Spatial–Temporal Evolution and Influencing Mechanism of Tourism Ecological Efficiency in China

Abstract

Although the development of tourism has a positive impact on local employment and economic growth, the high level of energy consumption and emissions generated by tourism have also attracted increasing attention. Based on the concept of tourism eco-efficiency, this article constructs a comprehensive evaluation system for tourism eco-efficiency in China, and the study concludes that (1) China’s tourism eco-efficiency as a whole was in a fluctuating upward trend from 2002 to 2018, but different provinces of tourism eco-efficiency varied greatly, and the range of tourism eco-efficiency on spatial spill-over increased significantly; (2) the impact of tourism economic development on tourism eco-efficiency was not significant over time, and the impact of tourism eco-efficiency on the sustainable development of the tourism economy in the western region was weaker; (3) the contribution of tourism economic development to tourism eco-efficiency was low at the national level, with regional levels ranging from large to small for the western region, the central region, and the eastern region. The contribution of tourism eco-efficiency to the tourism economy was also low at the national level, with the central, eastern, and western regions ranked in descending order at the regional level.

1. Introduction

With the development of China’s economy and the rise in consumption levels of the population, people’s material living standards are improving and their demand for tourism is becoming increasingly strong [1,2,3,4,5,6,7,8,9,10]. For local economic development, tourism is bringing a positive and significant impact on local employment, local fiscal revenue, local economic growth, and the tourism industry chain [11,12]. In terms of matching supply and demand, tourism has become one of the pillars of industries of China’s national economy, with total domestic tourism arrivals increasing from 2.103 billion to 6.006 billion between 2010 and 2019 and total domestic tourism revenue increasing from RMB 125.780 billion to RMB 5725.092 billion. However, contrary to the traditional perception of “tourism as a green industry”, the high energy consumption and emissions associated with the mass movement of people and tourism activities cannot be ignored [13]. The ecological aspects of tourism are a growing concern and sustainable tourism faces the challenge of balancing tourism growth with ecological protection [13,14]. According to the World Tourism Organisation (UNWTO), tourism carbon emissions from tourism traffic, tourism accommodation, and tourism-related activities account for 3.9% of the total global carbon emissions and 14% of the total global greenhouse benefits caused by the greenhouse effect.

Therefore, it is more urgent and necessary to promote the sustainable development of tourism. Tourism eco-efficiency can be defined as achieving better tourism benefits while consuming fewer natural resources, resulting in lower levels of waste and pollutant emissions [15]. Research on tourism eco-efficiency can help the tourism industry to mitigate its impact on climate change and is a realistic requirement for achieving the “double carbon” goal.

From the existing research, many scholars have identified the level of economic development as an important factor affecting tourism eco-efficiency [15,16], but the literature on the interaction between tourism eco-efficiency and tourism economy is still relatively scarce, and the interactive response between tourism eco-efficiency and tourism economy needs to be further explored [13,14,15]. This study took China as an example to explore the relationship between tourism eco-efficiency and economy. The differences in industrial structure, tourism specialisation, and tourism consumption patterns among Chinese provinces all contribute to the uneven development of regional tourism eco-efficiency in China, reinforcing spatial heterogeneity and making it increasingly difficult to coordinate governance and control among regions [11,16]. Little research has been conducted to examine the spatial and temporal characteristics and mechanisms of tourism eco-efficiency across provinces as well as the existence of convergent spatial and temporal characteristics [15,17]. In view of this, this paper used the models of epsilon-based measure (EBM) and data envelopment analysis (DEA) to measure and analyse the provincial tourism eco-efficiency in China from 2002 to 2018 to explore the temporal evolution of tourism eco-efficiency through the Malmquist–Luenberger (ML) index and to explore the spatial and regional differences in its distribution. On this basis, the interactive response mechanism between tourism eco-efficiency and tourism economic development was explored through the PAVR model.

The main contributions of this study are as follows: (1) combining the concepts of tourism eco-efficiency and the DEA model, a comprehensive evaluation system of tourism eco-efficiency in China was constructed; (2) by introducing the ML index, the static tourism eco-efficiency analysis was expanded into a dynamic analysis, and a cross-period comparative analysis was carried out; (3) the interactive response between tourism eco-efficiency and tourism economy is discussed to further enrich the current research on tourism eco-efficiency.

2. Literature Review

At present, academic research on sustainable tourism [16,17,18], low-carbon tourism [9,10], and green ecotourism focuses on tourism eco-efficiency [13,15], the interaction between tourism eco-efficiency and tourism economy [15,16,17,18,19,20], etc.

