The structure of globalized travel: A relational country-pair analysis

Abstract

Utilizing data from the UNWTO, IMF, World Bank, and UNESCO, this article analyzes the global structure of travel and its deep asymmetries, revealing that travel is not global but is highly concentrated among a handful of countries. Furthermore, I find that the effects of globalization are neither universal nor consistent but depend upon the identities of countries involved and their relationships with one another. This article conceptualizes travel as a result of the relationship between country attributes within a given country-pair. More specifically, it investigates the relationship between travel and relative inequalities, institutional connections, and cultural wealth. I find that measures of inequality and cultural wealth differ depending on the relationship between country-pairs while institutional connections are significant across models.

Keywords

Cultural wealth globalization inequality institutional connections travel

Globalization – the increasing interconnectedness among nation-states – alters economic and cultural systems worldwide; however, the nature of these changes remains disputed. Debates about the current era of globalization largely focus on whether it represents a unifying force rendering a ‘flat world’ (Friedman, 2006) or a dividing force that deepens structural inequality (Mahutga, 2006: 1864; Stiglitz, 2002), with researchers primarily assessing the effects of globalization on local and national communities (Bestor, 2000; Dreby, 2010; Raab and Milward, 2003; Scheper-Hughes, 2006; Smith, 2005; Watson, 2006). In order to fully understand globalization’s impact, we must analyze its underlying shape. Empirical work on the shape of globalization and the consequences of nation-states’ structural position in the global economy tends to focus on trade (e.g. Clark and Beckfield, 2009; Mahutga, 2006), international organizations (e.g. Beckfield, 2003, 2010), and currency (e.g. Centeno and Cohen, 2010: 69–71).

This article uses globalized travel as an indicator of global asymmetry. While previous empirical work lays the foundation for understanding globalization’s asymmetry, investigating new indicators and measures further reveals the breadth of globalization and the depth of its inequalities. This article proposes that individual preferences for particular travel destinations necessarily arise from particular relationships between two countries. It examines how travel relationships are associated with country-specific economic and cultural attributes and institutional connections within each country-pair. Examining global inequality through country-pairs can provide invaluable insight into global economic, social, and cultural activity, power structures, the reach of globalization, and the possible future of global integration.¹

This article makes three contributions to the sociological understanding of the structure and processes of globalization. Empirically, it ties together research on the macro travel industry with sociological theories related to globalization and macro-inequality, expanding our empirical knowledge of globalization’s unevenness. Second, while much of macro-oriented literature analyzes inequality vis-à-vis countries’ relative positions within a world system or by their individual resources, this article suggests that inequality between countries is structured by their relative position to one another. Finally, by using global travel as a case study, it seeks to understand how concepts from international political economy, institutionalism and cultural wealth may act as sources of attraction for documented international travel (DIT) and thus, how they expand or limit specific country-pair pathways of globalization.

Approaches to the structure of global inequality

Common approaches to globalization are concerned with the intersections of political and economic processes. For example, a fundamental assumption underlying international political economy and network approaches is that the world is interconnected – whether the focus is on the spread of ideas (e.g. modernization perspective and free-market ideology), in the ‘development of underdevelopment’ within dependency and world-systems perspectives, or the identification of ‘hubs’ in network analyses. Each of these perspectives is interested in the relationship among countries and offer opposing theories regarding the structure of globalization: modernization theorists highlight agency and individual country resources, while network, dependency, and world-systems researchers emphasize structure and countries’ positions in a global system.

Modernization theorists emphasize a given country’s agency and resources (Deudney and Ikeberry, 2009; Inglehart and Welzel, 2009; Przeworski and Limongi, 1997) and see structural inequality decreasing as countries move towards democracy, liberal economic expansion, and social and cultural development. The world is ‘flat’ (Friedman, 2006) in the sense that national leadership is the only barrier to developmental progress. These characterizations emphasize the aggregate benefits produced by increasing globalization and freer trade, including the creation of more jobs, declines in the gender wage gap (e.g. Dollar, 2005), and increased access to the market for the poor (Bhagwati, 2004; Friedman, 2006). In this view, increasing global integration and the spread of freer trade lifts all boats, including those at the bottom (Bhagwati, 2004). Countries not maximizing the resources in which they have a comparative advantage explain global inequality.

International institutions such as the World Bank, the United Nations and the International Monetary Fund tend to follow this perspective, and each of these organizations promotes tourism as a development strategy because theoretically it increases capital – which can be used to pay off foreign debt – builds infrastructure, boosts employment rates and builds a foundation for democracy (Goldstone, 2001). It is an industry that is dependent on exogenous and endogenous factors of a given country, including network formations between private and public sectors, and the multiplicity of actors involved in the negotiation, production, consumption and distribution of tourism goods and services in government, trade, finance, transportation and marketing sectors (Cornelissen, 2005; Milne and Ateljevic 2001). In line with modernization hypotheses, Mexican tourism policies have led to increased export revenues and the creation of regional employment opportunities (Clancy, 2001), while tourism employees tend to have higher employment rates, better-than-average incomes, and more updated amenities (such as electricity and water treatment) compared to non-tourism employees in certain regions in Mexico (Gullette, 2007). Kim et al. (2006) found that tourism and economic development in Taiwan mutually reinforce one another, while Rogerson and Visser (2006) and Yahya (2003) show how intra-regional tourist flows are increasingly becoming more important to the economic development of South Africa and Singapore, which can be seen as evidence of diversifying audiences and decreasing reliance on Western European and American tourists. Additionally, Lanfant (1980) suggests that travel prompts isomorphism across economies, social structures, and cultures (p. 34) thereby introducing new sets of relations at all levels of society; Ireland’s (1993) historical ethnography of Sennen (West Cornwall, UK), supports this by showing how late 19th-century women’s increased involvement in tourism lead to increased economic and social standing outside the home, altering social, familial, and gender relations.

