Where Have Carbon Emissions Gone? Evidence of Inbound Tourism in China

Abstract

Tourism is emerging as an important contributor sector to carbon emissions. As inbound tourism is the main component of tourism activities, measuring and evaluating the carbon footprint of inbound tourism can help achieve low-carbon development of the global tourism industry. Based on the carbon footprint theory, this study describes the distribution and transfer path of China’s inbound tourism carbon footprint by using the China multi regional input-output model embedded in the inbound tourism satellite account. The results reveal that embodied carbon emissions (ECE) of inbound tourism is higher than direct carbon emissions, and carbon emissions intensity is close to the average economic. Therefore, it is difficult for inbound tourism to achieve the dual goals of stimulating economic growth and emission mitigation. The ECE of inbound tourism has obvious inter-provincial transfer characteristics. Inner Mongolia and Guangdong are the largest provinces with net outflows/inflows of ECE, with net outflows of 1.47 million tons and inflows of 2.66 million tons. Besides, the ECE mainly flows from the northeast and northwest regions to the southern and eastern regions. From the perspective of industry transfer, 72.2% of ECE of inbound tourism comes from the power sector.

1. Introduction

Tourism is now an important sector that contributes to carbon emissions from human activities. Tourists will increase the demand for transportation, hotel accommodation and electricity consumption during their travel, thereby increasing CO₂ emissions [1]. Scholars have realized that tourism is one of the contributing sectors to climate change. In 2008, the United Nations Environment Programme (UNEP) pointed out that the carbon emissions from tourism transportation, accommodation and activities accounted for 4–6% of the total global carbon emissions. Without mitigation and response measures, tourism carbon emissions will increase by 1.5 times over the next 30 years. Lenzen et al. [1] also found that between 2009 and 2013, the global carbon footprint of tourism has increased from 3.9 Gt to 4.5 Gt CO₂, accounting for about 8% of global greenhouse gas emissions, among which the transportation, shopping and food sectors are important contributors. A report from Cambridge University’s sustainable development institute shows that tourism currently contributes about 5% to global greenhouse gas emissions. The World Tourism Organization (UNWTO) predicts that global tourism-related carbon dioxide emissions will increase by 25% in 2030, of which world tourism transportation emissions will increase by 45%, compared to 2016.

In research on the driving factors of global climate change, researchers often focus on the impact of energy-intensive sectors such as electricity and heavy industry, while ignoring the carbon emission characteristics of tourism activities. As a vital industry in the national economy, tourism will drive the demand for transportation, accommodation, catering, etc., which produce a large number of carbon emissions. Thus, it is of great significance to clarify the distribution characteristics and transfer paths of tourism carbon footprints in regions and industries for achieving global emission reduction and sustainable development goals. Carbon footprint theory is an important method to explore the direct and indirect carbon emissions of products or services, which refers to the number of greenhouse gases produced by consuming various types of energy to meet people’s production and living needs [2]. As a crucial connotation of deepening international tourism cooperation, inbound tourism is a major component of tourism activities. Discussing the carbon footprint of inbound tourism can provide a reference for the realization of low-carbon development of global tourism.

This paper selects China’s tourism industry as the research object, mainly because China has the world’s largest inbound tourism market, and carbon emissions from the tourism industry rank second worldwide, only behind the United States [3]. As depicted in Figure 1, there were 145 million inbound tourists from China in 2019, with a revenue of 131.3 billion dollars, an increase of 75% and 573% over 2000, respectively. The carbon emissions from accommodation, transportation, entertainment and other activities carried out by inbound tourists cannot be ignored. Han et al. [4] found that from 2007 to 2017, the carbon footprint of China’s inbound tourism showed a rapid upward trend, with the total amount rising from 5.62 million tons to 10.88 million tons. The transportation industries accounted for the most significant proportion. Therefore, the vigorous development of China’s inbound tourism provides a representative sample for this study.

The research motivation of the text has two points: on the one hand, the rapid growth of China’s inbound tourism in recent years makes it necessary to investigate the characteristics of the high energy consumption and high pollution that tourism-related upstream and downstream industries may have. On the other hand, the problem of carbon leakage among Chinese provinces causes carbon emissions generated by the development of tourism in one province to be transferred to other provinces, which will lead to a spatial mismatch between tourism value creation and carbon emissions. To solve the above problems, this paper measures and evaluates the carbon footprint of China’s inbound tourism based on the carbon footprint theory. Specifically, this paper first constructs the latest multi-region input-output model of China embodied with an inbound tourism satellite account, then, it quantitatively measures the actual and embodied carbon emission (ECE) intensity of the inbound tourism industry chain based on the value chain evaluation method, and finally, it depicts the distribution characteristics and transfer paths of China’s inbound tourism carbon footprint from two dimensions: province and industry. We expect that China’s inbound tourism carbon footprint will be huge, and because of the different resource endowments of various provinces, leading to different advantageous industries among provinces, and thus different positions in the tourism industry chain, the amount of hidden carbon in tourism will also be different among provinces. It is hoped that the conclusions of this paper can enrich the related research on carbon emission reduction in tourism, and provide insight for China’s provinces to achieve coordinated low-carbon tourism.