First, researchers introduced the concept of eco-efficiency into the framework of sustainable tourism [15,21,22]. As a completely new field, the role of tourism eco-efficiency is to enable the diversification of tourism products and services, reduce waste of resources, reduce pollutant emissions, provide core indicators for effective assessment, and advance the sustainable development of the tourism sector [17,18,19,20,21,22,23]. From the existing research results, researchers have measured tourism eco-efficiency using carbon dioxide emissions as a measure of tourism’s impact on the environment and the ecological footprint as a measure of tourism’s impact on the environment [14,15,24]. For example, Jayasinghe et al. (2021) examined the relationship between energy consumption, carbon dioxide emissions, GDP and international tourist arrivals in India [14]. Liu et al. (2022) evaluated the tourism eco-efficiency of China [17]. However, the aforementioned measures of tourism eco-efficiency were mainly focused on a certain period of time, lacking a measure of a continuous period of time in a certain region and, thus, cannot be studied comparatively in a time series. In conclusion, scholars at home and abroad have generally used indicators such as total tourism arrivals and total tourism revenue to assess the benefits of tourism development and indicators such as tourism carbon footprint and tourism carbon emissions to reflect the environmental impacts of tourism, which essentially further strengthen the need to ensure that resource inputs are minimised, environmental impacts are minimised, economic suspicions are maximised, and the relationship between economic value and environmental impacts is ensured. The relationship between economic value and environmental impact is balanced [14,15,16,17,25].

Second, there is the study of the interactive response of tourism eco-efficiency and tourism economy. Tourism eco-efficiency is currently an important quantitative tool to reflect the harmonious relationship between human and land relations and the degree of sustainable development in a region [14,26,27,28,29], and the process of its generation is closely related to, inseparable from, and interacts with and influences the local tourism economy, which therefore shows that the tourism economy provides a large number of bases for the development of local tourism, including a large amount of tourism infrastructure, a large tourism consumer market and a rich and powerful tourism development potential [30,31,32]. The development of a good tourism economy not only contributes to the development of local scenic areas, the construction of supporting facilities and the growth of tourism attractiveness [33,34] but also to the improvement of local human–land relations [35], such as the improvement of the local greenery, the attraction of ecological conservation funds, the introduction of green and innovative technologies and the increase in energy conservation and environmental awareness among tourism stakeholders [36,37,38], thus contributing to the improvement of the local ecological environment. It is therefore crucial to study the interactive effects of the level of the tourism economy on the ecological development of local tourism [14,33,39,40].

Again, the impact of economic level on tourism eco-efficiency was explored. Tourism eco-efficiency strategies were evaluated based on the data of economic efficiency, land use, employment, and other indicators [41,42], and it is proposed that there are two pathways affecting eco-efficiency: the economic scale growth is accompanied by the consumption of resources and the emission of pollutants, while the relative speed between the ecological environment change and economic growth determines the change of the eco-efficiency [29,43,44]. For example, Castilho et al. (2021) studied the overall ecological efficiency of 22 Latin American and Caribbean countries and found that the arrival of tourists reduces the ecological efficiency of these countries [45]. Based on the background of green development, the co-integration relationship between regional tourism income and economic and ecological effects was analysed, and the spatiotemporal evolutionary trajectory of tourism eco-efficiency and tourism economy was explored to analyse the interactive response relationship between tourism eco-efficiency and tourism economic development [46,47,48,49,50].

3. Methods and Data

3.1. Methods

3.1.1. EBM-ML Super Efficiency Model

Tourism eco-efficiency is an effective indicator of sustainable tourism development, the quality of tourism development, and the harmonious relationship between people and places in a region. Its intrinsic meaning is to maximise the benefits of economic and social output at the cost of minimal resource input and environmental damage. Traditional CCR (abbreviation of Charnes, Cooper, and Rhodes, and Charnes) [51], BCC (abbreviation of Banker, Charnes, and Cooper, and Banker) [52] and slack-based measure (SBM) models have strengths and weaknesses in the input and output ratio changes and radial and non-radial models, based on the ε parameter, i.e., EBM. The initial input-oriented EBM with constant returns to scale is expressed as follows:

$$\begin{array}{l}{\gamma }^{*}=\min \theta -{\epsilon }_x{\displaystyle \sum _{i=1}^m\frac{w_{\overline{s}}}{x_{i_0}}}\\ \theta X_0-X\lambda -\overline{S}=0\\ Y\lambda \ge 0;\lambda \ge 0;\overline{S}\ge 0\end{array}$$

(1)

$$\begin{array}{l}{\gamma }^{*}=\max u{y}_0\\ V{X}_0=1;V{X}_0=1\\ -vX+uY\le 0\\ vi\ge \frac{{\epsilon }_x{\overline{w}}_i}{x_{i0}}(i=1,\cdot \cdot \cdot ,m)\\ u\ge 0\end{array}$$

(2)

where $w_i$ refers to the weight (i.e., relative importance) of input $i$. ${{\displaystyle \sum }}_{i=1}^m{w}_i^{-}=1 \left(w^{-}\ge 0{\forall }_i\right). {\epsilon }_X$ is a key parameter that combines the radial measure $\theta$ and nonradial slack term $\mathrm{s}$. The parameters ${\epsilon }_X$ and $w^{-}$ must be made available before the measurement of the efficiency. According to $\frac{w_i^{-}{s}_i^{-}}{x_{io}}$ of the objective function of the EBM model, the unit of $\frac{s_i^{-}}{x_{io}}$ remains constant, and $w_i^{-}$, which reflects the relative importance of resource $i$, has an invariant unit.

