis suggests that the increase in mobile phone users at year-end will boost the redundancy of capital stock and energy consumption inputs, which will lead to inefciencies in capital inputs and energy consumption, while reducing the redundancy of employee inputs to make them more rational.
(3) Industry Market Scale. Te industry market scale represented by the number of registered legal entities throughout the logistics business has a high positive association with energy consumption slack variables, a negative correlation with capital slack variables, and no signifcant relationship with employee slack variables. Tis shows that as the number of registered legal persons increases, so will the number of frms, resulting in lower energy efciency and more rational capital allocation.
(4) Import and Export. Te provincial total imports and exports are signifcantly positively correlated with slack variables of employee and energy consumption, but negligible with capital stock. It implies that an increase in import and export will generate redundancy in the number of employees and energy consumption, causing irrational input of employees and inefcient use of energy in the logistics industry.

Te Tird
Stage. Te SBM model was applied to calculate the adjusted input variables instead of observed input variables to obtain accurate technical efciency, pure technical efciency, and scale efciency in the logistics industry, removing the impacts of environmental factors and statistical noise, as illustrated in Tables 7-9. Note. irs, drs, and "-" represent increasing, decreasing, and constant returns to scale, respectively.
Overall, the average technical efciency in the logistics sector increased from 0.710 in the frst phase to 0.734 in the third phase in 12 provinces along the new western land-sea corridor, indicating that, despite a lower average technical efciency, uncertainties such as external environment variables continue to underestimate the logistics industry's technical efciency. Te average technical efciency of the logistics sector in the southern area increased from 0.635 in the frst phase to 0.783 in the third phase, and the logistics sector's average technical efciency in the northern area decreased from 0.786 in the frst phase to 0.715 in the third phase, indicating the adjusted average technical efciency in the southern area is higher. Note. * , * * , and * * * are signifcant at 10%, 5%, and 1%, respectively.  Technical efciency in the logistics sector varies widely among provinces, both before and after adjustments. Te average technical efciency of Guangxi's logistics industry rose from 0.656 in the frst step to 0.904 in the third step, and that of Chongqing's logistics industry rose from 0.685 in the frst step to 0.917 in the third step, both of which showed that external environment variables had a great infuence. Among them, improving the pure technical effciency seems to be the key factor for improving the logistics sector's technical efciency in Guangxi and Chongqing, and increasing the logistics sector's scale effciency can only afect the technical efciency in Chongqing. Inner Mongolia and Shaanxi have much higher efciency levels in the logistics industry than other provinces. Inner Mongolia's adjusted technical efciency reached the production frontier, and Shaanxi's adjusted technical efciency was 0.969, mainly because Inner Mongolia and Shaanxi are close to central China, with relatively good transportation infrastructure conditions, relatively large labor and capital investment. Te logistic industry's technical efciency in Qinghai and Ningxia declined from 0.544 to 0.961 in the frst step to 0.259 and 0.697 in the third step, respectively. Te two provinces' pure technical efciency is at the production frontier. Te decline in technical efciency is mainly due to a fall in scale efciency, demonstrating that the high efciency of the frst step is infuenced by changes in the external environment. Although pure technical efciency has improved in Hainan, Guizhou, Gansu, and Xinjiang, scale efciency has decreased, resulting in a slight decrease in the logistics industry's technical efciency, indicating that external environmental variables have not signifcantly driven the logistics industry's development.

Malmquist-Luenberger Productivity Index. Tis paper uses
MaxDEA Pro8 to substitute the adjusted input and initial output variables into the Malmquist-Luenberger productivity index to calculate the changes of total factor productivity in 12 provinces along the new western land-sea corridor from 2010 to 2019, and the calculation results are shown in Tables 10 and 11.
Te Malmquist-Luenberger productivity index remained higher than 1 except in 2015-2016, 2016-2017, and 2018-2019. Over the last decade, the Malmquist-Luenberger productivity index for logistics businesses along the new western land-sea corridor has averaged 1.048, which indicates that the total factor productivity is increasing and that the logistics industry is developing quicker. On average, the technical change index increased by 6.6 percent, but the efciency change index decreased by 1.5 percent. Because the efciency change index has declined at a slower rate than the technical change index, the average annual growth of TFP has remained at 4.8 percent over the past decade, suggesting that technological progress is a key factor in increasing TFP. Te scale efciency change contributed − 1.6 percent to total factor productivity index change, indicating that change in the scale efciency has a detrimental impact on total factor productivity improvement. Te efciency change index was less than 1 in 7 out of 10 years, and the lowest being 0.937 in 2018-2019, indicating a downward trend. However, the technical change index was greater than 1 in 7 out of 10 years, and the highest value was 1.199 in 2014-2015, indicating an upward trend. Furthermore, the Malmquist-Luenberger index and the technical change index kept the same overall trends, with an apparent upward trend only in 2014-2015.
Only Qinghai had a Malmquist-Luenberger productivity index less than 1 among the 12 provinces, while the rest provinces had a Malmquist-Luenberger productivity index more than 1. Te average Malmquist-Luenberger productivity index was 1.049, indicating an overall increase in the total factor productivity, with 92 percent of provinces contributing to an increase in the logistics industry's total factor productivity. Technical change indexes of all provinces were greater than 1, with an average of 1.070, indicating an overall upward trend in the total factor productivity in the logistics sector. Te pure technical efciency change index was less than 1 in 5 provinces and greater than 1 in 7 provinces. Te overall mean of the pure technical efciency change index was 1.003, indicating an upward trend in 58.3 percent of the provinces in the logistics sector. Te average Malmquist-Luenberger index, as well as the average index of technical change in the southern and northern areas, were both more than 1, whereas the average index of scale efciency change was less than 1. Te acceleration of technological advancement is the reason for the growth of total factor productivity in the logistics sector in the 12 provinces of the new western land-sea corridor, and the primary decrease is the fall in scale efciency.