Dependency and world-systems theories, in contrast to modernization theories, highlight structural inequalities, and tend to categorize countries into two² main categories that reflect their global structural position: core countries that control and benefit from the world economy, and peripheral countries that are dependent on the core. Dependency theory is based on the ‘Prebisch-Singer hypothesis’, which suggests that peripheral countries experience a long-term decline in trade due to the relative prices of their primary exports, which are markedly lower than the prices for manufactured exports from core countries (Prebisch, 1950, 1959; Singer, 1949). Dependency theorists assert that market relations in the world economy are ‘unequal because development of parts of the system occurs at the expense of other parts’ (Dos Santos, 1970: 231). Peripheral countries served as satellites by channeling their profits to their respective metropolises in core countries; this left them incapable of self-sufficiency (Frank, 1973; Hubbell, 2008). World-systems analyses take into account history and power and suggest that the world-economy is more than the sum of its parts and should be studied as a series of systems, one that cuts across political, economic and cultural ties (Wallerstein, 2004: 17). Nations ‘act freely but their freedom is constrained by their biographies and the social prisons of which they are a part’ (p. 21); they are ‘a set of nested and overlapping interaction networks that link all units of social analysis’ (Chase-Dunn and Grimes, 1995: 389). Thus, inequality, as measured by position in the global economy, relies not only on history and structure but power and mobility.

Many tourism scholars draw upon these frameworks to suggest that tourism perpetuates developing nations’ dependency on industrialized nations (Crick, 1996: 28), because the industry’s benefits are unequally distributed among different sectors and regions (Cornelissen, 2005). For example, because of its vulnerability the travel industry is prone to travel fads, suffers from high leakage of profits, increases land value, food shortages, and malnutrition, and builds upon stereotypes, myths and fantasies rather than facilitating cross-national exchanges (Crick, 1996: 22–34). Furthermore, foreign industries reap the economic profits (Brown and Hall, 2008: 841) and control travel technology, product pricing, product design, and the image of the destination countries (Britton, 1996: 158), while the local labor force suffers from low wages, excessive hours, and job instability due to the seasonal and temporary nature of the industry (Brown and Hall, 2008: 841). It is an industry that is sensitive to positive and negative shocks in demand (Chan et al., 2005).

Sharing similarities with dependency and world systems analyses, a network approach to globalization focuses on the structural positions of countries, the number of ties, and the presence of hubs (e.g. Barabasi, 2003; Wasserman and Faust, 2009). For example, dependency, world systems and network approaches all locate countries within a global system, and some network scholars use world systems theories to inform their research. Network-based analyses show the persistence of hierarchical positions of countries in trade (e.g. Clark and Beckfield, 2009; Mahutga, 2006; Nemeth and Smith, 1985; Smith and White, 1992; Snyder and Kick, 1979), and the asymmetrical distribution of other types of global flows and ties, including the type of currency used in monetary transactions (Centeno and Cohen, 2010: 69–71), international telephone traffic (Louch et al., 1999), and the exclusive membership of international non-governmental organizations (Beckfield, 2003, 2010).

Network analyses within business management, physical sciences, and the social sciences also show uneven travel patterns when examining website connections – accommodations, agencies, transportation services, and activities – of a tourist destination (Baggio, 2007), firms (Erkus-Ozturk, 2009; Erkus-Ozturk and Eraydin, 2011), countries (Derudder et al., 2007; Greenbaum and Hultquist, 2006; Neal, 2010, 2012; Shih, 2006; Teixeira and Fernandes, 2012), regions (Borocz, 1996; Garin-Munoz, 2004; Van Nuffel et al., 2010), and globally (Derudder et al., 2007; Keeling, 1995; Miguens and Mendes, 2008; Rimmer, 1998; Smith and Timberlake, 2001; Taylor et al., 2007, 2009). Social science research on global travel tends to focus on world city networks (WCN), and in particular, the role of important cities in a national (Derudder et al., 2007; Neal, 2010, 2012; Van Nuffel et al., 2010) and international context (Derudder et al., 2007; Smith and Timberlake, 2001; Taylor et al., 2009). Rimmer (1998) examines the elite ‘top 25’ cities from 1984 to 1992 and finds that the network of airports have changed from a bifurcated system – one in the transatlantic, the other in the transpacific – to a single interconnected system where a ‘Main Street’ links together Europe, North America, and Asia, and other parts of the world are connected as if they were separate ‘cul-de-sacs’ within this system. Using airline passenger data, Smith and Timberlake (2001) map hierarchies of world cities across six time periods and find that although the importance of city airline hubs within Western Europe (e.g. London) and North America (e.g. New York) remain stable over time, the importance of cities within other regions of the world have changed; they identified the rise of importance in East Asian cities (e.g. Seoul) and the diminished importance of cities in Africa (e.g. Johannesburg), Latin America (e.g. Buenos Aires), and South Asia (e.g. Bombay). For a critique of airline passenger data, see Derudder and Witlox (2005). Other research supports the findings of European and North American cities as absolute hubs (Derudder et al., 2007), and the rise of Pacific Asian cores (Taylor et al., 2009) in the WCN.

