The Tourism Service Trade Network: Statistics from China and ASEAN Countries

Abstract

This study uses the social network analysis method to explore the structural changes of the China–ASEAN tourism services trade network and its influencing factors based on tourism services trade data of China and ASEAN countries in 2015 and 2018. The findings show that (i) the network density of the tourism services trade network increased by 30% in 2018 compared with 2015. (ii) China, Thailand, Singapore, and Malaysia rank highly in terms of degree, betweenness, and closeness centrality. (iii) The distance between countries and differences in GDP per capita significantly affect the tourism services trade network. The recommendation is that tourism services trade be developed in a vigorous and systematic manner in the China–ASEAN region. This approach would enhance overall stability and cooperation in the tourism services trade network and create a win–win situation for trade in tourism services between China and ASEAN countries.

1. Introduction

According to the 2020 World Trade Statistics published by the World Trade Organization, the global value of trade in services reached USD 58,980 billion in 2019, of which the total value of tourism services trade was USD 1416 billion. Currently, the ASEAN (Association of Southeast Asian Nations) region is one of the world’s most active regions in economic development. China has a history of friendly relations with ASEAN due to its geopolitical proximity, especially after the official signing of the ASEAN–China Free Trade Area (ACFTA) Agreement on Trade in Services on 14 January 2007, which further promoted the development of their economies. During that time, China and ASEAN entered a golden period of economic and trade cooperation. According to data released at the China–ASEAN Economic and Trade Cooperation Conference held by the State Council Information Office of the People’s Republic of China on 27 September 2020, the bilateral trade volume between China and ASEAN grew from USD 292.8 billion in 2010 to USD 641.5 billion in 2019. Two-way investments between China and ASEAN reached USD 223 billion by the end of 2019, making the ACFTA one of the world’s most dynamic free trade zones. According to data released by the General Administration of Customs of the People’s Republic of China, ASEAN overtook the EU as China’s top trading partner in goods. In another breakthrough, ASEAN overtook the United States as China’s second-largest trading partner in 2019. In addition, China has maintained its position as ASEAN’s top trading partner for 12 consecutive years, showing that the China–ASEAN economic and trade relations remain very stable.

Tourism services trade (also known as international trade in tourism services) is the remunerative flow and exchange of tourism services between economies. It is the exchange of various tourism services by economies to support international tourists and tourism activities [1]. There are no statistical indicators of the tourism services trade in existing international documents, agreements, and statistical manuals on trade, trade in services, and balance of payments. The Manual on Statistics of International Trade in Services 2010 (MSITS 2010), produced by agencies including the International Monetary Fund (IMF), the Organisation for Economic Co-operation and Development (OECD), the United Nations Conference on Trade and Development (UNCTAD), and the World Trade Organization (WTO), classifies trade in services into four modes according to the General Agreement on Trade in Services (GATS): cross-border supply, consumption abroad, commercial presence, and presence of natural people. The trade in services under these four modes encompasses both cross-border and non-cross-border transactions. Among them, cross-border supply, consumption abroad, and presence of natural people are considered cross-border transactions, while commercial presence is considered a non-cross-border transaction. A complete classification of tourism services trade should include all elements of cross-border and non-cross-border tourism services trade. However, at present, only a few countries, such as the United States, report data on trade in services that include the four modes; other countries do not have correspondingly clear statistics. The MSITS 2010, developed and published by the United Nations and other agencies, added a new section to measure the four modes of the supply of trade in services and expanded the scope of international trade in services statistics. It states that consumption abroad can include only service-related travel expenditures and that goods-related expenditures cannot be classified as any mode of supply [2]. In the current framework of the World Trade Statistical Review, statistics on outbound tourism expenditures are listed under mode 2 (consumption abroad) of international trade in services. Conversely, the United Nations World Tourism Organization (UNWTO) mainly counts the number of tourism trips and total or per-capita consumption reported by each country. The major difference between trade in tourism services and trade in other services is that tourists’ mobile transit consumption is part of trade in tourism services. Compared with physical trade, inbound tourism is an export of trade in tourism services, while outbound tourism is an import of trade in tourism [3]. Hence, we use the tourism consumption of inbound tourists in the destination country as an important indicator to measure the volume of tourism services trade in a country. We consider cross-border supply, consumption abroad, and presence of natural people overall; that is, we use the total consumption of tourists abroad as a proxy indicator for cross-border tourism services trade.

Scholars began to study international tourism services trade in the 1970s. Generally, they conducted qualitative studies on the basic theories of international trade in services and appropriately extended them to trade in tourism services. They explained how trade in tourism services followed the principles of comparative and competitive advantages [4]. Existing theoretical studies on tourism services trade do not form a complete theoretical framework. Scholars have focused mainly on two issues of tourism services trade: regarding tourism services trade as a subsector of trade services in researches of international tourism flows; academic study of international tourism flows is mainly aimed at tourism services trade for transit consumption. Rarely have extant studies discussed the four modes of tourism services trade under the GATS framework. Generally, scholars have used economic theories, such as externality, public goods, information asymmetry [5], comparative advantage, international competitive advantage, and service trade liberalization, for theoretical research on the competitiveness of tourism services trade from different perspectives. Chinese scholars research on tourism services trade has focused on two aspects: the relationship between outbound tourism, trade in services, and economic growth [6,7], and the competitiveness of tourism services trade. Scholars have measured the international competitiveness of China’s tourism services trade using the export market share index and Revealed Comparative Advantage Index, the comparative advantage index of trade in services [8], and the trade competitiveness index and the export rise advantage index. Models have been constructed using hierarchical analysis, the Delphi method [9], and international competitiveness theory to evaluate the level of international competitiveness in a country’s tourism services trade. In analyzing the development of the tourism services trade, most studies have focused on the direct or indirect economic impact of its development from a country’s perspective. In contrast, studies on the development of regional tourism services trade from the perspective of regional cooperation, such as between China and ASEAN, are relatively rare. Due to differences in research perspectives, current studies on China–ASEAN tourism services trade from the perspective of trade in services have focused mainly on topics such as comparing the competitiveness of tourism services trade between China and ASEAN countries [10], the legal protection system of China–ASEAN tourism services trade, and the China–ASEAN tourism services trade deficit. In addition, the extant studies have concentrated on qualitative research methods. Most quantitative studies have used methods such as gravity models, impulse responses, cointegration analysis, and Granger causality tests [11,12] to discuss the impact and spillover effects of trade and tourism. Additionally, generally, these research results have been from within China. Few studies have examined domestic tourism services trade.

