Spatial Spillover and Threshold Effects of High-Quality Tourism Development on Carbon Emission Efficiency of Tourism under the “Double Carbon” Target: Case Study of Jiangxi, China

Abstract

High-quality tourism development under the “double carbon” target (the peaking of carbon emissions and achievement of carbon neutrality) is an important path to achieving low-carbon emissions in the tourism industry and is vital for improving the industry’s carbon emissions efficiency. Using spatial and temporal panel data for 11 prefecture-level cities in Jiangxi Province from 2000 to 2020, a spatial Durbin model and a threshold model were constructed to assess the spatial spillover and threshold effects that high-quality tourism development has on the carbon emission efficiency of the tourism industry. The three key results were as follows. (1) There is a non-linear relationship between the carbon emission efficiency of tourism and the high-quality development trend of tourism, with differences in spatial distribution. (2) Coordinated development, green development, and open development all have significant positive direct effects on the carbon emission efficiency of tourism. Innovation-driven and coordinated development have a positive spillover effect on the carbon emission efficiency of tourism. In contrast, green development, open development, and shared results have a negative spatial spillover effect. (3) When the scale of the tourism economy crosses the first threshold in the second stage and the structure of tourism investment crosses the second threshold in the third stage, the ability of the tourism quality development to enhance the tourism carbon emission efficiency is the largest. When the tourism investment structure and tourism carbon emission intensity cross a single threshold, the role of the tourism quality development level in enhancing the tourism carbon emission efficiency decreases. Accordingly, three types of countermeasures are proposed: solving development problems, tapping into positive spillovers, and scientifically describing the impact of thresholds. The ultimate goal of this is to provide theoretical references and innovative ideas for promoting green, low-carbon, and high-quality development of tourism in Jiangxi Province and elsewhere.

1. Introduction

The rapid growth of the tourism economy has led to a series of problems around the world, including environmental pollution and carbon dioxide emissions [1]. China’s huge domestic and international tourism market has led to a significant incremental increase in total carbon emissions from the tourism industry between 1990 and 2019, with a slight decrease in 2020 due to the impact of the New Crown epidemic and a relatively sloppy development model for China’s tourism economy.

General Secretary Xi Jinping pointed out in his government report that “green water and green mountains are the silver mountains of gold [2]”, and, in 2020, China made efforts to achieve carbon peaks by 2030 and carbon neutrality before 2060 (referred to as the “double carbon” target) [3]. The introduction of the “double carbon” target has elevated China’s green development path to a new level, becoming one of the keynotes of China’s socio-economic development in the coming decades.

Based on the province’s rich tourism resources, Jiangxi has implemented a strategy to strengthen tourism and achieved rapid growth in its tourism economy. Early tourism development in Jiangxi Province built tourism facilities, hotels, and restaurants in a way that was at the expense of the ecological environment. The large number of tourists visiting Jiangxi has brought economic benefits but damaged its ecological environment [4]. Total carbon emissions from tourism in Jiangxi Province continue to rise and show “fluctuations” in carbon efficiency [5]. Uneven tourism development, low competitiveness in the same industry, tourism traffic restrictions, few inbound tourists, and many tourism gaps also constrain, to some extent, the quality development of tourism in Jiangxi Province.

In recent years, Jiangxi Province has been guided by the “double carbon” target and the “new development concept”, with the low carbon transformation and high-quality development of the tourism industry on the agenda. How to leverage tourism development opportunities to reduce carbon emissions and improve carbon efficiency in the tourism industry is important to achieve the goal of integrating “dual carbon” objectives into high-quality tourism development and low-carbon transformation.

2. Literature Review

2.1. Carbon Efficiency in the Tourism Sector

The carbon efficiency of the tourism sector is a reflection of the carbon emissions caused by the production of benefits from tourism production activities [6]. Increased carbon efficiency in the tourism sector indicates that higher economic efficiency is achieved while generating fewer carbon emissions, which creates the conditions for accelerating the achievement of carbon peaking and carbon neutrality.

Worldwide research on the carbon efficiency of tourism has been fruitful, with scholars focusing on the carbon efficiency measurement [7], spatial relationships [8], and impact factors [9]. For example, the carbon efficiency of China’s tourism industry increased significantly and with spatially significant differences between 2000 and 2019 [10]. The carbon emission efficiency of tourism in the Yangtze River Economic Zone shows an upward trend from 2011 to 2017 and has a significant positive effect on the level of economic development, specialization, urbanization, and energy consumption [11]. Li D et al. [12] demonstrate that the level of the tourism economy, environmental regulation, and ecological construction contribute to the efficiency of carbon emissions from forest park tourism in China.

In terms of research methodology, most scholars in the early years chose to evaluate the carbon efficiency of the tourism industry using single-factor indicators for tourism carbon emissions and tourism revenues [13]. The construction of all-factor indicators has become a popular method of research in recent years, the most common of which is data envelopment analysis (DEA). Fare et al. [14] were the first to evaluate environmental efficiency through undesired outputs and apply DEA models to allow for asymmetric treatment of desired and undesired outputs. Gossling et al. [7] are the first to analyze the interaction between economic benefits and environmental damage in the context of tourism to calculate the carbon efficiency of tourism. The SBM model is used to calculate the carbon efficiency of China’s tourism industry by including tourism carbon emissions in the non-desired output [15], and it demonstrates that the carbon efficiency of China’s inter-provincial tourism industry is “high in the east and low in the west” and has significant aggregation characteristics [16], which is a further improvement on the DEA model.

2.2. The Level of High-Quality Development of Tourism

The “double carbon” target should also be considered in the context of a strategy for quality development. High-quality tourism development is closely related to low-carbon emission reduction, and it is an important factor in achieving mutual harmony between “green mountains” and “golden mountains” [17]. That research on quality tourism development focused on horizontal measurement [18], spatial patterns [19], and influencing factors [20]. Studies find that quality tourism development and environmental sustainability go both ways [21]. Moreover, there is research to prove that technological innovation, and industry coordination, can contribute to high-quality tourism development [22] and that excessive openness and urbanization can inhibit tourism development [23].

Driven by the goal of “double carbon” and the promotion of an “ecological civilization”, the high-quality development of tourism needs to be driven by innovation as the main driving force, coordinated development as an endogenous feature, green development as a universal form, open development as the necessary path, and the sharing of results as the fundamental purpose [24]. The new development concept to construct an evaluation system for the high-quality development level of tourism has received widespread attention from Chinese scholars. The five dimensions of “innovation, coordination, green, openness and sharing” measure the level of quality tourism development at national [25], basin [26,27], and inter-provincial levels [28], respectively, proving that the overall level of tourism in China is on the rise.

2.3. Study on the Impact of High-Quality Tourism Development on the Carbon Efficiency of the Tourism Industry

High-quality tourism development and carbon emission reduction are closely related, and the dialectical relationship between tourism economic development and ecological environmental protection needs to be correctly grasped. More recently, scholars have combined research on high-quality development with research on carbon emissions and carbon efficiency in tourism. Research from one of the research perspectives, such as innovation-driven, coordinated the development, the green development, the open development, and the sharing of results.