3.1.2. Standard Deviation Elliptic Model

The standard deviation elliptical parametric model identifies spatially extended directions as well as centroid trajectories. The method characterises the spatially distributed elements by means of three elements: standard deviation of the long axis, standard deviation of the short axis, and the directional angle. The standard deviation of the long-axis direction describes the spatial extension of the geographic element, and the larger the indicator, the larger the spatial distribution of the element; the standard deviation of the short-axis direction reflects the spatial density of the geographic element, and the smaller the indicator, the stronger the spatial density of the geographic element; the azimuth angle reflects the deviation of the geographic element from the due north direction.

3.1.3. Panel Vector Autoregression (PVAR) Model

Vector autoregression is a method proposed by researchers to analyse how a particular shock event has an initial and subsequent impact on the outcome of an action. The main principle is to predict the time-series relationship of variables within a system and to explore the relationship between the system and a random disturbance term. As the focus of this paper was on the interaction between the level of tourism economic development and tourism eco-efficiency, the PVAR model in this paper focused on these two variables. In this paper, the level of tourism economic development is denoted as V₁, and the tourism eco-efficiency is denoted as V₂. Then, the PVAR expression for the interactive effect of the level of tourism economic development and tourism eco-efficiency in China can be expressed as

$${\left(\begin{array}{l}{V}_1\\ V_2\end{array}\right)}_{i,t}={\alpha }_i+{\beta }_0+{\displaystyle \sum _{j=1}^p{\beta }_j^{’}}{\left(\begin{array}{l}{V}_1\\ V_2\end{array}\right)}_{i,t-j}+v_{i,t}+{\epsilon }_{i,t}$$

(3)

where ${\alpha }_i$ refers to the individual fixed effect of a specific region, $v_{i,t}$ represents the time fixed effect, and ${\epsilon }_{i,t}$ is the disturbance in a normal distribution.

3.2. Evaluation Indicators and Data Sources

Tourism eco-efficiency-related indicators (

Table 1. Creation of indicators measuring the eco-efficiency of tourism and data sources.

Indicator Categories	Measurement Methods	Parameters	Data Source
Inputs
Personnel inputs in tourism	Directly acquired	Tourism-related employment (10,000 persons)	Yearbook of China Tourism Statistics
Capital inputs in tourism	Directly acquired	Original value of fixed assets in tourism (RMB 10,000)	Yearbook of China Tourism Statistics
Energy inputs in tourism	Bottom-up approach	Tourism transport—Passenger turnover (10,000 persons)	China’s regional database
		Tourism transport—Total number of tourists (10,000 persons)	China’s regional database
		Accommodation in tourism—Number of beds in star-rated hotels (10,000 persons)	Yearbook of China Tourism Statistics
		Accommodation in tourism—Average occupancy rate of star-rated hotels (%)	Yearbook of China Tourism Statistics
		Tourist activities—Total number of domestic tourist arrivals (10,000 persons)	Statistical yearbooks of provincial-level administrative divisions concerned
		Tourist activities—Total number of foreign tourist arrivals (10,000 persons)	Yearbook of China Tourism Statistics
		Tourist activities—Domestic tourist destinations (%)	Yearbook of China Tourism Statistics
		Tourist activities—Foreign tourist destinations (%)	Yearbook of China Tourism Statistics
Desirable outputs
Total number of tourist arrivals	Directly acquired	Domestic tourist arrivals (10,000 persons)	Yearbook of China Tourism Statistics
		Inbound tourist arrivals (10,000 persons)	Yearbook of China Tourism Statistics
Total tourism receipts	Directly acquired	Domestic tourism receipts (RMB 10,000)	Yearbook of China Tourism Statistics
		Foreign exchange earnings from international tourism (USD 10,000)	Yearbook of China Tourism Statistics
Undesirable outputs
Carbon dioxide emissions of tourism	Bottom-up approach	Tourism transport—Volume of passenger transportation (10,000 persons)	China’s regional database
		Tourism transport—Total number of tourists (10,000 persons)	China’s regional database
		Accommodation in tourism—Number of beds in star-rated hotels (10,000 persons)	Yearbook of China Tourism Statistics
		Accommodation in tourism—Average occupancy rate of star-rated hotels (%)	Yearbook of China Tourism Statistics
		Tourist activities—Total number of domestic tourist arrivals (10,000 persons)	Provincial-level statistical yearbooks
		Tourist activities—Total number of foreign tourist arrivals (10,000 persons)	Yearbook of China Tourism Statistics
		Tourist activities—Domestic tourist destinations (%)	Yearbook of China Tourism Statistics
		Tourist activities—Foreign tourist destinations (%)	Yearbook of China Tourism Statistics