The first study to examine trade patterns through the lens of networks was conducted by Snyder and Kick (1979). They analyzed the core-periphery structure of 118 sample countries by constructing networks using trade data. Succeeding researchers have explored the characteristics of international trade networks by constructing their structures using topological principles [13], while other researchers have used weights to construct trade network structures to study their nature and evolution. These studies have discovered that factors, such as international trade characteristics [14,15], geographic location, trade flows, and the degree of trade closeness [16], influence the services network trade structure. However, exploring international trade network patterns using the social analysis methods is a new research approach. Researchers have used import and export data between countries to construct international trade networks and discuss the structural characteristics of these networks. Some scholars believe that international trade networks have the characteristics of scale-free distribution with high clustering coefficients [17]. In contrast, others believe that the trade network is a negative matching network and that the rich-club phenomenon exists [18]. Based on the analysis of the structural characteristics of international trade networks, scholars have analyzed the factors influencing bilateral trade flows between countries as well. They have argued that a country’s level of economic development (represented by GDP), population, geographical distance [19], institutional distance, and cultural distance [20,21] play decisive roles in the structure of its international trade networks.

Thus, the extant research on tourism services trade has several characteristics. First, scholars have focused on the competitiveness of the tourism services trade, while few have focused on the structure of the tourism services trade. Second, researchers have adopted quantitative methods based on economic theories and analyzed the main factors affecting competitiveness by establishing econometric models. They have evaluated tourism services trade from service trade indices such as the export market share index and the RCA index. Third, the results on tourism services trade network structures are scarce, which may be due to the lack of statistical data and the difficulty of identifying their scope. The most common research methodologies used to study tourism services trade are evaluation indicators and the gravity model, impulse response, cointegration tests, and the Granger causality test. Although prior studies have applied social network analysis to trade in tourism, their focus has been on the integrity of tourism services trade relationships and the relationship between the network nodes as research objects [22]. Few studies have captured the development laws and characteristics of the regional tourism services trade network as a whole.

The purpose of this study is as follows: (i) use the pattern of the China–ASEAN tourism service trade network as the research object, and make use of SNA method to measure the overall characteristics of the network and understand the status of China–ASEAN tourism service trade; (ii) use the QAP (Quadratic Assignment Procedure) method to explore the influencing factors of the structure of the tourism service trade network, and explore the evolution law and influencing factors of the inter-regional China–Asean tourism service trade network; (iii) provide policy suggestions for win–win development of inter-regional trade in tourism services between China and ASEAN countries.

To some extent, this study enriches and complements the existing research on the ASEAN tourism service trade. Although many studies have focused on the evolution characteristics of trade networks, there are relatively few studies that apply the social network analysis method to the field of tourism trade, taking the overall pattern of tourism service trade relations and the relationship between network nodes as the research object, and grasp the characteristics of regional tourism service trade network pattern as a whole. This paper introduces the social network analysis method into the study of the tourism service trade network, measures the overall pattern and influencing factors of the tourism service trade network between China and ASEAN countries, and reveals the current situation of China–Asean tourism service trade. In terms of practical application, this paper explores the evolution rules and influencing factors of the regional China–Asean tourism service trade network, so as to enrich the research results of tourism service trade relations between China and ASEAN countries, and provide a reference basis for the win–win development of regional tourism service trade.

2. Data and Methodology

2.1. Data Sources

According to the concept of tourism services trade discussed above, we chose outbound tourism consumption as an important indicator to measure the amount of tourism services trade. The nodes of the tourism services trade network are the 11 China–ASEAN countries, beginning with the source country of outbound tourists and ending with the destination country that provides inbound tourism reception services. Tourist spending in the tourism destination country represents the establishment of a tourism services trade relationship between countries. As there are no official statistics on the mutual exchanges between China and ASEAN countries, we used indirect data with the following calculation method. We used the world tourism statistics published by the UNWTO, the total inbound tourism revenue and the number of inbound tourists in the 11 China–ASEAN countries to calculate the per-capita tourism consumption of inbound tourists in a country, which we then multiplied by the number of inbound tourists to obtain the total inbound tourism consumption of a country, which represents the source country’s total tourism services trade exports. Using China as an example, we multiplied the number of tourists traveling from China to a certain ASEAN country (represented as country j) by the per-capita tourism consumption of inbound tourists in country j in that year to obtain the total inbound tourism expenditure from China to country j, which is the total tourism services trade exports from China to country j. We repeated this calculation to obtain the tourism services trade exports between China and each ASEAN country, thereby constructing data on the bilateral tourism services trade between China and ASEAN countries. Then, we applied social network analysis to construct a network model to explore the structural characteristics of the China–ASEAN tourism services trade network. Our data are from 2015 and 2018. We describe indicator data, such as geographical and cultural distance, in Indicator Selection in Section 3.3.1.

