How far will we travel? A global distance pattern of international travel from both demand and supply perspectives

Abstract

The travel distance of international journeys critically determines our reliance on different transportation modes and the associated carbon intensity. This study quantified the influence of macrolevel determinants to the inbound and outbound average distance per visitor from a panel data of 152 countries using spatial econometric analysis. Results confirmed that national development and transport capacity assisted the spatial expansion of outbound travel, while tourism competitiveness, geographic attributes, and institutional arrangements regarding people’s mobility facilitate inbound visits from distant source markets. A high level of heterogeneity was found across five continents where the distance friction effect through geographic barrier, transport accessibility, and the freedom of people’s movement exhibited a different level of influences. To manage the spatial expansion of international travels for a sustainable transport future, a strong geopolitical integration system across countries within the region and adjustments to the aviation capacity to disfavor long-haul flights have been proposed.

Keywords

connectivity distance decay effect international travel distance national development tourism competitiveness

Introduction

The time and space expansion of global travel activities is directly related to our reliance on different transportation modes and the substantial energy demand. Especially, the average distance that a person travels during international journeys critically determines our usage of aviation and the associated carbon intensity of the trip (Scott et al., 2012). Based on the UN World Tourism Organization (UNWTO) (2017), it has been estimated that international tourist arrivals have increased from 278 million globally in 1980 to 1235 million in 2016, representing a 4% annual growth rate over the past four decades (UNWTO, 2017). Economic activity of this scale generates substantial environmental burdens on global carbon emissions (Lenzen et al., 2018). In response, policies have been proposed that aim at changing travel behavior at departure points by encouraging visitors to choose destinations that are close at home, travel less but with an extended length of time per trip, and use less aviation (WTO-UNEP, 2008).

With respect to these objectives, statistics however have shown some pessimistic results as international tourist arrivals remain strong and is expected to increase by an annual rate of 3.3% over the period of 2010 to 2030 (UNWTO, 2011). The aviation sector has experienced an even faster expansion as total passenger-kilometers increased by 7.9% annually (International Civil Aviation Organization (ICAO), 2018). The sheer increase in the number of visits and the flight mileages however are not equally allocated across different destinations. For the past decade, emerging economies have reported an average of 4.5% annual growth in the number of arrivals while the advanced economies have been slower with 3.5% expansion (UNWTO, 2017). In terms of geographic regions, the strongest records of inbound visits were found in South Asia (10.7%), sub-Saharan Africa (10.5%), and Southeast Asia (7.9%), while the Middle East (3.7%) and Western Europe (0.0%) experienced declines or stagnation from 2010 to 2016. The uneven growth of international visits signals the shifting patterns of origin–destination (O-D) pairs, which leads to a modification of average distance per journey under this changing tourism demand and supply geography. Using a global forecast model, Peeters et al. (2018) profiled that the business as usual context in 2100 compared to the one in 2005, in which demand for trips will increase by 340%, the number of guest-nights by 170%, distance traveled by 880%, and carbon dioxide emissions by 370%. Among all travel-related attributes, “distance traveled” has shown the most striking growth, by a factor of almost 10. Therefore, one important question that is highly relevant to the magnitude of tourism carbon footprint is: What are the key drivers at the macrolevel that will influence the organic changes of the distance component?

On the aspect of travel distance research, the majority of the literature has focused on the micro factors that motivate travelers to choose short-, medium-, or long-distance destinations. It has examined this from (1) an inbound arrivals perspective—that is, enriching the knowledge of a host country with respect to sociodemographic patterns from different market segments (Ahn and McKercher, 2013; Ho and McKercher, 2012; Mishra and Bansal, 2017; Qian et al., 2017), or (2) through an outbound departure perspective—profiling citizen’s foreign travel behaviors and persistence in distance distribution across time, demographics, and economic cycles (Lee et al., 2012; Lim et al., 2008; McKercher and Lew, 2003; Wong et al., 2016). While these researchers have provided important observations about the key drivers that are associated with an individual’s trip distribution, they are mostly based on a single country through either a supply or demand perspective. The seminal work of McKercher et al. (2008) is one of the few exceptions that concurrently address the international trip distribution pattern from demand, supply, and relationship characteristics. Their research confirms the applicability of the distance decay theory to international outbound travel on a global scale, as 80% of all international travel in 2002 occurred to countries within 1000 km from the source market.

While the distance decay effect is acknowledged for the majority of countries, there exists limited discussions on the key determining factors and their dynamics to the aggregated distance attributes with respect to the total distances and average distances traveled per visitor by countries. In addition, no country would experience identical distance distribution curves between inbound and outbound travels—implying that there are different mechanisms in play between the supply and demand for international journeys. To systematically analyze the distance attributes thus requires a comprehensive approach—which addresses demand, supply, and connecting characteristics at the macrolevel.

In this article, we propose that there are three important determinants with respect to international travel distance, that include national development at the departure point, tourism competitiveness at the source destination, and the connecting factors with respect to the geographic attributes, transport capacity, and the institutional arrangement regarding people’s movement between countries. The purpose of this study is to identify these three sets of determinants with respect to the travel distance, namely, inbound and outbound average distance, using a panel of 152 countries, which are studied as both tourist origins and destinations. Spatial econometric models which take into account the global spillover effects of visits, have been employed to validate the quantitative influences from both the demand and supply factors. This study aims to establish insights regarding the demand and supply mechanism that drives travel distance at the macrolevel as well as providing a foundation for future scenario analysis for tourism expansion, transportation demand, and carbon emissions.

Conceptual framework

Distance decay effect and the gravity model

To understand the international trip distribution pattern on a global scale, the distance decay effect serves as an important framework. Distance is recognized as an important factor in the distribution of ideas, technology, population, time, and the interaction of various components in geography (Eldridge and Jones, 1991). This key law of geography proposes that the volume of international journeys, either outbound departures or inbound arrivals, decreases along with the distance between the O-D points. In other words, the majority of international travels occurs over a short distance while the long-haul visits account for a small share of the aggregated total. This idea has been examined in many tourism settings and generally found this pattern to be applicable, with the distribution of visit volume and trip distance approximated using a negative binominal or Poisson distribution with a long tail to the right-hand side (Mishra and Bansal, 2017; Yang et al., 2018).

The distribution of travel distance with respect to visitor volume then determines the total trip distances and average distance per visitor. While the majority of countries may find their inbound or outbound visits following the distance decay effect, the average distance for all aggregated international journeys is drastically different across countries. This is because after standardizing the distribution pattern—which is to plot the percentages of international visitors against distance, “peaks of the distribution” and “tails of the distribution” vary. Peaks of the distribution describe the most significant segment from the distance cohort, and the tail of the distribution evaluates the share of travelers who are willing to engage with long-haul cross-continent travels. The number and location of “peaks” in distance and the thickness of the “trail” simultaneously determine the average distance traveled, independent from the overall travel volume. For example, in 2013, Germany, the largest tourism outbound source market in Europe, leads a smaller average outbound distance (1516 km) than its nearby island neighbor, the United Kingdom (2622 km). Their difference in distance attribute is contributed to by: (1) the peak of Germany outbound tourism is within the “less 1000 km” category (62% of visits) while UK visitors preferred destinations at ranges of “less 1000 km” (32%) and “1000–2000 km” (33%), and (2) a higher percentage of British visitors (4.5%) have chosen destinations that are located more than 10,000 km away than for Germany (0.9%). Overall, the peak of German outbound travel is closer to the point of origin and their tail distribution to distant destinations is less significant, leading to a smaller average distance traveled than its counterpart.

