Determinant of the tourism economy in Chinese cities: from an urban centrality perspective

Abstract

This paper aims to analyze and explain how urban centrality influences tourism development in a city. Based on the panel data of 295 prefecture-level cities in China from 2005 to 2018, the paper develops explanatory mechanisms and discusses the influence theoretically and empirically. To advance the analysis, this paper constructs a new index for urban centrality. Our empirical findings are as follows: (1) urban centrality promotes tourism income significantly. (2) Mechanism analysis illustrates that urban centrality fosters tourism through the agglomeration and industrial structure effects. (3) Heterogenous analysis suggests that the influence of urban centrality on tourism varies with city sizes and locations.

Keywords

urban centrality tourism economy agglomeration effect industrial structure effect network analysis

Introduction

The spatial pattern of China’s economic development has changed dramatically in recent years. Central cities and urban agglomerations have become the primary spatial forms that possess resources and production factors and are the main engines for general economic growth (You and Chen, 2021). Traditional analysis based on the geographic location of a city can no longer meet the requirement of the current economic development (Pereira et al., 2013). As a result, the study of the centrality of a city in the urban agglomeration or urban network has gained important theoretical and practical significance. In addition, the service industry is becoming the dominant form of China’s economy, where tourism is the most growing sector of the service industry. Tourism’s development has contributed to higher economic output (Zhang et al., 2021). The tourism economy is considered a “Cgreen” driver of regional economic growth and industrial upgrading. Considering that the hierarchical position of a city in an urban network is closely related to its accessibility to various resources needed for the city’s economy, it is reasonable to assume that the change in a city’s urban centrality might influence tourism development in the city. This paper investigates such influence and the main mechanisms through which urban centrality affect the development of the regional tourism economy.

Literature has described how urban systems or networks form and organize (Neal, 2013). The “Globalization and World Cities Study Group and Network” (GaWC), led by Taylor (2004), is dedicated to exploring the formation and evolution mechanism of world city networks and the interactive relationship between network structure and economic globalization. The GaWC team regards cities as nodes in the network and believes that the city’s role and status depend on the city node’s interaction with other nodes in the network (Derudder and Taylor, 2021). The closer a city interacts with other cities in the network, the stronger the “space-time compression” effect it produces. This effect can be reflected in the changes in regional spatial accessibility. It can intensify the agglomeration effect, which causes the resources to gather in a particular area and accelerate the area’s economic growth. It leads finally to a “central area” with a higher income level, while the relatively backward surrounding areas are called “peripheral areas”. This is precisely the “central area and peripheral area” argument in the unbalanced growth theory of regional development (Hirschman, 1958). The urban network argument has considered the dynamic interconnection between cities and thus complemented the insufficiency of previous research treating the urban system as a static and relatively independent system (Taylor et al., 2014). With the widespread application of multi-source data, the urban network gradually becomes a new paradigm for global urban research. As a result, urban centrality has become an important topic in the study of urban development and urban system (Pereira et al., 2013).

Scholars have investigated urban centrality from various perspectives. Some researchers have focused on developing the concept of urban centrality and proposed using different research methods and theoretical perspectives to measure it. For example, Pereira et al. (2013) introduced an extension to the spatial separation index to construct a new measure of urban centrality. Curado et al. (2020) analyzed and compared several measures of centrality applied to urban networks and presented a new measure (called Adapted PageRank Algorithm) with real data on the city of Rome. Other studies related the concept of urban centrality or network centrality to issues in transportation infrastructure (Ma et al., 2019), energy consumption and pollution (Rüttenauer, 2019), industrial structure, etc.

Although some researchers have related the network or centrality concept to the tourism industry, they focused mostly on the network of the tourism economy (e.g. Gan et al., 2021) or transportation network (e.g. Chen et al., 2020). Moreover, many studies were only case studies or related to a specific urban destination (e.g. Aranburu et al., 2016; Sugimoto et al., 2019). Even when there were studies involving centrality, the measurement relied on the traditional social network analysis method (e.g., Luo et al., 2020; Ledesma González et al., 2021). Unlike previous literature, this paper links the concept of urban centrality with tourism development and explicitly discusses them in one framework.

Benefiting from its specialties, the tourism industry has a strong comprehensive driving effect on other industries in a city. Tourism is an intangible and unmovable trade product. The development of the tourism industry mainly relies on the flow of people and thus depends on transportation infrastructure. Good tourist attractions need to satisfy a variety of demands, such as hotels, restaurants, leisure centers, and shopping centers in the local area. It means that the development of the tourism industry depends on the local industrial structure. Therefore, the tourism industry is more sensitive to the city’s centrality than other industries.

Based on the above analysis, this paper explores the influence of urban centrality on the tourism industry and investigates the mechanism of this influence. By employing theories from agglomeration economy, network science theory, and urban study, we discuss the agglomeration effect and industrial structure effect as transmission mechanisms. This paper applies data about 295 prefectural administrative cities in China from 2005 to 2018 for empirical analysis, overcoming the problems caused by insufficient samples, and conducts a large number of tests to ensure robustness and reliability. The empirical results show that urban centrality has a significant positive effect on tourism income. The results are important for the sustainable and healthy development of the tourism industry and optimizing the industrial layout of cities and urban agglomerations.

Our study contributes to the agglomeration economy and network science theory, tourism research, and urban study. First, by including the concept of “urban centrality” from network science and urban study, this paper fills a research gap in the study of determinants of tourism development. Although previous tourism literature has adopted a social network perspective in the analysis, most focused on the network structure of the tourism economy but not the position or the status of a city in the urban network. Second, this paper elaborates on two transmission mechanisms through which urban centrality influences tourism: the agglomeration effect and structural effects. It complements the theoretical and empirical discussion about the intersection of agglomeration economy and network theory. Third, this paper constructs the urban centrality index based on Taylor’s GaWC interlocking city network model (2001) and the data from China’s leading producer service company (PS). Considering the spatial location of cities in urban networks and the service-oriented economic environment of modern cities, this approach overcomes the limitation of geographic centrality in traditional network analysis. In this regard, it provides new theoretical insights for network analysis in future research. Network analysis needs to consider the role of non-geographic factors such as spatial and economic structures. Fourthly, the data used in this paper ensures a comprehensive heterogeneity analysis. Considering the imbalanced situation in the economy, institutions, and culture in Chinese regions, the results of this paper can provide reasonable implications for policymakers.

The remainder of the study is arranged as follows: The literature review section presents a literature review. The mechanism analysis section theoretically explores the transmission mechanism of urban centrality to the development of the tourism economy. The methods and data section introduces econometric models and data. The empirical results and analysis section reports the empirical results and additional tests. The last section concludes with theoretical and managerial implications, limitations, and future research suggestions.

Literature review

This study is based on two main strands of literature. The first strand of literature concerns factors affecting the development of the tourism industry. The second lies in the research field of urban centrality and the development of the tourism economy.

Literature on factors affecting the development of the tourism industry

Existing literature has investigated the determinants of tourism’s development from various perspectives, such as culture, language, or other mental and emotional characteristics (e.g., Ritchie & Zins,1978; López et al., 2018). Researchers usually take a tourism destination such as a country or region as a case study. For example, López et al. (2018) found that residents’ perceived benefits significantly affect tourism sustainability in Trujillo, the third-largest city in Peru. Similarly, a study from Sharma and Nayak (2020) confirmed that experience quality is a dominant construct of interest for India’s tourism industry’s success.