2.2. Research Methods

The gravity model is a traditional tool for empirical research to measure bilateral trade flows. However, the gravity model must address independent assumptions between variables, as required by traditional econometric methods. The social network analysis method, as a non-parametric estimation method, does not require the independent variables to be independent to establish the model and is more robust in dealing with relational data [23]. The advantage of using network analysis methods to study the characteristics of tourism services trade networks and their influencing factors is that the connections between the constructed network nodes are clearer and more intuitive, clearly presenting the structural characteristics of tourism services trade networks. The results are also more objective than simply analyzing attribute data. Second, network analysis methods are not only applicable to individual relationships at the micro-level but can also be applied to macro social relationships and organizational structure analysis. Social network analysis is one of the most convincing structural analysis methods and is, therefore, more applicable to examine the characteristics and influencing factors of network structures. In this study, we use the three dimensions of overall network characteristics, individual network characteristics, and network influencing factors to explore the structural characteristics of the China–ASEAN tourism services trade network by using indicators such as network density and centrality. We use QAP regression analysis to explore the factors affecting the China–ASEAN tourism services trade.

2.3. Key Indicators for Social Network Analysis

(1)

Network density indicator

Network density is an important indicator of a network’s overall condition. Network density indicates the closeness of the connections between individual nodes and is the ratio of the number of linkages that exist in the network to the number of linkages that could exist theoretically. The more countries connected between nodes, the greater the network density, the higher the level of network trade, and the greater the openness and access of the node countries to resources in the overall network and its nodes. The formula for network density is

D = m/n(n − 1),

where D is the network density with values in the range of [0, 1], m represents the number of relationships that exist, and n represents the number of node countries in the network.

(2)

Centrality indicators

Three indicators—degree, betweenness, and closeness centrality—are commonly used to characterize the role and position of each node country in network trade.

Degree centrality is the number of points directly related to the focal point. The larger the value of degree centrality, the stronger the node centrality, indicating that the node in the network has a stronger ability to interrelate with other nodes and has more resources and power. The degree centrality of the network is divided into outdegree and indegree, which indicate the ability of that node to issue and receive ties, respectively; that is, the ability to export and import tourism services trade between China–ASEAN countries. The formulas for outdegree and indegree are

$$C_{0,i}={\displaystyle \sum }_{j=1,j\ne i}^N{l}_{ij}/\left(N-1\right)$$

$$C_{I,i}={\displaystyle \sum }_{j=1,j\ne i}^N{l}_{ji}/\left(N-1\right)$$

where $C_{0,i}$ and $C_{I,i}$ denote the outdegree and indegree of node i, respectively, and $l_{ij}$ is the connection strength between node i(j) and node j(i).

Betweenness centrality measures the extent to which a node is located in the middle of other nodes and reflects the ability of the node country to control resources in the network. The higher the value of betweenness centrality, the more centrally located the country is and the more structural advantage it has. The formula for betweenness centrality is

$$C_B=[{{\displaystyle \sum }}_{j<k}{a}_{jk}(c_j)/a_{jk}]/(n-1)(n-2),$$

where $a_{ij}$($c_j$) represents the number of shortcuts between two nodes, and $a_{jk}$ represents the number of shortcuts between country $c_j$ and country $c_i$.

Closeness centrality is divided into outgoing and incoming closeness centrality, which reflect the degree to which a node is not controlled by other nodes when issuing and receiving connections. $C{C}_{0,i}$ and $C{C}_{I,i}$ denote outgoing and incoming closeness centrality, and $a_{ij}$ $a_{ji}$ denotes the average length of the shortest path of node i(j) to node j(i). The formulas for outgoing and incoming closeness centrality are

$$C{C}_{0,i}=\left(N-1\right)/{\displaystyle \sum }_{j=1,j\ne i}^N{a}_{ij}$$

$$C{C}_{I,i}=\left(N-1\right)/{\displaystyle \sum }_{j=1,j\ne i}^N{a}_{ij}$$

2.4. QAP Analysis

The QAP analysis, a method of measuring relations between relations, is a non-parametric estimation method that includes QAP correlation analysis and QAP regression analysis [24]. QAP correlation analysis focuses mainly on whether the relations between matrices are significantly correlated. QAP regression analysis examines the relationship of the relations between multiple matrices and one relationship matrix by converting the matrix into a long vector for multiple regression analysis. The explanatory variables and the explained variables contain relational data, which usually have autocorrelation. Simple regressions may cause problems, such as multicollinearity, whereas network analysis methods can effectively avoid these problems. QAP regression analysis, as a non-parametric method, does not require the independent variables to be independent, is more robust in dealing with relational data, and can effectively test the relationship between relational data. Therefore, we believe that applying the QAP method in this study will offer better results than the OLS method.

3. Results

3.1. Overall Network Structure and Centrality Measurements

3.1.1. Measurement Methodology

The network structure characteristics of international trade can be measured in two ways: weights and binary networks. The latter method sets a cut-off value to determine whether a trade relationship exists, and when the bilateral trade flow is greater than this cut-off value, the relationship is established and assigned a value of 1; otherwise, its value is 0.

In this study, we use the binary network method with set cut-off values to measure the structural characteristics of the China–ASEAN tourism services trade network. The measurement process is as follows. First, we represented the bilateral trade in services relationship between China–ASEAN countries by trade matrix M. The element $m_{ij}$ in row i and column j of matrix M is the trade volume of tourism services exported from country i to country j; that is, the number of inbound tourists multiplied by the per-capita tourism expenditure of that country on arrival. Next, we obtained the mean of all the values in matrix M, which we calculated to be USD 75,202,876.12. Finally, we set USD 75,200,000 as the cut-off value for binarization. Countries above or equal to the cut-off value were given a value of 1, while countries below the cut-off value were given a value of 0. The result is matrix G, which is a binary network representing the tourism services trade between China and ASEAN countries.

3.1.2. Visualization of the Network Structure

Based on matrix G obtained above, we used NetDraw, a program integrated into Ucinet 6.0, to visualize the spatial structures of the China–ASEAN tourism services trade network in 2015 and 2018, as illustrated in Figure 1 and Figure 2, respectively.