In order to examine and explain the spatial distribution of journeys, the gravity model provides a sound theoretical underpinning, indicating that visitor volume from the origin country i to destination j is influenced by three sets of factors: attributes at the departure country, attractiveness of the destination country, and connecting factors that capture the distance friction effect between each O-D pair (Morley et al., 2014). For factors that are not directly associated with country i and j, they are then captured by the error term. The formula of a general gravity model can be expressed as

N_{i j} = f (O_{i}^{s}, D_{j}^{p}, {OD}_{i j}^{r})

where $N_{i j}$ denotes the number of tourist numbers between origin i and destination j; $O_{i}^{s}$ is a vector of S variables that represent the push factors from origin i; $D_{j}^{p}$ is a vector of P variables that represent the pull factors from destination j; and ${OD}_{i j}^{r}$ is a vector of R variables that determines the distance-influenced components with respect to physical, psychological, social, and economic dimensions.

To examine the proposed distance attributes, we first calculate the inbound average distance (IAD) of a destination country (j), which is to sum up the trip distances of inbound visitors from all departure points, divided by the total inbound visitor number to destination j. It can be expressed as follows

{IAD}_{j} = \frac{\sum_{i = 1}^{n} N_{i j} \times d_{i j}}{\sum_{i = 1}^{n} N_{i j}}

where $d_{i j}$ is the distance between origin i and destination j. Since that distance for each country pair is constant, IAD _j is determined by the composition of N_ij from different distance cohorts. To model IAD is essentially to model a country’s capacity to attract visitors from short-, medium-, and long-haul source markets. This can then be tied to the gravity model to understand the influences exhibited by individual factors to the average distance attribute.

In equation (2), the summation of all inbound visitors to calculate IAD _j implies that all push factors from all source markets are aggregated. Translate this summation to the gravity model entails the influence of push factors to become a constant $(\sum_{i = 1}^{I} O_{i}^{s} = β)$ , which will be estimated through the intercept. The vector of R variables that determines the distance-related components between each ij pair in the gravity model, after the process of aggregating visits across all source markets, will also become a destination (j) dependent attribute $(\sum_{i = 1}^{I} {OD}_{i j}^{r} = {OD}_{.j}^{r})$ . In other words, modeling IAD through the gravity model can be approximated and simplified with the consideration of only two factors: the attractiveness attributes at a destination j $(D_{j}^{p})$ , and the aggregated connecting factors at a destination j $({OD}_{.j}^{r})$ . We can then rewrite the gravity model as

{IAD}_{j} = f (N_{j}) \sim f (D_{j}^{p}, O D_{.j}^{r})

Similarly, the OAD can be expressed as the weighted trip distance in relation to the composition of outbound journeys to different destinations

{OAD}_{i} = \frac{\sum_{j = 1}^{n} N_{i j} \times d_{i j}}{\sum_{j = 1}^{n} N_{i j}}

The summation process of residents to all destinations for the numerator implies that, through the gravity model, the pull factors from destinations will become a constant $(\sum_{j = 1}^{J} D_{j}^{p} = β)$ . In addition, the vector of R variables that determines the distance-related components between each ij pair, after the process of aggregation, will become a source market (i) dependent attribute $(\sum_{j = 1}^{J} {OD}_{i j}^{r} = {OD}_{i.}^{r})$ . To model the OAD through the gravity model is to understand this distance attribute through the influences of the following factors: the push variables at departure point i $(O_{i}^{s})$ , and the aggregated connecting factors at departure point i $({OD}_{i.}^{r})$

{OAD}_{i} = f (N_{i}) \sim f (O_{i}^{s}, {OD}_{i.}^{r})

Influencing factors

When choosing destinations that are either close or further away from home implicitly reflects many personal considerations and external contextual factors. At the microlevel, the distance decay effect reflects our travel choices over the considerations of time availability, cost, travel budget, available transportation modes, culture distance, market access, and personal motivations (Lee et al., 2012; McKercher, 2018; Yang et al., 2018). Hypermobility travelers may even associate distance for social status and network capital (Cohen and Gössling, 2015; Pappas, 2014). In aggregation, these micro-considerations reflect some fundamental characteristics from both the demand (point of departure) and supply (point of destination) perspectives.

For the source market, distance traveled is highly related to the economic status of a country. Based on Burton’s four phases of tourism participation, the affluent level of a nation which is hypothesized not only to determine the volume of domestic and international visits but also to reflect on the destinations of the different journeys (Weaver and Oppermann, 2014). Once countries reach a high affluence level, massive participation in the long-haul destinations will be expected than for countries that are industrializing or are almost industrialized. Higher income, in essence, not only represents higher purchasing power and a capacity to overcome the monetary costs associated with distance but also simultaneously reflects the fundamental characteristics of a society in terms of the advance in welfare of human beings. With an increase in affluence, economic growth has typically led to, and supported by women’s empowerment, improved education, increased longevity, an aging population, and greater urbanization (World Bank and International Monetary Fund, 2016). These sociodemographic changes are found to be the underlying contributing factors for behaviors that desired more unique and diversified travel experiences which are manifested by journeys to new and distant countries (Czepkiewicz et al., 2018; Lee et al., 2012; McKercher and Lew, 2003; Wong et al., 2016; Yan, 2015). Therefore, economic expansion combined with sociodemographic changes has empowered the capacity of outbound travelers to overcome distance in physical, financial, and psychological dimensions. We now propose our first hypothesis:

H1: For developed countries, a weaker distance decay effect is expected, which is, a relatively higher proportion of residents will engage in long-haul travel, than for developing countries. As a result, the national development status of economic and social wellbeing will have a positive impact on the average distance traveled for outbound visits.

On the other hand, from the perspective of the destination, McKercher et al. (2008) indicated that distance is a valid proxy variable that reflects the attractiveness or unattractiveness of a country to potential travelers. Based on this argument, tourism attractiveness of the host nation not only reflects its sheer arrival volume but is also based on distance traveled. The more attractive the country is, the higher possibility for it to appeal to visitors across the globe, producing a counter effect on the friction of distance (McKercher et al., 2008; Prideaux, 2005). This is also supported by the tourists’ viewpoints that exotic or faraway destinations are associated with happiness (Ram et al., 2013), while for others, distance can be an important factor for social status and network capital (Cohen and Gössling, 2015; Pappas, 2014). In other words, the tourism competitiveness of a nation, in comparison to regional and global competitors, can be a proxy that influences the geographic origination of their inbound visitors. If the majority of inbound tourists come from nearby regions with only a few from distant countries, a strong distance decay pattern is presented as the distribution of trip distance which skews strongly toward the left (close to the origin) and declines sharply after a short threshold. In this context, the host nation only exhibits its tourism attractiveness at the regional level, while the average distance traveled by the collective inbound travelers will be relatively low. On the other hand, for destinations that demonstrate strong tourism competitiveness, a relatively “fat tail” in distance distribution of inbound visits is expected as the attractiveness may produce a distorting effect over the barriers of time, cost, and effort from distance.

Traditionally, tourism attractiveness is found to be highly associated with visitor numbers, market share, tourist expenditure, employment, or tourism gross domestic product (TGDP) (Blanke and Chiesa, 2013; Craigwell, 2007; Crouch and Ritchie, 1999; Dwyer and Kim, 2003). The geographic distribution of inbound tourism has not been considered in the literature. In this article, we now propose the second hypothesis as follows:

H2: For countries with strong tourism competitiveness, a weaker distance decay effect is expected, which is, a relatively higher proportion of their inbound arrivals from their distant source markets, compared to low tourism competitiveness states. As a result, tourism competitiveness has a positive impact on the total and average distance traveled for inbound visits.