The study found that technological innovation in tourism [29] and the promotion of renewable energy [30] can significantly improve carbon efficiency. Coupled economic-tourism–environmental coordination can promote coupled coordination in the western provinces and cities in the region and in neighboring regions [31]. Green development has a “push-back” effect on the efficiency of carbon emissions from tourism [32], as well as a “green paradox” effect [33]. The environmental impact of open development is equally mixed, according to Mahrinasari et al. [34], suggesting a positive correlation between carbon emissions and trade liberalization. However, Ling et al. [35] demonstrate the impact of trade openness on environmental degradation in five Asian countries. Under the role of new urbanization, tourism development can curb the growth of carbon emissions from tourism [36]. However, studies have shown that higher populations and incomes in developing and emerging countries have a negative impact on environmental performance [37]. Tourism energy technology upgrading, tourism economic development, tourism green development, tourism trade opening, and tourism urbanization are all inseparable from the scientific guidance and guarantee of high-quality tourism development.

The research mainly covers the interaction between tourism quality development and tourism carbon efficiency [38], the evaluation of coupled and coordinated development [39], the analysis of influencing factors [40], and theoretical values [41]. For example, tourism development in Belt and Road countries [42] and African countries [43] can enhance carbon performance. Other focuses have included weak decoupling and negative decoupling between tourism economic growth and tourism carbon efficiency in China [44]. Some scholars have also explored achieving high-quality tourism economic development while reducing carbon emissions and have provided countermeasures on how to achieve sustainable tourism development [45].

The main research methods are macro qualitative descriptions [46], coupled coordination models [47], LMDI decomposition methods [48,49], and decoupling models [50]. In addition, the most common quantitative research methods include spatial econometrics [51,52] and threshold models [53]. It has been demonstrated that there are positive direct and negative indirect effects of tourism development on the carbon efficiency of tourism in 70 tourism countries around the world [52]. The spatial spillover effect of tourism development on urban carbon emissions in China is characterized by a “U” and “M” pattern [54]. Studies have been carried out on carbon emissions in Mediterranean countries using “a number of tourists” and “tourism regime” as threshold variables [55]. Chinese scholars have used “environmental regulation” [56] and “urbanization” [36] as threshold variables to explore the efficiency of carbon emissions from tourism in China.

Although studies have helped to analyze the interrelationship between the high quality of tourism and the carbon efficiency of tourism, there are still some limitations. Existing research perspectives are biased toward studying the spatial spillover effects of carbon efficiency in tourism from a single dimension of tourism quality. Secondly, there is more analysis of the factors directly affecting the level of quality development and the carbon efficiency of tourism, and insufficient attention has been paid to the study of heterogeneity under different threshold variables. These research gaps leave a series of open questions: Do the five dimensions of quality tourism development push the region’s tourism industry to become more carbon efficient? Does high-quality tourism development in the region affect the carbon efficiency of tourism in neighboring regions? Does high-quality tourism development under different threshold variables promote carbon efficiency in tourism?

Therefore, from the perspective of “double carbon”, the text uses the non-expected output Super-SBM model and entropy method to construct an evaluation system to scientifically measure the carbon emission efficiency and the level of quality tourism development in Jiangxi Province. Using the spatial Durbin model, the innovative analysis of the spatial spillover effect between the five dimensions of the level of quality tourism development and the carbon efficiency of tourism was conducted. By constructing a threshold model, the heterogeneous impact of the level of quality tourism development on the carbon efficiency of the tourism industry at different stages is analyzed to provide a scientific basis for the sustainable development of tourism in Jiangxi Province and the early achievement of the “double carbon” target.

3. Study area, Methods, and Variable Selection

3.1. Study Area

Jiangxi Province has a good ecological background and is one of the first ecological civilization pilot zones in China, with a forest coverage rate of 63.350%, ranking second in China. Poyang Lake basin is a first-class basin of the Yangtze River, and 96.6% of the province of Jiangxi belongs to the Poyang Lake water system. The Poyang Lake wetland in the territory is the number one freshwater lake ecological wetland in China. Jiangxi is rich in red, ecological, rural, and recreational tourism resources, with the title of “Red Cradle and Green Home”, and is ranked third in the list of China’s comprehensive competitiveness in county tourism. The provincial government attaches great importance to the development of tourism, earnestly implements the national strategy of the Poyang Lake Ecological Economic Zone, and accelerates the construction of the Poyang Lake Ecological Tourism Demonstration Zone, which is of great strategic significance to promote sustainable development. Jiangxi’s high-quality ecological background, complete study basin, rich tourism resources, and important strategic position make it stand out among the case sites for studying the low-carbon transformation and the high-quality development of tourism.

3.2. Methods

3.2.1. Measurement of Carbon Efficiency in the Tourism Sector

The “bottom-up” approach requires the analysis of data on visitors arriving at a destination to cascade upwards in terms of energy and emissions. China does not have statistics on energy consumption in the tourism or service sectors and has not established a dedicated database of tourism satellite carbon accounts [57]. Therefore, through surveys and literature research, the energy consumption and carbon emission factors of tourism accommodation, tourism transportation, tourism activities, tourism catering, and tourism waste in Jiangxi Province were determined. The total carbon emissions from tourism in Jiangxi Province can be obtained by summing up the carbon emissions from each segment measured separately using the “bottom-up” method.

The non-desired output Super-SBM model is used to calculate the efficiency of carbon emissions from tourism in Jiangxi Province. This addresses the problem of including only the proportion of equivalent increases or decreases in inputs and outputs, without considering the slack improvement part of the efficiency bias. The approach also considers the impact of non-desired outputs (CO₂ emissions) so that the problem of “bad outputs” such as pollution emissions and ecological damage in economic activities can be effectively analyzed. This fits the objective of increased desired outputs and reduced inputs and undesired outputs under green development.

The associated model calculation is as follows:

$$\min \mathsf{\rho }=\frac{1-\frac{1}{\mathrm{C}}{\sum }_{\mathrm{c}=1}^{\mathrm{C}}\frac{{\mathrm{x}}_{\mathrm{c}}^{\prime }}{{\mathrm{x}}_{\mathrm{ck}}}}{1+1/\left[\left(\mathrm{B}+\mathrm{D}\right)\left({\sum }_{\mathrm{b}=1}^{\mathrm{B}}\frac{{\mathsf{\gamma }}_{\mathrm{b}}^{\prime }}{{\mathsf{\gamma }}_{\mathrm{bk}}}+{\sum }_{\mathrm{d}=1}^{\mathrm{D}}\frac{{\mathsf{\delta }}_{\mathrm{d}}^{\prime }}{{\mathsf{\delta }}_{\mathrm{dk}}}\right)\right]}$$

(1)

In the formula, $\mathsf{\rho }$ is the objective function; the number of decision-making units (DMUs) is N; x is an input factor; the number associated with C, $\mathsf{\gamma }$ is the desired output; the number associated with B, $\mathsf{\delta }$ is a non-desired output; and the numbers associated with D; ${\mathrm{x}}_{\mathrm{c}}^{\prime }$, and ${\mathsf{\delta }}_{\mathrm{d}}^{\prime }$ are the slack variables, i.e., the distance of the input or output from the frontier. The term $\mathsf{\mu }$ is used for weighting. When a DMU is on the frontier, DEA determines that its input–output mix is the most efficient and sets its efficiency indicator at 1.