Figure 1 and Figure 2 reflect the interconnected China–ASEAN tourism services trade network and depict the environment and position of each node country in the whole network. In 2015, four countries—Thailand, China, Singapore, and Vietnam—were at the center of the network, more closely connected with other countries, and with a greater radiating effect and influence on other countries. In 2018, Malaysia and the Philippines entered the center of the network; six countries—China, Thailand, Malaysia, Singapore, Vietnam, and the Philippines—were at the center of the network, more closely connected with other countries, and with a greater radiating effect and influence on other countries. The results show that the Malaysia’s tourism industry grew rapidly in the three years, with Singapore, Indonesia, China, and Thailand being the main tourism sources for Malaysia. In that year, Malaysia’s tourism revenue accounted for 15.2% of its GDP and 23.5% of all jobs. According to the Philippines Tourism Infrastructure and Enterprise Zone Authority, international tourist arrivals reached 7.12 million in 2018, a 7.6% increase from 2017 and the highest ever recorded in the Philippines. Data from the WTO shows that tourism in the Philippines grew faster than the average for Southeast Asian countries (6%) and that tourism became one of the country’s pillar industries. Thus, our results corroborate with the actual tourism flows published by these countries.

3.1.3. Measurement Results

Table 1. Overall China–ASEAN tourism services trade network summary for 2015 and 2018.

Year	Web Density	Inter-Nodal Number of Connections	Out-Centralization	In-Centralization	Web Diameter	Shortest Average Rails
2015	0.264	29	0.370	0.260	4	1.758
2018	0.564	62	0.370	0.370	3	1.400

shows the overall characteristics of the China–ASEAN tourism services trade network in 2015 and 2018 in terms of network density, the number of links between nodes, centrality, network paths, and average shortest path. The closer the network density value is to 1, the greater the density of the network and the greater the closeness of trade ties among countries in the network. Conversely, the closer the network density value is to 0, the sparser the network. Based on the results, the network density value of the China–ASEAN tourism services trade network in 2015 was 0.264. Its degree of tourism services trade connection between network node countries was relatively low and not tightly connected, especially for Brunei, Laos, and Cambodia. The network density value of the China–ASEAN tourism services trade network in 2018 was 0.564. Its degree of tourism services trade network connection improved. Among the bilateral tourism services trade links between countries, half of the countries established mutual links. The overall network structure was relatively close, 30% greater than its network density in 2015. With stronger tourism exchange between China and ASEAN countries, there was a gradual development of beneficial two-way exchanges, complementary advantages, and mutual benefits, improving the overall degree of trade networking between China and ASEAN. Centrality indicates the degree of network nodes converging to the center. The closer the value is to 1, the greater the centrality and the closer the overall structure is to a star. In 2015, the outgoing centrality of the China–ASEAN tourism services trade network was 0.370, while the incoming centrality was 0.260, which is at a medium to low level. The average shortest path was about 4, which indicates that, on average, indirect trade relations needed to be established between each pair of trading countries through two other countries. In 2018, the outgoing centrality of China–ASEAN tourism services trade network was 0.370, while its incoming centrality was 0.370, which is at a medium level. The average shortest path was about 3, indicating that, on average, indirect trade relations needed to be established between each pair of trading countries through one other country. The cooperation path between each pair of trading countries shortened, and the efficiency of cooperation improved. Overall, the network density of tourism services trade links between China and ASEAN increased in 2018 compared to 2015, and the degree of networking gradually deepened.

3.2. Network Centrality Analysis

3.2.1. Overall Centrality

Degree centrality reflects how much a node is connected to other nodes and measures the ability of a country to exchange goods or services with other countries. In this study, degree centrality mainly measures which countries are central to bilateral tourism services trade and the degree of the trade links with other countries. Closeness centrality, in contrast, reflects the smoothness of the flow of goods and services between countries. Betweenness centrality reflects the degree of control and dependence of a country on other countries in trade relations. We summarize the measurement results of China–ASEAN tourism services trade centrality in

Table 2. Individual characteristics of the China–ASEAN tourism services trade network for 2015 and 2018.

Out Degree Centrality			In Degree Centrality			Betweenness			Out Closeness Centrality			In Closeness Centrality
Ranking	Country	Numerical Value	Ranking	Country	Numerical Value	Ranking	Country	Numerical Value	Ranking	Country	Numerical Value	Ranking	Country	Numerical Value
2015
1	CHN	6.000	1	CHN	5.000	1	CHN	19.670	1	CHN	0.435	1	CHN	0.476
1	MYS	6.000	1	THA	5.000	2	MYS	11.167	2	MYS	0.435	2	THA	0.455
2	SGP	5.000	2	VNM	4.000	3	THA	8.500	3	SGP	0.417	2	PHL	0.455
3	THA	4.000	2	SGP	4.000	4	VNM	8.000	4	THA	0.400	3	VNM	0.435
4	LAO	2.000	2	MYS	4.000	5	SGP	3.167	4	MMR	0.400	3	SGP	0.435
4	MMR	2.000	3	PHL	3.000	6	PHL	0.000	5	LAO	0.370	3	MYS	0.435
4	IDN	2.000	3	IDN	3.000	6	LAO	0.000	6	IDN	0.357	4	IDN	0.400
5	VNM	1.000	4	LAO	2.000	6	MMR	0.000	7	VNM	0.345	5	LAO	0.370
5	KHM	1.000	5	BRN	1.000	6	IDN	0.000	8	KHM	0.313	5	BRN	0.370
6	PHL	0.000	6	MMR	0.000	6	KHM	0.000	9	PHL	0.200	6	MMR	0.200
6	BRN	0.000	6	KHM	0.000	6	BRN	0.000	9	BRN	0.200	6	KHM	0.200
2018
ranking	country	numerical value	ranking	country	numerical value	ranking	country	numerical value	ranking	country	numerical value	ranking	country	numerical value
1	THA	9.000	1	CHN	9.000	1	THA	12.567	1	THA	0.909	1	CHN	0.769
1	SGP	9.000	1	THA	9.000	2	CHN	10.233	1	SGP	0.909	1	THA	0.769
2	CHN	8.000	2	VNM	7.000	3	MYS	6.567	2	CHN	0.833	2	VNM	0.667
2	MYS	8.000	2	MYS	7.000	4	SGP	6.067	2	MYS	0.833	2	MYS	0.667
3	VNM	7.000	3	SGP	6.000	5	VNM	3.667	3	VNM	0.769	3	SGP	0.625
4	IDN	6.000	3	PHL	6.000	6	KHM	0.500	4	IDN	0.714	3	PHL	0.625
5	PHL	5.000	4	IDN	5.000	7	IDN	0.200	5	PHL	0.667	4	IDN	0.588
5	KHM	5.000	5	KHM	4.000	7	PHL	0.200	5	KHM	0.667	5	LAO	0.556
6	LAO	3.000	6	LAO	3.000	8	LAO	0.000	5	LAO	0.556	5	KHM	0.556
7	MMR	2.000	6	MMR	3.000	8	MMR	0.000	7	MMR	0.526	6	MMR	0.526
8	BRN	0.000	7	BRN	2.000	8	BRN	0.000	8	BRN	0.250	7	BRN	0.500