World-systems approaches and network analyses are based upon relational dynamics by detailing the position of countries in relation to all other countries in the world. World-systems approaches tend to privilege economic inequality over other forms of inequality (e.g. Wallerstein, 2004), while network approaches are not necessarily concerned with the content of ties between countries. This article uses a different relational approach by assuming that economic decisions are shaped by the relationship between two or more parties (e.g. Bandelj, 2002, 2008, 2009; Zelizer, 2005, 2009, 2010). Rather than viewing globalization and global inequality as a cause or consequence of a dichotomous or trichotomous world or city system, this article highlights how globalization is structured by countries’ relative positions to one another vis-à-vis country-pairs. The difference between an international political economy or a network approach and the approach used in this article is the unit of analysis. It is not individual countries or their position in the world system as it relates to others. Rather it focuses on the relationship between country-pairs and how these relationships change depending on the actors in the origin and the destination. For example, this approach assumes that inequality in the relationship between the United States and Mexico is fundamentally different to the inequality in the relationship between France and Mexico or France and Algeria because of the differences in connections between countries within the respective pairs.

Hypotheses

Drawing upon other common approaches in globalization literature that emphasize culture and institutions, this article also emphasizes the variability of what is meant by inequality – it is not just economic, but also institutional and cultural inequalities that shape the structure of globalization. I construct inequality as relational so that a destination has more or less economic or cultural inequality compared to a given destination, and institutional inequalities are based on different types of institutional connections. Although previous work has shown the majority of travel is within and between core countries (e.g. Smith and Timberlake, 2001), and therefore overall travel may not be explained by economic incentives, this article examines four different types of travel: between core countries, between peripheral countries, from core to peripheral, and finally from peripheral to core countries. In analyzing travel in this way, I examine how and if economic incentives are important for certain types of travel over others. I do not use ‘core’ and ‘periphery’ terms in the way that the world-systems literature does; instead, I define ‘core’ and ‘peripheral’ vis-à-vis the number of travel flows, where ‘core’ status correlates with a high volume of flows and ‘peripheral’ countries are those with a relatively low volume of travel flows. These statuses are calculated in two ways – for incoming and outgoing travel flows. Therefore a country may have a ‘core’ status relating to outgoing travel flows, but may also have a ‘peripheral’ status in regards to incoming travel flows.

I expect travel to differ depending on the (a) economic and (b) cultural country attributes of each country and (c) the institutional connections within each type of country-pair. Assuming economic inequality is one reason to travel, I hypothesize (1) that travel between country-pairs that occupy similar core or peripheral positions in the globalized structure of travel are not influenced by relative inequality measures. Since the degree of economic inequality will be relatively similar for each country in their respective country-pair, there is little economic incentive to travel, as previous work suggests. However, I do expect country-pairs of dissimilar positions to have different relationships with inequality measures depending on the structural position of the destination and the origin, and on the type of economic inequality. Differing types of economic inequality can facilitate the amount of documented international travel (DIT) between each country-pair; for example, if we assume that economic measures for within-country inequality relate to the relative price of items, I hypothesize (2) that those traveling from core to peripheral countries will go to countries with higher within-country inequality and industrial development but relatively lower wealth and standards of living because the priority for this type of travel is the ability to visit locales that have luxury goods that are inexpensive when compared to the origin country. Destination countries that are industrially developed and have high amounts of within country inequality could account for this. In contrast, those traveling from peripheral to core countries will have different economic motives. I hypothesize (3) that they will go to relatively wealthier, more industrially developed destinations that do not necessarily have higher standards of living or within-country inequality. I suspect that these types of travelers are the relatively wealthy from the periphery and desire the status that comes with traveling to a wealthier, more developed core country.