Traditional studies have believed that resource endowment in the tourism industry, such as reception facilities, natural and cultural landscape, and personnel qualifications has a significant and direct influence on regional tourism development (Xiang et al., 2012). By analyzing the correlation between tourism competitiveness in Southeast Europe countries and the level of competitiveness of tourism infrastructure, Jovanović and Ivana (2016) pointed out that tourism infrastructure is a significantly important factor in the tourism sector.

Some literature has evaluated the regional differences in tourism development from the perspective of the macro environment. For example, Ramkumba et al. (2012) found that macro policies regarding market guidance, staff training, and financial support stimulate related enterprises or corporations to improve tourism products and thus promote local tourism development. The study by Liu et al. (2018) showed that the evaluation activity performed by the Ministry of Housing and Urban-Rural Development in China had enhanced the role of the National Scenic Spot in the local tourism economy.

In recent years, researchers have begun to study regional tourism development with the application of spatial economy. Factors like spatial dependence and location accessibility resulting from agglomeration significantly affect the regional tourism economy. For example, Yang and Fik (2014) examined and identified the influence of two types of spatial effects, spatial spillover effects and cross-city competition effects, on regional tourism.

Urban centrality and the development of the tourism economy

As mentioned above, the existing literature rarely discussed urban centrality and tourism economy specifically in one explanatory framework. Most relevant studies have used the concept of centrality from the perspective of the transportation network, the network of business or non-business tourism organizations, or the network constructed through tourist behavior and mobility (Stienmetz and Fesenmaier, 2015; Liu et al., 2017). In addition, they have applied social network analysis to discuss tourism-related issues. For example, a study by Wang et al. (2016) used social network analysis to explore the effect of high-speed rail (HSR) on the spatial structure of regional tourist flows. The research pointed out that HSR can strengthen the aggregation effect, and the tourism nodes in the network can take advantage of the endowment of tourism resources, hospitality capacity, tourist transportation network density, etc. In addition, Xie et al. (2021) constructed the spatial structure of the European Union (EU) by the modified gravity model and social network method. They analyzed the effects on the EU tourism economy. One of their main findings was that the improvements in the complete network connectedness and a reduction in graph efficiency can significantly reduce differences in EU tourism economic development levels and improve spatial equity. With the Chinese empirical setting, Kong and Li (2021) proposed that the city’s centrality, intermediary, connectivity, and other network characteristics in the HSR network affect the development of tourism to a different degree. Many scholars have discussed the relationship between spatial network structure and the tourism economy. Some examined the characteristics of the spatial network structure in tourist destinations by adopting the tourism economic gravity model and social network analysis (e.g. Gan et al., 2021). Some tourism literature insisted that tourism and leisure significantly influence urban space when considering tourism’s spatial organization (e.g. Derek, 2018). Conversely, others thought regional spatial structure changes affect the tourism economy (e.g. Fang et al., 2021).

Some researchers have applied networks of the tourism economy to construct indicators of centrality in tourism networks (e.g. Shih, 2006; Zhang et al., 2015a). However, urban centrality differs from traditional network centrality (Neal, 2011). It results from regional economic links in urban networks constituted by productive services networks, population flow networks, and others. The concept of urban centrality is reflected in the difference in a city’s spatial pattern, location advantage, and economic status (Luo et al., 2020; Neal et al., 2020). The variation in the city’s hierarchical structure in the urban network represents unequal city functions. As cities and urban agglomerations play more important roles in the development of the modern economy, the upgrading and improvement of the urban centrality of a city can promote the regional economy as well as the development of the tourism industry.

Altogether, the literature review shows that most previous studies on determinants of the tourism economy have paid attention to the influence of some absolute indicators such as urbanization or integration degree while very few have underlined the impact of changes in the relative position of cities in the urban network. To fill this research gap, this paper explores the influence of a city’s urban centrality on tourism development.

Mechanism analysis

Agglomeration effect

The theory of new economic geography proposed that the agglomeration effect exists with the opening of the high-tech service industry. With the establishment of high-level production service industry networks, improving urban status in areas with high urban centrality will stimulate the accumulation of human and social resources. It ultimately results in an agglomeration economy. A larger local market resulting from an agglomeration economy expands the supply and demand for regional tourism with the improvement of accessibility. Many researchers have studied the effect of agglomerations in tourism, such as cooperation and competition among tourism-related firms or visitor attractions (e.g. Denicolai et al., 2010), labor pool effect (e.g. Kim et al., 2021), and knowledge diffusion and spillover (e.g. Majewska, 2015).

One important feature of an agglomeration economy is the co-existence of cooperation and competition effects (Sosnovskikh, 2020). From the supply perspective, the agglomeration effect facilitates travel and tourism marketing firms to share resources, risks, and costs. In addition, it helps firms gain advantages from the mutual learning and exchange of information. Literature has documented that collaborative behavior via agglomeration economies enhances tourism competence and favors sustainable tourism development (e.g. Denicolai et al., 2010). From the demand side, tourism is very susceptible to concentration. For example, the spatial concentration of tourist attractions can induce tourism flow. The study from Kalnins and Chung (2004) focused on demand-side agglomeration and found that agglomeration increases demand in the US lodging industry. In addition, the competition effect strengthens tourism firms’ need to be innovative constantly to maintain their position in the tourism industry. Moreover, competition effects in tourism promote productivity spillover, which drives, in turn, intense competition because the tourism firms locate in the relatively homogeneous visitor market segment (Kim et al., 2021).

The agglomeration effect influences regional tourism in different ways (Wang et al., 2016). Labor market pooling and knowledge spillover exist in the tourism context (Kim et al., 2021). Benefiting from the agglomeration economy due to high urban centrality, the cost of labor force transfer and movement is greatly reduced, resulting in a more efficient allocation of labor (Lin et al., 2021a). Moreover, a large labor pool makes it much less expensive for tourism companies to search and find workforces, thus improving labor productivity. The matching friction between the labor force and the job position can be reduced, thus improving the matching efficiency (Liu and Wen, 2015).

The movement of the labor force in the tourism context can generate significant knowledge spillover. Unlike other sectors, knowledge in tourism is more tacit, and tourism employees are more likely to share incremental knowledge (Shaw and Williams, 2009). Moreover, products in the tourism industry are mostly similar, and knowledge sharing is more critical than radical knowledge (Zhang et al., 2015b). Knowledge accumulation and sharing results in knowledge spillover occurs along with labor mobility. This process stimulates learning and innovation, ultimately contributing to tourism growth (Kim et al., 2021). In addition, the higher concentration of the tourism industry in a specific area encourages the diffusion of knowledge among enterprises inside the tourism industry, which is conducive to the enterprises’ innovation activities and thus promotes the development of the tourism industry (Boschma & Ter Wal, 2007). Moreover, the attractiveness of tourism products needs to be complemented by other industry sectors, such as the transportation and service sectors. The spillover of complementary knowledge across diverse firms and economic sectors becomes much easier because of a high degree of agglomeration (Kim et al., 2021). The diversification of industries within the region benefits the tourism sector (Marshall, 1890; Jacobs, 1969).