The contributions of this paper focus on the following three aspects: (1) constructing the latest version of inbound tourism satellite account, which is helpful to define the reasonable boundary of tourism system and overcome the underestimation of tourism’s actual economic contribution and carbon emissions caused by the traditional statistical system; (2) measuring the carbon footprint of China’s inbound tourism, and clarifying the distribution characteristics of carbon footprint, which is conducive to clearly analyze the economic benefits and carbon emission impacts brought by the development of China’s inbound tourism; and (3) analyzing the transfer path of inbound tourism carbon emissions between industries and regions, contributing to strengthening the regional cooperation in inbound tourism carbon emission mitigation among provinces and regions, and providing decision-making reference for related industries to reduce indirect carbon emissions caused by inbound tourism.

2. Literature Review

According to Eurostat’s definition, tourism is considered as a person’s non-residential travel activities of not more than one year (including business, leisure and other personal purposes). Tourism can be further divided into domestic tourism, inbound tourism and outbound tourism. Among them, inbound tourism refers to the activities of tourists visiting another country (https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Glossary:Tourism, Visit Date: 14 August 2022). Inbound tourism is a good indicator to assess whether a region is tourism-competitive because it is an important part of a country’s economic system [5].

Many studies have shown that tourism can significantly stimulate economic growth, regardless of whether the country is rich or poor [6], even though the effects are regionally heterogeneous [7]. The heterogeneity of tourism on economic growth is also significant in different regions within a country. For example, Zuo and Huang [8] took 31 provinces in China as research objects and found that there is an inverted U-shaped relationship or an N-shaped relationship between tourism and economic growth in different provinces. Inbound tourism has been widely proven to have a stimulating effect on the economy, which is one of the important forms of tourism [9]. Brida et al. [10] discovered that countries can be divided into high tourism performance countries and low tourism performance countries according to the degree of inbound tourism’s effect on economic growth. Rasool et al. [11] argued that there is a long-term cointegration relationship within inbound tourism, financial development and economic growth in the BRICS countries. If the unit tourism income of the countries increases by 1%, the unit GDP of the BRICS will increase by an average of 0.31%. They further used the Granger causality test to prove that there is a two-way causality between inbound tourism and economic growth. However, Anser et al. [12] insisted that there is only a correlation between inbound tourism and economic growth, not a causal relationship. For China, Liu et al. [13] recruited the dynamic spatial Durbin model (SDM) to explore the direct and spillover effects from domestic and inbound tourism perspectives. They concluded that the direct promotion effect of the above two is significant but the spatial spillover effect only exists in domestic travel, and the long-term impact is more significant. Li et al. [14] proposed that the development of tourism in rural areas is a necessary way to realize the development of rural green economy.

Tourism consumption also has a significant impact on both the social and ecological environment. Tourism’s impact on society is generally positive. Rational tourism consumption can achieve two international goals of gender equality [15] and poverty reduction [16,17]. However, the role of tourism in poverty eradication is limited and may only occur in the developing countries [18]. Furthermore, inbound tourism has a limited role in reducing poverty compared to domestic tourism [19]. The impact of tourism on the ecological environment is mostly negative. The development of tourism can lead to changes in land use, extinction of wild species, spread of diseases, shifts in environmental perceptions, increased energy use and pollutant emissions [20] and rises the water footprint [21]. Qureshi et al. [22] examined tourism’s ecological footprint of 35 countries during 1995–2016, and concluded that there is a U-shaped relationship between inbound tourism and national NOx emissions, and that inbound tourism leads to biodiversity loss. Outbound travel, on the other hand, has a beneficial “rebound effect” on the country’s ecosystem and air pollutants. Anser et al. [12] comprehensively assessed the ecological footprint of inbound tourism, population growth, and international trade. They observed that inbound tourism was a major contributor to global environmental degradation, but this effect exhibited an “inverted U-shape”, different from Qureshi et al. [22].

The intensification of climate change makes more and more scholars pay attention to the impact of tourism on climate change. On the surface, tourism appears to be low-carbon, but from a lifecycle perspective, it is actually a high-carbon industry [23]. [1] calculated the total emissions of the global tourism industry, which is 8% of the anthropogenic carbon emissions. Mancini et al. [24] believe that the high carbon emissions of tourism are mainly caused by transportation. Geng et al. [25] pointed out that the impact of inbound tourism on climate change depends on planning, that is, inbound tourism can mitigate climate change if the planning is reasonable; otherwise, it will make climate change more severe. To accurately reflect the impact of tourism development on emissions and achieve the goal of sustainable tourism, it is necessary to measure the carbon footprint of tourism [26]. Lenzen et al. [1] measured the carbon footprint of tourism among 160 countries around the world and found that the global carbon footprint of tourism increased from 3900 MtCO₂-eq in 2009 to 4500 MtCO₂-eq in 2013, accounting for about 8% of global greenhouse gas emissions. Transport, shopping and food are significant contributors, and the bulk of tourism’s carbon footprint is generated by and within high-income countries.