Besides the source and destination factors, one key component that links the travel flow between pairs rests on the connectivity between countries, a concept that embraces infrastructure, logistics, regulation, and people mobility, which collectively influences the time, cost, and comfortability required in international travel (ASEAN, 2016b). Connectivity between countries can be interpreted from multiple angles, which exhibit collectivity influences on decisions with respect to the incremental change of travel distance. First, geographic connectivity is determined by geographical attributes with respect to opportunities to access neighbors, typically measured through the length of land border to neighboring countries or number of countries with mutual borders. These geographic features then influence the physical connectivity between countries through different modes of transport such as air, land, or water movement as well as the frequency of services, which inherently have a direct relationship to the cost and time of travel. Transport capacity is therefore highly valued as part of the national tourism competitiveness and has been found to strongly influence inbound/outbound volume and receipts (Khan et al., 2017; Prideaux, 2005). This factor grows significantly once inbound visitors come from distant countries rather than those from neighboring regions. In addition, visitors who are not familiar with destination countries further value transport capacity as they expect efficient, reliable, and safe traveling disregarding how different the destinations are to their home country with respect to economics, political, cultural, or physical infrastructure dimensions (Khadaroo and Seetanah, 2008).

The next level of connectivity concerns institutional policies which determines the ease of movement across borders through the control of freedom of movement and transport liberalization. Intercountry tourism movement is regulated by visa policy which describes the criteria and length of stay permitted for visitors from a specific source market. Tourist visa policy can range from strict regulation requiring a face-to-face interview, to presenting the visa upon arrival and visa-free agreements which generally provides a few months stay in the destination country. Visa policy is not only used as a means of developing international diplomatic relationships but also reflects the source and type of international visitors that a destination prefers (Prideaux, 2005). From the country’s perspective, visa-waving policy is generally complemented with an increased international visitor flow (Balli et al., 2013), and from a traveler’s perspective, the more visa-free countries that one passport is entitled to also provides a subjective ranking for the traveling power of residents to specific countries (Passport Index, 2018).

Transport liberation is another important institutional policy that relates to the connectivity between countries through the control of transport freedom rights, typically discussed from an international aviation perspective. The granting of aviation freedom to foreign carriers through open-skies agreements have been gradually adopted among the major regions such as the European Union (EU) and the United States in 2008, and the Association of Southeast Asian Nations (ASEAN) consisting of 10 countries in 2016 (ASEAN, 2016a; European Commissions, 2015). Transport liberation supports a market structure that is free from governmental regulation and where airlines, routes, and frequencies are up for the market competition (Pitfield, 2009). This typically provides increased direct flights between subregions from bilateral travel and reduced airfares, which collectively improves the customer’s welfare (Winston and Yan, 2015).

The last component of connectivity relates to people to people relationship, where a firm interaction can be formed through migration and international trade. Immigration ties between countries have been found to create strong linkages with respect to people movement on inbound and outbound travel, that has been empirically supported by cases in Australia (Seetaram, 2012), New Zealand (Feng and Page, 2000), the European region (Massidda et al., 2014), and across the Organization for Economic Co-operation and Development countries (Balli et al., 2016). Immigration ties not only forms an important travel motivation of visiting friends and relatives but also supports a strong migration–tourism nexus that goes beyond and expands travel flow for business and leisure journeys. This is because migration and international trade reinforces each other as people’s intercountry movement increases demand for imported goods and services from the home country where international trade further exposes the opportunity for linkages of local residents to the outside world where a migration path is created. A similar two-way causality between international travel and trade is also found where trade openness and business travel are strongly linked (Shan and Wilson, 2001). While migration and international trade patterns are regulated by national policy for its own preferences for an open system that facilitates integration versus isolation, both activities are the key to establish people’s relationship between countries, which intangibly navigates the barriers over distance traveled.

In the tourism demand model, physical distance between countries is typically endogenously adopted as a given variable to model the distance-related barrier (Morley et al., 2014; Peng et al., 2014). In this study, we explicitly addressed the distance effect for the aggregated connectivity of a country to the world through geographic barriers, transport capacity, and institutional arrangement regarding people’s connection, and have proposed the third hypothesis as follows.

H3: For countries with a strong connectivity in transport, institutional policy and people’s connectivity, a weaker distance decay effect is expected for both inbound and outbound travel. That is a relatively higher proportion of inbound arrivals or outbound departures are from/for distant countries, compared to states with an inferior connectivity. As a result, connectivity has a positive impact on the average distance traveled for inbound and outbound visits.

Method

Data

Dependent variable

To understand the global inbound and outbound trip distribution patterns, it is necessary to establish an O-D country pattern for international travels. In other words, for each country, the inbound visits by source markets and the outbound visits to individual destinations are needed, and this forms a bilateral travel flow for an N × N matrix where N is the number of country. To compile this N × N matrix, the origin- and destination-specific data from the World Tourism Organization (UNWTO, 2009-2013) is adopted from “Arrivals of non-resident visitors at national borders by country of residence” and “Outbound tourism—departures of visitors”. The latest and the most comprehensive UNWTO data contain 193 states (N) and is only available for year 2013 at the time of writing. This is adopted as the analysis data for the study.

When compiling the bilateral travel database from the UNWTO statistics, we found that some countries reported their international arrivals or departures in a less disaggregated manner and grouped some visits into “others.” This has produced missing values for the matrix due to difficulties with classification. Therefore, official inbound/outbound data published by individual tourism authorities are manually searched online for important destinations¹ across the five continents. Even with this effort, some countries, especially small and developing states in Africa, still contain limited information regarding their international visits by states. Consistent with McKercher et al. (2008), we set a minimum criterion that only countries with at least 70% of total inbound or outbound visits identified by O-D pairs will be entered for analysis so that the average distance associated with international travel will not be distorted. At the end, our bilateral travel database contains 152 countries and identifies country level O-D volume for 1.28 billion international visits, about 94.0% of the total 1.37 billion international visits during that year. In comparison to McKercher et al. (2008), our data set has two advantages as (1) each country is studied as both departure country and destination country, revealing different factors that navigate the inbound and outbound travel distance, and (2) it includes countries with a wider range of characteristics with different levels of economic status, social development, and tourism significance. A better representation to the global travel pattern is provided.

Total inbound/outbound travel distance by states is then calculated by multiplying inbound/outbound visitor numbers with distances between O-D pairs. Distances between countries are calculated as the Euclidean distance between capital cities. Once total distance per country is computed, average distance is derived by dividing total inbound/outbound distance with respective visitor numbers for that state. Overall, two dependent variables are generated for each of 152 countries, including OAD per visitor and IAD per visitor.

We acknowledge that taking the Euclidean distance between capitals leads to estimation errors for the actual trip distances, especially for journeys to those large geographic territories with multiple significant subnational tourism regions, such as United States, Canada, Australia, or Russia. For example, a Canadian resident in Vancouver can incur a long distance trip to Florida or a short cross-border visit to Seattle, which is not directly relevant to the direct distance between two capital cities. However, this procedure is the primary treatment when calculating country to country distance (Huang and Peng, 2012; Khadaroo and Seetanah, 2008). If there are no substantial changes in tourism regions within the country, a consistent treatment of distance over time still reveals a valid pattern for the organic trip distribution.