Although culture is generally neglected in macro-analyses of global inequality, there are two exceptions. One is Huntington’s (1996) ‘clash of civilizations’ hypothesis which states that deep divisions based on culture will be/is the source of conflict among nation-states; the other is a world systems approach that suggests that ‘culture’ (defined as either a set of characteristics that distinguishes one group from another or as a way of differentiation within any one group) is used as a way to justify and perpetuate existing inequalities (Wallerstein, 1990: 33). What is missing in both these analyses is that culture can also be used as a tool for manipulation and attraction. For example, Rivera (2008) examines how government officials, in order to increase international tourism and improve their economy, reframed Croatia’s culture and history by characterizing it as Western European and removing references to war in travel brochures and reviews. Rivera (2008) uses the emerging perspective of cultural wealth, which combines Bourdieu’s (1984 [1979]) idea of cultural distinction with world-systems theories of inequality. It is defined as the historically and socially constructed ‘intangible qualities of products and services emanating in part from the perceived cultural heritage of the people engaged in their production’ (Centeno et al., 2011: 26). This is by definition relational – it is the judgment from one or more countries about another. Other scholars using this approach show how public narratives shape indigenous participation in the Costa Rican (Wherry, 2007) and Ecuadorian tourism markets (Colloredo-Mansfeld, 2002) and how South African marketing affects within-country travel destinations (Cornellissen, 2005: 110).

Assuming that having cultural wealth attracts tourists, I expect that the relationship between country-pairs and travel to differ depending on the relative amount of cultural wealth for origins and destinations. I expect that core countries have more cultural wealth than peripheral countries, and that higher levels of cultural wealth are associated with higher levels of DIT. In particular, I hypothesize (4) that DIT between countries with core positions to be characterized by movement between places of similar high cultural wealth. I also (5) expect DIT to and from peripheral countries to not be influenced by cultural wealth, since both countries within the country-pair presumably have similar low levels of cultural wealth. However, for country-pairs characterized by unequal positions and presumably unequal amounts of cultural wealth, I expect DIT to be characterized by movement to destinations that have relatively high levels of cultural wealth. More specifically, that (6) travel from the peripheral to the core will be marked by movement to destinations with higher levels of cultural wealth, controlling for all other factors, while (7) DIT from core to peripheral countries will be characterized by travel to those peripheral countries that have relatively higher amounts of cultural wealth than peer peripheral countries.

Cultural wealth and its measures are not value-free. They are helped or hindered by institutional contexts of organizations determining the cultural wealth of countries (e.g. Kowalski, 2011) or the legal, political, and economic context of individual countries (e.g. Gaggio, 2011). Institutions are the ‘formal and informal rules and practices that surround us as we go about our daily lives’ (Duina, 2011: 3), and scholars from an institutionalist tradition emphasize the how historical patterns of social organization affect countries’ competitiveness and opportunities within the global market (Biggart and Guillen, 1999: 723). It focuses on how resources channel individuals and countries into certain pathways and emphasizes how institutions shape behavior (Pierson and Skocpol, 2002: 707; Thelen, 1999: 379), are themselves a legacy of historical processes, are outcomes of negotiation and conflict between social forces (Korzeniewicz and Moran, 2009: 51; Thelen, 1999: 382), and how differentials embedded in organizations and institutions (Pierson and Skocpol, 2002: 699) result in the privileging of some actors and relationships at the expense of others (Biggart and Gullien, 1999: 722). The work on institutions in travel studies is often descriptive in nature – observing or theorizing patterns of colonial legacies as characterized by travelers of previous colonial cores to their respective peripheral sites (e.g. Britton, 1996). Yet less is known about what actually occurs on the macro-level and how non-travel institutional connections between countries affect travel and the general shape of globalization. Similar to the economic and cultural wealth measures used in this article, I expect that the type of institutional connections between country-pairs – not institutions in and of themselves – matters. Using two institutional connections, one that emphasizes historical power relations and one that reflects a more contemporary and theoretically equal relationship, I hypothesize (8) that flows from peripheral countries to core countries and flows from core to peripheral countries are characterized by historical power-based ties and will show how history begets familiarity and institutionalizes patterns of travel. In contrast, I hypothesize (9) that the presence of a theoretically more equal institution will capture contemporary ties among countries with similar structural position, both core-to-core and peripheral-to-peripheral travel.

Data

Country-pair travel flows

DIT flows between country-pairs, which measure the volume of arrivals, serve as my dependent variable, and indicators of economic, institutional and cultural wealth will serve as independent variables. Data on 2006 DIT flows derive from the Tourism Factbook, published by the United Nations World Tourism Organization (UNWTO). The UNWTO is a specialized agency of the United Nations dedicated to the field of DIT. At the time of data collection, the 2006 data represented the most complete information for each country. The UNWTO collects statistics on key trends in travel, including international arrivals and outbound tourism (UNWTO, 2008: 7). The technical definition of an international tourist is ‘an international visitor who stays at least one night in a collective or private accommodation in the country visited’ (UNWTO, 1995: 13), and whose activities are ‘outside their usual environment for not more than one consecutive year for leisure, business and other purposes’ (UNWTO, 1995: 12). It is any person ‘whose main purpose of [the] trip is other than the exercise of an activity remunerated from within the place visited’ (UNWTO, 1995: 17).³ Diplomats, members of the armed forces, nomads, refugees, prisoners, international transit, and same-day visitors crossing borders^{4
5} are excluded from these statistics (UNWTO, 1995: 81–83).