From a national perspective, Munday et al. [27] examined the carbon footprint of tourism in the UK and found that the average total carbon footprint of all tourists was 12.72 tons per thousand tourists per day, lower than that of day-trip tourists (15.87 tons), but higher than EU tourists (10.64 tons). Cadarso et al. [28] found that the ECE of the Spanish tourism industry in 2007 accounted for 10.6% of the country’s total emissions, of which inbound tourism accounted for 47%, domestic tourism accounted for 36%, business tourism accounted for 14%, and public administration expenditure accounted for 3%. El Hanandeh [29] investigated pilgrims’ daily greenhouse gas emissions (religious tourism behavior) in Saudi Arabia to be 60.5 kg CO₂-eq. Rico et al. [30] measured the tourism carbon footprint of Barcelona (including arrival and departure transportation, intra-city transportation, accommodation, leisure activities) and obtained a total carbon emission of about 9.6 MtCO₂-eq every year, equivalent to each tourist’s daily emission being 96.9 kg CO₂-eq. Kitamura et al. [31] adopted the input-output method to estimate the total annual emissions of tourism consumption in Japan to be 136 mt CO₂. Among them, the transportation sector accounted for the largest share, accounting for 56.3%, but there was an unexpected result: souvenirs were the second largest carbon footprint industry, accounting for 23.2%.

As one of the emerging developing countries, China’s carbon emissions will continue to increase in the future [32]. However, compared with the countries mentioned above, the carbon footprint research of China’s tourism industry started relatively late [33]. Meng et al. [33] combined the input-output method with the tourism satellite account and found that the total carbon emissions of China’s tourism industry in 2002, 2005, 2007 and 2010 were 111.49, 141.88, 169.76 and 208.4 Mt, respectively, which account for 2.489%, 2.425%, 2.439% and 2.447% of China’s total carbon emissions, respectively. The embodied carbon emissions of other tourism sectors excluding the transportation industry are 2–3 times higher than the direct. By comparing 2009 and 2015, Luo et al. [34] found that although the travel distance and travel time of tourists in Zhangjiajie, China, have shortened in the past few years, the increase in the number of trips has resulted in a significant increase in carbon emissions, while from the perspective of carbon intensity, it is declining. However, their study only considered the transportation sector’s carbon footprint, which is not a comprehensive concept of tourism’s carbon footprint. Luo et al. [35] examined the driving factors of tourism carbon emissions on the basis of constructing China’s economic environment accounts to estimate tourism carbon emissions, and found that domestic travel (rather than inbound tourism) is the main driving force of tourism carbon emissions. In their sample period, the direct carbon emissions of domestic tourism increased by 140%, and the total carbon emissions including indirect carbon emissions increased by 263%. Han et al. [4] examined the carbon footprint of inbound tourism in 31 provinces in China from the provincial level. They found that from 2007 to 2017, the carbon footprint of China’s inbound tourism continued to rise (from 5.62 Mt to 10.88 Mt). They also found that there is temporal and spatial heterogeneity in the change of carbon footprint. However, in their analysis, the carbon emission coefficient of the IPCC is used to specify the emission boundary and industry scope of the tourism industry, which makes their research results possibly inaccurate and incomparable with other studies.

The measurement methods of carbon footprint, including tourism, are mainly divided into two categories: the bottom-up life cycle method; and the top-down input-output method. The life cycle approach calculates the total direct or indirect carbon emissions by recording the entire life cycle of a product or service from production, transportation, distribution to consumption, disposal and recycling [36,37]. The advantage of this method is that it can show the flow of carbon emissions from production to consumption, but the disadvantage is that it is difficult to obtain research data and cannot analyze the linkages between industrial sectors [26]. Using the life cycle approach, Puig et al. [38] measured the carbon footprint of tourism in Spain. Focusing on religious tourism, Campos et al. [39] examined the carbon footprint of the Lebanese pilgrimage and found that the carbon footprint of the three-day pilgrimage was 13.69 kg CO₂-eq, of which accommodation and services accounted for 71.47% and food accounted for 17.08%, and waste treatment accounted for 11.45%. The top-down input-output method mainly decomposes the relationship between the intermediate sector and the final demand based on the input-output table when calculating the carbon footprint. The advantage of this method is that it can intuitively measure tourism between different countries, and the relationship between tourism and other industries [40]. However, the disadvantage is that it is limited by the time when the state publishes the input-output table, which means the timeliness is poor. On this basis, we can formulate tourism emission reduction measures at the macro level and analyze the spillover effects of tourism on other sectors [26]. In addition, the input-output method can be combined with tourism satellite accounts, thus making research conclusions comparable [41]. Patterson and McDonald [42] were among the first scholars to calculate the carbon footprint of tourism using the input-output method. In the actual calculation, due to the spatial spillover effect of carbon emissions, it is more suitable to use a multi-regional input-output model for computation. Yu et al. [43] combined an environmental multi-regional input-output model with a tourism satellite account using Beijing, China, as a sample. They observed that between 2007 and 2012, Beijing’s inbound tours’ carbon footprint declined steadily, while domestic tours’ continued to rise. Sun et al. [44] further summarized three tourism carbon accounting methods based on the multi-regional input-output model. From the above analysis, we believe that the two carbon footprint accounting methods have their own merits, and the key lies in the data that the researcher has obtained and the purpose of their study.