): Tourism eco-efficiency is mainly composed of input and output indicators. Among them, tourism input indicators include tourism capital input, tourism human input, and tourism energy input. The desired tourism outputs include total tourism arrivals and total tourism revenues. The undesired outputs of tourism include tourism carbon emissions. In this paper, the number of people employed in the tourism industry in 30 provinces and regions from 2002 to 2018 was selected from the China Tourism Statistical Yearbook as the tourism personnel input of the region. The original value of the fixed assets in the tourism industry was taken as the asset input of the tourism industry in the region, and the raw data were calculated twice using the depreciation rate and perpetual inventory method. A “bottom-up” approach was used to construct the energy consumption of the tourism industry.

Thirty Chinese provinces and regions were selected for their tourism revenue (domestic tourism revenue and foreign exchange tourism revenue), tourism attractiveness (domestic tourism arrivals and inbound tourism arrivals), tourism structural development (tourism revenue as a proportion of GDP and tourism revenue as a proportion of tertiary industry), and tourism development potential (total tourism revenue growth rate and total tourism arrivals growth rate) to measure the level of economic development of tourism in China’s provinces and regions in four areas. The data involved were obtained from the China Tourism Database, the China Regional Database, the China City Database, as well as the China Statistical Yearbook and the statistical yearbooks of each province. Some of the missing data were filled in by interpolation (

Table 2. Selected indicators measuring the economic gains of tourism and data source.

Indicators	Subindicators	Database
Tourism receipts	Domestic tourism receipts	China’s state-level tourism database
	Foreign exchange earnings from international tourism	China’s state-level tourism database
Destination attractiveness	Domestic tourist arrivals	China’s state-level tourism database
	Inbound tourist arrivals	China’s state-level tourism database
Structural development of tourism	Percentage of tourism receipts in GDP	China’s regional-level database
	Percentage of tourism receipts in added value of the tertiary industry	China’s regional-level database
Tourism potential	Growth rate of total tourism receipts	China’s state-level tourism database
	Growth rate of total number of tourist arrivals	China’s state-level tourism database

).

4. Analysis of the Results

4.1. Spatial and Temporal Patterns of Tourism Eco-Efficiency

4.1.1. The Time-Series Evolution of Tourism Eco-Efficiency

This section measures tourism eco-efficiency in a temporal dimension. The overall level of the EBM-ML index of tourism eco-efficiency in China from 2002 to 2018 showed a slightly fluctuating upward trend (Figure 1). It gradually rose from a historical low of 0.886 in 2003 to 1.063 in 2005, before falling sharply in 2006. This was followed by a gradual rise from 2008 to 2010 to a peak of 1.129 in 2010, after which the gradual decline levelled off. The year 2017 saw a rapid decline to 0.963 followed by a slow increase to 1 in 2018. The EBM-ML index of the tourism eco-efficiency in China from 2004 to 2018 was greater than 1 overall, indicating that China’s tourism eco-efficiency was in an upward trend, which also indicates that China’s tourism eco-effect was growing and the quality of the tourism eco-environment was somewhat protected.

4.1.2. Analysis of the Trajectory of the Centre of Gravity of Tourism Eco-Efficiency

This section analyses the changes in the trajectory of the centre of gravity of tourism eco-efficiency and reveals its spatiotemporal evolution pattern (Figure 2). Compared to the visualised tourism eco-efficiency distribution map, the standard deviation elliptical parametric model can effectively capture the spatial shift of the regional tourism eco-efficiency, resolve the trend of tourism eco-efficiency changes in different years, and provide a more adequate explanation for the trajectory changes of tourism eco-efficiency in China from 2002 to 2018. Overall, the provinces with a higher tourism eco-efficiency in China were widely distributed across a vast region centred on Henan, north to Beijing–Tianjin–Hebei, east to Shanghai, south to Guangdong and Guangxi, and west to parts of Sichuan, Guizhou and Gansu, reflecting the declining tourism eco-efficiency in China’s northeast and northwest regions. This is largely consistent with the current state of China’s tourism economy as well as its ecological protection.