Inequality

Previous studies have shown that structures of globalization are related to economic status (e.g. Clark and Beckfield, 2009). This article analyzes the relationship among economic, institutional, and cultural wealth inequalities and travel flows. National inequality is measured in four ways: (1) GDP per capita in US dollars to reflect the wealth of a nation’s population, (2) the Gini coefficient to compare within-country inequality, (3) the Human Development Index (HDI), which takes into account health, knowledge (education), and standard of living, and (4) industrial development, measured by percent of manufacturing products in GDP.⁶ Although the HDI includes the gross national income per capita, I use it as a separate measure of inequality because the measure also accounts for life expectancy at birth, mean years of schooling and expected years of schooling. A variance inflation factor (VIF) of less than ten and a mean VIF of less than 5 suggests that multicollinearity between GDP per capita and HDI is low enough to be unproblematic and that each indicator represents a distinct aspect of inequality. These measures are lagged one year in order to be able to predict 2006 travel flows. See Table 1 for descriptive statistics.

Table 1.

Descriptive statistics of tourist arrivals, economic inequalities, institutional connections, and cultural wealth measures.

Variable	Obs	Mean	Std. Dev.	Min	Max
Arrivals	41987	18623.91	274173.40	0.00	17500000.00
Origin characteristics
GDP ($1000s) per capita	34226	9149.10	13782.33	106.81	73147.06
HDI	31974	0.63	0.18	0.16	0.93
Gini	30740	40.86	9.89	24.70	74.61
% Manufacturing	32381	14.01	7.32	0.48	40.00
Cultural sites	41985	3.40	6.24	0.00	42.00
Travel guides	41987	2.93	13.73	0.00	186.00
Population (millions)	40762	32.16	127.75	0.00	1314.48
% Tourism	29921	7.46	10.37	0.03	71.35
High out degree centrality	41987	0.07	0.26	0.00	1.00
Low out degree centrality	41987	0.93	0.26	0.00	1.00
Destination characteristics
GDP ($1000s) per capita	34234	9116.28	13755.50	106.81	73147.06
HDI	31990	0.63	0.18	0.16	0.93
Gini	30766	40.83	9.87	24.70	74.61
% Manufacturing	32399	14.00	7.30	0.48	40.00
Cultural sites	41987	3.39	6.23	0.00	42.00
Travel guides	41987	2.92	13.70	0.00	186.00
Population (millions)	40763	31.98	127.30	0.00	1314.48
% Tourism	29952	7.45	10.35	0.03	71.35
High in degree centrality	41987	0.09	0.29	0.00	1.00
Low in degree centrality	41987	0.91	0.29	0.00	1.00
Country pair connections
Embassies	30790	0.23	0.42	0.00	1.00
Region – same	41987	0.11	0.31	0.00	1.00
Colonial ties	41987	0.18	0.39	0.00	1.00
Type of country pairs
High out degree centrality–high in degree centrality	41987	0.01	0.08	0.00	1.00
High out degree centrality–low in degree centrality	41987	0.07	0.25	0.00	1.00
Low out degree centrality–high in degree centrality	41987	0.09	0.28	0.00	1.00
Low out degree centrality–low in degree centrality	41987	0.84	0.37	0.00	1.00

Source: Data on travel flows come from UN World Tourism Organization, Yearbook of Tourism Statistics, 2006, data on inequality measures are from the International Monetary Fund (GDP per capita), Human Development Indicators (United Nation’s Development Programme), and the World Bank’s World Development Indicators Index (gini coefficient and percent manufacturing of GDP). Institutional links derive from the Mapping Globalization Project (empire and embassy ties), while cultural wealth measures derive from the United Nations World Heritage Center (World Heritage cultural sites) and 4 travel websites (Foders, Lonely Planet, Michelin and Fromers).

Institutional connections

This article examines two specific institutional connections: colonial and embassy ties. Among others, world-systems scholars have shown that the legacies of colonization continue to influence contemporary societies (e.g. Chase-Dunn and Grimes, 1995; Wallerstein, 2004). Since the formal and informal rules and practices of colonialism continue to have influence, I use colonial ties as a measure of institutional connections between two countries. Colonial influence is not one-way, from the colonizer to the colony; rather, Steinmetz (2008) shows how colonial states competed for and constructed ethnographic capital – ‘a reciprocally recognized talent for making judgments of the colonized’ (p. 596). Although the colonizer’s influence penetrated the colonized more deeply, both countries have knowledge and familiarity with one another. The act of colonialism created a bridge between the two countries. In contrast to colonial ties, embassies may reflect contemporary political and economic affairs, and capture the informal and formal rules of state relationships, one distinct from colonization and more contemporary; it serves as another useful indicator for institutional connections between countries. Data on colonial links between 18th- and 19th-century empire cores and their respective colonial outposts, and on embassy ties derive from Princeton University’s Mapping Globalization Project (see Table 1 for descriptive statistics). The measure for colonial ties is not directional; it just measures whether there is a colonial link between the two. The measure for embassies is directional, and is coded as ‘1’ if the origin country has an embassy in the destination country.