Industrial structure effect

Urban centrality has a positive influence on the urban industrial composition. The continuous development of urban centrality is also the process of larger market formation. Benefiting from larger markets and more supporting infrastructure, areas with high urban centrality are more likely to produce synergy among different industries, possess diversified products, and thus shorten the spatial distance of products in cities. Therefore, such cities are more inclined to produce high-complexity products (Sun et al., 2021). The tourism sector in such cities is more able to add value to both diversified tourism products and tourists’ experiences (Majewska, 2015). Since tourism cannot be developed without the complement of other service industries, changes in tourism products and attractions further promote the adjustment and upgrading of the industrial structure in the whole city.

The industrial re-composition includes two aspects. One is the head effect. Because of its higher position of centrality in the regional network, the city can attract the agglomeration of industries with regional barriers. In addition, the city can meet the development of industries with high demand thresholds due to the shared labor pool, services, and infrastructure. As a result, industrial transformation leads to the dominance of tourism firms which provides high-level tourism brands and image. The other is the long-tail effect. Cities with high urban centrality choose to raise their entry barriers and attract higher-end industries to enter. As a result, it crowds out the lower end industries to regions with lower centrality. The transformation of the regional industrial structure saves time and communication costs for the spatial transfer of production factors required for economic activities in the tourism industry. It also makes the transaction of tourism products and the supply of services more convenient.

Methods and data

To examine the influence of urban centrality on tourism growth, this paper adopts an econometric panel regression method by controlling other factors that may also affect tourism’s development. Considering the potential impact of provincial effect, time trend, and time effect on the model results, this paper uses a fixed-effect model for regression analysis. The empirical model of this paper is as follows:

{T o u r i s m}_{i t} = α + β_{1} U {C e n t r a l i t y}_{i t} + γ X_{i t} + δ_{c} + η_{t} + ε_{i t}

(1)

Among them, ${T o u r i s m}_{i t}$ is the main explained variable representing the tourism income of city $i$ in year $t$ . In addition, we apply the number of tourists ( ${T o u r i s t}_{i t}$ ) instead of tourism income as explained variable for the following robustness test. ${U C e n t r a l i t y}_{i t}$ is the main explanatory variable representing the city centrality index of city $i$ in year $t$ , $X_{i t}$ represents a series of control variables, $δ_{c}$ is the province fixed effect used to control the province-level characteristics that do not change with time, $η_{t}$ denotes the year fixed effect, and $ε_{i t}$ is the error term. Due to the significant difference in the magnitude of various variables, heteroscedasticity impacts estimation results. We take natural logarithms for some variables in empirical tests to avoid it.

Explained variable

As introduced above, this paper uses the tourism income ( ${T o u r i s m}_{i t}$ ) as the main explained variable. Figure 1 reflects the distribution of tourism income in 295 cities in the years 2005, 2009, 2013, and 2018, respectively. The tourism incomes of almost all cities have increased significantly during these years. However, the imbalance among regions has increased as well. A great discrepancy exists in different cities in China.

Figure 1.

City-level distribution of tourism income in China. Figure (a), (b), (c), and (d) are the income distribution in 2005, 2009, 2013, and 2018, respectively.

Explanatory variable: construction of urban centrality index

This paper selects the urban centrality $({U C e n t r a l i t y}_{i t})$ of each city as the main explanatory variable. The construction of the urban centrality index is based on the GaWC interlocking city network model by Taylor (2001). The interlocking city network model can overcome the limitations of using traditional urban geographic attributes to assess centrality (Deruddera et al., 2018). Referring to the calculation method for district core or city from Hennemann and Derudder (2014), we apply an evaluation model of the city network centrality based on China’s leading producer service company (PS). The data comes from the “Qixinbao” Web site (http://www.qixin.com/). Data on this Web site is updated synchronously with the Web site of the State Administration for Industry and Commerce in China. The Web site provides information such as the family tree of a company. Referring to the available lists of Chinese service industry companies, combined with the annual operating revenue of the related companies, we collect the top 100 companies inside China for each sector, including management consulting, accounting, law firms, advertising, banking, securities, insurance, and other industries. In addition, we get the data set of foreign multinational companies in the same sectors with branches in China. Moreover, we exclude companies with less than two subsidiaries.

We assume there are j high-level producer service companies with branches distributed in n cities. The service value $v_{m i}$ reflects the relative importance of the branches located in city m in the branch network of company i, which can form a j×n service value matrix V. Because we have information such as the company’s registered address and industry sector, its investment holding companies, and its organization-level data, we then assign a value of 0, 1, 2, 3, 4, 5 to the different level of branches, including the headquarters, first-level, second-level, and third-level subsidiaries. For example, 5 means the city where the headquarters is located, and 0 means that the company has no branches or offices in this city. The larger service value means that enterprise i is more important in the city m. In this way, the city’s total service value can be used to measure the city’s status or position, expressed by the city’s service value $C_{m}$ (Lin et al., 2021b).

C_{m} = \sum_{i}^{j} v_{m i}

(2)

The same method is used to measure the strength of an enterprise’s service ability, expressed by the enterprise’s service value $F_{i}$ :

F_{i} = \sum_{m}^{n} v_{m i}

(3)

Although the variable $C_{m}$ intuitively reflects the city’s accumulated professional knowledge and service level, it cannot indicate the “power” possessed by the city. That means, it is not enough to use the city’s service value $C_{m}$ alone to measure the status of a city. Therefore, we need the transformation matrix V by enterprises to understand the interaction among cities.

In network analysis, the service value matrix V is usually called a two-mode network (Liu and Derudder, 2012). In the one-mode network, the main body is directly connected, while the two-mode network is connected through two separate node data sets. The service value matrix V is a two-mode network that connects cities and enterprises. There is no direct correlation between nodes of the same type. We can use a “projection function” to map a two-mode network to a one-mode network. The chained city network model is essentially a projection function, which infers the association among cities based on the hierarchical information flow of enterprises co-existing in different cities, and transforms the service value matrix V into the correlation matrix R reflecting interactions among cities.

The key to the chained city network model’s projection function is to define the intercity network’s connectivity. The model believes that the connections between internal branches of a company are greater than those with other companies in the same industry. Moreover, the importance of a branch leads to its more connections to other branches, which eventually produces a multiplicative effect on the relationships among cities. Therefore, based on each pair of cities and companies in matrix V, we can calculate the connectivity of company i between city a and city b.

r_{a b} = \sum_{i = 1}^{j} r_{a b i}

(4)

The total network connectivity of city a is,

{U C e n t r a l i t y}_{a} = \sum_{m = 1}^{n} r_{a m}

(5)

A city’s higher total network connectivity indicates a higher degree of the city’s integration into the entire producer service industry network. In network analysis, “network connectivity” is usually regarded as the centrality of a city. Based on the above method, we get each city’s urban centrality from 2005 to 2018. Figure 2 describes the centrality distribution in the years 2005, 2009, 2013, and 2018, respectively. We find that urban centrality has generally increased during these years. However, the degree of agglomeration in some cities has become higher. As a result, the concentration of urban agglomerations has emerged.

Figure 2.

The urban centrality map of each city in China. Figure (a), (b), (c), and (d) are the distribution in 2005, 2009, 2013, and 2018, respectively.