In terms of the trajectory and direction of the movement of the centre of gravity, the four years of 2004, 2008, 2014, and 2018 were used as the study’s cross-section: (112.226, 34.005), (112.171, 34.288), (112.059, 33.770), and (112.101, 33.754). In addition, from 2004 to 2008, the centre of gravity of China’s regional tourism eco-efficiency shifted towards the northwest, indicating an increase in the tourism eco-efficiency in the northwestern region and a decrease in the tourism eco-efficiency in the southeastern coastal region; from 2008 to 2014, China’s tourism eco-efficiency gradually shifted towards the southwest, indicating that at this time the northeast’s tourism eco-efficiency declined while that of the southwest increased; from 2014 to 2018, China’s tourism eco-efficiency underwent a slight shift towards the southeast. Overall, over the 17-year period of 2002–2018, the centre of gravity of the tourism eco-efficiency in China showed a shift from the northeast to the southwest.

In terms of the long-axis standard deviation, the long-axis standard deviation at the four points in time showed a general upward trend, decreasing from 1007 km in 2004 to 985 km in 2008, then extending to 1010.8 km in 2014, and finally reaching 1010.7 km in 2018. This trend indicates that more areas on the northeast and southwest trend increased the ecological efficiency of the tourism efficiency, resulting in a larger and larger area. This cannot be separated from the green and sustainable development strategy that China has been adhering to for a long time. In addition, the standard deviation of the short axis did not change significantly over the four time points, with the standard deviation of the short axis being 1134 km, 1182 km, 1157 km, and 1159 km in each of the four years in the western provinces, such as Shaanxi and Gansu.

The change in the azimuthal angle shows that the azimuthal angle also tended to be stable at the four time points, with angles of 29.469°, 37.305°, 28.799°, and 25.547°, respectively. This indicator reflects a 3.922° deflection in the spatial pattern of China’s regional tourism eco-efficiency from a due east–south angle to a northeast–southwest angle (

Table 3. Parameters of the standard deviational ellipse of the eco-efficiency of tourism in China from 2002 to 2018.

Year	Geographic Coordinates of Centre Point		Length of Long Axis	Length of Short Axis	Rotation Angle
	Longitude	Latitude
2004	112.226	34.005	1,007,205.941 m	1,134,766.96 m	29.469°
2008	112.171	34.288	985,885.404 m	1,182,140.371 m	37.305°
2014	112.059	33.770	1,010,879.632 m	1,157,689.127 m	28.799°
2018	112.101	33.754	1,010,710.972 m	1,159,984.179 m	25.547°

).

4.2. Interactive Response Mechanisms of the Tourism Eco-Efficiency and Tourism Economic Development

4.2.1. Description of the Tourism Eco-Efficiency Variables and Smoothness Tests

The following is the descriptive analysis of the tourism eco-efficiency and economic development. The descriptive statistics of the domestic tourism revenue, foreign exchange tourism revenue, domestic tourism arrivals, inbound tourism arrivals, tourism revenue as a proportion of GDP, tourism revenue as a proportion of the tertiary sector, and the growth rate of total tourism revenue are shown in the table (

Table 4. Descriptive statistics of the indicators of economic gains of tourism in China.

Variable	Abbreviation	Observed Value	Mean	Min.	Max.	Variance
Domestic tourism receipts	lnDOMECO	510	15.98	6.980	18.67	1.460
Foreign exchange earnings from international tourism	lnFORECO	510	10.92	4.430	14.53	1.750
Domestic tourist arrivals	lnDOMPEO	510	9.140	5.580	11.36	1.170
Inbound tourist arrivals	lnFORPEO	510	4.780	−1.190	9.260	1.540
Percentage of tourism receipts in GDP	TOU GDP	510	0.100	0.0200	0.460	0.0500
Percentage of tourism receipts in added value of the tertiary industry	TOU THREE	510	0.230	0.0500	1	0.100
Growth rate of total tourism receipts	INPEO	510	0.170	−0.880	11.54	0.530
Growth rate of total number of tourist arrivals	INECO	510	0.180	−0.310	1.250	0.150

).

As nonstationary panel data usually lead to the occurrence of pseudo-regressions, which affect the validity and unbiasedness of the estimates, in order to avoid this phenomenon, PVAR models usually need to be tested for unit roots prior to testing. Common methods include those consistent with previous panel unit roots, such as the Levin–Lin–Chu (LLC) and Phillips–Perron (PP)–Fisher methods. In this paper, we proposed to adopt the aforementioned augmented Dickey–Fuller test (ADF), LLC, and PP to conduct a smoothness test on the level of tourism economic development and tourism eco-efficiency using Stata 17.0 software. The results are shown below in

Table 5. Variance decomposition of the eco-efficiency and economic gains of tourism.