Independent variable

Economic and sociodemographic indicators have been compiled for departure countries from the World Bank Database (2018), UNDP (2018) and United Nations (2018), including population, land size, GDP per capita (current US$), purchasing power parity (PPT),² population ages 65 and above, education index, and urbanization level (Table 1). Economic indicators are used to proxy the purchasing power of the departure country to counter the economic distance (cost distance) (Hall, 2005). At the same time, sociodemographic factors (age, education, and urbanization) are used to represent the capacity to embrace the perceived distance as segments with older aged travelers, urban dwellers with a higher education level generally reporting themselves as experienced travelers (Ahn and McKercher, 2013; Wong et al., 2016).

Table 1.

Variables considered and definition.

Notation	Description	Source
OAD	Outbound average distance per visitor (kilometer)
IAD	Inbound average distance per visitor (kilometer)
Push factor
gdpp	GDP per capita (current US$)	World Bank
ppp	Purchase power disparity	World Bank
pop65	Population ages 65 and above (% of total)	World Bank
edu	Education index	World Bank
Urban	Pct of urban population	World Bank
Pull factor
TCI	Tourism competitiveness index	WTTC
EPI	Environmental performance index	Yale University
Connectivity
Island	Island country (1 = island, 0 = otherwise)	Wikipedia
Neighbor	Number of land neighbors	Wikipedia
Border	Total length of land borders (km)	Wikipedia
Airport	Number of international airports	ICAO
Port	Number of sea ports	World Port Source
Road	Roadways length	CIA
Connect	Number of countries connected with international flights	ICAO
Flight	Number of international flights	ICAO
Visa	Number of visa-free countries	Passport Index
Migrant	Pct of migrations to the population	United Nation
Trade	International trade	World Bank
Geographic areas	Four dummy variables to represent Africa, America, Europe, and Oceania. Baseline = Asia.	UNWTO

Source: Blanke and Chiesa (2013), ICAO (2018), Passport Index (2018), United Nations (2017), UNWTO (2009-2013), World Bank (2018), and Yale Center for Environmental Law and Policy (2018).

Note: GDP: gross domestic product.

For the destination country, the tourism attractiveness indicator has been compiled. It was first sourced from the Travel and Tourism Competitiveness Report, published by World Economic Forum (WEF) (Blanke and Chiesa, 2013). This WEF index is a composite indicator containing both qualitative and quantitative information from four dimensions: enabling environment, policy and enabling conditions, infrastructure, and natural and cultural resources. The index reports a normalized 1 to 7 point score. Key quantitative variables that were not adopted in year 2013 but later identified as important to individual competitive pillars in 2017 have also added to reflect a more comprehensive consideration. This additional variable is the environmental performance index (EPI), a composite indicator across 10 categories covering environmental health and ecosystem vitality (Yale Center for Environmental Law and Policy, 2018). We felt that the addition of the EPI reflects very well many emerging criteria, such as biodiversity, forest, fishery, air pollution, and climate change, that have been perceived as highly relevant to the tourism competitiveness of a country (Crotti and Misrahi, 2017).

The third set of key variables is to measure the connectivity of one country to the rest of the world. For physical connectivity, geographic differences are captured through island status, number of land neighbors, and border length with land neighbor countries. Transport attributes are related to a country’s capacity to reach another state which is critical to minimize the network distance when air capacity is explicitly measured (Bieger and Wittmer, 2006). Transport connectivity is measured through transport infrastructures, including the number of international airports, number of ports, length of roadways, and service frequency. Due to data availability, only one attribute is available to represent transport frequency, which is “the number of international departure flights” from ICAO (2018). Institutional connectivity is represented using two indicators: passport index (Passport Index, 2018), which describes the number of visa-free countries, and number of countries with direct flight services (ICAO, 2018), to capture the political relationships with others on visa agreements and air transport liberalization. Finally, people’s connectivity is measured through the prevalence of immigration and international trade, captured through the percentage of population that are international migrants, and international trade.

There are other important independent variables associated with the socioeconomic dimension and the transport capacity dimension. These variables have not been considered due to limited data availability and the nature of the aggregate analysis in this study. The criteria are that if the variable contains values for less than 100 countries (approximately two-thirds of 152 countries in the database), this variable will be excluded, such as income inequality index and the road density ratio. Some of the other variables (e.g. exchange rate, cultural distance, colonial ties, and local price level), although available for each country, have not been considered in this study because this information is only workable when O-D pairs are analyzed individually. This study considered aggregated distance by countries for 152 countries where the bilateral travel pattern (a total of 151 × 151 combinations) have not been currently considered.

Analysis

There were in total 18 independent variables identified for the global data set, where a high level of correlation was observed for several pairs of dependent variables. To avoid multicollinearity problems, exploratory factor analysis (EFA) was performed before running the multiple regression statistics. To analyze outbound travel distribution, independent variables associated with push factors from source markets (GDP per capita, PPT, population ages 65 and above, education index, and urbanization) and connectivity variables (island country, number of land neighbors, length of land border, international airports, number of seaports, road length, number of international flights, passport index, number of countries connected with direct flights, immigrant ratio, and international trade volume) were used to extract key components that were seen as important for analyzing the outbound journey. Similarly, EFA was performed to analyze components that are relevant to tourism attractiveness of individual countries (tourism competitiveness index and EPI) and connectivity variables, respectively.

The EFA analysis identified one factor to represent national development, one factor to capture destination tourism attractiveness, and three factors that characterized the connectivity of the country to the world (Table 2). Specially, the connectivity of a country can be evaluated as (1) the “transport capacity” which considers transport infrastructures, transport network density, and transport service frequency, (2) “geographic attributes” which captures the border length and number of land neighboring countries, and (3) “institutional arrangement on people’s movement” that incorporates the number of visa-free countries that a passport is entitled to, and the proportion of international migrants to local population. These two variables in an essence capture the freedom of the cross-border movement, temporally as travelers and permanently as migrants.

Table 2.

Factor loading score for the pull component, push component, and connectivity component.

Factor loading	Component
Pull factor	Factor 1: destinationattractiveness
EPI index	0.961
TCI index	0.961
Push factor	Factor 2: nationaldevelopment
Education level	0.869
GDP per capita	0.865
Purchase power parity	0.847
Pct of population 65+	0.816
Urbanization level	0.768
Connectivity factor	Factor 3: transportcapacity	Factor 4: topographicattribute	Factor 5: peoplemovement
No. of airport	0.890	0.119	0.081
No. of sea port	0.883	−0.039	0.165
Road length (km)	0.871	0.146	−0.190
Trade value	0.867	0.208	0.063
No. of international flight	0.824	0.153	0.425
No. of countries connected with flights	0.592	0.386	0.559
No. of land neighbors	0.185	0.901	−0.128
Island country	0.012	−0.854	−0.097
Total length of land borders (km)	0.471	0.644	−0.405
No. of visa-free country	0.251	−0.088	0.701
Pct of migrants to population	−0.067	−0.063	0.698

Note: EPI: environmental performance index; TCI: tourism competitiveness index; GDP: gross domestic product.

In order to ascertain how travel distance relates to the macrolevel determinants, we modeled two distance variables as a function of the explanatory factors. Two equations were specified in the following. For outbound travel, four EFA factor scores representing national development and national connectivity from the origin state were used. For inbound travel, four EFA factor scores representing tourism attractiveness and national connectivity for the destination country were also used. In this study, we also included four dummies in the equation to account for the clustering effect by five continents where Asia is the baseline. The regional effect through dummies is assumed to be strong because not only is international travel highly influenced by regional geomorphological, socioeconomic, and political factors but also reflected upon the effective tourism exclusion zones (ETFZs), proposed by McKercher and Lew (2003). The size of ETFZ is different for individual countries. However, regions nearby may face a similar size of ETFZ as their travels are constrained to a large extent by geographic barriers (oceans or deserts), or destinations that are less appealing.