Note: The names of the countries are abbreviations in the above table. CHN: the People’s Republic of China; MYS: Malaysia; SGP: Republic of Singapore; THA: The Kingdom of Thailand; LAO: The Lao People’s Democratic Republic; Laos; MMR: The Republic of the Union of Myanmar; IDN: Republic of Indonesia; VNM: Socialist Republic of Vietnam; KHM: The Kingdom of Cambodia; PHL: Republic of the Philippines; BRN: Negara Brunei Darussalam.

.

Table 2 presents several results: (1) Comparing the rankings of each centrality indicator in 2015 and 2018, China and Thailand have a high degree of centrality and are at the core of network trade. Malaysia, Singapore, and Vietnam also have a high degree of centrality and are large players in tourism services trade. Laos, Myanmar, and Brunei are small players in tourism services trade. (2) Thailand, China, and Malaysia have high betweenness centrality and act as distribution centers for tourism goods and services, indicating that these three countries connect large and small trading countries in the network. In 2018, China was ranked in the top two of the trade network, especially in the indegree centrality and incoming closeness centrality rankings, indicating that China is in a central position in the China–ASEAN tourism services trade network and is the central node of the trade network.

3.2.2. Analysis of the Centrality Measures for Different Countries

(1)

Most centrally located countries in the network

As Table 2 shows, China and Thailand are the most centrally located countries of the network. Comparing the change in China’s position in the network between 2015 and 2018 China’s betweenness centrality and outgoing closeness centrality dropped to second place in 2018. Since the financial crisis in 2008, China’s inbound tourism market entered a state of turmoil. In addition to the slowdown in the number of tourists arriving in China, tourism consumption income declined significantly compared with previous years, possibly because few countries have visa waiver programs with China. Another possibility lies in the low number of destinations in or flights to China. According to a questionnaire survey conducted by the China Tourism Academy (Data Center of Ministry of Culture and Tourism) of inbound tourists, tourists rated “whether the price of tourism is reasonable” and “whether the quality of tourism matches the price” with scores close to 8 out of 10, indicating that inbound tourists generally believed that the price of travel to China was reasonable and cost-effective and that inbound tourists had a high overall satisfaction level. Although inbound tourists’ consumption evaluations were generally positive, their consumption level was still low. Hence, there was a shortage of tourism consumption and currently only a single product offering for inbound tourists, mainly tourist attractions. Similarly, tourist supply cannot meet the consumption needs of inbound tourists, which may cause a continuous decline in China’s inbound tourism consumption market.

The rise of Thailand to first place in 2018 in terms of betweenness centrality and outgoing closeness centrality indicates that Thailand plays an important role in the China–ASEAN tourism services trade network. Thailand has not only come out on top in the traditional tourism market for sightseeing purposes but is also a leader in new types of tourism, especially in the medical tourism industry, in which it is one of the most developed countries in the world. Compared with other ASEAN countries, Thailand’s tourism service providers are more professional and provide better travel experiences for tourists, which is why tourists from many countries prefer traveling to Thailand [25]. Additionally, the Thai government has adopted several encouraging and supportive policies and measures, such as developing tourism resources and infrastructure, simplifying entry procedures, relaxing restrictions on tourists’ length of stay, investing in overseas promotion, striving to improve the quality of tourism services, and actively participating in the ASEAN regional tourism integration. These efforts have made Thailand one of the world’s most advanced countries in the tourism industry.

(2)