	ADF		LLC		PP		Conclusion
	Statistical Value	Significance Level	Statistical Value	Significance Level	Statistical Value	Significance Level
Economic gains of tourism
The whole of China	−4.1948	0.0000	−8.5029	0.0000	−5.1046	0.0000	Stationary
East of China	−6.2676	0.0000	−3.9974	0.0000	−2.6540	0.0051	Stationary
Middle of China	−6.1792	0.0000	−2.7614	0.0000	−3.707	0.0003	Stationary
West of China	−1.8528	0.0320	−0.6126	0.2701	−5.0565	0.0000	Nonstationary
Eco-efficiency of tourism
The whole of China	−11.2259	0.0000	−8.3394	0.0000	−16.747	0.0000	Stationary
East of China	−10.5526	0.0000	−7.1695	0.0000	−9.8705	0.0000	Stationary
Middle of China	−3.4269	0.0003	−3.8578	0.0001	−11.280	0.0000	Stationary
West of China	−3.5806	0.0002	−3.3421	0.0004	−9.8149	0.0000	Stationary

. It can be seen that most of the indicators of the level of tourism economic development were smooth. For nonstationary data, a lagged first-order approach was used to deal with the data.

4.2.2. Analysis of the Impulse Impact of Economic Development on Tourism Eco-Efficiency

In order to better estimate the PAVR model of tourism economic development on tourism eco-efficiency, this study used the Akaike information criterion (AIC), Bayesian information criterion (BIC), and Hannan–Quinn information criterion (HQIC) methods to discern the optimal lag order of the model. The results show that for the nationwide and eastern regions only, each indicator showed the most appropriate lag term with the third order, while in the central and western regions, all four indicators were significant at the first-order lag term. Therefore, in this paper, the third-order lag term was used to calculate the relationship between tourism economic development and tourism eco-efficiency at the national level and in the eastern region, while the first-order lag term was used to calculate the relationship between tourism economic development and tourism eco-efficiency in the central and western regions.

An impulse response analysis was conducted on the relationship between tourism economic development and tourism eco-efficiency. Specifically, the paper used a Monte Carlo method to simulate 200 analyses and select 10 period dynamics for each variable to be plotted, as shown below in Figure 3. The horizontal axis is the number of periods of dynamic response to a shock, and the vertical axis is the degree of impact of the shock on the variable. The solid line indicates that the effect of the shock on the variable was zero, and the grey areas above and below indicate 95% confidence intervals.

The relationship between tourism economic development and the shock response of tourism eco-efficiency was analysed through an impulse response model. The results show that at the national level, the development of the tourism economy had a significant impact on the tourism eco-efficiency in the early stages, but over time, the impact of tourism economic development on the tourism eco-efficiency was no longer significant. In fact, the greatest impact of the development of the tourism economy on the tourism eco-efficiency was in the improvement of the tourism infrastructure and the expansion of the tourism consumer market, thus increasing its economic efficiency, but this effect was evident in the early stages, and after a certain level of economic development, the stimulating effect of the infrastructure and the consumer market on the tourism eco-efficiency gradually decreased, giving way to the impact of the improvement of the technological level.

In the eastern region, the development of the tourism economy had a dampening effect on the tourism eco-efficiency followed by a boosting effect. One reason for the negative impact of economic development on the tourism eco-efficiency may be that the economic development of the eastern region contributed to the influx of tourists, but the lack of effective green management and energy conservation measures reduced tourism eco-efficiency. As the attractiveness of economic development to tourists diminished, the eco-efficiency was restored. The left-hand graph in the second row shows that the tourism eco-efficiency in the eastern region had an M-shaped effect on the tourism economy, i.e., it increased, then decreased, then increased again and then decreased and plateaued. As can be seen, tourism eco-efficiency promoted the human–land relationship which, in turn, contributed to a stronger growth momentum in the region’s economy but then caused cost increases and economic disincentives; however, this disincentive disappeared in the medium term.

In the central region, the development of the tourism economy had a catalytic and then inhibitory effect on the tourism eco-efficiency. This may be due to the fact that economic inputs in the central provinces promoted the development of local tourism infrastructure and the expansion of tourism consumer markets, thus contributing to the impact of the economic effects of tourism eco-efficiency. Subsequently, however, the shock began to decrease and gave way to technological progress. The left panel in the second row shows that tourism eco-efficiency in the central region had a decreasing and then increasing impact on the tourism economy, i.e., from period 1, the tourism eco-efficiency decreased and then fluctuated upwards and levelled off. As can be seen, the tourism eco-efficiency promoted the relationship between people and places, thus contributing to a stronger economic growth momentum in the region but then caused a cost increase and economic disincentive; however, this disincentive disappeared in the medium term, similar to the effect in the east.