Outbound travel

OAD = β_{0} + β_{1} national development + β_{2} transport capacity + β_{3} geography + β_{4} people movement + β_{5} Afirca + β_{6} American + β_{7} Europe + β_{8} Ocenia + ε_{i}

Inbound travel

IAD = β_{0} + β_{1} tourism attractiveness + β_{2} transport capacity + β_{3} geography + β_{4} people movement + β_{5} Africa + β_{6} American + β_{7} Europe + β_{8} Ocenia + ε_{i}

For the international travel distance analysis, spatial dependences between national development, tourism attractiveness, and three connectivity factors were considered. For example, with the development of globalization, the interconnection of economic activities across countries has resulted in a high correlation for economic fluctuation within the neighboring countries. When one country thrives, it has a positive impact on the nearby regions due to bilateral trades. Similarly, not only do independent variables have spatially dependences, it is very likely the dependent variables, distance attributes, will be spatially correlated. This is because regional countries may exert a competitive or collaborative effect on visitors’ destination choice, indirectly reflecting on the distance attributes (Marrocu and Paci, 2013).

To measure spatial correlations, the first step is to establish a W matrix (spatial weights) that specifies the neighboring relationship with each other. One way to define if countries are neighbors is (1) if they share a common land board, or (2) if their minimum distance is 200 or fewer kilometers (Gleditsch and Ward, 2001). The neighbors’ weights are assumed to be equal and proportional to 1 over the total number of connectivity. The matrix W is nonnegative and describes the spatial structure of observed countries with diagonal elements being 0.

Moran’s statistic gives a formal indication on the degree of linear association between a value and a weighted average of its neighbors (Moran, 1950a; Moran, 1950b). Higher Moran’s I statistic indicates that the geographical clustering is stronger, that is, the tourism competitiveness of neighboring countries is similar. These geographical dependences can be examined for dependent variables as well as for independent variables. The general model assumed here is a spatial model containing a spatially lagged-dependent variable model, as well as lagged-independent variables in the following form

Y = ρ W Y + X β + W X θ + ε, ε \sim N (0, σ^{2} I)

where W is an given weight matrix, X is a matrix of explanatory variables, ρ, $β = {[\begin{matrix} β_{0} & β_{1} β_{2} \end{matrix}]}^{T}$ , and $θ = {[\begin{matrix} θ_{1} & θ_{2} \end{matrix}]}^{T}$ are unknown parameters and need to be estimated. The spatial lag ρ associated with a positive parameter value indicates that countries are expected to have a higher distance attribute (e.g. OAD) if, on average, their neighbors report a high distance attribute (OAD) as well. This model is referred to as the spatial Durbin model (SDM) (Anselin, 1988; Arbia, 2006). If ρ and ϑ are not significantly different from zero, this reduces the SDM equation back to the original one that can be directly estimated through the ordinary least squares (OLS) method.

Results

Total and average trip distances by inbound and outbound perspectives

Global trip distance patterns of inbound total distance (ITD), IAD, outbound total distance (OTD), and OAD are first plotted on a world map to showcase and contrast the country and regional differences (Figures 1 to 4). Countries are marked with a light yellow to a dark brown that is a direct ratio to the calculated number. In terms of total distances traveled, regions that record the largest ITD associated with their inbound visits are North America, East Asia, Australia and New Zealand, and West Europe, and the top four subregions of OTD for outbound visits are North America, Australia and New Zealand, West Europe, and East Asia. For the country specific pattern, inbound tourists to China, United States, Hong Kong SAR, Italy, and France generated the largest aggregated trip distances, and the top five countries whose residents incurred the most overseas distances were the United States, China, United Kingdom, Germany, and Hong Kong. A symmetric pattern seems to exist as those regions/countries whose inbound arrival distances were substantial are also the ones that produce a high outbound travel distance impact. In addition, a geographic cluster pattern was observed. By and large, most of the countries with large ITD were located in East Asia and West Europe, while countries with a smaller ITD were found in the Middle East and South Africa.

Figure 1.

Inbound total distance by country, 2013.

Figure 2.

Outbound total distance by country, 2013.

Figure 3.

Inbound average distance by country, 2013.

Figure 4.

Outbound average distance by country, 2013.

In term of average distance traveled, big players are shifting to countries with island status in isolated geographic areas that are far away from major tourism destinations/origins. The top four subregions with highest distance engaged per inbound visit were Australia and New Zealand, Polynesia, South Asia, and Melanesia. For countries that are located in Oceania, the location makes them only accessible through long-distance flights. This is especially true for international visits to Australia and New Zealand which have incurred about threefold the distance than the global average. For South Asia, the reason is not directly linked with its location but rather on its demand structure as they embrace a predominantly inbound market from North America and West Europe. For the OAD, Australia and New Zealand, Polynesia, Melanesia, and Western Africa report the largest values. Countries within the Western Africa region are generally poorly developed states that incur very limited outbound visits, but within these states a large proportion of foreign trips are to China (from Guinea and Mali), and to the United States (from Cabo Verde). The “big fat trail” on the trip distribution curve makes them rank high on average distance traveled.

When plotting total distances against average distance, some interesting patterns have emerged (Figure 5). For North America and East Asia, although they receive a high volume of total inbound distances, their visitors travel on an average of 3200 km, very close to the global average per journey. Countries within Europe, on the other hand, attract foreign visitors predominately from the neighboring EU region, leading to a very short average distance per count even though these regions harbor a high volume of international visits annually. In the meantime, a worrying pattern was identified for Australia and New Zealand as currently both countries are high up in the ranking for absolute total distances and average distance per visit, reflecting a strong tourism demand (double-digit growth in recent years) and an unavoidably long journey to these destinations. A similar pattern also exists for outbound journeys from Australia and New Zealand.

Figure 5.

The distribution of total travel distance and average travel distance by 18 geographic regions.

Key drivers for the average travel distance globally

SDM is adopted to analyze the determinants for OAD and IAD as the Jarque–Bera test, Shapiro–Wilk normality test, and Breusch–Pagan test for the OLS model were all significant, leading to the rejection of the two assumptions of normality and homoscedasticity of the OLS model. In other words, there exists spatial dependence observations among distance attributes, and the error term is no longer independent of one another and, as a result, the standard error of βs may be underestimated (Bivand et al., 2013).

For OAD, national development and transport capacity have positive and significant influences after controlling for regional differences. This result, in essence, confirms that national development (hypothesis 1) and transport capacity (hypothesis 3) not only facilitates the departure volume but also the average distance per journey, leading to a global expansion for total outbound distances (Table 3). The average outbound distance is highly region specific. Countries in Oceania and Africa reported a much higher foreign departure distance than countries located in Europe, which is about 1300-km shorter than the baseline for Asian travelers. However, there was still a spatial interdependence among the error term in the SDM model, implying that factors identified do not fully capture the spatial autocorrelation among OAD.

Table 3.

Regression results for outbound total distance and outbound average distance.