Sub-central countries in the network

As Table 2 shows, Malaysia, Singapore, Vietnam, and the Philippines are sub-central countries in the network. Malaysia maintained its ranking in the top three for each centrality, indicating that Malaysia’s participation in the tourism services trade network is increasing, as it becomes a bridge and a middleman, just as China and Thailand are. Malaysia’s tourism industry grew rapidly from 2015 to 2018, with the country investing heavily in image promotion, product development, and international cooperation. The tourism industry has become an important pillar industry of the Malaysian national economy [26]. The Malaysian government cultivated tourism as a major driver of the Malaysian economy by introducing guidelines for tourism development plans and related policies; accelerating the development of special types of tourism, such as medical tourism, educational tourism, and ecotourism; adjusting the industrial structure; and realizing economic transformation. Singapore has a high degree of centrality and closeness centrality, indicating that the country has the shortest distance between other countries in the network in terms of export trade. Singapore has the most efficient tourism services trade. It is strategically located in the Straits of Malacca and traditionally has served as a regional service center. Its government sees tourism development as very important. Singapore insists on developing and implementing tourism plans tailored to local conditions, mainly focusing on developing city tourism and meetings, incentives, conferences, and exhibitions (MICE), along with various other types of tourism such as shopping, medical, cruises, and theme parks. Singapore also helps tourism companies provide quality tourism services and comprehensive advertising. It set up a USD 2 billion tourism development fund to support the development of new attractions, improve tourism infrastructure, and enhance the capabilities of its tourism industry to support tourism development. Vietnam has a high level of incoming closeness centrality, indicating a relatively high level of independence in importing tourism services. Vietnam implemented a visa-free policy for tourism in 2016, catered especially for Southeast Asian tourists to enter Vietnam without a visa. The Vietnamese government has taken the initiative to strengthen tourism destination management, ensure public order and food hygiene safety, and deal with zero-dollar tourists. It has also implemented a special policy to attract domestic and international companies to invest in developing tourist attractions and building high-class hotels and resorts and focused on developing high-class and high-value tourism products to promote tourism development. The Philippines has a high degree of indegree centrality and incoming closeness centrality, with the other indicators ranking relatively low. The country has a high level of independence in importing tourism services trade. The Philippines is an export-oriented economy with a high dependence on external markets. Its tertiary sector holds a prominent position in the national economy, with the services sector accounting for over 59% of Manila’s economic system. Although the Philippines moved to the network’s sub-center group of countries in 2018, its tourism development is relatively behind the countries mentioned above, as it has yet to fully develop its tourism potential.

(3)

Countries at the periphery of the network

As Table 2 reports, Indonesia, Cambodia, Laos, Myanmar, and Brunei are at the periphery of the network. Indonesia ranks relatively low on various network indicators and is at the periphery of the tourism services trade network. As the largest archipelagic country in the world, Indonesia is purportedly a favorite tourist destination [27]. However, Indonesia’s tourism industry is underdeveloped compared with its neighbors, Thailand and Singapore. There is still a gap compared with Malaysia, mainly due to poor transportation and support facilities and insufficient and low-quality human resources. Compared with other ASEAN countries, Cambodia’s tourism resource endowment is outstanding, but its infrastructure is relatively poor, security problems are prominent, and the overall level of tourism is low. Laos’ economy has been underdeveloped, with one-third of its nationals living below the world poverty line of USD 1.25 per day. Poor infrastructure and inadequate transportation services have hindered tourism development. In addition, Laos’ tourism industry started late, its overall level of service quality is low, and tourism cost is relatively high, all of which has caused a decline in the number of inbound tourists annually from 2015 to 2018. Myanmar’s tourism industry is experiencing slow development due to ethnic conflicts and an unstable regime, resulting in underdeveloped infrastructure such as transportation, communication, and electricity. Other factors contributing to the slow tourism development include a shortage of tourist reception facilities, such as guest rooms, and safety factors such as drugs and AIDS. Brunei, in contrast, faces difficulties such as inadequate tourism infrastructure, a shortage of skilled labor in the tourism industry, unattractive tourism resources, and fierce competition from neighboring countries. These factors have hindered its tourism development due to its small size, small population, and tourism capacity constraints.

Overall, China, Thailand, Singapore, and Malaysia rank high in terms of degree, betweenness, and closeness centrality, indicating that these countries are centrally or sub-centrally located in the network. These findings suggest that these countries have a mature tourism industry, strong tourism appeal, a more active outbound and inbound tourism market, meet the increasingly diverse needs of tourists, and have more tourism services trade links with other countries. Cambodia, Laos, Myanmar, and Brunei have the lowest degree, betweenness, and closeness centrality. They are located at the periphery of the network, with more room for development.

3.3. Analysis of Factors Affecting the China–ASEAN Tourism Services Trade Network

3.3.1. Indicator Selection and Model Construction

Indicator Selection

Although level of economic development, tariffs, and exchange rates affect changes in trade, they also have some impact on the trade network. We analyzed the impact of variables, such as geographical distance, economic distance, institutional distance, and cultural differences, on the China–ASEAN tourism services trade network.

(1)

Geographic distance (bod): The distance between countries is the main factor affecting the scale of trade in services. The closer the distance between two countries, the lower the transportation cost of trade and the greater the possibility of trade. We constructed bod as follows: for two countries bordering each other, the corresponding matrix element takes the value of 1; otherwise, it takes the value of 0, forming a two-valued symmetric matrix.

(2)

Economic distance (gdp): The impact of the economic development level on trade relations differed, with countries that tended to have similar levels of economic development trading more often. We obtained data from the GDP per capita of each country in the World Development Indicators (WDI) database. We divided our data into pairs to form a variable matrix.

(3)

Institutional distance (ins): Institutional differences at the national level generated institutional distance. The greater the institutional distance between two countries, the less institutional similarity between them and the more obstacles and frictions there were in trade exchanges. We used the Worldwide Governance Indicators (WGI) to measure institutional distance [28]. It contains six indicators: voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption. We measured institutional distance as the absolute value of the difference between the composite scores of two countries. The greater the institutional distance, the smaller the institutional similarity between the two countries, and the greater the number of obstacles and frictions in the mutual trade of tourism services. These data were purchased from the Home of Economics and Management.

(4)

Cultural differences (cul): Cultural awareness is also a major factor influencing inter-country trade. Usually, countries with similar cultural backgrounds belong to the same sector and influence the formation of inter-country trade networks. We constructed cul as follows: for countries with the same official language, the corresponding matrix element took the value of 1; otherwise, it took the value of 0, forming a binary symmetric matrix.

Model Construction

$$\mathrm{W}=f(bod,gdp,ins,cul)$$

(1)

The explanatory variables in Equation (1) contain data of different magnitudes, so we used the QAP method to conduct the test, which included the two steps of QAP correlation analysis and QAP regression analysis. First, we determined the correlation between the dependent and independent variables through a QAP correlation analysis and removed insignificant independent variables. We then performed the regression analysis using the QAP method on the dependent and independent variables to observe the correlation coefficients and interpret them.