3.2.2. Space Durbin Model

Tourism is an integrated industry with a very strong linkage drive, so the improvement of carbon efficiency in tourism is inevitably influenced by the level of quality tourism development in the region and the sector. The spatial Durbin model encompasses the spatially lagged terms of the independent and dependent variables. As a result, the coefficient estimation of the explanatory and error terms do not affect the results due to the omission of variables.

The model is calculated as follows:

$${\mathrm{TCEP}}_{\mathrm{it}}={\mathsf{\alpha }}_{\mathrm{i}}+\mathsf{\rho }{\displaystyle \sum }_{\mathrm{t}=1}^{\mathrm{N}}{\mathrm{W}}_{\mathrm{it}}\times {\mathrm{C}}_{\mathrm{it}}+{\mathsf{\beta }\mathrm{X}}_{\mathrm{it}}+\mathsf{\phi }{\displaystyle \sum }_{\mathrm{j}=1}^{\mathrm{N}}{\mathrm{W}}_{\mathrm{it}}{\mathrm{X}}_{\mathrm{it}}+{\mathsf{\theta }\mathrm{C}}_{\mathrm{it}}$$

(2)

In the formula, ${\mathrm{TCEP}}_{\mathrm{it}}$ is the explanatory variable, indicating the efficiency of tourism carbon emissions in year t of the prefecture-level city i; ${\mathrm{X}}_{\mathrm{it}}$ denotes the explanatory variables, ${\mathrm{X}}_{\mathrm{it}}$=[ln(ID),ln(CD),ln(GD),ln(OD),ln(RS)]; ${\mathrm{C}}_{\mathrm{it}}$ Indicates a control variable, ${\mathrm{C}}_{\mathrm{it}}$=[ln(EDL),ln(ER),ln(GI),ln(OPEN)]; ${\mathrm{W}}_{\mathrm{it}}$ is a 21$\times$21 order economic geography weighting matrix; $\mathsf{\beta }\mathrm{and}\mathsf{\phi }$ denote the regression coefficient variables and spatial regression coefficients of the core explanatory variables, respectively; $\mathsf{\rho }$ represents the spatial regression coefficients of the explanatory variables; $\mathsf{\theta }$ is the regression coefficient variables that represent the control variables; and $\mathsf{\sigma }$ is the regression coefficient representing the spatial error term.

3.2.3. Threshold Model

There are multiple institutional constraints on the level of quality tourism development and the efficiency of tourism carbon emissions. Therefore, the heterogeneity of the relationship between the level of quality tourism development at different stages on the carbon emission efficiency of the tourism industry was further investigated with the threshold variables of tourism economic scale, tourism industry structure, tourism investment structure, and tourism carbon emission intensity. A non-linear threshold model was used, and the data were logarithmically processed to find the threshold intervals and to conduct significance tests.

The model is calculated as follows:

$${\mathrm{lnTCEP}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{lnHQDT}}_{\mathrm{it}}\times \mathrm{I}\left(\mathrm{th}\le \mathsf{\gamma }\right)+{\mathsf{\beta }}_2{\mathrm{lnHQDT}}_{\mathrm{it}}\times \mathrm{I}(\mathrm{th}>\mathsf{\gamma })+{\mathsf{\mu }\mathrm{X}}_{\mathrm{it}}+{\mathsf{\epsilon }}_{\mathrm{it}}$$

(3)

In the formula, ${\mathsf{\beta }}_0$ denotes the coefficient to be estimated for each variable; $\mathrm{I}(·)$ is an indicative function; $\mathsf{\gamma }$ is the threshold value; and $\mathrm{th}$ is the threshold variable [TES, TIS, TISS, TCEI]. The double and triple thresholds are expanded based on Equation (3).

3.3. Variable Selection

A spatial Durbin model is constructed with the level of economic development, environmental regulation, government intervention, and openness to the outside world as control variables, and an economic geography weight matrix is constructed to focus on the spatial spillover effects of the level of high-quality tourism development as the core explanatory variable on the carbon emission efficiency of tourism as the explanatory variable. A threshold model was constructed to study the threshold effect of tourism quality development level on tourism carbon emission efficiency under different threshold variables, using tourism economic scale, tourism industry structure, tourism investment structure, and tourism carbon emission intensity as threshold variables, while keeping the above variables consistent.

3.3.1. Explained Variable

The evaluation index system of carbon emission efficiency of tourism is constructed based on the concept of “economic–social–environmental” coordinated development and draws on research by Tang C et al. [58] (1). In the selection of input indicators, capital, labor, and land are the most basic factors in economics, as the land factor is less binding on the carbon efficiency of tourism, while energy consumption is an important input indicator in the tourism production process, so this paper selects the amount of fixed asset investment in tourism, tourism workers, and tourism energy consumption to represent. The indicator system is shown in

Table 1. Tourism Carbon Emission Efficiency Evaluation Index System.

Input Indicators	Desired Output Indicators	Non-Desired Output Indicators
Tour Operators	Total Tourism Revenue	Carbon emissions from tourism
Fixed Asset Investment in Tourism
Energy consumption in the tourism sector

.

The desired output reflects the economic returns to the tourism industry and is expressed in terms of total tourism receipts. Non-desired outputs are the “bad outputs” of tourism development and are characterized by the carbon dioxide emissions generated during the production of the tourism industry.

3.3.2. Core Explanatory Variable

Research on the high-quality development of regional tourism to help build a strong tourism-oriented province has contemporary implications [24]. Combining the “double carbon“ goal, the characteristics of tourism quality development, and the “new development concept“, from the innovation-driven, coordinated development, green development, open development results that share five dimensions, the indicators are logarithmic processing, and we can build a tourism quality development level evaluation index system containing 5 variables layers and 14 indicators layers seen in

Table 2. Evaluation index system of high-quality tourism development level.

System Level	Variable Layer		Indicator Layer	Properties	Weight (%)
The level of high-quality development of tourism	Innovation driven	ID	Science and technology expenditure as a proportion of fiscal expenditure (%)	+	10.474
	Coordinated development	CD	Tourism Industry Aggregation (%)	+	4.286
			Total tourism revenue as a percentage of social fixed asset investment in tourism (%)	+	16.315
			Degree of integration of tourism with the primary sector (%)	+	1.043
			Degree of integration of tourism with the secondary sector (%)	+	1.744
			Integration of tourism with the tertiary sector (%)	+	2.100
	Green development	GD	Greenery coverage in built-up areas (%)	+	2.216
			Green space per capita (m2/ person)	+	10.784
			Energy consumption per unit of tourism revenue (MJ/10,000 yuan)	−	14.880
	Open development	OD	Foreign exchange earnings from tourism as a proportion of total tourism receipts (%)	+	9.383
			Inbound tourist arrivals as a proportion of total tourism arrivals (%)	+	6.621
			The ratio of road miles to regional land area (km/km2)	+	11.531
	Results sharing	RS	Number of star-rated hotels (pcs)	+	5.920
			Share of tertiary sector employment (%)	+	2.703

.