Control variables

To control other features that may influence the development of the tourism economy, this paper adds eight control variables.

The paper considers the influence of population size and economic development level on the tourism industry. It takes the population density ( ${P d e n s i t y}_{i t}$ ) and per capita GDP ( ${P g d p}_{i t}$ ) of each city as indicators. We also apply the PM2.5 data of each city as a proxy variable $({p m 2.5}_{i t})$ of environmental quality because the urban environment is also an important factor in attracting tourists.

In addition, previous literature has shown that level of transportation infrastructure influences the development of the tourism industry as well (Khadaroo and Seetanah, 2008). We use the highway mileage of each city $({h i g h w a y}_{i t})$ , dummy variable about the existence of HSR station in city $({h s r}_{i t})$ , and dummy variable about the presence of an airport in city $({a i r p o r t}_{i t})$ to reflect the level of transportation infrastructure in each region. The dummy variables ${h s r}_{i t}$ and ${a i r}_{i t}$ are calculated by the city dummy variable (City) × the year dummy variable (Year). For cities with HSR, City equals 1. Otherwise, it is 0. Year is 1 in the opening year and later; otherwise, it is 0. The same rule applies for the variable ${a i r}_{i t}$ .

Moreover, the development of the tourism industry is based on the construction and sufficient supply of tourism-related infrastructure. The adequate supply of tourism infrastructure benefits the free flow of production factors and the reasonable allocation of resources, thereby promoting the agglomeration effects in the tourism economy. In this paper, we use the number of star-rated hotels $({h o t e l}_{i t})$ and the number of 5A-level tourist attraction $({a t t r a c t i o n}_{i t})$ to reflect the level of hospitality facilities in the tourism industry.

Instrumental variable

The development of a city’s tourism industry could also be influenced by some unobservable factors. Missing variables might result in an endogeneity problem. In addition, a causal relationship might exist between urban centrality and tourism development if one questions that areas with a high level of tourism development have a larger flow of people, which may, in turn, lead to an increase in urban centrality. This paper adopts the city’s geographic slope ( $i v {s l o p e}_{i t}$ ) as an instrumental variable (IV) to tackle the endogeneity problem. An ideal instrument needs to satisfy two requirements: relevance and exogeneity. The geographical slope is related to urban centrality. A region with a lower average slope usually has a smoother and more frequent flow of various resources. It is more likely to become an area with higher urban centrality. Moreover, the geographical slope is a natural condition formed in the long-term history of the area and has no direct relationship with the current tourism income.

Mechanism variables

The widely accepted measure of the agglomeration effect is the location quotient (LQ) (Majewska, 2015). In this study, we employ the LQ to measure the degree of specialized agglomeration in a city. The equation is shown as follows:

L Q = \frac{\frac{e_{i m}}{e_{i}}}{\frac{e_{a m}}{e_{a}}}

where $e_{i m}$ is the number of employees of the tertiary (service) industry in city $i$ , $e_{i}$ represents the number of employees of all industries in the city $i$ , $e_{a m}$ is the number of employees of the tertiary industry in all cities, and $e_{a}$ means the number of employees of all industries in all cities. A large location quotient indicates an area’s high level of industrial agglomeration.

In addition, we use the indicator “the ratio of the tertiary industry to GDP” (structure) to reflect the structure effect because industrial structure or composition effect accompanied by the change of the city’s position in the urban network encourages industry structure upgrading from the primary and secondary industry to the tertiary industry, namely service industry.

Data sources and descriptive statistics

The data set for this paper is panel data covering 295 cities from 2005 to 2018. This paper contains two major data sources. The first is the China Premium Database (https://insights.ceicdata.com), the industry’s leading time-series economic database. It includes national macroeconomic data, regional economic data, industry economic data, and thematic data. The second is the China City Statistical Yearbook coedited by NBS. It comprehensively reflects the annual social and economic development of the prefecture-level administrative city in China. In addition, we get the data about the 5A-level tourist attractions from Baidu Baike (https://baike.baidu.com). The slope data is based on SRTM (Shuttle Radar Topography Mission) elevation data extracted from the 2017 administrative division map. The PM2.5 data comes from the China Air Quality Online Monitoring and Analysis Platform (www.aqistudy.cn). The descriptive statistics of variables are in Table 1. With data from all observed years, the scatter plot in Figure 3 intuitively illustrates the correlation relationship between the two main variables.

Table 1.

Descriptive statistics.

	(1)	(2)	(3)	(4)	(5)
Variables	N	Mean	Sd	min	Max
LnTourism	4130	9.072	1.488	2.725	13.23
LnUCentrality	4116	10.79	0.524	5.037	12.12
LnPdensity	4015	5.711	0.949	0.483	7.887
LnPgdp	4002	10.32	0.770	4.595	15.68
lnhighway	4130	9.141	0.722	6.181	11.97
Hsr	4130	0.355	0.479	0	1
Airport	4130	0.432	0.495	0	1
Lnhotel	4130	3.319	0.869	0	6.729
attraction	4130	0.403	0.827	0	7
lnpm2.5	4004	3.652	0.514	1.141	4.702
lnivslope	4130	9.108	1.058	6.567	10.81
lnTourist	4128	9.349	1.223	4.248	12.74
Structure	4130	224.9	14.73	109.3	280.6
lnLQ	3990	−0.468	0.755	−2.037	2.647

Figure 3.

Scatter plot (2005–2018).

Empirical results and analysis

Results of baseline model regression

As stated above, this paper applies a fixed effect model for regression analysis. Table 2 reports the results of the baseline model regression. Column (1) shows the results without consideration of control variables. Columns (2) to (9) are the results after the stepwise addition of control variables. The results in all columns show that the coefficients of the main variable

(\ln U C {e n t r a l i t y}_{i t})

are always significantly positive at the 1% confidence level, whether or not the influence of control variables is considered. It means that a significant positive correlation exists between urban network centrality and tourism income. In addition, all coefficients of the control variables are significant at the 1% confidence level. Therefore, the empirical results are consistent with our expectations. More specifically, estimates on the level of tourism reception facilities (

{\ln h o t e l}_{i t}

;

{a t t r a c t i o n}_{i t}

) are positive and quite significant. It indicates that tourism incomes tend to be higher in cities with high-level tourist attractions and supporting facilities. It is intuitive since travel is a consuming behavior of leisure and enjoyment, and people tend to go to famous tourist attractions. In addition, the significantly positive coefficients of control variables

({r o a d}_{i t}, {h s r}_{i t}, {a i r}_{i t})

indicate that the transportation infrastructure has a positive role in promoting tourism income. It reflects the reality in China that HSR has developed vigorously in recent years, increasing the accessibility and connectivity of various cities, thus making their tourism resources more accessible. Also, the positive and statistically significant coefficients of variables (

{l n P d e n s i t y}_{i t}

;

{l n P g d p}_{i t}

) suggest that cities having more population and higher income promote its tourism industry as well. It means that residents in rich areas are more willing to travel. Developed areas may also attract more tourists to increase tourism revenue because of better infrastructure and cultural environment. In addition, the significant negative coefficient of the variable (lnpm2.5) shows that poor environmental quality has a negative impact on tourism. Generally speaking, higher urban centrality leads to an improvement in regional tourism income, thereby contributing to the development of the tourism industry in a city (Table 3).