	B	Standard error	Standardized Beta	Sig.
DV: Outbound average distance
(Constant)	2320.3	2320.3		0.000
national development	457.5	172.8	0.28	0.009
transport capacity	467.5	104.5	0.29	0.000
geographic attribute	−138.3	113.5	−0.09	0.225
people movement	−172.7	142.1	−0.11	0.226
W_national development	107.9	256.7	0.04	0.675
W_transport capacity	−163.4	201.0	−0.05	0.417
W_topographic attribute	−31.5	169.1	−0.01	0.852
W_people movement	−178.7	266.8	−0.06	0.504
Africa	988.9	327.0	0.24	0.003
America	19.4	294.0	0.01	0.948
Oceania	2861.2	455.8	0.44	0.000
Adjusted R ² = 0.540, F test = 13.801**, Wald test = 0.048, LR test = 0.038, and LM test = 2.9842
DV: Inbound average distance
(Constant)	3436.8	250.5		0.000
destination competitiveness	466.3	251.0	0.23	0.065
transport capacity	176.8	146.0	0.09	0.228
geographic attribute	−465.2	153.4	−0.23	0.003
people movement	−379.7	208.3	−0.19	0.070
W_destination competitiveness	383.1	414.3	0.11	0.357
W_transport capacity	−595.9	292.9	−0.16	0.044
W_topographic attribute	5.9	227.9	0.00	0.980
W_people movement	−625.5	401.9	−0.18	0.122
Africa	965.9	450.0	0.19	0.034
America	−374.8	384.1	−0.08	0.331
Oceania	1455.9	614.8	0.18	0.019
Europe	−1715.4	411.6	−0.37	0.000
Adjusted R ² = 0.427, F test = 9.558*, Wald test = 0.054, LR test = 0.045, and LM test = 3.9693

Note: W_variable captures the spatial inter-relationship among neighboring countries of that variable; ***: 0.001; **: 0.01; *: 0.1; LR test: Likelihood Ratio test; LM test: Lagrange Multiplier test.

For IAD, destination competitiveness, geographic attributes, and institutional arrangement regarding people’s movement are significant factors. Destinations with strong tourism competitiveness tend to receive a larger IAD, implying a weaker distance decay effect, which is a relatively higher proportion of their inbound arrivals from distant source markets, compared to low tourism competitiveness states. This confirms hypothesis 2. With respect to the connectivity factors, geographic attributes, such as land country with more neighbors in the meantime, will decrease IAD, as strong bilateral travel is supported with easy accessibility across borders. People movement, especially the level of migrant in the country reduces IAD, implying that migration ties are formed among countries closer to each other which then encourages shorter distance journeys.

The SDM model also indicates that the space-dependent transport capacity (−595.9) is significant. The negative signs from this SDM model supports the hypothesis that a country’s expected IAD would be lower if its neighbors have a higher transport capacity. From the transport perspective, a competitive effect between destinations is detected here as when one country’s transport capacity is strengthened and specific source markets are targeted, it reduces the share from the same source markets received from its neighboring destinations.

Overall, three proposed factors were found significant in influencing the two distance attributes that were evaluated. When comparing the standardized beta values across factors, tourism competitiveness and geographic attributes are most influential on the IAD while national development and transport capacity share equal weights based on their strong influences over OAD.

Key drivers for the average travel distance by five continents

We further analyzed the data, individually, across five continents to identify the key factors that influence IAD and OAD at the regional level. When continents were considered separately, distant factors began to exhibit their influences, explaining the unique contextual factors based on geographic areas (Table 4). In terms of the determinants of IAD, transport capacity exhibits positive influences on inbound travel distances for Europe and Africa. In addition, a competitive pattern existed when one country enhances its transport capacity, it decreased IAD for the neighboring country, which was evident for both Asia and African countries. People movement concerning migration policy had a negative influence over IAD in Asia but has a positive influence in Oceania. This is well explained by the global migration pattern that around 90% of migrants from Asian countries stay in the region while more than half of immigrants to Oceania (mainly Australia and New Zealand) are from Asia (United Nations, 2017). The intracontinent migration pattern decreases IAD while intercontinent migration boosts IAD.

Table 4.

Factors that influence average travel distance across five continents.

Variables	Asia	Africa	America	Europe	Oceania
Outbound travel distance
national development				+	+
transport capacity	+	+	+	+
geographic attribute	−		+	−
people movement				−
w_national development
w_transport capacity		−		−
w_geographic attribute
w_people movement		−
Inbound travel distance
tourism competitiveness
transport capacity		+		+
geographic attribute	−
people movement	−				+
w_tourism competitiveness
w_transport capacity	−	−
w_geographic attribute		+
w_people movement		−

Note: Factors that have positive and significant influences on average travel distance are marked using the “+” symbol while those with negative and significant influences are noted with the “−” symbol.

In terms of the determinants of OAD, the connectivity factor of the transport capacity is dominant and consistently observed for the majority of countries across Asia, Africa, North America, and Europe. An enhancement of transport capacity undoubtedly leads to spatial expansion in OAD, enabling residents to reach further destinations. However, a competitive pattern also emerged when one country enhances its transport capacity, it decreases OAD for neighboring countries which was evident for European and African countries. This may be explained by the fleet allocation choices for each air route, and which decision is constrained by the fleet capacity at a given time period. In other words, one country’s enhancement in its international flight frequency which are offered by domestic and foreign carriers through increased connection to various destinations globally, is at the expense of other countries, as they are only able to secure less resources over international routes, fleets, and aviation infrastructures at the same time.

Discussion and conclusion

A sustainable tourism transport future requires management and intervention on the distance attribute in order to deliver a climate and social desirable future (Peeters et al., 2018). While the traditional tourism demand model focuses primarily on the income and price influences, few researchers has explicitly examined the underlying mechanism over the aggregated distance attribute in travel (Morley et al., 2014; Song and Li, 2008). This article contributes to the literature by providing a better understanding regarding the macrolevel factors that directly influence the distance attributes of international travels from origin, destination, and connectivity perspectives. An analysis of the global database of 152 countries in 2013 confirms the positive influences of “national development of economic and social wellbeing” on outbound travel distance and “tourism competitiveness” on the inbound travel distance. In addition, transport capacity critically determines the aggregated travel distance for both inbound and outbound journeys while geographic attributes and the freedom of people’s movement via visa and migration control is only significant for selected subregions.

In terms of future outlook, we expect to see unavoidable longer distances for international journeys under the current course of action for the next decade. The International Monetary Fund (IMF, 2018) has projected that the world’s average per capita GDP to increase by 4.9% annually with accompanied social changes on aging, women’s empowerment, urbanization, and the provision of a better education. All economic and social factors portray a global society that desires unique and different travel destinations with increasing affordability. At the same time, 22,000 new aircrafts will enter the market by the year 2036, doubling the current fleet size of 20,500 in 2017 (Airbus, 2017). Compounded with the increasing size of the aircraft and a better connectivity across regions due to open-skies policies, aviation capacity is expected to strengthen as countries that are connected directly with international flights will increase across the globe (Airbus, 2017). When national development and aviation capacity are progressing, both factors affect the visits (a volume problem) and average distance traveled per international journey (an intensity problem).

An interesting observation from the results is that institutional arrangement regarding people’s movement over visa and migration openness of a country is significantly determining its inbound distance traveled, especially in Asia and Oceania. For countries in Asia, more migrants accepted will attract a higher volume of inbound visits nearby while the migration pattern in Oceanic will induce cross-continental travels, leading to a higher IAD. This pattern supports the linkage between travel distance and the regional geopolitical system. In addition, a spatial interdependent and competitive pattern of transport capacity is observed within continents. When one country enhances its transport capacity, it creates negative influences on its neighboring countries for their ability to attract visitors from distant source markets. This implies, on global averages, destinations are substitutes of each other, instead of being complementary, with respect to securing inbound visitors.