3.3.3. Threshold Variables

This section discusses the threshold variables used in this study. (1) The tourism economic scale variable reflects the overall scale and development of the tourism industry in each prefecture-level city in Jiangxi Province [59]. (2) The tourism industry structure is a variable representing the sophistication, rationalization, and coordination of the industry structure, which influences the carbon efficiency of the tourism industry [60]. (3) The structure of tourism investment is an important component of the development of the tourism economy. Four sectors have been chosen to fully represent the amount of social fixed asset investment in tourism: transport, storage, and postal services; accommodation and catering; wholesale and retail trade; and culture, sport, and entertainment, which are closely linked to tourism development. The scale of their investment affects the carbon efficiency of tourism and the level of quality of tourism development. (4) The carbon intensity of tourism is an important indicator of sustainable tourism development and the soundness of the energy mix. Greater levels of carbon intensity indicate greater total carbon emissions and greater pressure to reduce them [61].

3.3.4. Control Variables

The control variables are as follows. (1) Level of economic development represents the development of local tourism attractions, which attract tourists and expand the tourism market. Changes in tourism inputs and outputs, in turn, affect the efficiency of tourism carbon emissions [62]. (2) Environmental regulation is also a control variable as it may lead to higher costs for environmental investments but may also provide incentives for tourism enterprises to optimize resource allocation and reduce undesired outputs, thus increasing the efficiency of tourism carbon emissions [63]. (3) Government intervention is a variable because improving the efficiency of carbon emissions on tourism days requires a reasonable determination of the buffer for incremental carbon emissions from key tourism construction projects, intending to curb the unreasonable growth of carbon emissions at the source. (4) Openness to the public is another variable that may introduce advanced technologies into the tourism industry, contributing to technological advances in tourism’s carbon efficiency, but it may also result in more carbon emissions from tourism [64].

3.4. Data Sources

This study uses panel data for eleven prefecture-level cities in Jiangxi Province from 2000–2020. Data were obtained from the China Urban Statistical Yearbook, Tourism Sample Survey Information, Jiangxi Provincial Statistical Yearbook, the statistical yearbooks of each prefecture-level city, and the statistical bulletin on national economic and social development. The indicator system is shown in

Table 3. Descriptive statistical characteristics and correlation coefficients of the variables.

Classification	Definition	Calculation Method	Symbols	Average Value	Standard Deviation	Minimum Value	Maximum Value
Explained variable	Carbon efficiency in the tourism sector	Carbon Emission Efficiency Evaluation Index System for Tourism	lnTCEP	−0.313	0.679	−1.543	1.106
Core explanatory variable	The level of high-quality development of tourism	Tourism quality development level evaluation index system	lnHQDT	2.075	0.199	1.265	2.495
Threshold variables	Tourism economy scale	Tourism revenue as a share of GDP (%)	TES	16.966	14.093	2.331	77.591
	Tourism Industry Structure	Tourism revenue as a share of the tertiary sector (%)	TIS	44.626	31.407	7.389	158.796
	Tourism Investment Structure	Amount of fixed asset investment in tourism as a percentage of GDP (%)	TISS	0.001	0.001	0.000	0.008
	Tourism Carbon Emission Intensity	Tourism carbon emissions per unit of tourism revenue (kg/yuan)	TCEI	0.021	0.019	0.002	0.132
Control variables	Level of economic development	GDP per capita (yuan/One person)	lnEDL	9.904	0.954	7.972	11.579
	Environmental regulation	Investment in fixed assets in water, environment, and public facilities management (10,000 yuan)	lnER	12.840	1.590	8.389	15.833
	Government intervention	Fiscal expenditure as a percentage of GDP (%)	lnGI	−3.141	0.883	−5.565	−1.146
	Open to the public	The actual amount of foreign capital used in the year (USD million)	lnOPEN	10.303	1.431	6.271	13.614
Spatial weighting matrix	Economic Geography Weighting Matrix	The inverse of the difference between the annual average GDP of the two regions	W

.

4. Empirical Results and Analysis

4.1. Carbon Efficiency in Tourism and the Level of Quality Tourism Development

Using the natural breakpoint method, the spatial and temporal evolution trends of carbon emission efficiency and the level of high-quality tourism development in Jiangxi Province from 2000 to 2005 were divided into five levels, as shown in Figure 1.

During the study period, the average value of tourism carbon emission efficiency in Jiangxi Province was 0.911, which is a high overall level. The high- and higher-efficiency levels concerning tourism carbon emissions were mainly located in Xinyu, Ji’an, Jiujiang, and Yingtan cities in the early period. In contrast, more prefecture-level cities in the southern and central regions were less efficient, with a clear two-tier division pattern. In the later part of the study period, the excellent- and good-efficiency zones were mainly located in the northern part of Jiangxi Province and Xinyu; the average-efficiency zones were located near the high-efficiency zones and show a trend of attachment; and the fair- and poor-efficiency zones were mainly located in the southern part of Jiangxi Province, showing an overall relatively stable spatial pattern.

Given the influence of COVID-19, the carbon emission rates from tourism in Nanchang, Ganzhou, Ji’an, and other prefecture-level cities showed different degrees of decline in 2020. Further, while non-desired output decreases significantly in 2020 compared to 2019, there were also significant decreases in employment, investments in tourism-related fixed assets, and income. The result is that it leads to a reduction in the efficiency of carbon emissions from tourism.

Figure 1 shows that the overall high-quality level of tourism in Jiangxi Province generally experienced an upward trend. The spatial pattern analysis shows that excellent- and good-efficiency level areas were represented by Jingdezhen, Jiujiang, Nanchang, etc.; average efficiency areas were more distributed in the northeastern part of Jiangxi Province; and fair- and poor- efficiency level areas were mostly located in the south-central part, represented by Yichun, Shangrao, and Fuzhou. During the period between 2000 and 2010, the high-quality tourism coefficients were higher than average in Nanchang, Jingdezhen, Jiujiang, and Ji’an. In 2019, the high-quality tourism coefficient was higher than average in Nanchang, Pingxiang, Jiujiang, Xinyu, and Yingtan. In 2020, except for Fuzhou, the rest of the prefecture-level cities experienced significant decreases in high-quality tourism development levels, indicating the influence of COVID-19 on tourism supply-side challenges. This exposed imbalances in the industrial structure and other problems, with systemic negative impacts.

4.2. Study on the Spatial Spillover Effect of Tourism Quality Level on Tourism Carbon Efficiency

4.2.1. Spatial Correlation Tests and Model Tests

The Moran’s I index was used, and an economic geography weighting matrix was constructed to test the relevance of carbon efficiency in tourism.

Table 4. Global Moran’s I statistic for carbon emissions efficiency in tourism.

Year	I-Value	Z-Value	Year	I-Value	Z-Value
2000	−0.178	−1.114	2011	−0.249 **	−2.114
2001	−0.263 **	−2.306	2012	−0.249 **	−2.171
2002	−0.218	−1.636	2013	−0.252 **	−2.293
2003	−0.201	−1.407	2014	−0.260	−2.352
2004	−0.244 **	−2.023	2015	−0.242 **	−2.054
2005	−0.272 **	−2.481	2016	−0.192	−1.384
2006	−0.257 **	−2.230	2017	−0.179	−1.095
2007	−0.241 **	−2.027	2018	−0.171	−0.994
2008	−0.238 *	−1.911	2019	−0.178	−1.089
2009	−0.238 *	−1.893	2020	−0.103	−0.039
2010	−0.253 **	−2.198

Note: **, p < 0.05, * p < 0.1.

shows that many years during the study period are negatively significant and suitable for spatial econometric analysis, with the overall I-value fluctuating in a “W” pattern within the range of [−0.272, −0.103]. The negative characteristics of the carbon emission efficiency of tourism in Jiangxi Province decreased in the later years of the study period, with a gradual decrease in spatial variability.