Table 2.

The results of the baseline regression.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
lnUCentrality	1.596***	0.661***	0.633***	0.639***	0.599***	0.597***	0.650***	0.488***	0.546***
lnUCentrality	(46.18)	(21.15)	(20.55)	(20.71)	(19.21)	(19.20)	(16.30)	(11.62)	(11.42)
Lnhotel		0.790***	0.742***	0.730***	0.723***	0.696***	0.687***	0.657***	0.637***
Lnhotel		(54.96)	(50.37)	(48.10)	(47.83)	(43.89)	(42.32)	(40.39)	(38.87)
attraction			0.153***	0.156***	0.145***	0.143***	0.140***	0.138***	0.135***
attraction			(11.56)	(11.75)	(10.92)	(10.79)	(10.84)	(10.76)	(10.68)
lnhighway				0.049***	0.050***	0.049***	0.079***	0.179***	0.181***
lnhighway				(3.35)	(3.42)	(3.37)	(5.32)	(10.32)	(10.25)
Hsr					0.148***	0.151***	0.113***	0.093***	0.098***
Hsr					(6.83)	(6.99)	(5.30)	(4.39)	(4.69)
Airport						0.110***	0.113***	0.099***	0.104***
Airport						(5.62)	(5.76)	(5.11)	(5.35)
lnPdensity							0.113***	0.145***	0.146***
lnPdensity							(6.82)	(8.75)	(7.69)
lnPgdp								0.224***	0.220***
lnPgdp								(10.86)	(10.46)
lnpm2.5									−0.096***
lnpm2.5									0.546***
Constant	−8.635***	−1.584***	−1.159***	−1.587***	−1.190***	−1.110***	−2.483***	−3.865***	−4.041***
Constant	(−24.41)	(−5.33)	(−3.93)	(−4.94)	(−3.67)	(−3.43)	(−6.82)	(−10.13)	(−10.01)
Observations	4992	4992	4992	4992	4992	4850	4830	4795
R-squared	0.786	0.874	0.876	0.877	0.877	0.881	0.885	0.887
Province FE	YES	YES	YES	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES	YES	YES	YES

t-statistics in parentheses. *** p < .01, ** p < .05, * p < .1.

Table 3.

The results of robustness test.

	(1)	(2)	(3)	(4)
Variables	lntourist	lnTourism_w	GMM	lnToursism_s
lnUCentrality	0.292***	0.470***	0.420***	1.002***
lnUCentrality	(7.38)	(10.27)	(7.34)	(21.82)
Lnhotel	0.499***	0.611***	0.638***	0.364***
Lnhotel	(36.75)	(38.97)	(33.22)	(23.08)
attraction	0.094***	0.098***	0.103***	0.188***
attraction	(8.97)	(8.08)	(7.07)	(13.12)
lnhighway	0.239***	0.155***	0.249***	0.357***
lnhighway	(16.28)	(9.17)	(12.55)	(19.10)
Hsr	0.136***	0.127***	0.128***	0.214***
Hsr	(7.81)	(6.34)	(5.61)	(9.21)
Airport	0.068***	0.097***	0.097***	0.103***
Airport	(4.24)	(5.24)	(4.51)	(4.65)
lnPdensity	0.180***	0.150***	0.351***	0.160***
lnPdensity	(11.44)	(8.23)	(21.24)	(7.64)
lnPgdp	0.196***	0.220***	0.295***	0.509***
lnPgdp	(11.27)	(10.94)	(11.17)	(23.75)
lnpm2.5	−0.082***	−0.130***	−0.137***	−0.488***
lnpm2.5	(−2.85)	(−3.92)	(−4.25)	(−13.89)
Constant	−0.898***	−2.783***	−5.056***	−10.772***
Constant	(−2.68)	(−7.20)	(−12.01)	(−34.08)
Observations	3961	3963	3963	3963
R-squared	0.881	0.880	0.832	0.850
Province FE	YES	YES	YES	YES
Time FE	YES	YES	YES	YES

t-statistics in parentheses. *** p < .01, ** p < .05, * p < .1.

Robustness test

To ensure the reliability of the regression results, this paper carries out a robustness test through four methods. (1) Replace the explained variable (Column (1)). We use the number of tourists ( ${l n t o u r i s t}_{i t}$ ) instead of tourism income to measure the level of urban tourism development. (2) Smooth outliers (Column (2)). We remove outlier observations by smoothing the samples' top and bottom 5%. (3) Change different econometric methods (Column (3)). Here we adopt the generalized moment estimation (GMM) method. (4) Exclude the four municipalities (Peking, Shanghai, Tianjin, Chongqing) directly under the Central Government of China. Because these four cities have more priority in formulating city policy, they possess better development levels in aspects such as economy, culture, environment, and transportation infrastructure. Therefore, we remove their samples and get results in Column (4). All the robustness tests show that the signs and significance of the coefficients of main variables are consistent with those in the baseline regression, implying that the estimation is robust and reliable.

Endogenous test

To deal with the endogenous problem, this paper employs the slope of each city ( ${i v s l o p e}_{i t}$ ) as an instrumental variable. However, the slope data as the cross-sectional data does not change with time, which increases the difficulty in the panel data regression. This paper draws on the method from Hu et al. (2021) and uses the multiplication term of the slope data and the year to construct a composite variable ( ${l n i v s l o p e}_{i t}$ ) as instrumental variables. In addition, we apply the two-stage least squares method (2SLS) and limited-information maximum likelihood estimation (LIML) to do the test.

The results are shown in Table 4. Column (1) is the first stage of the 2SLS model, representing the influence of slope on urban centrality. The coefficient of the IV is significantly negative at the 1% confidence level, indicating that cities with larger geographical slopes tend to have lower network centrality. Column (2) is listed as the second stage of the 2SLS model. The coefficient of the main variable in Column (2) is significantly positive at the 1% confidence level. The result is consistent with that in the baseline model. Column (3) uses the LIML regression, which is relatively insensitive to weak IV. The sign and the significance of all variables in Column (3) are close to that in 2SLS.

Table 4.

The results of endogenous test.

	(1) first stage	(2)2SLS	(3) LIML
Variables	lnUCentrality	lnTourism	lnToursim
lnivslope	−0.013***
lnivslope	(-3.79)
lnUCentrality		1.702***	3.085***
lnUCentrality		(17.59)	(11.58)
Lnhotel	0.118***	0.340***	0.259***
Lnhotel	(17.82)	(19.43)	(10.25)
attraction	0.013***	0.132***	0.094***
attraction	(3.26)	(7.37)	(4.48)
lnhighway	0.112***	0.229***	−0.030
lnhighway	(14.28)	(8.83)	(-0.53)
Hsr	0.061***	0.205***	0.093***
Hsr	(9.05)	(7.81)	(2.63)
Airport	0.055***	0.045*	−0.031
Airport	(7.85)	(1.85)	(-0.96)
lnPdensity	0.146***	0.220***	0.027
lnPdensity	(26.51)	(9.88)	(0.62)
lnPgdp	0.195***	0.251***	−0.193**
lnPgdp	(18.27)	(6.93)	(-2.15)
lnpm2.5	−0.057***	−0.281***	−0.222***
lnpm2.5	(−5.39)	(−7.65)	(−4.93)
Constant	6.019***	−15.500***	−22.272***
Constant	(40.16)	(−26.77)	(−16.52)
Observations	3963	3963	3963
R-squared	0.774	0.774	0.681
Province FE	YES	YES	YES
Time FE	YES	YES	YES

t-statistics in parentheses. *** p < .01, ** p < .05, * p < .1.