Carbon intensity is an important indicator to measure a country or region’s carbon emissions in addition to the carbon footprint. Measuring the carbon intensity of the whole industry chain based on the tourism industry’s value chain is meaningful. With the integration of global trade, the global value chain concept has gradually been widely discussed in academic circles. The decomposition of the worldwide value chain based on the multi-regional input-output model has been created and started to be used by scholars to track the carbon footprint of tourism, but related research is still very lacking. Xia et al. [45] combined environmental expansion input-output analysis and data envelopment analysis, and observed that tourism produces a large amount of indirect carbon emissions in the entire supply chain, and the most important contributor is tourism hotels.

By combing the existing literature, we can see that the combination of the environmental expansion input-output model and the tourism satellite account is the main method to measure the carbon footprint from a macro perspective, which can make the research results scientific and comparable. In addition, since tourism has carbon leakage between regions, the multi-regional input-output model is adopted to study the carbon footprint of tourism. From the research perspective, the literature that focused on the carbon footprint of China’s inbound tourism at the provincial level is insufficient. On the other hand, the literature exploring the carbon intensity of the inbound tourism industry chain is still limited. This study attempts to fill these gaps.

3. Methodology and Data

3.1. MRIO Model

The MRIO model has been widely adopted to track commodity flows can carbon footprint between regions’ major economic sectors [46]. MRIO analysis makes it possible to track the value-chain and environment footprint from different perspectives (region to region, sector to sector, and sector to region, etc.). Supposing an MRIO table with $m$ regions and $n$ sectors, the total output of sector $i$ in region $j$ is $X_i^j$, which complies with the horizontal accounting balance as Equation (1). Where $x_{ik}^{jt}$ denotes the intermediate input from sector $k$ in region $t$ to sector $i$ in region $j$, $y_i^{jt}$ represents the final use of region $t$ for commodities in sector $i$ from region $j$.

$$X_i^j={\displaystyle \sum }_t{\displaystyle \sum }_k{x}_{ik}^{jt}+{\displaystyle \sum }_t{y}_i^{jt}$$

(1)

The balance relationship can be represented as the matrix form (Equation (2)), where $A$ stands for the direct consumption matrix.

$$X=AX+{\displaystyle \sum }_t{Y}^t$$

(2)

$L^{i,j}$ is the Leontief inverse matrix, which is a n × n matrix, which denotes the total requirement from region $j$ to region $t$. In our study, we focus on the impact of inbound tourism. The final use requirement caused by inbound tourism can be calculated by Equation (3), where $YIT$ stands for the demand of inbound tourism.

$$X=L\ast YIT$$

(3)

The direct emission triggered by the inbound tourism $DE$ can be estimated according to the carbon emission factor vector $CMF={\left[\begin{array}{cc}cm{f}_1^1& cm{f}_2^1\end{array}\begin{array}{cc}\cdots & cm{f}_m^n\end{array}& \right]}^T$.

$$DE=CM{F}^T\ast YIT$$

(4)

The embodied emission contains the direct emission and the emission embodied in the intermediate input, which can be illustrated by Equation (5).

$$EE=CM{F}^T\ast L\ast YIT$$

(5)

We denote $YI{T}_i^j={\left[\begin{array}{ccccc}0& \cdots & y_i^j& \cdots & 0\end{array}\right]}^T$ as the final use of inbound tourism for sector $i$ in region $j$. The carbon footprint transferring from sector $i$ in region $j$ to other sectors can be described by Equation (6), where the symbol $°$ means the Hadamard product, $E{E}_i^j$ is a $n\times m$ column vector, which represents the embodied emission caused by sector $i$ in region $j$.

$$E{E}_i^j=CM{F}^T°L\ast YI{T}_i^j$$

(6)

If we replace the carbon emission factor vector with the value-added vector $VAF={\left[\begin{array}{cc}v{a}_1^1& v{a}_2^2\end{array}& \begin{array}{cc}\cdots & v{a}_m^n\end{array}& \right]}^T$. The value chain of inbound tourism can be evaluated. The formulas are listed as following, where $DVA$ denotes the direct value-added created by inbound tourism, $EVA$ denotes the embodied value-added caused by inbound tourism, and $EV{A}_i^j$ means the value-added created by the inbound tourism consumption of sector $i$ in region $j$.

$$DVA=VA{F}^T\ast YIT$$

(7)

$$EVA=VA{F}^T\ast L\ast YIT$$

(8)

$$EV{A}_i^j=CM{F}^T°L\ast YI{T}_i^j$$

(9)

3.2. Tourism Satellite Account

The Tourism Satellite Account (TSA) is a dataset of measuring the direct economic impacts of tourism consumption to a national economy [47]. Since 1993, the World Tourism Organization (UNWTO) began to develop the TSA that observed the principles of SNA. Some countries such as New Zealand, Canada and Sweden now release the TSA to provides a picture of the impact of tourism activity.

Tourism-related activities are scattered in different product sectors, and the existing national statistical account system given by the National Bureau of Statistics of China cannot treat tourism as a sector alone. Referring to Frechtling [47] on the division rules of tourism satellite accounts, this study classifies the tourism activities into the corresponding sectors. The inbound tourism expenditure activities include long-distance transportation, tour, accommodation, food and beverage, shopping, entertainment, post and telecommunications, urban transportation, and other services. These nine activities can correspond to the sector of the input-output table, which can be treated as the ESA, and are listed in

Table 1. ESA setting.