Cultural wealth

I draw upon theories of cultural wealth that suggest there are national cultural hierarchies (e.g. Rivera, 2008; Wherry, 2007) and apply them at an international level. Two measures are used: the number of United Nations-certified cultural World Heritage sites and the existence of a major trade travel guidebook. UN World Heritage cultural sites represent locations around the globe internationally recognized as having significant cultural value, and although these sites may represent political agendas, in theory, they are the democratic outcome of a global debate on cultural wealth. The number of travel guidebooks is a continuous measure. Guidebooks included are country-specific (including cities and regions within a single country). Regional or multi-country guides are not counted. Four travel guides were coded from their websites: Fodors, Lonely Planet, Michelin (Green Guides), and Frommers⁷ (Complete Guides). Although these are Western European- or North American-based, they represent some of the most popular travel guides in the world. However, I recognize the limitation of this approach because popular regional travel guides are not included. This article analyzes the association, rather than the causal link, between travel guidebooks and travel flows. Travel guidebooks are an important indicator of cultural wealth because travelers use them to pinpoint and assess travel destinations. These guidebooks also direct tourism to certain destinations over others and therefore reflect global locations ‘worthy’ of visiting (see Table 1 for descriptive statistics). Examining the cultural wealth of the origin country is also crucial, since this provides a basis of comparison. Do travelers from countries with more cultural wealth travel to countries with less or equal amounts of cultural wealth? Does the inverse hold true?

Controls

For control variables, I use population and percentage of tourism of GDP per capita for both the origin and the destination, and whether or not the origin and the destination are both located in the same region. Data are collected from the International Monetary Fund, the Mapping Globalization Project, and the World Tourism Organization. The population measure derives from 2006 data to ensure that the phenomena observed are not simply due to population density. The region variable is used in order to account for travel due to geographical ease and proximity. I control for the effect of geographic proximity by using a dichotomous regional indicator that reflects whether or not countries within a given pairs are located within the same region. These regions include: North America and the Caribbean,⁸ Central America,⁹ South America,¹⁰ Oceania,¹¹ East and Southeast Asia,¹² Central Asia,¹³ Middle East,¹⁴ North Africa and the Horn,¹⁵ Southern Africa,¹⁶ Western Europe and the Mediterranean,¹⁷ and Eastern Europe.¹⁸ Drawing upon modernization theories that emphasize internal state factors, the extent of a given country’s tourism infrastructure may indicate whether or not the state has the resources to support inbound international travel. To account for this infrastructure, I use the percentage of tourism per GDP per capita in 2006, since it measures goods and services produced within countries.

I calculated variance inflation factors (VIF) in order to check for multicollinearity between variables. The results of VIF<10 and mean VIF<5 indicate that each indicator is measuring a distinct component of each concept. As one of the first sociological studies to look at the global structure of DIT using the UNWTO tourist data, the results presented are derived from the raw numbers of arrivals and indicators.

Methods

Centrality divide

To identify different connections of globalized travel and disentangle the structural position of each country within the country-pairs, I use UCINET (Borgatti et al., 2002) to inductively divide countries by degree centrality. There are two ways to calculate degree centrality: through the number of ties or the value of ties. In this article, degree centrality reflects the value of ties. Those countries with high degree centrality have a high volume of travel flows. Three other popular measures of centrality include closeness (how close an actor is to all others), betweenness (whether an actor is between links that connect other actors), and information (the amount of information within all paths that originate from an actor). In order to focus on the broad structure of travel I use degree centrality to identify those ‘most active’ in the network – those with the highest volume of travel flows, when compared to other actors (Wasserman and Faust, 2009). I use dyads and distinguish ties between origin and destination, and identify out- and in-degree centrality for each country. For this analysis, those with high in- or out-degree centrality are considered core countries while those with low in- or out-degree centrality are considered peripheral countries. Actors with high out-degree centrality have ties to many other actors and are said to be ‘influential’ while actors with high in-degree centrality are ‘prestigious’ because they receive many ties (Hanneman and Riddle, 2005). In this article, I use the terms ‘influential’ and ‘prestigious’ to refer to the volume of ties.

To examine the different patterns of globalized travel, countries are divided into two categories: high or low in-degree centrality (HIDC or LIDC) and high or low out-degree centrality (HODC or LODC). Next, each country-pair is evaluated on whether the origin is a ‘HIDC’ or a ‘LIDC’ and whether the destination is a ‘HODC’ or a ‘LODC’. For example a HODC-LIDC country-pair would indicate travel from a core country (HODC) to a peripheral country (LIDC). Both ‘high’ out- and in-degree centrality is defined by a country having travel flows equal to or higher than one standard deviation above the mean (>= .40 and >= .34 respectively). Fifteen countries have a high out-degree centrality, while nineteen countries have high in-degree centrality.¹⁹ All other countries are classified as low out- or in-degree centrality respectively. The designation of one, rather than two, standard deviations above the mean is because very few countries have higher degrees of centrality. Therefore, the results in this article are a conservative estimate as they encompass many countries that may otherwise be considered as having low out- and in-degree centrality. Additionally, these designations do not correspond to the qualitative notions of core and periphery found in the world-systems literature; instead ‘core’ and ‘periphery’ refer to the volume of incoming and/or outgoing travel flows.