Mechanism test

As stated above, we carry out the mechanism test to examine the channels through which urban centrality influences tourism’s development in a city. Column (1) in Table 5 shows the baseline model regression results. Column (2) shows that urban centrality can indeed promote the scale of agglomeration, while the coefficient is significantly positive at the 1% confidence level. Column (3) shows that after adding the agglomeration effect variable (lnLQ), the significance of the main explanatory variable remains unchanged, and the coefficient becomes from 0.546 to 0.478. As stated above, urban centrality positively influences the tourism industry’s development through the agglomeration effect. The regression results of the mechanism test about the industrial structure effect are shown in Columns (4) and (5). The coefficient of urban centrality affecting the industrial structure effect variable (structure) in Column (4) is 12.822, which is significantly positive at the level of 1%. After the structural variable is added to the model in Column (5), the significance remains unchanged, but the coefficient of urban centrality decreases. In addition, the coefficient of urban centrality is significantly positive at 1%. It means that urban centrality induces the industrial structure effect, eventually benefiting the tourism industry. As a whole, our analysis here indicates that urban centrality influences tourism income through both agglomeration and industrial structure effects.

Table 5.

The results of the mechanism test.

	(1)	(2)	(3)	(4)	(5)
Variables	lnTourism	LnLQ	LnToursim	structure	lnTourism
lnUCentrality	0.546***	0.986***	0.478***	12.822***	0.384***
lnUCentrality	(11.42)	(28.96)	(9.08)	(16.58)	(7.93)
lnLQ			0.071***
lnLQ			(3.16)
Structure					0.013***
Structure					(13.05)
Lnhotel	0.637***	0.181**	0.625***	5.305***	0.570***
Lnhotel	(38.87)	(2.45)	(37.01)	(20.01)	(33.84)
attraction	0.135***	0.058***	0.133***	0.289	0.132***
attraction	(10.68)	(-3.75)	(10.41)	(1.41)	(10.61)
lnhighway	0.181***	−0.105***	0.207***	−4.585***	0.239***
lnhighway	(10.25)	(-6.47)	(10.59)	(-16.04)	(13.38)
Hsr	0.098***	0.014	0.097***	1.062***	0.085***
Hsr	(4.69)	(0.71)	(4.64)	(3.13)	(4.13)
Airport	0.104***	−0.102***	0.095***	0.360	0.099***
Airport	(5.35)	(−5.72)	(4.52)	(1.15)	(5.22)
lnPdensity	0.146***	0.017	0.154***	2.749***	0.112***
lnPdensity	(7.69)	(0.98)	(8.04)	(8.94)	(5.93)
lnPgdp	0.220***	0.307***	0.192***	2.318***	0.190***
lnPgdp	(10.46)	(15.96)	(8.52)	(6.83)	(9.21)
lnpm2.5	−0.096***	0.106***	−0.100***	−2.362***	−0.066*
lnpm2.5	(-2.77)	(3.32)	(-2.88)	(-4.21)	(-1.95)
Constant	−4.041***	−6.743***	−3.753***	81.397***	−5.069***
Constant	(-10.01)	(18.25)	(-8.93)	(12.47)	(-12.58)
Observations	3963	3961	3961	3963	3963
R-squared	0.886	0.503	0.884	0.715	0.890
Province FE	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES

t-statistics in parentheses. *** p < .01, ** p < .05, * p < .1.

Heterogeneity analysis

The heterogeneous characteristics of various cities may lead to differences in the intensity of the influence on the tourism industry. To explore the influence more deeply, this paper employs two indicators to classify our samples. One is the size of the city represented by the number of permanent residents in a city. The other is the location of the cities.

We classify the samples into three sub-samples with residents’ populations of more than 5 million as large-scale cities, those between 2 million and 5 million as medium-sized cities, and those of less than 2 million as small-scale cities. The results in Table 6 show that urban network centrality has a significant positive effect on tourism income in both large and medium-sized cities. However, the regression coefficient of urban network centrality is not significant in small-sized cities, indicating that the improvement of urban network centrality does not significantly increase tourism development in small-sized cities. Moreover, although significantly positive effects exist in the first two types of cities, the medium-sized cities experience a weaker effect than the large-sized cities in terms of the intensity of the influence. It is consistent with our expectation that cities with a larger population generally have higher centrality due to the agglomeration effect and thus have an overall higher economic development level in various sectors, especially the service industry, such as the tourism sector.

Table 6.

Heterogeneity test by city size.

	(1) large-scale	(2) middle-scale	(3) small-scale
Variables	(1) large-scale	(2) middle-scale	(3) small-scale
lnUCentrality	0.702***	0.478***	0.225
lnUCentrality	(7.21)	(7.17)	(1.47)
Lnhotel	0.568***	0.602***	0.574***
Lnhotel	(18.02)	(24.70)	(13.34)
attraction	0.112***	0.120***	0.231***
attraction	(6.71)	(5.32)	(5.23)
lnhighway	0.228***	0.048	0.256***
lnhighway	(5.73)	(1.36)	(4.67)
Hsr	0.094***	0.053*	0.153**
Hsr	(2.91)	(1.90)	(2.20)
Airport	0.189***	0.102***	0.255***
Airport	(6.21)	(3.49)	(4.60)
lnPdensity	0.183***	0.102***	−0.047
lnPdensity	(3.89)	(3.10)	(−0.79)
lnPgdp	0.342***	0.246***	0.262***
lnPgdp	(8.71)	(8.28)	(4.24)
lnpm2.5	−0.126*	−0.113**	0.420***
lnpm2.5	(-1.71)	(-2.34)	(4.62)
Constant	−6.922***	−2.056***	−2.867**
Constant	(−7.55)	(−2.92)	(−2.11)
Observations	1456	1803	704
R-squared	0.892	0.878	0.900
Province FE	YES	YES	YES
Time FE	YES	YES	YES

t-statistics in parentheses. *** p<0.01, ** p<0.05, * p<0.1.

Considering the huge differences in climatic conditions, tourism resources, and urban development levels among various cities in China, this paper categorizes cities into three geographical parts: the eastern, central, and western regions. Table 7 shows that urban centrality has the greatest influence on the tourism industry in western region cities than in the other two regions. It is reasonable when considering the relative underdevelopment in this region in the aspects of economics, infrastructure, city management, and others.

Table 7.

Heterogeneity test by city’s location.