Based on the Elhorst [65] study,

Table 5. Model selection test.

Test Methods	Model	Eigenvalue
LR test	Spatial lag	14.410 ***
	Spatial error	9.454 ***
Wald test	Spatial lag	82.870 ***
	Spatial error	69.570 ***
Hausman test		412.72 ***

Note: *** p < 0.01.

shows that the validation results passed the significance test through LR, Wald, and Hausman tests, the model fit was good, the confidence level was high, and the final model was selected as the double fixed spatial Durbin model (SDM) for spatial econometric analysis.

4.2.2. Analysis of Spillover Results

(1)

Spatial Durbin model analysis

Equation (2) is used to calculate the spatial spillover effect evaluation results shown in

Table 6. Evaluation results of spatial spillover effects of carbon emission efficiency in tourism.

Variables	SDM	Direct Effects	Indirect Effects	Total Effect
ln(ID)	0.028 (0.800)	0.022 (0.680)	0.050 (0.510)	0.072 (0.680)
ln(CD)	0.866 *** (6.450)	0.620 *** (5.330)	1.707 *** (4.150)	2.327 *** (5.080)
ln(GD)	0.343 *** (3.480)	0.478 *** (4.490)	−0.887 * (−1.820)	−0.409 (−0.900)
ln(OD)	0.116 (1.640)	0.185 ** (2.580)	−0.485 *** (−3.470)	−0.300 ** (−2.190)
ln(AS)	−0.190 * (−1.860)	−0.127 (−1.540)	−0.440 (−1.390)	−0.568 * (−1.680)
ln(EDL)	−0.449 (−1.350)	−0.575 ** (−2.110)	0.902 (1.440)	0.327 (0.410)
ln(ER)	0.060 (0.830)	0.026 (0.410)	0.238 (1.070)	0.264 (1.070)
ln(GI)	−0.160 *** (−2.750)	−0.059 (−1.120)	−0.729 *** (−2.830)	−0.788 *** (−3.030)
ln(OPEN)	0.226 *** (3.040)	0.124 * (1.840)	0.765 *** (3.810)	0.889 *** (4.050)
W × ln(ID)	0.103 (0.550)
W × ln(CD)	3.720 *** (5.520)
W × ln(GD)	−1.204 (−1.510)
W × ln(OD)	−0.689 *** (−2.710)
W × ln(AS)	−0.938 (−1.610)
W × ln(EDL)	1.117 (0.800)
W × ln(ER)	0.446 (1.100)
W × ln(GI)	−1.404 *** (−3.350)
W × ln(OPEN)	1.504 *** (4.330)
Spatial	−0.969 *** (−4.750)

Note: *** p < 0.01, **, p < 0.05, * p < 0.1. Values in brackets are z-values. The total effect is the sum of the direct and indirect effects.

. Tourism’s carbon emission efficiency itself has a negative spatial spillover effect, with a spatial autoregressive coefficient value of −0.969, and it is significant. Further analysis across the five dimensions shows that coordinated development and its spatial lag term have a significant positive impact on the tourism carbon efficiency sector. Coordinated development requires tourism and related industries to change their development models and implement supply-side structural reforms. Tourism enterprises should improve their management efficiency to achieve the goal of improving the carbon efficiency of the regional tourism industry. There is a positive correlation between the innovation drive and the spatial lag term on the carbon emission efficiency of tourism. Relying on innovative institutional mechanisms and invoking advanced technologies, the efficiency of tourism energy use in the region has improved; however, the spatial layout of tourism innovation is a long-term process, and imbalances remain. As a result, the positive spillover effect is not significant.

The spatial lag term of open development harms the carbon emission efficiency of tourism at a 1% significance level. Tourism open development can accelerate the development of the tourism economy and expand the tourism market. However, the level of openness of each region in Jiangxi Province is not high enough to facilitate this process, and the level of openness in the north, middle, and south of Jiangxi Province varies significantly. This highlights the need to promote the development of tourism with a high level of openness, maximize the advantages of Jiangxi tourism resources and markets, and fully utilize the two major international and domestic markets. This should help achieve the refined and characteristically open development of tourism.

Both green development and shared outcomes have a non-significant negative spatial spillover effect on the carbon efficiency of tourism. There remain contradictions between the coordinated development of economic and social development and resources and environment in some areas of Jiangxi. The environmental carrying capacity restricts the growth rate of the tourism economy, and there remains a lag in sharing the benefits of tourism development. This inhibits green development, and shared outcomes show a dampening effect on the overall regional tourism carbon emissions efficiency.

(2)

Analysis of direct effects

Among the direct effects, coordinated development, green development, and open development, all have a positive and significant effect on the carbon efficiency of the tourism industry. Neither the positive innovation driven nor the negative achievement sharing effects are significant.

Although the direct impact of the innovation drive is not statistically significant, it does show a certain “technology dividend”. The proportion of fiscal expenditure on science and technology increased from 0.472% in 2000 to 3.009% in 2020, and the mechanism of using technology to reduce emissions has taken shape. Financial support for green technology research and development should be further strengthened to reinforce the “green” bias of technological progress, with innovation drive being a key enhancement factor.

For every 1% increase in coordinated development, the carbon efficiency of the region’s tourism industry increases by 0.620%. Jiangxi has implemented a strategy to strengthen the tourism industry and promote its high-quality development. Coordinated development requires tourism economic growth to be synchronized with carbon emission reduction development efforts. The most significant effect of coordinated development on tourism carbon emission efficiency is due to two factors: the increased agglomeration of tourism industries in Jiangxi Province and the increased dynamism of economic growth based on tourism.

Every 1% increase in green development increases the carbon efficiency of tourism in the region by 0.478%. Increased energy efficiency is in line with the “sustainable development concept” of increasing the efficiency of energy use in tourism without a loss in economic growth. The improved quality of the ecological environment encourages a greater awareness of environmental protection in the region, forcing the local tourism industry to upgrade its structure, improve the way tourism enterprises operate, and increase resource allocation efficiency. Therefore, green development has a significant incremental effect.

Every 1% increase in open development is associated with a 0.185% increase in the carbon efficiency of the local tourism industry. The expansion and opening of tourism will attract more foreign tourists, which will lead to an increase in carbon emissions in the region. However, it also encourages tourism production and operational activities in the province to adapt to international standards and the laws of the market economy, thus promoting the quality and speed of tourism.

Every 1% increase in results sharing decreases the carbon efficiency of the region’s tourism industry by 0.127%. While benefiting from the economic dividends of tourism sharing, carbon emissions from the tourism industry are expected to increase. As a result, the tourism industry should supplement its focus on “industry incrementalism” to include a focus on the revitalization of the “extra-industry stock”, to serve, enrich, benefit, and support the enjoyment of the people. The tourism industry should be broadened from a sole focus on “industry growth” to include a focus on the revitalization of “non-industry stock” as well.