Inbound Tourism Expenditure	Corresponding Sector
Long-distance transportation	Transport, warehousing and post
Tour	Culture, sports and entertainment
Accommodation	Accommodation and food service
Food and beverage	Accommodation and food service
Shopping	Wholesale and retail trade
Entertainment	Culture, sports and entertainment
Post and telecommunications	Telecommunications, computer programming and information services
Urban transportation	Transport, warehousing and post
Other services	Resident services, repairs and other services

.

3.3. Data Source

The multi-region input-output table is the prerequisite to construct the MRIO model. In this study, we recruit China’s Regional Input-Output Table (CRIOT) 2017, released by the National Bureau of Statistics of China, as the original dataset. Due to the CRIOT 2017, this only includes the input-output relationship within each province, as well as the total inter-provincial flow of each province. However, the commodity flow relationship between provinces and sectors is not given. We refer to Greaney and Kiyota [48] and adopt the Gravity model to measure the inter-provincial commodity flow. It should be pointed out that the input-output table of China’s provincial level is only updated to 2017, which limits the research of this paper. In other words, using the method of combining the input-output table of environmental expansion, with the environmental satellite account to calculate the carbon footprint, will make us limited by data.

The expenditures of different provincial region are acquired from the “China Tourism Statistics Yearbook” published by Ministry of Culture and Tourism, and the data released by each provincial government. The descriptive statistic of sample is shown in

Table 2. Descriptive statistics of sample.

Region		Mean	Max	Min	Std
Number of Tourists (Million)		3.74	36.55	0.07	6.34
Expenditure (Million CNY)	Long-distance transportation	5975.41	41,070.95	49.11	7934.06
	Tour	674.28	4174.42	10.98	837.68
	Accommodation	2750.58	19,660.19	22.80	4046.29
	Food and beverage	1435.16	13,196.57	11.68	2406.21
	Shopping	3503.45	30,971.53	23.92	5491.53
	Entertainment	868.28	7675.55	6.61	1408.85
	Post and telecommunications	360.92	2423.86	2.53	450.99
	Urban transportation	398.19	2962.49	4.22	568.02
	Other services	1806.41	12,388.61	8.73	2317.48

.

4. Results and Analysis

4.1. Comparison of Direct and Embodied Carbon Emissions from Inbound Tourism

This section firstly calculates the direct and embodied carbon emissions of China’s inbound tourism, which are shown in Figure 2. The spatial distribution of inbound tourism carbon emissions is higher in the east and lower in the west. The total carbon emissions of coastal provinces are significantly higher than those of other provinces. This is mainly because the coastal provinces have better transportation conditions, so the inbound tourism activities are more intensive. Meanwhile, the carbon emissions caused by inbound tourism vary widely among provinces. In general, the carbon emissions of inbound tourism show a decreasing trend from the southeastern provinces to the northwestern provinces. As for direct carbon emissions from inbound tourism, Guangdong, Yunnan and Shanghai ranked the top three, with 4.04 million tons, 2.5 million tons and 1.98 million tons, respectively. This is mainly due to the fact that Guangdong and Yunnan provinces are traditional popular tourism areas, so the number of inbound tourists remains high, resulting in massive carbon emissions from inbound tourism. These provinces need to focus more on promoting low-carbon and clean consumption in the inbound tourism industry in the future. Compared with the above two provinces, Shanghai is an economically developed area with obvious geographical advantages and close international exchanges, with relatively high carbon emissions of inbound tourism. Shanghai should make greater efforts to reduce carbon emissions from inbound tourism in the future. After considering the ECE, the emissions in Guangdong, Yunnan and Shanghai rose to 7.69 million tons, 3.36 million tons and 4.4 million tons, respectively, which indicates that from the perspective of whole tourism industry chain, there are still a large number of indirect carbon emissions generated by the intermediate process that have been ignored. This also reflects the necessity of quantifying the carbon footprint of inbound tourism.

By comparing the direct and the ECE of China’s inbound tourism (see Figure 2), it can be found that the ECE of inbound tourism are higher than direct carbon emissions, especially for western provinces such as Ningxia and Gansu. Ningxia, Gansu, Shaanxi and Inner Mongolia have direct carbon emissions of 0.01 million tons, 0.008 million tons, 0.078 million tons and 0.22 million tons, while the embodied carbon emissions are 0.44 million tons, 0.25 million tons, 1.37 million tons and 2.57 million tons. Obviously, the direct carbon emissions caused by inbound tourism in these provinces are not high, but the ECE scale cannot be neglected. In Ningxia province, the proportion of ECE in inbound tourism to direct carbon emissions even reached 44 times. The reason is that the industries leading to direct emissions from inbound tourism are mainly the power and transportation sectors, and intermediate inputs need to be consumed in the process of providing service in these industries. Due to the pollution transfer to the western region, the embodied carbon intensity of the local industrial chain is relatively high, resulting in the high embodied carbon intensity of inbound tourism in the above provinces. For a long time, due to differences in regional low-carbon policy intensity, developed areas such as eastern China tended to import high-carbon products or transfer high-polluting enterprises from less-developed areas, such as western China, resulting in the phenomenon of “pollution havens”. Considering the low economic development level and fragile ecological environment in the western region, the government should introduce targeted policies according to local conditions in the future, to help regions effectively reduce the carbon footprint of inbound tourism. In addition, in order to achieve China’s carbon neutrality goal, the national level should promote the low-carbon of national inbound tourism as a whole.