	(1)	(2)	(3)
Variables	Eastern region	Central region	Western region
lnUCentrality	0.441***	0.670***	0.729***
lnUCentrality	(4.76)	(8.41)	(8.62)
Lnhotel	0.543***	0.506***	0.801***
Lnhotel	(17.97)	(19.36)	(26.05)
attraction	0.101***	0.237***	0.089***
attraction	(5.83)	(9.60)	(3.68)
lnhighway	0.175***	0.178***	0.122***
lnhighway	(5.99)	(5.99)	(3.76)
Hsr	0.189***	0.100***	0.109**
Hsr	(6.29)	(2.86)	(2.36)
Airport	0.072**	0.272***	−0.022
Airport	(2.47)	(7.68)	(−0.64)
lnPdensity	0.172***	−0.084*	0.186***
lnPdensity	(5.79)	(-1.93)	(6.01)
lnPgdp	0.261***	0.413***	0.062*
lnPgdp	(6.85)	(11.41)	(1.67)
lnpm2.5	−0.320***	0.287***	−0.229***
lnpm2.5	(-6.40)	(4.02)	(-3.64)
Constant	−2.045***	−6.958***	−4.207***
Constant	(−2.67)	(−9.15)	(−6.84)
Observations	1406	1391	1166
R-squared	0.879	0.874	0.891
Province FE	YES	YES	YES
Time FE	YES	YES	YES

t-statistics in parentheses. *** p < .01, ** p < .05, * p < .1.

Conclusions

The tourism industry in China has developed rapidly in the recent 20 years and contributed substantially to the whole economic growth. Previous literature has discussed the determinants of the tourism industry from various perspectives. However, very few studies have explored it from the perspective of urban centrality. As an index to measure the relative importance of a city in the urban network, urban centrality is a vital factor that cannot be ignored in the study of urban tourism. This paper tries to fill the research gap by introducing the concept of urban centrality in studying tourism’s determinants. In addition, we identify two mechanisms through which urban centrality influences a city’s tourism development. To facilitate the analysis, this paper uses panel data of 295 cities from 2005 to 2018 in China to study the influence of urban centrality on tourism income. By controlling a series of related variables, this paper uses fixed effect regression methods to testify the influence. Endogeneity test and robustness test are carried out to ensure reliable and robust results.

Empirical results show that urban centrality has a significantly positive influence on tourism income. The results, in general, are in line with our expectations. It highlights the importance of a city’s centrality in the urban network for economic growth. Moreover, considering the huge imbalance in Chinese cities, we employ the heterogeneity analysis. Results show the existence of regional diversification as well. The magnitude of the influence is particularly prominent in large-scale and medium-scale cities. In addition, the effect becomes stronger in cities in the Western region than those in other regions. Moreover, theoretical and empirical mechanism analysis shows that agglomeration and structural effects are the two main channels through which a city’s centrality affects the development of its tourism industry. It is one of the newest insights of this study.

Theoretically, this paper contributes to the agglomeration economy, network science theory, and tourism research. Although some scholars have discussed the network or centrality concepts as the determinants of tourism, they mostly emphasized the transportation network (e.g. Chen et al., 2020) or the network of the tourism economy (e.g. Gan et al., 2021). Our study pays attention to a city’s centrality in the urban network, originally from network science and urban study. Consequently, our study enriches the existing discussion about the determinants of tourism development. In addition, we identify the two mechanisms, agglomeration, and industrial structure effect, thereby expanding theoretically and empirically the interdisciplinary of network theory and agglomeration economics. Moreover, the econometric analysis method in this paper provides a more comprehensive complementary to the tourism literature, which usually employs case studies.

The analysis of this paper provides several implications for national and regional policymakers, especially in China. Considering the importance of the central position in the city network or urban agglomeration, cities need to break through the constraints and bottlenecks due to geographical location. Cities can formulate relevant industrial policies to attract the entry of high-end production resources and exploit their competitive advantages. Cities with high centrality can be considered strategic cities for transforming and upgrading the regional tourism industry. Their development will have a spillover effect on other cities in the network. Moreover, small- and medium-scale cities could strengthen exchanges and cooperation with other cities in the network to realize a synergy effect. In addition, tourism is currently one of the industries most damaged by COVID-19. Our analysis shows that tourism development is closely related to other sectors and the city’s status. The industrial upgrading and supporting facilities within the city are necessary, but the distribution of transportation infrastructure and the increased urban centrality brought about by improved transportation facilities also need to be underscored.

This paper also has its limitations. We use the data on China’s leading producer service company to construct the urban centrality index as our explanatory variable. Literature has provided several methods for constructing this index. It might be more persuasive to employ other comparable methods or data for the construction in future analysis. Furthermore, urban tourism development depends on natural and cultural landscapes, as well as urban amenities and facilities. Future related research can include more indicators to represent the heterogeneity of tourism’s development, which might help yield more implications in practice.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Shanghai Philosophy and Social Science Planning Projec (No. 2019BJB012) and National Natural Science Foundation of China (No. 72003141).

ORCID iDs

Jieping Chen

Chao Ma

Author biographies

Jieping Chen works as Post-doc in School of Economics and Management, Tongji University. She received her PhD from The Graduate School of Economic and Social Sciences in the University of Mannheim in Germany. Her research area includes Innovation and Entrepreneurship, Enterprise innovation and regional economy, urban economy and economic geography.

Chao Ma is a doctoral student at the Department of Economics and Management of Tongji University. His research interest is in applied econometrics, with a particular focus on time-series analysis (including forecasting) and panel-data analysis. He specific information technology-related interests include big data analysis, blockchain, and online travel information search. His current primary fields of thematic interest are the economics of tourism and of sustainability.

Shijun Chen is a doctoral student at the Department of Economics and Management of Tongji University. His research area includes regional economic development, tourism economy and leisure economy.

References

Aranburu

Plaza

Esteban

(2016) Sustainable cultural tourism in urban destinations: does space matter? Sustainability 8(8): 699.

Boschma

Ter Wal

ALJ

(2007) Knowledge networks and innovative performance in an industrial district: the case of a footwear district in the south of Italy. Industry & Innovation 14(2): 177–199.

Chen

Liu

, et al. (2020) Analysis of Complex Transportation Network and its Tourism Utilization Potential: A Case Study of Guizhou Expressways: Complexity. https://doi.org/10.1155/2020/1042506

Curado

Tortosa

Vicent

, et al. (2020) Analysis and comparison of centrality measures applied to urban networks with data. Journal of Computational Science 43: 101127.

Denicolai

Cioccarelli

Zucchella

(2010) Resource-based local development and networked core-competencies for tourism excellence. Tourism Management 31(2): 260–266.

Derek

(2018) Spatial structure of tourism in a city after transition: the case of Warsaw, Poland. In: Tourism in Transitions. Cham: Springer, pp. 157–171.

Derudder

Zhan

Xingjian

, et al. (2018) Changing connectivity of Chinese cities in the world city network 2010–2016. Chinese Geographical Science 28(02): 183–201.

Derudder

Taylor

(2021) The GaWC perspective on global-scale urban networks. In: Handbook of Cities and Networks. Massachusetts, USA: Edward Elgar Publishing.

Fang

Cheng

(2021) The spatial spillover effect of integration progress on tourism economy: a case study of the Yangtze River Delta urban agglomeration. Scientia Geographica Sinica 41(9): 1546–1555. (in Chinese).

10.

Gan

Voda

Wang

, et al. (2021) Spatial network structure of the tourism economy in urban agglomeration: a social network analysis. Journal of Hospitality and Tourism Management 47: 124–133.

11.

Hennemann

Derudder

(2014) An alternative approach to the calculation and analysis of connectivity in the world city network. Environment and Planning B: Planning and Design 41(3): 392–412.

12.

Hirschman

(1958) The Strategy of Economic Development. New Haven: Yale University Press.