(3)

Analysis of indirect effects

When considering the order of the intensity of indirect effects, the positive spillover effects are greatest for coordinated development, followed in descending order by the variable “innovation-driven”. The negative spillover effects of green development are greater than those associated with open development and sharing of results, in descending order.

For every 1% increase in innovation drive, the carbon intensity of the neighborhood’s tourism industry increases by 0.05%. However, the positive spillover effect is not significant, likely due to the large difference in the level of science and technology investment among prefecture-level cities in Jiangxi Province and the low innovation capacity in some prefecture-l0evel cities. Thus, the positive spatial spillover effect of the innovation drive is not significant.

For every 1% increase in the level of coordinated development, the carbon emission efficiency of neighborhood tourism increases by 1.707%. Due to the economic linkage effect and the realistic needs of regional tourism development, the coordinated development of tourism in the region may also expand opportunities to develop tourism in neighboring regions, build a new system of the tourism industry, and cultivate regional tourism growth poles. This may, in turn effectively promote improvements in the carbon emission efficiency of tourism in neighboring regions.

For every 1% increase in green development, the carbon emission efficiency of neighborhood tourism is expected to decrease by 0.887%. This may be due to the insufficient combination of ecology and tourism, the poor alignment of resources and products, and a lack of unity in protection and development when developing all-area tourism. This may result in neighborhood tourism enterprises being aware of environmental protection but not acting strongly enough, leading to a significant negative spatial spillover effect of green development on neighboring regions.

For every 1% increase in open development, the carbon emission efficiency of neighboring regions’ tourism industry decreases by 0.485%; the significance of this result passes the 1% significance test. The tourism development space is divided by geographical, political, and economic factors and has not yet formed a comprehensive open development pattern. In addition, COVID-19 heavily depressed the major domestic, outbound, and inbound tourism markets. This led to an imbalance in the supply and demand of tourism markets, in which open development significantly inhibited improvements in tourism’s carbon emission efficiency.

For every 1% increase in shared outcomes, the carbon efficiency of neighborhood tourism decreases by 0.440%. However, the test does not pass the significance test. Tourism outcome sharing is not about tourism development everywhere and building scenic spots everywhere, and hospitality services are not judged by the simple number of hotels and restaurants. Instead, rational layouts and conservation development are what are expected to improve the public’s experience of tourism outcomes. In addition, tourism was the first industry to experience the crisis of the COVID-19 outbreak and uncertainties in the economic development environment. This included employment difficulties. Non-desired outputs are reduced, but inputs and desired outputs are more frustrated, and the sharing of results has a negative effect on the efficiency of neighborhood tourism carbon emissions.

4.3. Study on the Effect of the Level of Quality Tourism Development on the Carbon Efficiency Threshold of the Tourism Industry

4.3.1. Threshold Model Determination

Using Equation (3), the explanatory variables, core explanatory variables, threshold variables, and control variables are brought into the panel threshold model, and each variable is estimated using “self-sampling” 300 times. The results of the self-sampling test in

Table 7. Threshold effect test of carbon emission efficiency in tourism.

Model	Types of Thresholds	F-Value	p-Value	Threshold Estimates	95% Confidence Interval	Number of BS
TES	Single threshold	22.700 ***	0.000	5.496	[5.388, 5.5243]	300
	Double Threshold	12.460	0.333			300
	Three-fold threshold	5.920	0.667			300
TIS	Single threshold	35.840 ***	0.000	14.938	[14.736, 15.154]	300
	Double Threshold	16.230 ***	0.000	42.317	[38.581, 42.340]	300
	Three-fold threshold	6.940	0.667			300
TISS	Single threshold	14.930 ***	0.000	0.003	[0.003, 0.004]	300
	Double Threshold	7.390	0.333			300
	Three-fold threshold	8.180	1.000			300
TCEI	Single threshold	25.460 ***	0.000	0.025	[0.024, 0.026]	300
	Double Threshold	12.590	0.667			300
	Three-fold threshold	8.690	0.667			300

Note: *** p < 0.01.

show that when the threshold variables are tourism economic scale, tourism investment structure, and tourism carbon emission intensity, there is a single threshold effect that is significant at the 1% level; the associated threshold estimates are 5.496, 0.003, and 0.025, respectively. The single threshold and double threshold estimates of the tourism industry structure were 14.938 and 42.317, and both passed the significance test.

4.3.2. Analysis of Threshold Results

As shown in

Table 8. Tourism carbon efficiency threshold return results.

Variable	TES	TIS	TISS	TCEI
HODT(Stage 1)	0.702 ***	0.647 ***	0.957 ***	0.850 ***
	(2.990)	(3.120)	(4.430)	(4.000)
HODT(Stage 2)	0.867 ***	0.923 ***	0.741 ***	0.638 ***
	(3.950)	(4.700)	(3.150)	(2.820)
HODT(Stage 3)		1.032 ***
		(5.220)
ln(EDL)	−0.156 ***	−0.165 **	−0.107	−0.218 ***
	(−2.010)	(−2.280)	(−1.350)	(−2.820)
ln(ER)	0.022	−0.020	0.006	0.012
	(0.540)	(−0.510)	(0.140)	(0.300)
ln(GI)	−0.042	−0.034	−0.023	−0.041
	(−0.920)	(−0.790)	(−0.510)	(−0.920)
ln(OPEN)	0.044	0.025	0.002	0.016
	(1.340)	(0.820)	(0.050)	(0.490)

Note: *** p < 0.01, ** p < 0.05. t values in brackets.

. The effect of a high level of quality tourism development on the carbon efficiency of the tourism industry is enhanced when the size of the tourism economy crosses a single threshold (5.496). When TES ≤ 5.496, the coefficient is 0.702, indicating that the level of high-quality tourism development is significantly and positively correlated to tourism carbon emission efficiency. When TES > 5.496, the coefficient is still positively and significantly correlated, and the coefficient increases to 0.867. This indicates that the strong scale of the tourism economy in Jiangxi Province has led to a more pronounced increase in carbon emission efficiency as the combined contribution of tourism revenue increases and the value of new tourism output expands. Combined with the raw data, the sample size of the tourism economy scale above the threshold (5.496) accounted for 84.416% of the total sample. All the prefecture-level cities in Jiangxi Province crossed the threshold in 2012, entering the second stage earlier.

When the structure of the tourism industry crosses the double threshold (14.938, 42.317), the effect of a high-quality level of tourism development on the carbon efficiency of the tourism industry increases. The process for this improvement can be described in three stages. When TIS ≤ 14.938, the sample size is 9.524%, with more prefecture-level cities entering the second stage of enhancement at the early stage of the study. Nanchang was the latest (2011) to cross over to the second stage. This has led to a low level of carbon efficiency in the tourism industry in Nanchang in the early stages.