Figure 3 further compares each province’s economic contribution and emission share of inbound tourism. It can be found that the overall emission contribution of inbound tourism in each province is slightly lower than the economic contribution, indicating that carbon emissions’ intensity is close to the average economic carbon intensity. In other words, inbound tourism is not a low-carbon and clean industry. From the above analysis, we can see that China’s inbound tourism cannot simultaneously achieve economic growth and carbon emission reduction goals. This differs from previous scholars’ perception that tourism is a clean industry.

Figure 3 shows the contribution of inbound tourism in different provinces to the economy and carbon emissions. We can clearly see that the dependence of the economy and the carbon emissions of different provinces on inbound tourism is very different. The direct economic contribution of inbound tourism in Yunnan, Fujian and Guangdong reaches 0.95%, 0.84% and 0.77%, respectively. After accounting for the embodied economic contribution, this value rises to 1.59%, 1.55% and 1.46%, respectively. In contrast, the direct economic contribution of inbound tourism in western provinces such as Gansu and Ningxia are negligible, only 0.01% and 0.04%, and embodied economic contribution is also low. Meanwhile, in first-tier cities such as Beijing and Shanghai, the ECE of inbound tourism is much larger than the direct carbon emissions, with a difference of 1.36% and 0.75%, respectively. As China’s economic and political centers, Shanghai and Beijing have convenient transportation and highly mature tourism development. The number of inbound tourists has always been at the forefront of the country, so the ECE of inbound tourism is relatively high. Besides, an interesting phenomenon is that the difference between the ECE contribution and the embodied GDP contribution of Shanghai and Beijing reached 1.21% and 1.19%, respectively, indicating that the inbound tourism activities in these key cities have exacerbated the carbon emission intensity of the local economy. In the process of promoting carbon emission reduction, China’s provinces should not only focus on manufacturing, but also attach importance to the important role of tourism.

After investigating the contribution of inbound tourism to the economy and carbon emissions, we want to compare further the direct carbon emission intensity of inbound tourism with the specific carbon emission intensity, which enables us to find out the efficiency of inbound tourism more directly (direct carbon intensity is defined as “direct emissions/direct GDP impact”, while embodied carbon intensity refers to “ embodied emissions/embodied GDP impact”). Figure 4 shows that on average across the country, the direct and the embodied carbon emission intensity of inbound tourism are 68 g/CNY and 89 g/CNY, respectively, which are different to some extent. This indicates that a large number of carbon emissions generated by intermediate tourism links are transferred through spillover effects between provinces and industries, which will also be analyzed in detail in Section 4.2 and Section 4.3. It suggests that local governments should not only consider the local tourists, but also take into account the carbon footprint left by non-local tourists in the local area when designing low-carbon tourism policies. In terms of provinces, there is a big difference between the direct and embodied carbon emission intensity of inbound tourism. Tibet and Yunnan have the highest direct carbon emission intensity, with 186 g/CNY and 163 g/CNY, respectively. As for the embodied carbon emission intensity of inbound tourism, Ningxia, Shaanxi and Inner Mongolia rank among the top three in China, with 764 g/CNY, 304 g/CNY and 293 g/CNY, correspondingly. Due to the geographical disadvantages and inconvenient transportation, these three provinces have few inbound tourists, and inbound tourism has a limited impetus to the local economy. At the same time, because of the phenomenon of “pollution havens”, it bears the carbon emissions transferred through other provinces and the tourism industry chain, so the embodied carbon emissions’ intensity of inbound tourism is far higher than that of other provinces.

Figure 5 describes the unit carbon emissions of tourists in the process of inbound tourism. It is obvious that Ningxia, Shaanxi and Inner Mongolia are at the forefront of the country, which is mainly due to the high ECE of inbound tourists and lower inbound tourists in these provinces. In contrast, the per capita ECE of inbound tourists in Guangdong and Zhejiang are at a relatively low level, which is mainly caused by two reasons. On the one hand, these regions are rich in tourism resources, convenient in transportation and superior geographical location, which attract more inbound tourists. On the other hand, these economically developed regions also transfer part of the tourism carbon emissions to some western provinces through industrial chain transmission.