This strong spatial-dependent effect of travel distance has implications for both modeling and policy formulation. For future studies on travel distance, spatial econometric models are useful to identify dependences in travel patterns and to untangle influencing factors, echoed the call from Marrocu and Paci (2013) and Supak et al. (2015). In addition, our study supports the validity of “connectivity of a country,” as traditionally these attributes are easily ignored by the travel demand model (Morley et al., 2014). This implies that modeling average trip distance from the panel data set is the most desirable as it takes a systematic perspective to consider interdependent relationship between infrastructure, logistics, policy, and people mobility for any given geographic cluster.

From a policy perspective, we find opportunities that exist to rein in the spatial expansion of international travel by the regions. First, Europe provides a role model to be benchmarked. Their international trip distribution, both inbound and outbound, is highly left-skewed as more than 65% of their international travel is within the range of 1000 km with only a small proportion of long-distance journeys. In contrast, the trip distributions for Asia, American, Africa, and Oceania travelers are more widely spread geographically: a relatively small first peak within the range of 1000–2000 km and a fat tail with an obvious preference for the “lure of distance.” The desired performance of relatively short-distance travel by European residents is contributed to by the cohesive transportation network, cultural similarity, integrated business system, and liberalization in mobility across borders among all member states (Cardoso and Ferreira, 2000; Euromonitor Research, 2014). This model of an efficient integration among EU members provides a learning lesson that a cohesive geopolitical system will facilitate short-haul travels, which is a sustainable travel behavior. We argue that this European model can be further developed in Southeast Asia, a region that has been forecast to experience the largest tourism growth over the next two decades (UNWTO, 2017). With the rise of the ASEAN, this region has the best opportunity to integrate infrastructure, technology, logistics, regulation, and mobility to facilitate intra-region travel, instead of intercontinental journeys.

For other regions, policy intervention for aviation capacity can be thought of because our results indicate that transport capacity exhibits a much stronger influence on the spatial expansion of journeys than others. A carrot policy can be pursued by subsidizing and liberalizing airline operations for short-distance O-D pairs in the continent while a stick policy that aims at taxing long-haul routes and control of the freedoms of air will discourage flight volume and direct international routes. A distance-dependent tax system has been implemented in United Kingdom where a departure tax for flights over 6000 miles is about seven times the tax for distances less than 2000 miles (Daniel et al., 2012). Similarly, in Germany, passengers are charged €8 for flights up to 2500 km, €25 for flights of 2500–6000 km, and €45 for flights longer than 6000 km. By adjusting attributes of aviation capacity, this provides a restructure opportunity to facilitate short-distance travels while discouraging cross-continental journeys.

There are several areas worth further exploration. First, a detailed analysis regarding the substitution of destinations through different transportation modes is needed. Especially, the potential for leveraging low cost carriers (LCCs) as a menu to effectively reduce overall trip distances should be evaluated. LCC are found to intensify travel volumes but they are generally positioned for short-distance destinations. The trade-off between travel frequency and average distance per trip with respect to total trip distances annually and their impact on transport emissions deserves a global inspection. Second, specific attention should be given to different policy designs that can constrain the continuing expansion of international aviation in revenue passenger-kilometers, especially for hot spots. In this research, four hot spots were identified, including the United States and China for their enormous inbound and outbound travel volumes, and Australia and New Zealand for their distant location with blossoming tourism demand. These are all large countries with geographic or political limitations from forming regional geopolitical systems that will facilitate short-haul travel. Mechanisms toward regulating transport capacity may become the next feasible route. Thus, the implementation of the Carbon Offsetting and Reduction Scheme for International Aviation for international aviation of these countries should be closely monitored to see whether this market approach is an effective measure. Finally, we need to acknowledge that several important region-specific aspects in play are not currently captured in the proposed model due to data limitations. These include cultural similarity and differences (Ahn and McKercher, 2013; Yang et al., 2018), colonial ties (Etzo et al., 2014), or open-skies policies (Airbus, 2017). Taking these factors into consideration in the future further advances our understanding of relevant policy aspects. In addition, the observed results are supported for year 2013 in which the global economy continued to expand at a slow pace despite improved global financial conditions and reduced short-term risks. Once economic, political, and societal factors change at the macrolevels, they may lead to different travel distance patterns.

Footnotes

Acknowledgements

The authors would like to thank anonymous reviewers and Professor David Weaver and Professor Christian Laesser for their helpful comments to the previous version of this article.

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) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ya-Yen Sun

Notes

References

Ahn

McKercher

(2013) The effect of cultural distance on tourism: a study of international visitors to Hong Kong. Asia Pacific Journal of Tourism Research 20(1): 94–113.

Airbus (2017) Growing Horizon—year 2017–2036. Blagnac Cedex: Airbus.

Anselin

(1988) Spatial Econometrics: Methods and Models. Berlin: Springer Science & Business Media.

Arbia

(2006) Spatial Econometrics: Statistical Foundations and Applications to Regional Convergence. Heidelberg: Springer Science & Business Media.

ASEAN (2016a) ASEAN open skies a reality. Available at: http://asean.org/asean-open-skies-reality/ (accessed 6 July 2016).

ASEAN (2016b) Master Plan on ASEAN Connectivity. Jakarta: ASEAN.

Balli

Cebeci

(2013) Impacts of exported Turkish soap operas and visa-free entry on inbound tourism to Turkey. Tourism Management 37: 186–192.

Balli

Jean Louis

(2016) The impacts of immigrants and institutions on bilateral tourism flows. Tourism Management 52: 221–229.

Bieger

Wittmer

(2006) Air transport and tourism—perspectives and challenges for destinations, airlines and governments. Journal of Air Transport Management 12(1): 40–46.

10.

Bivand

Pebesma

Gómez-Rubio

(2013) Hello World: Introducing Spatial Data. Applied Spatial Data Analysis With R. New York: Springer New York, pp. 1–16.

11.

Blanke

Chiesa

(2013) The Travel & Tourism Competitiveness Report 2013. Geneva: World Economic Forum.

12.

Cardoso

Ferreira

(2000) The effects of European economic integration on tourism: challenges and opportunities for Portuguese tourism development. International Journal of Contemporary Hospitality Management 12(7): 401–409.

13.

Cohen

Gössling

(2015) A darker side of hypermobility. Environment and Planning A: Economy and Space 47(8): 1661–1679.

14.

Craigwell

(2007) Tourism Competitiveness in Small Island Developing States. Bridgetown, Barbados: Research paper, UNU-WIDER, United Nations University (UNU).

15.

Crotti

Misrahi

(2017) The Travel & Tourism Competitiveness Report 2017. Geneva: World Economic Forum.

16.

Crouch

Ritchie

JRB

(1999) Tourism, competitiveness, and societal prosperity. Journal of Business Research 44: 137–152.

17.

Czepkiewicz

Heinonen

Ottelin

(2018) Why do urbanites travel more than do others? A review of associations between urban form and long-distance leisure travel. Environmental Research Letters 13(7): 073001.

18.

Daniel

Hall

Gossling

(2012) Tourism and Climate Change: Impacts, Adaptation and Mitigation. Milton Park: Taylor and Francis.

19.

Dwyer

Kim

(2003) Destination competitiveness: determinants and indicators. Current Issues in Tourism 6(5): 369–414.

20.

Eldridge

Jones

(1991) Warped space: a geography of distance decay. The Professional Geographer 43(4): 500–511.

21.

Etzo

Massidda

Piras

(2014) Migration and outbound tourism: evidence from Italy. Annals of Tourism Research 48: 235–249.

22.

Euromonitor Research (2014) Datagraphic: impact of EU integration on tourism. Available at: https://blog.euromonitor.com/2014/02/datagraphic-impact-of-eu-integration-on-tourism.html (accessed 1 July 2018).

23.