When 14.938 < TIS ≤ 42.317, the sample size is 51.082%, and more prefecture-level cities are in the second stage of enhancement from 2000 to 2016. Further, more are distributed in south-central Jiangxi Province, and overall, the structure of the tourism industry in Jiangxi province varies considerably. When TIS > 42.317, the coefficient is 1.032. At this level, Jingdezhen, Jiujiang, and Yingtan maximize their tourism resources to improve the overall, functional, dynamic, and related nature of the tourism industry. They lead the sample when entering the third stage in 2012. The province’s prefecture-level cities cross the second threshold in 2017 and enter the third stage of enhancement. In 2020, COVID-19 significantly impacts the tourism industry structure, with the six cities of Nanchang, Xinyu, Ganzhou, Ji’an, Yichun, and Shangrao experiencing a decline in tourism industry structure coefficients, which led to a decline in the role of promotion to the second stage.

The contribution of a high level of quality tourism development to the carbon efficiency of the tourism industry decreases when the tourism investment mix crosses a single threshold (0.003). When TISS ≤ 0.003, the coefficient was 0.957 and significant. The province’s tourism investment in the first stage results in high and steady progress; the tourism industry transformation and upgrade speed is accelerated; and the high-quality tourism development level has a higher effect on enhancing the tourism industry’s carbon emission efficiency than during the second stage. When TISS > 0.003, the proportion of investment in tourism fixed assets increases, but the effect of tourism quality level on tourism carbon emission efficiency decreases. This indicates that raising the proportion of investment does not have a positive effect on tourism carbon emission efficiency. It is necessary to revitalize stock assets, improve the vitality of tourism investments, strengthen the coordination and co-ordination ability of each prefecture-level city, and create more high-quality projects to encourage tourism quality development and carbon emission efficiency.

The contribution of a high level of quality tourism development to the efficiency of tourism carbon emissions decreases when the tourism carbon intensity crosses a single threshold (0.025). When TCEL ≤ 0.025, the coefficient is 0.850 and significant. This indicates that the tourism quality development level effectively contributes to tourism carbon efficiency improvement when tourism carbon emission intensity is low. When TCEI > 0.025, the coefficient is 0.638 and significant. However, the effect of the high-quality level of tourism on the increase of carbon emission efficiency of tourism decreases. Overall, the carbon emission intensity of tourism has a significant “push-back” effect, which helps improve the carbon emission efficiency of the tourism industry. However, some prefecture-level municipalities have relaxed their environmental oversight to maximize tourism revenue and have not implemented Jiangxi Province’s regulations related to the environmental protection of construction projects and environmental pollution prevention. This has led to some areas falling into the “green paradox” trap.

5. Conclusions and Discussions

5.1. Conclusions

(1) The carbon emission efficiency of tourism in Jiangxi Province is highly consistent with the spatial and temporal characteristics of high-quality tourism development. Spatially, there are spatial differences between the high north and low south, and, temporally, there are fluctuating trends in the efficiency of carbon emissions from tourism, with an overall upward trend in the level of quality tourism development.

(2) The different dimensions of the level of quality tourism development have different impacts on the carbon efficiency of tourism in the region and neighboring regions. Coordinated development, green development, and open development all contribute significantly and directly to the carbon efficiency of tourism, while innovation-driven has a positive effect and results-sharing has a negative but insignificant effect. Neighboring regions’ innovation-driven and coordinated development can all improve the carbon efficiency of local tourism to varying degrees, but the positive spillover effect of innovation-driven is not significant. Green development and open development both show significant negative spatial spillover effects, while shared outcomes show insignificant negative spatial spillover effects.

(3) The enhancement of tourism’s carbon efficiency by the level of quality tourism development increases when the size of the tourism economy crosses a single threshold and the structure of the tourism industry crosses a double threshold. This enhancement effect is weakened when tourism investment structure and tourism carbon emission intensity cross a single threshold.

5.2. Discussions

This paper completes the research framework of carbon efficiency in tourism from the perspective of “double carbon” and enriches the research content of tourism quality development level. It fills the research gap of the relationship between carbon efficiency and quality development in tourism from the perspective of spatial spillover and threshold effect. Both the construction of indicators and the analysis of results help to enrich the theoretical study of sustainable and green development. At the practical level, it can provide feasible suggestions and methods for low-carbon transformation and high-quality development of tourism for the surrounding areas and even the whole country.

(1) During the study period, the average value of tourism carbon emission efficiency in Jiangxi Province was 0.911. Its “fluctuating” development trend converges with that of the Yangtze River Economic Belt Jiangxi Province is one of the first ecological civilization pilot zones in China, and its tourism resources consist of mostly green, revolutionary sites, and rural tourism sites [66]. The area has few high-energy tourism products [67] and a high environmental carrying capacity, making the province’s overall tourism industry more efficient concerning carbon emissions.

Government policy support is an important reason for the overall upward trend in the level of quality tourism development in Jiangxi Province. Covered by national strategies, such as the Central China Rising Plan, the Poyang Lake Ecological, and the Economic Development Zone and the first batch of ecological civilization pilot zones in China, Jiangxi Province has distinct advantages in regional strategies [68].

The quality of tourism development is influenced by economic developments. In 2020, the cities of Ji’an, Fuzhou, Ganzhou, and Shangrao, with lower GDP per capita levels in the province, did not invest much in tourism and invested even less in low-carbon tourism due to poor economic performance. The gap between the low-value areas in the south and the high-value economic areas in the north creates a “high north and low south” difference.

(2) The study analyzes the direct and spatial spillovers from the level of quality tourism development to the carbon efficiency of the tourism industry. Concerning the study by Dong and Balsalobre-Lorente, this study concludes that the tourism innovation drive in Jiangxi Province has a positive effect on the carbon efficiency of tourism concerning both direct and spatial spillovers. However, compared to developed countries, there remains a gap in Jiangxi Province’s innovation capacity, making the positive effect less significant.

The results of the study by Zhang F et al. demonstrate that the coupled economic–tourism–environmental coordination system optimizes the degree of coupled coordination in the region and in the neighborhood. They found that the ETE subsystem promotes improved regional coordination and optimizes the coordination of neighboring regions. Similarly, this study demonstrates that the coordinated development of tourism in Jiangxi Province has significantly increased the carbon efficiency of tourism in the region and neighboring regions.

Green development has a “reverse emissions reduction” effect and a “green paradox” effect. The ecological quality of Jiangxi Province has gradually improved, and the green development of tourism is contributing significantly to the carbon efficiency of the region’s tourism industry. However, the uncoordinated green development between regions has led to negative spatial spillover effects, confirming the existence of a “green paradox”.

The results are equally mixed concerning the impact of open development on the environment. Mahrinasari believes that trade liberalization will reduce carbon emissions, but Ling takes the opposite view. The opening up of Jiangxi Province to the outside world can expand the international tourism market, and the favorable impact of the desired outputs compensates for the unfavorable impact of the undesired output. This results in a positive and significant impact on the carbon efficiency of tourism in the region. However, a fully open tourism development pattern has not been formed, so the spatial spillover effect has a negative significant impact.

Bernard J’s findings suggest that population and income in developing and emerging countries can inhibit environmental performance. China remains the world’s largest developing country, and the direct and spatial spillovers from the sharing of tourism outcomes in Jiangxi to the broader carbon efficiency of the tourism industry have not reached a positive effect.