Count models

Using zero-inflated negative binomial regression models, I estimate the relationships between inequality, institutional connections, cultural wealth, and travel arrivals per country-pair in 2006. I first run the models on the entire dataset, with the omitted category being low out-degree to low in-degree travel flows since travel within this type of flow account for the vast majority of country-pairs. Regression results will differ depending on the omitted category. I then subdivide the data by type of travel flow (high out-degree to high in-degree, high out-degree to low in-degree, low out-degree to high in-degree, and low out-degree to low in-degree centrality) and run the models on each type. Generating data subdivisions, rather than including an interaction variable, allowed for an interaction effect with each variable. This allowed me to take a relational, country-pair approach to examine which variables are significant for particular types of flows, and to examine LODC-LIDC pairs, which were previously used as the omitted category. For both regression analyses I conducted a Wald test, which is one way to test the goodness of fit of regression models. The Wald test produces a chi-squared value, included in the regression tables, and a p-value associated with the chi-squared. P-values for both types of regression analyses were 0.000. Since the subdivided data allows for more information and has a p-value of 0.000, the Wald test suggests that doing so significantly improves the fit of the model to the data.

The sample is limited to countries with data on both sending and receiving tourists. I assume that blanks within the dataset are zeros. As I cannot separate empirical zeros from structural zeros (missing data) and because the count data is non-linear with a significant amount of overdispersion, I use zero-inflated negative binomial regression models. This approach takes into account excess zeros and assumes that they can be modeled independently. As a robustness check, I calculated the results using multiple imputation on my independent variables, which yielded similar patterns. Additionally, I disaggregated data that were aggregated by region and continent level flows by population and ran the models with both raw numbers and multiple imputation. I also calculate the results by dropping all zeros and using negative binomial regression models, which also takes into account overdispersion but does not model excess zeros. The results all exhibit similar patterns (results available upon request).

To account for standard errors and because travel arrivals are not independent, I cluster the models by country of origin. By exponentiating the coefficients, the outputs can be interpreted as the percent change in travel between country-pairs associated with a one unit increase in the independent variable. In a sample, the p-values represent whether the results reflect a random distribution, with a value of p < .05 suggesting 95% confidence that the independent variable has a relationship with the dependent measure. I have the full population of travel flows so significance values are used to indicate the strength of an association.

Structure of globalized travel

As with other forms of cross-national exchange (e.g. Centeno and Cohen, 2010: 69–71; Louch et al., 1999), there is a concentration of flows among a small number of countries. I find that only two country-pairs account for almost 5 percent of DIT, five country-pairs for almost 10 percent, 92 for 50 percent, and 948 country-pairs for 90 percent.²⁰ Since there are 41,987 country-pairs in the dataset, this means that 2.26 percent of all country-pairs account for 90 percent of global travel.²¹ Clearly the distribution of travel does not conform to a normal distribution, but reflects a deep asymmetry, where travel is concentrated among a small number of countries. Figure 1 shows the wide gulf between the small number of country-pairs that account for a disproportionate share of travel and the overwhelming majority of country-pairs that have relatively few travelers.

Figure 1.

Distribution of documented international travel (DIT) flows by country–pair.

Examining the top 18 country-pairs that account for over 25 percent of world tourism, I find that global tourism is in fact an overwhelmingly high out-degree to high in-degree centrality phenomenon, and these countries also correspond to the World Bank’s classification of high-income economies²² (see Table 2). Additionally, when examining overall travel by each type of flow (HODC to HIDC, HODC to LIDC, LODC to HIDC and LODC to LIDC) I find that all HODC to HIDC flows account for 43 percent (335,000,000) of DIT while LODC to LIDC accounts for only 18 percent (141,000,000). Travel flows between countries that occupy dissimilar positions (HODC to LIDC and LODC to HIDC) have roughly similar low numbers of DIT (19% or 151,000,000 and 20% or 155,000,000 respectively). Globalized travel is concentrated among a handful of countries, with almost one-half of all travel occurring between countries that occupy core positions in both influence (out-degree centrality) and prestige (in-degree centrality).

Table 2.

List of country–pairs that comprise the top 25% of world tourism.

Origin	Destination	Arrivals	Percent	Cumulative percent
United States	Mexico	17512050	2.24%	2.24%
United Kingdom	Spain	16210327	2.07%	4.31%
Canada	United States	15992242	2.05%	6.36%
Germany	France	13878000	1.77%	8.13%
United States	Canada	13855468	1.77%	9.90%
Mexico	United States	13317000	1.70%	11.61%
China	Macao	11985655	1.53%	13.14%
Germany	Spain	10139586	1.30%	14.44%
Germany	Austria	10101194	1.29%	15.73%
Singapore	Malaysia	9656251	1.23%	16.96%
France	Spain	9081957	1.16%	18.12%
Germany	Italy	8982845	1.15%	19.27%
Romania	Hungary	8650871	1.11%	20.38%
China	Hong Kong	8434300	1.08%	21.46%
Slovakia	Hungary	7967608	1.02%	22.48%
Italy	France	7682000	0.98%	23.46%
Netherlands	France	7554000	0.97%	24.43%
Hong Kong	Macao	6935586	0.89%	25.31%

Source: UN World Tourism Organization, Yearbook of Tourism Statistics (2006).