3.3.2. Measuring Influencing Factors

QPA Correlation Analysis

We conducted the QAP correlation analysis in Equation (1) to test the correlation between the China–ASEAN tourism services trade network matrix and the influencing factors using 5000 random permutations. The results are shown in

Table 3. QAP correlation analysis of the tourism services trade network and the influencing factors.

Variable Name	Actual Correlation Coefficient	Significance Level	Mean Value of Correlation Coefficient	(Statistics) Standard Deviation	Minimum Value	Maximum Value	p ≥ 0	p ≤ 0
bod	0.346	0.012	0.004	0.142	−0.454	0.472	0.012	0.996
gdp	0.496	0.009	0.007	0.245	−0.731	0.660	0.009	0.991
ins	0.173	0.221	−0.001	0.291	−0.571	0.483	0.221	0.779
cul	0.190	0.150	0.003	0.143	−0.398	0.308	0.150	0.948

. The significance levels in the table reflect the significance levels of the actual correlation coefficients. The minimum and maximum values pertain to the minimum and maximum values of the correlation coefficients obtained from the random permutation calculation results. p ≥ 0 and p ≤ 0 represent the probability that the correlation coefficients obtained from the random permutations are greater or less than the actual values, respectively.

As Table 3 shows, bod and gdp are significant at the 5% level, while ins and cul are not significant, indicating that both the proximity of two countries and GDP per capita significantly affect the composition of the China–ASEAN tourism services trade network. The correlation coefficient between network matrix G and bod is 0.346, indicating that if more countries are adjacent to a particular country, the more favorable it is to carry out inbound and outbound tourism, which, in turn, positively affects the flow of tourism services trade. The correlation coefficient between network matrix G and gdp is 0.496, which indicates that the higher the GDP per capita of the country, the more favorable it is to developing tourism services trade. Institutional distance and cultural differences have no significant impact on tourism services trade between China–ASEAN countries. One possible reason is that institutional and cultural differences do not play a decisive role in tourism services trade and are not important factors affecting its development. Different political, economic, legal, and social institutions in different countries also are not influential factors for tourism. Cross-border movement of tourists can occur as long as normal and positive national relations are established between such countries. According to tourism psychology, the pursuit of cultural identity and the search for cultural differences is one of the main incentives for tourists to travel [29]; thus, it is not similarity of cultural backgrounds that facilitates tourism activities.

QAP Regression Analysis

We conducted the regression analysis of the factors influencing the structure of the China–ASEAN tourism services trade network using the QAP method. The adjusted R2 calculated after 2000 random permutations was 0.281, which shows that the two matrix variables can explain 28.1% of the changes in the structure of the China–ASEAN tourism services trade network. It also passed the 10% significance level test. The results are shown in

Table 4. QAP regression analysis of the tourism services trade network and its influencing factors.

Variable	Unstandardized Regression Coefficients	Standardized Regression Coefficients	Probability of Significance	p ≥ 0	p ≤ 0
Intercept	0.580	0.000	0.000	0.000	0.000
Bod	0.261	0.256	0.075	0.925	0.152
gdp	0.000	0.427	0.054	0.054	0.947

.

The standardized regression coefficient of bod is 0.256, as shown in Table 4, passing the 10% significance test. Thus, when other influencing factors remain unchanged, the geographical location between countries significantly affects tourism services trade between China and ASEAN. The standardized regression coefficient of gdp is 0.427, which passed the 10% significance test, indicating that if other influencing factors remain unchanged, then economic level has a greater positive impact on the tourism services trade network. The higher the GDP per capita, the stronger the tourism services trade linkages between countries.

4. Discussion

Tourism factor flows such as tourist flow, information flow and capital flow influence and interact with each other, and the spatial relationship formed by them constitutes the tourism economic flow network, which affects the coordinated development of tourism economic flow. The research of Chinese and foreign scholars mainly focuses on analyzing the network structure characteristics among tourism destinations, and then analyzing the temporal, spatial, and demographic characteristics of tourism flows [30,31,32]. Scholars have applied the social network analysis method to the study of tourism flows, studying tourism destinations at different administrative levels such as national, provincial, and municipal levels, as well as different types of tourism flows such as package tours and self-guided tours. The construction methods and evaluation index system of tourism flow networks are improving year by year [33,34,35,36,37,38,39,40,41,42]. China and ASEAN countries mainly source countries and tourist destinations to each other. It is of great practical significance to study the distribution and rules of tourist flows between China and ASEAN countries to promote the economic growth of tourism in the China–Asean region. At present, as network analysis methods such as complex network theory and social network theory have become popular in international trade research; these methods are increasingly used to analyze the local trade relations among multiple countries. Chinese and foreign scholars have conducted empirical analysis on the structural characteristics of international trade networks based on the import and export data among countries. At present, existing related studies pay more attention to the trade pattern itself, but less to the factors affecting the trade network pattern. The degree of openness of a country and the overall pattern of trade network are both important factors that affect the benefits of foreign trade. At present, the research of the national trade network mainly focuses on commodity trade and lacks the analysis of the service trade network. As the most dynamic region in the world, there are few studies on the trade network of China and ASEAN region. If the social network analysis method is applied to tourism, and with the integrity of network structure and the relationship between the network nodes in the field of trade in tourism services trade as the research object, we will have an overall grasp of the features of the regional tourism service trade network pattern, providing theoretical research and empirical analysis of coordinated development for tourism service trade in countries around the region. Based on the method of social network analysis, this paper analyzes the structure characteristics and influencing factors of the China–ASEAN tourism service trade network. It draws the following conclusions:

First, regarding the overall characteristics of the network, the network density of the China–ASEAN tourism services trade network in 2018 was 0.564, which is at a medium level. This indicates that in the bilateral tourism services trade, ties between China and ASEAN, half of the countries have established mutual relations. The network is relatively tightly connected, but there is still more room for improvement.