In the west, the development of the tourism economy had a catalytic and then inhibitory effect on the tourism eco-efficiency. Economic investment in the western provinces contributed to the development of local tourism infrastructure and the expansion of the tourism consumer market, but this effect was very short lived. For the western region, increasing the tourism economy was not a good strategy to improve the tourism eco-efficiency in the long term. The left panel in the second row shows that the tourism eco-efficiency in the west had a decreasing and then increasing effect on the tourism economy, which also lasted very briefly. In other words, the contribution of tourism eco-efficiency in the west to the sustainability of the tourism economy was weak.

4.2.3. Analysis of the Interaction Mechanisms between Tourism Eco-Efficiency and Economic Development

Through an equation decomposition analysis, the PVAR model can explore the value of the contribution of one variable to the impact of another variable. Therefore, this paper explored the impact of tourism economic development on the tourism eco-efficiency through equation decomposition. As previously mentioned, this paper explored the contribution values of the impact of tourism economic development on the tourism eco-efficiency at the national level, in the eastern region, in the central region, and in the western region, as shown below in

Table 6. Variance decomposition of the eco-efficiency of tourism and economic gains of tourism.

		The whole of China		East of China		Middle of China		West of China
		(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Economic gains of tourism, V1	lag order	V1	V2	V1	V2	V1	V2	V1	V2
	1	1	0	1	0	1	0	1	0
	2	0.9997	0.0003	0.9967	0.0033	0.9941	0.0059	0.9977	0.0023
	3	0.9992	0.0008	0.9962	0.0038	0.9924	0.0076	0.9932	0.0068
	4	0.9985	0.0015	0.9951	0.0049	0.9929	0.0071	0.9920	0.0080
	5	0.9985	0.0015	0.9950	0.0050	0.9921	0.0079	0.9919	0.0081
	6	0.9985	0.0015	0.9950	0.0050	0.9919	0.0081	0.9917	0.0083
	7	0.9985	0.0015	0.9950	0.0050	0.9920	0.0080	0.9916	0.0084
	8	0.9985	0.0015	0.9950	0.0050	0.9919	0.0081	0.9916	0.0084
	9	0.9985	0.0015	0.9950	0.0050	0.9919	0.0081	0.9916	0.0084
	10	0.9985	0.0015	0.9950	0.0050	0.9919	0.0081	0.9916	0.0084
Eco-efficiency of tourism, V2	lag order
	1	0.0003	0.9997	0.0030	0.9970	0.0031	0.9969	0.0007	0.9993
	2	0.0074	0.9926	0.0346	0.9654	0.0081	0.9919	0.0010	0.9990
	3	0.0080	0.9920	0.0602	0.9398	0.0124	0.9876	0.0050	0.9950
	4	0.0214	0.9786	0.0654	0.9346	0.1235	0.8765	0.0049	0.9951
	5	0.0223	0.9777	0.0654	0.9346	0.1376	0.8624	0.0049	0.9951
	6	0.0224	0.9776	0.0655	0.9345	0.1392	0.8608	0.0051	0.9949
	7	0.0227	0.9773	0.0657	0.9343	0.1465	0.8535	0.0051	0.9949
	8	0.0227	0.9773	0.0657	0.9343	0.1477	0.8523	0.0051	0.9949
	9	0.0227	0.9773	0.0657	0.9343	0.1482	0.8518	0.0051	0.9949
	10	0.0228	0.9772	0.0657	0.9343	0.1486	0.8514	0.0051	0.9949

. As can be seen from column 2 of the table, at the national level, the highest value of the contribution of tourism economic development to the tourism eco-efficiency was 0.0015, implying that the contribution of the tourism economy to the tourism eco-efficiency was low. At a regional level, the contribution of tourism economic development to the tourism eco-efficiency was in descending order in the western region (column 4, 0.0084), the central region (column 6, 0.0081), and the eastern region (column 8, 0.0050). In other words, the impact of tourism economic development on the tourism eco-efficiency was greater in the western region than in the central and eastern regions. This may be due to the fact that the eastern and central regions had a stronger economic base and were at a higher level themselves, with the economic effect on the tourism eco-efficiency being at a very low marginal benefit, and the increase in eco-efficiency relying more on technological advances and the introduction of new management techniques. In contrast, the increase in the tourism eco-efficiency was more closely linked to the economic development of tourism in the western region, where the economic development of tourism was more closely linked to the economic development of tourism, as the impulse impact model in the previous section showed that the upfront impact of tourism economic development on the tourism eco-efficiency was greater. In addition, the national contribution of the tourism eco-efficiency to the impact of the tourism economy was relatively low at 0.0228.