4.2. Transfer Path of Inter-Provincial Embodied Carbon Emission of Inbound Tourism

Based on the above analysis, there exists inter-provincial transfer in the ECE of inbound tourism. Thus, this section further clarifies the inter-provincial transfer path, with results shown in Figure 6. From the perspective of net outflow of ECE from inbound tourism, Inner Mongolia, Shanxi and Xinjiang occupy the top three in China, with net ECE outflows of 1.47 million tons, 1.07 million tons and 1.02 million tons to other provinces. Taking Inner Mongolia as an example, its ECE mainly flows to Guangdong, Shanghai and Beijing, accounting for 24%, 14% and 13% of the total emissions, respectively. Guangdong province has the largest net inflow of carbon emissions from inbound tourism in China, with 2.66 million tons of carbon emissions from other provinces. Inner Mongolia, Xinjiang and Shanxi are the main inflow provinces, accounting for 12%, 8% and 7% of the total inflow, respectively. Benefiting from the obvious location advantages and tourism resources, Guangdong and Beijing accept the most ECE transferred from other provinces, which leads to the contradiction between tourism economic development and emission reduction in these areas, and is not conducive to the realization of carbon neutrality goal. On the contrary, due to the long travel distance and incomplete infrastructure, the transfer characteristics of the ECE of inbound tourism in Tibet, Qinghai and Shaanxi are not obvious. In the future, with the in-depth implementation of “the Belt and Road Initiative”, the western region will not only usher in new opportunities for tourism development, but also face a new constraint of energy conservation and emission reduction. This shows us that government should take into account the inter-provincial transfer path of ECE for inbound tourism when formulating low-carbon tourism policies.

From the perspective of regional transfer, Figure 7 shows that the ECE of inbound tourism mainly flows from the northeast and northwest regions to the southern and eastern regions. The northeast, northwest and northern regions are the top three areas with a net outflow of ECE from inbound tourism, with outflows of 1.65 million tons, 1.62 million tons and 0.83 million tons, respectively. Taking northeast China as an example, its main outflow regions are the eastern, southern and northern, accounting for 33%, 25% and 21% of the total ECE outflow, correspondingly. The south is China’s largest net inflow region of ECE from inbound tourism, and its main inflow regions are the north, east and northwest, accounting for 28%, 25% and 17% of the total ECE inflow. The formation of a regional ECE transfer path is mainly due to the huge differences in economic development level, technological level, industrial scale and structure of tourism in various regions.

4.3. Transfer Path of Inter-Industry Embodied Carbon Emission of Inbound Tourism

This section will focus on the industry perspective to explore the industrial transfer path of ECE in inbound tourism. Compared with the flow of ECE between regions, the relationship of ECE flow between sectors is more concentrated, and the impact on a key sector may quickly spread to other sectors in the network, resulting in changes in the entire industrial system. This means that grasping the core embodied carbon flow path of the inbound tourism industry chain, and identifying key sectors in the flow of ECE, is conducive to promoting the coordinated emission reduction of related departments in the inbound tourism industry. Therefore, it is necessary to explore the industrial transfer path of ECE in inbound tourism. As can be seen from Figure 8, the sectors directly related to inbound tourism are Wholesale and Retail Trade (TRD), Transport, Warehousing and Post (TWP), Accommodation and Food Service (AFS), Telecommunications, Computer Programming and Information Services (TEL), Resident Services, Repairs and Other Services (RRO), Culture, Sports and Entertainment (CSE). Among them, TWP, which are closely related to tourism, have the highest ECE of 11.32 million tons. Residents need corresponding accommodation services during the travel process, which will drive the development of hotels, catering and other related industries. Thus, the ECE of AFS brought by tourism activities ranked second, with 6.58 million tons. In addition, since the travel process is often accompanied by the consumption of goods, the ECE of TRD increased by 3.86 million tons. It can be seen that the sectors directly related to inbound tourism have a high concentration of ECE, with the top three industries accounting for about 82.5% of total emissions.

Further analysis of the upstream industries of tourism reveals that the ECE of inbound tourism ultimately comes from the production process of Production and Distribution of Electricity (ELE), Casting of Metals (CME), and Coke and Refined Petroleum Products (ROC). Among them, the ELE contributed the most to the ECE of inbound tourism, accounting for 17.61 million tons, with 72.2% of the total emissions. Electricity and heat, as the basic energy produced by enterprises, are closely related to national economic activities. As can be seen from Figure 8, the TWP, AFS, and TRD all drive the production and carbon emissions of ELE. Since the consumption of TWP is mainly supported by ELE, this also implies a key sectoral transfer path from “ELE- TWP- inbound tourism”. Due to the ELE, are emission-intensive and high-emission industries, this indicates that industrial restructuring is crucial to curbing the transfer of ECE from inbound tourism between industries.

5. Conclusions and Policy Implications

5.1. Conclusions

Tourism is emerging as an important contributor to climate change as inbound tourism is the main component of tourism activities. It is of great significance to clarify the distribution characteristics and transfer paths of tourism carbon footprints in regions and industries for achieving global emission reduction and sustainable development goals. This paper constructs the multi-region input-output model of China’s embodied with inbound tourism satellite account, then quantitatively measures the actual and embodied carbon emission intensity of the inbound tourism industry chain, based on the value chain evaluation method, and finally it depicts the distribution characteristics and transfer paths of China’s inbound tourism carbon footprint from two dimensions, province and industry. The main research conclusions are as follows:

(1)

The carbon emissions of inbound tourism show a decreasing trend from southeastern provinces to the northwestern provinces. Besides, the ECE of inbound tourism are higher than direct carbon emissions. In Ningxia, the proportion of ECE in inbound tourism to direct carbon emissions even reached 44 times. The reason is that the industries leading to direct emissions from inbound tourism are mainly the power and transportation sectors, and intermediate inputs need to be consumed in the process of providing service in these industries. Due to pollution transfer to the western region, the embodied carbon intensity of the local industrial chain is relatively high, resulting in the high embodied carbon intensity of inbound tourism in the western provinces. It can be found that the overall emission contribution of inbound tourism in each province is slightly lower than the economic contribution, indicating that the carbon emissions intensity is close to the average economic carbon intensity. In other words, inbound tourism is not a low-carbon and clean industry. Thus, it is difficult for inbound tourism to achieve the dual goals of stimulating economic growth and reducing emission.