European Commissions (2015) International aviation. Available at: http://ec.europa.eu/transport/modes/air/international_aviation/country_index/united_states_en.htm (accessed 1 July 2018).

24.

Feng

Page

(2000) An exploratory study of the tourism, migration–immigration nexus: travel experiences of Chinese residents in New Zealand. Current Issues in Tourism 3(3): 246–281.

25.

Gleditsch

Ward

(2001) Measuring space: a minimum-distance database and applications to international studies. Journal of Peace Research 38(6): 739–758.

26.

Hall

(2005) Tourism: Rethinking the Social Science of Mobility. London: Pearson Education.

27.

McKercher

(2012) A comparison of long-haul and short-haul business tourists of Hong Kong. Asia Pacific Journal of Tourism Research 19(3): 342–355.

28.

Huang

Peng

(2012) Fuzzy Rasch model in TOPSIS: a new approach for generating fuzzy numbers to assess the competitiveness of the tourism industries in Asian countries. Tourism Management 33(2): 456–465.

29.

IndexMundi (2018) Price level ratio of PPP conversion factor (GDP) to market exchange rate. Available at: https://www.indexmundi.com/facts/indicators/PA.NUS.PPPC.RF (accessed 1 July 2018).

30.

International Civil Aviation Organization (ICAO) (2018) API data service. Available at: https://www.icao.int/safety/iStars/Pages/API-Data-Service.aspx (accessed 1 May 2018).

31.

International Monetary Fund (IMF) (2018) IMF Data Mapper. Available at: http://www.imf.org/external/datamapper/NGDPDPC@WEO/OEMDC/ADVEC/WEOWORLD (accessed 1 May 2018).

32.

Khadaroo

Seetanah

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

33.

Khan

SAR

Qianli

SongBo

, et al. (2017) Travel and tourism competitiveness index: the impact of air transportation, railways transportation, travel and transport services on international inbound and outbound tourism. Journal of Air Transport Management 58: 125–134.

34.

Lee

Guillet

Law

, et al. (2012) Robustness of distance decay for international pleasure travelers: a longitudinal approach. International Journal of Tourism Research 14(5): 409–420.

35.

Lenzen

Sun

Faturay

, et al. (2018) The carbon footprint of global tourism. Nature Climate Change 8(6): 522–528.

36.

Lim

Min

JCH

McAleer

(2008) Modelling income effects on long and short haul international travel from Japan. Tourism Management 29(6): 1099–1109.

37.

Marrocu

Paci

(2013) Different tourists to different destinations. Evidence from spatial interaction models. Tourism Management 39: 71–83.

38.

Massidda

Etzo

Piras

(2014) Migration and inbound tourism: an Italian perspective. Current Issues in Tourism 18(12): 1152–1171.

39.

McKercher

(2018) The impact of distance on tourism: a tourism geography law. Tourism Geographies 1–5. DOI: 10.1080/14616688.2018.1434813.

40.

McKercher

Lew

(2003) Distance decay and the impact of effective tourism exclusion zones on international travel flows. Journal of Travel Research 42(2): 159–165.

41.

McKercher

Chan

Lam

(2008) The impact of distance on international tourist movements. Journal of Travel Research 47(2): 208–224.

42.

Mishra

Bansal

(2017) Role of source-destination proximity in international inbound tourist arrival: empirical evidences from India. Asia Pacific Journal of Tourism Research 22(5): 540–553.

43.

Moran

(1950a) Notes on continuous stochastic phenomena. Biometrika 37(1/2): 17–23.

44.

Moran

(1950b) A test for the serial independence of residuals. Biometrika 37(1/2): 178–181.

45.

Morley

Rosselló

Santana-Gallego

(2014) Gravity models for tourism demand: theory and use. Annals of Tourism Research 48: 1–10.

46.

Pappas

(2014) The effect of distance, expenditure and culture on the expression of social status through tourism. Tourism Planning & Development 11(4): 387–404.

47.

Passport Index (2018) Global passport power rank 2018. Available at: https://www.passportindex.org/byRank.php (accessed 1 July 2018).

48.

Peeters

Higham

Cohen

, et al. (2018) Desirable tourism transport futures. Journal of Sustainable Tourism 1–16. DOI: 10.1080/09669582.2018.1477785.

49.

Peng

Song

Crouch

(2014) A meta-analysis of international tourism demand forecasting and implications for practice. Tourism Management 45: 181–193.

50.

Pitfield

(2009) The assessment of the EU–US Open Skies agreement: the counterfactual and other difficulties. Journal of Air Transport Management 15(6): 308–314.

51.

Prideaux

(2005) Factors affecting bilateral tourism flows. Annals of Tourism Research 32(3): 780–801.

52.

Qian

Law

Wei

(2017) Effect of cultural distance on tourism: a study of pleasure visitors in Hong Kong. Journal of Quality Assurance in Hospitality & Tourism 19(2): 269–284.

53.

Ram

Nawijn

Peeters

(2013) Happiness and limits to sustainable tourism mobility: a new conceptual model. Journal of Sustainable Tourism 21(7): 1017–1035.

54.

Scott

Gossling

Hall

(2012) International tourism and climate change. WIREs Climate Change 3: 213–232.

55.

Seetaram

(2012) Immigration and international inbound tourism: empirical evidence from Australia. Tourism Management 33(6): 1535–1543.

56.

Shan

Wilson

(2001) Causality between trade and tourism: empirical evidence from China. Applied Economics Letters 8(4): 279–283.

57.

Song

(2008) Tourism demand modelling and forecasting—a review of recent research. Tourism Management 29(2): 203–220.

58.

Supak

Brothers

Bohnenstiehl

, et al. (2015) Geospatial analytics for federally managed tourism destinations and their demand markets. Journal of Destination Marketing & Management 4(3): 173–186.

59.

UNDP (2018) Education index. Available at: http://hdr.undp.org/en/content/education-index (accessed 1 July 2018).

60.

United Nations (2017) International Migration Report. New York: United Nations.

61.

UNWTO (2009–2013) Tourism Statistics. Madrid: World Tourism Organization.

62.

UNWTO (2011) Tourism Towards 2030 - Global Overview. Madrid: UNWTO.

63.

UNWTO (2017) Tourism Highlights 2017. Madrid: World Tourism Organization.

64.

Weaver

Oppermann

(2014) Tourism Management. Hoboken: John Wiley and Sons.

65.

Winston

Yan

(2015) Open Skies: estimating travelers’ benefits from free trade in airline services. American Economic Journal: Economic Policy 7(2): 370–414.

66.

Wong

Fong

LHN

Law

(2016) A longitudinal multilevel model of tourist outbound travel behavior and the dual-cycle model. Journal of Travel Research 55(7): 957–970.

67.

World Bank (2018) World Bank open data. Available at: https://data.worldbank.org/ (accessed 1 July 2018).

68.

World Bank and International Monetary Fund (2016) Global Monitoring Report – Development Goals in an Era of Demographic Change. Washington, DC: World Bank.

69.

WTO-UNEP (2008) Climate Change and Tourism: Responding to Global Challenges. Spain: World Tourism Organization.

70.

Yale Center for Environmental Law and Policy (2018) Environmental Information Index. Available at: https://epi.envirocenter.yale.edu/ (accessed 1 July 2018).

71.

Yan

(2015) Uneven distance decay: a study of the tourism market segments of Hong Kong. International Journal of Tourism Sciences 11(1): 95–112.

72.

Yang

Liu

(2018) The world is flatter? Examining the relationship between cultural distance and international tourist flows. Journal of Travel Research 58(2): 224–240.