13.

Nie

Shi

(2021) Can you have your cake and eat it? -- The influence of innovation-oriented city pilot policy urban industrial agglomeration. Industrial economic research 1: 128–142. (In Chinese).

14.

Jacobs

(1969) The Economy of Cities. New York: Random House.

15.

Jovanović

Ivana

ILI Ć

(2016) Infrastructure as important determinant of tourism development in the countries of Southeast Europe. Ecoforum journal 5(1): 288–294.

16.

Kalnins

Chung

(2004) Resource-seeking agglomeration: a study of market entry in the lodging industry. Strategic Management Journal 25(7): 689–699.

17.

Khadaroo

Seetanah

(2008) The role of transport infrastructure in international tourism development: a gravity model approach. Tourism Management 29(5): 831–840.

18.

Kim

Williams

Park

, et al. (2021) Spatial spillovers of agglomeration economies and productivity in the tourism industry: the case of the UK. Tourism Management 82: 104201.

19.

Kong

(2021) Study on high-speed railway opening, network centrality and tourism economy development. Industrial economic research 5: 113–127. (In Chinese).

20.

Ledesma González

Merinero-Rodríguez

Pulido-Fernández

(2021) Tourist destination development and social network analysis: what does degree centrality contribute? International Journal of Tourism Research 23(4): 652–666.

21.

Lin

Chen

(2021a) Rapid transportation and green technology innovation in cities—from the view of the industrial collaborative agglomeration. Applied Sciences 11(17): 8110.

22.

Lin

Yan

, et al. (2021b) Do prevention and control measures work? Evidence from the outbreak of COVID-19 in China. Cities 118: 103347.

23.

Liu

Derudder

(2012) Two-mode networks and the interlocking world city network model: a reply to Neal. Geographical Analysis 44(2): 171–173.

24.

Liu

Wen

(2015) The effect of labor pool effect and knowledge spillover effect on labor income share – based on Marshall’s Industrial agglomeration theory. Journal of Yunnan University of Finance and Economics 31(06): 3–12. (In Chinese).

25.

Liu

Huang

(2017) An application of network analysis on tourist attractions: the case of Xinjiang, China. Tourism Management 58: 132–141.

26.

Liu

Kang

, et al. (2018) Scenic spot selection, government public service supply and regional tourism economic development. Chinese industrial economy. 2: 118–136. (In Chinese).

27.

López

MFB

Virto

Manzano

, et al. (2018) Residents’ attitude as determinant of tourism sustainability: the case of Trujillo. Journal of Hospitality and Tourism Management 35: 36–45.

28.

Luo

Chen

Sun

, et al. (2020) Analysis of city centrality based on entropy weight TOPSIS and population mobility: a case study of cities in the Yangtze River Economic Belt. Journal of Geographical Sciences 30(4): 515–534.

29.

Ren

Yuen

, et al. (2019) The spatial coupling effect between urban public transport and commercial complexes: a network centrality perspective. Sustainable Cities and Society 50: 101645.

30.

Majewska

(2015) Inter-regional agglomeration effects in tourism in Poland. Tourism Geographies 17(3): 408–436.

31.

Marshall

(1890) Principles of Economics. London: Macmillan and co. Limited.

32.

Neal

(2011) Differentiating centrality and power in the world city network. Urban Studies 48(13): 2733–2748.

33.

Neal

(2013) The Connected City: How Networks Are Shaping the Modern Metropolis. New York: Routledge.

34.

Neal

Derudder

Taylor

(2020) Forecasting the world city network. Habitat International 106: 102146.

35.

Pereira

RHM

Nadalin

Monasterio

, et al. (2013) Urban centrality: a simple index. Geographical Analysis 45(1): 77–89.

36.

Ramukumba

Mmbengwa

Mwamayi

, et al. (2012) Analysis of local economic development (LED) initiated partnership and support services for emerging tourism entrepreneurs in George municipality, Western Cape Province, RSA. Tourism Management Perspectives 2: 7–12.

37.

Ritchie

Zins

(1978) Culture as determinant of the attractiveness of a tourism region. Annals of Tourism Research 5(2): 252–267.

38.

Rüttenauer

(2019) Bringing urban space back in: a multilevel analysis of environmental inequality in Germany. Urban Studies 56(12): 2549–2567.

39.

Sharma

Nayak

(2020) Examining experience quality as the determinant of tourist behavior in niche tourism: an analytical approach. Journal of Heritage Tourism 15(1): 76–92.

40.

Shaw

Williams

(2009) Knowledge transfer and management in tourism organisations: an emerging research agenda. Tourism Management 30(3): 325–335.

41.

Shih

(2006) Network characteristics of drive tourism destinations: an application of network analysis in tourism. Tourism Management 27(5): 1029–1039.

42.

Sosnovskikh

(2020) What agglomeration externalities impact the development of the hi-tech industry sector? Evidence from the Russian regions. International Journal of Technology 11: 1091.

43.

Stienmetz

Fesenmaier

(2015) Estimating value in Baltimore, Maryland: an attractions network analysis. Tourism Management 50: 238–252.

44.

Sugimoto

Ota

Suzuki

(2019) Visitor mobility and spatial structure in a local urban tourism destination: GPS tracking and network analysis. Sustainability 11(3): 919.

45.

Sun

Qin

Wang

(2021) Does the Performance of Urban Center Promote the Upgrading of Urban Export Structure? Evidence from Chinese Cities and Urban Agglomerations. World Economic Studies 134(01): 17–31. (In Chinese).

46.

Taylor

(2001) Specification of the world city network. Geographical Analysis 33(2): 181–194.

47.

Taylor

(2004) Regionality in the world city network. International Social Science Journal 56(181): 361–372.

48.

Taylor

Derudder

Hoyler

, et al. (2014) City-dyad analyses of China’s integration into the world city network. Urban Studies 51(5): 868–882.

49.

Wang

Chen

, et al. (2016) HSR mechanisms and effects on the spatial structure of regional tourism in China. Journal of Geographical Sciences 26(12): 1725–1753.

50.

Xiang

Zheng

Wang

(2012) Spatial econometric study on factors of tourism economic growth. Economic Geography 32(06): 162–166. (In Chinese).

51.

Xie

Yin

(2021) Research on the spatial structure of the European union’s tourism economy and its effects. International Journal of Environmental Research and Public Health 18(4): 1389.

52.

Yang

Fik

(2014) Spatial effects in regional tourism growth. Annals of Tourism Research 46: 144–162.

53.

You

Chen

(2021) Regional integration degree and its effect on a city’s green growth in the Yangtze River Delta: research based on a single-city regional integration index. Clean Technologies and Environmental Policy 23: 1–13.

54.

Zhang

Wang

, et al. (2015a) Village network centrality in rural tourism destination: a case from Yesanpo tourism area, China. Journal of Mountain Science 12(3): 759–768.

55.

Zhang

Xiao

Gursoy

, et al. (2015b) Tacit knowledge spillover and sustainability in destination development. Journal of Sustainable Tourism 23(7): 1029–1048.

56.

Zhang

Chen

Guan

, et al. (2021) Asymmetric Impacts of Technology Innovation and Environmental Quality on Tourism Development in Emerging Economies. Economic Research-Ekonomska Istraživanja, pp. 1–17.