(2)

There are large distinctions in the dependence of the provincial economy on inbound tourism. The direct economic contribution of inbound tourism in Yunnan and Fujian reaches 0.95%, and 0.84%, respectively. In contrast, the direct economic contribution of inbound tourism in western provinces such as Gansu and Ningxia are negligible, only 0.01% and 0.04%. Besides, an interesting phenomenon is that the difference between the ECE contribution and the embodied GDP contribution of Shanghai and Beijing reached 1.21% and 1.19%, respectively, indicating that the inbound tourism activities in these key cities have exacerbated the carbon emission intensity of the local economy. As for embodied carbon emission intensity of inbound tourism, Ningxia, Shaanxi and Inner Mongolia rank among the top three in China, with 764 g/CNY, 304 g/CNY and 293 g/CNY, correspondingly. Due to the geographical disadvantages and inconvenient transportation, these three provinces have few inbound tourists, and inbound tourism has limited impetus to the local economy. Meanwhile, because of the phenomenon of “pollution havens”, these regions bear the carbon emissions transferred through other provinces and the tourism industry chain, so the ECE intensity of inbound tourism in these provinces are far higher than that of others.

(3)

There exists inter-provincial transfer in the ECE of inbound tourism. From the perspective of net outflow of the ECE from inbound tourism, Inner Mongolia, Shanxi and Xinjiang occupied the top three in China, with net ECE outflows of 1.47 million tons, 1.07 million tons and 1.02 million tons to other provinces. Taking Inner Mongolia as an example, its ECE mainly flows to Guangdong, Shanghai and Beijing, accounting for 24%, 14% and 13% of the total emissions, respectively. Guangdong province has the largest net inflow of carbon emissions from inbound tourism in China, with 2.66 million tons of carbon emissions from other provinces. Inner Mongolia, Xinjiang and Shanxi are the main inflow provinces, accounting for 12%, 8% and 7% of the total inflow, respectively. From the perspective of regional transfer, the ECE of inbound tourism mainly flows from the northeast and northwest regions to the southern and eastern regions. The northeast, northwest and northern regions are the top three areas with net outflow of ECE from inbound tourism, with outflows of 1.65 million tons, 1.62 million tons and 0.83 million tons, respectively. Taking northeast China as an example, its main outflow regions are the eastern, southern and northern, accounting for 33%, 25% and 21% of the total ECE outflow, correspondingly. The south is China’s largest net inflow region of ECE from inbound tourism, and its main inflow regions are the north, east and northwest, accounting for 28%, 25% and 17% of the total ECE inflow, respectively.

(4)

There exists industry transfer in the ECE of inbound tourism. Further analysis of the upstream industries of tourism reveals that the ECE of inbound tourism ultimately come from the production process of Production and Distribution of Electricity (ELE), Casting of Metals (CME) and Coke and Refined Petroleum Products (ROC). Among them, the ELE contributed the most to the ECE of tourism, accounting for 17.61 million tons, with 72.2% of the total emissions. Since the consumption of TWP is mainly supported by ELE, this also implies a key sectoral transfer path from “ELE- TWP- inbound tourism”.

5.2. Policy Implications

Based on the relevant research conclusions, this paper puts forward some targeted policy recommendations to improve the government’s mechanism design.

Firstly, when designing low-carbon tourism policies, local governments should not only consider local tourists, but should also take into account the carbon footprint left by non-local tourists in the local area when designing low-carbon tourism policies. In addition, each province should specify the amount of carbon emissions from inbound tourism within its region. In the future, the carbon footprint of inbound tourism should be considered in carbon emission management. All provinces should improve green tourism planning.

Secondly, it should play a leadership role in reducing carbon emissions from inbound tourism without compromising tourism economic growth. Green travel should be fully encouraged, or carbon sinks should be adopted to neutralize carbon emissions during inbound tourism. Considering the low economic development level and fragile ecological environment in the western region, the government should pay more attention to the transfer of ECE from inbound tourism in these provinces, such as Tibet, Qinghai and Ningxia. In the future, targeted policies according to local endowment should be introduced to effectively reduce the carbon footprint of inbound tourism. However, there are more green carbon sinks in western China. Through the sale of carbon sinks, the benefits of green travel can be increased.

Finally, as inbound tourism mainly involves energy-intensive and high-emission industries, industrial restructuring is crucial to curbing the transfer of ECE from inbound tourism between industries. The government should strengthen the emission reduction responsibilities of various industries in the tourism industry chain, innovate low-carbon technologies to optimize the industrial structure and promote low-carbon inbound tourism. Additionally, the government may improve carbon emission accounting methods and commitment mechanisms.

However, this study only analyzes the carbon footprint of China’s inbound tourism in 2017. In the future, we can further analyze the dynamic evolution, or we can carry out structural decomposition analysis in combination with other years.