. Conclusions and Policy Proposals
Tis article considers the relationship between logistics sector expansion and environmental protection in a systematic manner, using panel data from the logistics industry along China's new western land-sea corridor between 2010 and 2019, and measures actual logistics industry efciency after excluding the impact of external environmental variables as well as statistical noise by employing nonradial,  nonoriented three-stage SBM with undesirable output. Finally, the Malmquist-Luenberger productivity index is used to study the dynamic evolution of logistics business efciency, and the following conclusions are drawn: (1) External environmental variables have a great infuence on the efciency of the logistics sector. Economic conditions, informatization development, industrial market scale, and import and export all have an infuence on the logistics sector's efciency. Te efciency of the logistics business varies between the frst and third phases, and the efect of each external environment variable on the logistics sector efciency is diferent. (2) In general, the overall logistics sector efciency in the new western land-sea corridor remains low, despite ongoing improvement. Te logistics industry's resource allocation is insufciently optimized, and the development rate is rather modest. Te pure technical efciency is relatively high, and it is the key factor infuencing the expansion of the technical efciency, while the scale of the logistics sector does not correspond to the current level of industrial development. (3) Te technical efciency in the logistics sector in the southern and northern areas of the new western land-sea corridor both have development potential. On average, the logistics industry's technical efciency in the southern area is greater than that in the northern area, while it is substantially higher in the provinces near central China than in the remote hinterland areas. (4) Te efciency of the logistics business varies greatly by province. Except for Inner Mongolia, where the adjusted efciency of the logistics sector is on the production frontier, the technical efciency of other provinces has changed signifcantly in the past 10 years. Shaanxi and Chongqing continue to have the highest adjusted technical efciency, while Qinghai, Xinjiang, and Hainan have the lowest. (5) Over the 10 year period, the total factor productivity of the logistics industry along the new western landsea corridor has improved by approximately 4.8 percent each year, while the technological change index increased by an average of 6.6 percent every year, indicating an overall improvement. Te primary cause of the improvement is technical development, whereas the primary cause of the drop is the decrease in scale efciency. Te total factor productivity of the logistics industry in 92% of the provinces in the new western land-sea corridor has improved.
Tis research makes the following recommendations relying on the measurement results: Firstly, it is essential to continually improve the external environment in order to help the expansion of the logistics frms. Given the variety of external environmental variables in each province along the new western land-sea corridor, environmental adjustments have resulted in signifcant variances in logistics sector efciency in various provinces. Each province should pool its resources and develop regulations to increase logistics business efciency, as well as encourage logistics enterprises to grow larger and stronger through government investment, form a market operation mode guided by the government but dominated by enterprises, and accelerate resource gathering.
Secondly, it is necessary to enhance both pure technical efciency and scale efciency. Te technical efciency is mostly governed by the pure technical efciency of the logistics sector in each province, and the average scale effciency of most provinces is declining. Each province should take full advantage of the opportunity of "new infrastructure" construction, upgrade and promote logistics information technology and the connection between railway freight transport and port hubs, promote the construction of a perfect multimodal transport system and enhance air-rail and other modes of transportation, integrate cargo capacity with Internet platforms, and promote cross-regional cooperation in the logistics industry.
Tirdly, it is crucial to enhance the current operating efciency of the new western land-sea corridor to launch the new western land-sea corridor's close connection with the Yangtze River Economic Belt, ensure smooth southbound rail-sea and international railway intermodal transport, strengthen synergistic and joint development inside and outside the corridor, and encourage low costs and service enhancement in the logistics sector.

Data Availability
Te experimental data used to support the fndings of this study are available from the corresponding author upon request.