At a regional level, the contribution of tourism eco-efficiency to the tourism economy was ranked from the largest to smallest in the central region (column 5, 0.1486), the eastern region (column 3, 0.0657), and the western region (column 7, 0.0051). In general, the impact of the tourism eco-efficiency on tourism economic development was higher in the central region than in the eastern and western regions. In fact, the main role of the tourism eco-efficiency is to improve the relationship between people and land and to increase the income of the local tourism economy from the perspective of sustainable development. However, the increase in the tourism eco-efficiency also means that more money needs to be invested in technology development, which can limit further economic development. As the western region itself has a relatively weak tourism economy and limited technology, improving tourism eco-efficiency is not very significant for local economic development but rather detrimental to economic development and should be focused on improving local tourism infrastructure and tourism consumer markets. In contrast, in the central–eastern region, it is more beneficial to the tourism economy to improve the concept of sustainable tourism.

5. Conclusions and Policy Implications

Based on the concept of tourism eco-efficiency, this article constructed a comprehensive evaluation system of the tourism eco-efficiency in China and integrated the BM-ML super-efficiency model, the standard elliptical model and the PVAR model to measure the mechanism of the role between tourism eco-efficiency and tourism economic development. This study drew the following main conclusions and contributions.

First, based on a dynamic analysis of the tourism eco-efficiency in China from 2002 to 2018, this paper constructed a comprehensive evaluation system of China’s tourism eco-efficiency and expanded previous studies on tourism eco-efficiency [14,15,16]. From the dynamic evolution of time, the regional tourism eco-efficiency as a whole was in a fluctuating upward trend, which shows the contribution of China’s policies to tourism development and environmental protection. This result is consistent with the finding of Zha [15]. In terms of the spatial distribution characteristics, the tourism eco-efficiency of the different provinces varied greatly, and the spatial shift effect between provinces increased, the scope of tourism eco-efficiency in terms of spatial spill-over increased significantly and the linkage effect of tourism economic synergy and tourism carbon emissions between different regions gradually improved. In terms of the spatial trajectory shifts, the centre of gravity of the tourism eco-efficiency in China showed an overall shift towards the southwest during the 17-year period from 2002 to 2018. Through a time and space analysis, the original static tourism eco-efficiency analysis was further extended to a dynamic analysis, and a cross-regional and cross-period comparative analysis was carried out to make the conclusion more robust [15,16].

Second, in response to the research by Peng et al. (2020) and Zha et al. (2020), this study explored the relationship between tourism economic development and tourism eco-efficiency through the PVAR model and provides policy suggestions for the sustainable development of tourism [16,21]. The results of the impulse response model show that at the national level, the development of tourism economy had a significant impact on the tourism eco-efficiency in the early stage but not significantly over time. There were also regional differences in the interaction mechanism between the development of tourism economy and tourism eco-efficiency; for example, the tourism eco-efficiency had a weaker impact on the sustainable development of the tourism economy in the western region. A variance decomposition model was used to explore the contribution of tourism economic development and tourism eco-efficiency to each other. The results showed that the contribution of tourism economic development to the tourism eco-efficiency was low at the national level, with regional levels ranging from large to small in the western, central and eastern regions. The contribution of tourism eco-efficiency to the tourism economy was also low at the national level, with the central, eastern, and western regions ranked in descending order at the regional level.

Through valid indicators, such as tourism development quality, harmonious relationship between people and places within the region, and sustainable tourism development, this study evaluated the tourism eco-efficiency of China and found that value and problems of the sustainable development of China’s tourism, which provides policy enlightenment for promoting the sound development of China’s tourism industry [14,15,16]. Based on the above findings, this paper proposed that the collaborative effect of tourism between regions should be strengthened, policy barriers to tourism resource elements should be broken down, the resource sharing of inputs and outputs of regional industries should be promoted, unbalanced spatial development should be reduced, and the goal of synergistic and optimal development should be reached, in addition to optimising the spatial structure of tourism and industrial upgrading, effectively optimising the regional tourism spatial network, expanding the service scope of the tourism industry, realising the sharing of tourism industry outputs, and promoting the interaction between regional tourism networks. Furthermore, due to the regional differences in the contribution rate of tourism eco-efficiency to tourism economy, the Chinese government should improve the tourism economy of different regions from the perspective of sustainable development so as to narrow the development gap between regions.

Although this paper analysed the interactive response relationship between tourism eco-efficiency and tourism economy, there are still some problems worthy of further study. This paper only quantified the direct impact between tourism eco-efficiency and tourism economy. However, tourism economy may indirectly affect tourism eco-efficiency through the allocation of tourism resources and human capital [16,50]. Therefore, future scholars can explore the intermediary mechanism between the two from the above perspective. In addition, due to the absence of some data, the time span of this study only covered 2002–2018 and has not been updated to the latest data (such as 2019–2021). Future studies can add the latest data to consider the impact of COVID-19.