Second, the results of the network centrality analysis show that four countries, namely, China, Thailand, Singapore, and Malaysia, are at the center of the network and hold important nodal core positions and control ability. They are the intermediaries and bridges of the whole network. Conversely, countries such as Cambodia, Laos, Myanmar, and Brunei are less centrally located and are in disadvantaged positions, as they are being controlled, playing the role of followers in the trade network. The network shows a significant core-periphery structure, with China, Thailand, Vietnam, Singapore, Malaysia, and the Philippines at the network’s core. At the same time, Laos, Myanmar, Indonesia, Cambodia, and Brunei are at the network’s periphery, vulnerable to control by other countries and lacking opportunities to participate in tourism services trade.

Third, geographical proximity and GDP per capita are the main factors affecting the structure of the China–ASEAN tourism services trade network. The QAP correlation analysis results show that neighboring countries and countries with high GDP per capita have a closer tourism services trade relationship. The QAP regression analysis indicates that the factors of proximity and GDP per capita can explain 28.1% of the China–ASEAN tourism services trade relationship, which is a high explanatory power for factors influencing the trade network. Geographical proximity strongly influences the overall network; the more geographically close countries are, the more likely tourism services trade occurs. It also has a significant positive correlation with GDP per capita, indicating that countries with higher GDP per capita are more likely to have tourism services trade. Apart from geographical distance, the current development of China–ASEAN tourism services trade is also limited by countries’ economic distance, a similarly important determining factor.

The four countries, i.e., China, Thailand, Singapore, and Malaysia, have a higher tourism service trade development level. They play central roles in the whole trade network, while Laos, Myanmar, Indonesia, Cambodia, Brunei, and other countries have a relatively low level of tourism service trade. They have less tourism services trade with other countries and are the edge position in the trade networks. The main reason for this result is that, on the one hand, there are differences in economic level, tourism resource endowment, geographical location and cultural background among countries. These basic conditions affect the level of domestic tourism service trade to a certain extent. On the other hand, China and ASEAN countries began to deepen their cooperation and exchanges in the past decade due to the establishment of the ASEAN Free Trade Area. It was a relatively late start. In the whole process, there are both opportunities and challenges, which require mutual understanding and coordination among countries. Recently, due to the slowdown of global economic growth and sluggish international trade, tourism service trade cooperation between China and ASEAN countries has been affected to varying degrees. It will take some time for tourism service trade among countries to enter into healthy competition and collaborative cooperation.

5. Conclusions

Through the above research, we think there are conclusions as well as a theoretical contribution and practical value as follows. First, the overall trade in tourism services between China and ASEAN shows a positive trend of development, but there are obvious gaps between countries. Therefore, China, Thailand, Singapore and Malaysia, as the core countries, should improve the cooperation ability among countries, have the responsibility to play a pivotal role, actively promote the cooperation of tourism service trade between China and ASEAN countries, and further enhance the connectivity of the regional tourism service trade network. Countries lagging behind in tourism service trade should be guided to give play to their own characteristic tourism resources and increase their participation in the trade network.

Second, the results show that the proximity of countries and GDP per capita are the main influencing factors in the evolution of the China–ASEAN tourism service trade network. Therefore, it is suggested to further strengthen the tourism service trade interaction between China and ASEAN countries. Countries can shorten the geographical distance by constructing inter-country transportation systems, such as opening direct flights. In addition to increasing the number of regular international routes, countries can also promote cooperation among travel agencies and airline charter businesses while simultaneously constructing high-speed rail between countries as early as possible to reduce the impact of geographical distance. For economic distance, the export value of tourism services trade will rise with the growth of GDP per capita. China, Thailand, Singapore, Malaysia, and other countries with more developed tourism industries should play a leading role in promoting tourism trade cooperation in the free trade area. These countries should also strengthen economic and trade cooperation between their countries and countries with less economic development, take the initiative to connect with neighboring countries at the periphery of the network for further tourism interaction, promote a more closely linked tourism economy, realize the synergistic development of countries in the core and countries at the periphery, and construct a network space suitable for cooperation.

Third, the theoretical significance of this study lies in using a cutting-edge social network analysis method to analyze and visualize the characteristics of the China–ASEAN tourism service trade network, and then using a QAP (Quadratic Assignment Procedure) method. To examine the influencing factors of the tourism service trade network, and to explore the characteristics and influencing factors of the regional pattern of the China–ASEAN tourism service trade network. Most of the existing literature studies international tourism flow based on the tourism demand framework, mainly measuring and comparing the competitiveness of tourism destinations, and most of the research methods choose the gravity model, impulse response, cointegration analysis, Granger causality test, etc. The practical significance of this study is to study and solve the problems in the development of the China–ASEAN tourism and service trade, so as to provide a reference for the coordinated development of tourism and service trade between China and ASEAN countries, and for the complementarity and cross-border cooperation of tourism and service trade between China and ASEAN countries. However, due to the complex countries within the ASEAN Free Trade Area, tourism service trade involves the various industries and many enterprises, parts of the world tourism organization of data are missing or delayed, there are statistical problems and issues with the use of data with each country’s own master data, there is a certain deviation for more in-depth analysis of tourism trade in services network set-up obstacles. Moreover, although the social network analysis method can better analyze the various relationship structures between countries, this relationship is only based on some quantitative values, and the actual situation is far more complex, so there will be some deviation in the measurement results. At the same time, the tourism service trade network of China and ASEAN countries is regarded as a closed group in this paper. However, these countries will also be in contact with other countries outside the group, which can be added into the subsequent analysis for auxiliary analysis.