(3) Many studies have used the size of the tourism economy [69] and the structure of the tourism industry as threshold variables. Few studies have examined tourism investment structure and tourism carbon emission intensity as threshold variables. Based on previous research, this study identifies a single threshold effect for tourism economic scale, tourism investment structure, and tourism carbon emission intensity, and it identifies a double threshold effect for tourism industry structure. When the scale and the structure of the tourism industry cross their respective thresholds, there is a jump in the quality of tourism development concerning tourism carbon emissions efficiency. When tourism investment structure and tourism carbon intensity cross the threshold, the effect of tourism quality development on tourism carbon emission efficiency decreases.

6. Recommendations and Future Research

6.1. Recommendations

(1) Accelerating the low carbon transition and developing high quality tourism can address the “high north, low south” development dilemma.

The “high in the north and low in the south” is a development dilemma for Jiangxi’s tourism industry in terms of both carbon efficiency and the level of quality tourism development. To enhance the overall tourism industry carbon efficiency in Jiangxi Province, the northern high-efficiency areas should maximize their positive demonstration effect of tourism carbon emission efficiency management and realize the green low-carbon tourism experience spillover. For southern regions at low and medium efficiency levels, the tourism economic income is not positively related to the carbon efficiency of tourism. This highlights the need to improve the input perspective, strengthen learning related to low-carbon technology and enterprise management from the high-efficiency regions, optimize the investment in tourism fixed assets, and promote regional energy savings and emission reduction.

To enhance the level of quality development of tourism in Jiangxi, five dimensions of integrated development need to be achieved. First, we must adhere to innovation-driven, high-quality development; increase scientific and technological support; widely apply advanced technology; and promote the innovation and enhancement of tourism, service delivery, consumption patterns, and management tools. Secondly, we should adhere to coordinated development and optimize the layout; strengthen tourism infrastructure in weak areas such as the south and upgrade tourism elements; and comprehensively promote the integration of cross-regional resource elements, accelerate the development of tourism industry clusters, and enhance the integration of tourism with the primary, secondary, and tertiary industries. Thirdly, we should adhere to green development and ecological priority; hold fast to the ecological bottom line, improving the green coverage rate of the built-up area and the area of park green space per capita, to create conditions for enhancing the city’s carbon sink; and reduce energy consumption per unit of tourism revenue to improve the ecological benefits of tourism. Fourth, we should adhere to open development that is tourism-driven. Effectively increasing reform and opening up to be more tourism-driven, the southern region can use quality tourism resources to attract foreign tourists, thus promoting the growth of the tourism economy. The fifth dimension is to adhere to the sharing of results and tourism for the people. Low-value areas can improve tourism reception capacity, play a comprehensive role in tourism drive, create more employment and entrepreneurship opportunities, and better serve economic and social development.

(2) Tapping the positive spillover effects of variables to improve the efficiency of carbon emissions from tourism is also important.

Under the spatial demonstration effect, neighboring regions can continue to strengthen regional tourism industry coordination and deepen low-carbon technology exchanges. This should further maximize the positive spatial spillover effects of coordinated development and innovation drive. For the negative spatial spillover effects associated with green and open development as well as shared results, Jiangxi Province should build an environmental regulation system for low-carbon tourism development according to local conditions, eliminate administrative divisions and system obstacles, and transform from the current approach to a two-pronged approach of focusing on “industry incrementalism” and revitalizing “extra-industry stock”.

(3) The scientific grasp of the non-linear threshold impacts accelerate the green and low carbon transformation of tourism.

This study indicates that the tourism economy needs to change in the direction of scale and efficiency, with a focus on conservation and development, moderate and scientific development of ecological tourism resources, and orderly guidance and promotion of low-carbon tourism products, such as rural, revolutionary sites, and leisure tours. It is important to continue to promote the group development of tourism enterprises and work towards the early introduction of a new stage of the tourism industry structure in the post-COVID-19 era. Other recommendations include optimizing the tourism investment structure and promoting the green transformation of tourism with investment-based tourism quality development. Local governments should dynamically adjust their policy intensity to local conditions to curb excessively high-polluting and energy-intensive projects and actively encourage innovative and diversified markets to take root in low carbon emission intensity regions.

6.2. Future Research

This paper presents an index system for evaluating the carbon emission efficiency and the high-quality tourism development level based on tourism inputs and outputs. As informed by related studies, the research is consistent with scientific concepts and reasonable principles; however, as with all studies, there remain limitations. First, due to the study’s long period, the conversion factors of carbon dioxide and energy consumption may differ in different periods, and the use of uniform carbon emission coefficients may lead to biased results. Second, the variables in the model need to be further enriched. The indicators of quality tourism development all follow the existing research results, while tourism is a comprehensive industry with a strong correlation, how to comprehensively and scientifically reflect the level of quality tourism development is the next step to improve the research.

Compared with previous studies, however, this paper is innovative in several ways. First, previous studies have focused primarily on the measurement of carbon emissions from tourism, which does not fully reflect the efficiency relationship between pollution and output. To improve upon that approach, this paper measures the carbon emission efficiency of tourism in Jiangxi Province, which is more relevant to low carbon economic growth than total carbon emissions and carbon intensity.

Secondly, in the selection of indicators for the explanatory variables, energy consumption and carbon emissions from tourism, which are associated to achieve “peak carbon”, are included in the evaluation system. In terms of the selection of core explanatory variables, previous studies have mainly used economic and green variables, with a single evaluation dimension. This approach does not fully consider the comprehensive nature of tourism development. This paper combines the new development concept and emphasizes the five major development concepts as a guide in the development process to promote and achieve efficient, equitable, and green sustainable development of tourism in Jiangxi Province. In the green development latitude, the variables of built-up area green coverage and per capita park green space, which are closely related to the “carbon neutrality” target, are included in the evaluation system.

The selection of tourism economic scale, tourism industry structure, tourism investment structure, and tourism carbon emission intensity as threshold variables helps to fully understand the impact of tourism high-value development on tourism carbon emission efficiency at different thresholds and within different threshold stages. In the selection of control variables, in addition to economic indicators, environmental regulation and government intervention are also considered to make the study more scientific.

Third, in terms of research methodology, the spatial Durbin model was used to quantify the spatial and temporal relationship between the carbon emission efficiency of Jiangxi’s tourism industry and the level of quality tourism development. This provides empirical data that can inform recommendations for the coordinated development of prefecture-level cities in Jiangxi Province. The study also uses a threshold model to test the non-linear impact of the level of high-quality tourism development on the carbon emission efficiency of the tourism industry under different constraints. The four aspects of the economic scale, tourism industry structure, tourism investment structure, and tourism carbon emission intensity are used to identify better ways to improve the carbon emission efficiency of tourism in Jiangxi Province.

This study highlights the following opportunities for future research. This study identifies the non-linear effect that different dimensions have on quality tourism development in Jiangxi Province from 2000 to 2020 concerning carbon efficiency. The study also identifies the different degrees of impact variables have at different threshold stages. However, the study scope is limited to eleven prefecture-level cities in Jiangxi Province. Future studies could expand the sample size to provide a theoretical basis and innovative ideas for coordinated development in all provinces and cities in China. Future studies could also expand the set of variables and consider whether high-quality tourism development could be achieved through supply-side structural reforms and narrowing regional development differences. Future research could also ask whether supply-side reform and regional coordinated development could also be incorporated into the factors influencing the carbon efficiency of the tourism industry. This could further inform future directions in the post-COVID-19 era concerning high-quality tourism development and low-carbon transition.