Note: All figures are raw numbers.

Table 3 displays the actual and expected figures of travel flows by the centrality divide, normalized by population of the origin country. Although empirical research demonstrates that global flows are not randomly distributed (e.g. Beckfield, 2003, 2010; Centeno and Cohen, 2010; Clark and Beckfield, 2009; Louch et al., 1999; Smith and Timberlake, 2001), a comparison between theoretical randomness and the actual data can illustrate the extent to which travel is concentrated. If travel is random or has a high degree of dispersion then I would expect approximately 71 million travelers to countries with a high number of in-degree ties (HIDC) since their population represents 9 percent of the global population. Likewise, country-pairs whose destination has a low number of in-degree ties (LIDC) should have approximately 711 million travelers since they comprise over 91 percent of the total population. What I find, however, is that DIT to countries with high in-degree centrality accounts for almost two-thirds of total arrivals, and almost seven times the expected amount of DIT. This concentration bear repeating: only 9 percent of country-pairs account for two-thirds of globalized travel.

Table 3.

Expected versus actual travel by centrality divide, normalized by population of origin.

	Expected travel	% Expected travel	Actual travel	% of actual travel	Actual/expected
High in–degree centrality	70,945,036	9%	490,000,000	63%	6.91
Low in–degree centrality	711,055,025	91%	292,000,000	37%	0.41

Note: The expected numbers of travel by destination were obtained by generating the proportion of the population of origin countries (1,310,910) multiplied by the total number of arrivals (782,000,000).

Source: UN World Tourism Organization, Yearbook of Tourist Statistics (2006).

These simple exercises demonstrate the structural asymmetry of travel flows. These flows are not global but highly concentrated within countries that occupy a core position in the globalized travel structure, a pattern similar to that found in trade and other types of global flows. Although global travel in and of itself may be limited to those with high income, an examination of the structure of its patterns and uncovering the degree to which it is concentrated within a handful of countries is important.

Explaining the volume of travel flows

This article goes beyond describing the asymmetry that exists in DIT flows by examining processes that underlie this unequal structure of attraction. It is important to investigate the systemic ways these country attributes and institutional connections help shape travel preferences and how they differ depending on country-pair.

Table 4 presents the results from a more traditional analysis of integrating economic, institutional and cultural inequalities and their relationship with globalized travel by including centrality measures as separate variables into the regressions. Results show that different types of economic inequalities are important; however, colonial ties, embassy ties, and cultural wealth are all significant for globalized travel. Model 1 is run on origin characteristics, Model 2 adds destination variables, Model 3 incorporates institutional connections, and Model 4 includes the types of flows. Overall, results suggest that inequality by itself is only marginally related to DIT flows on a country-pair level; however, as there is an unequal distribution of flows, many of these findings may be driven by a single case. After including the relational and flow indicators, the only significant inequality measures for the origin country are HDI (+) and Gini coefficient (-), while only percent manufacturing (+) is significant for the destination (model 4). Since the Gini coefficient represents the degree of within-country inequality, the HDI coefficient indicates human, or social, development, and percent manufacturing indicates industrial development, these numbers suggest that overall, people who travel tend to come from countries with relatively lower within-country inequality as well as higher levels of education, life expectancy, and standards of living. They choose locations that have relatively high industrial development, and thus, infrastructure. In examining globalized travel as a whole, I find that those who benefit from the social and cultural interactions and experiences that derive from travel are those from relatively privileged countries.

Table 4.

Unstandardized coefficients from zero–inflated negative binomial regressions of 2006 tourist flows on relative inequalities, institutional links and cultural wealth.

	(1)	(2)	(3)	(4)
	Origin	Destination	Relational	Flows
Origin characteristics
GDP ($1000s) per capita	0.000	–0.000	–0.000	–0.000
	(0.000)	(0.000)	(0.000)	(0.000)
HDI	3.528^***	1.276	2.389^*	2.361^**
	(0.939)	(0.925)	(0.969)	(0.908)
Gini	–0.005	–0.010	–0.034^***	–0.028^**
	(0.012)	(0.012)	(0.010)	(0.009)
% Manufacturing	0.037^**	0.047^***	0.011	0.013
	(0.014)	(0.011)	(0.012)	(0.011)
Cultural sites	0.031^**	0.030^*	0.031^***	0.021^**
	(0.011)	(0.012)	(0.007)	(0.007)
Travel guides	0.007^***	0.010^***	0.010^***	0.008^***
	(0.002)	(0.001)	(0.002)	(0.002)
Population (millions)	0.000	–0.000	0.000	0.000
	(0.000)	(0.000)	(0.000)	(0.000)
% Tourism	–0.080^***	–0.094^***	–0.075^***	–0.078^***
	(0.016)	(0.016)	(0.015)	(0.013)
Destination characteristics
GDP ($1000s) per capita		0.000	–0.000	0.000
		(0.000)	(0.000)	(0.000)
HDI		1.849	0.367	0.275
		(1.055)	(0.604)	(0.592)
Gini		0.013	0.004	0.009
		(0.012)	(0.008)	(0.007)
% Manufacturing		0.054^***	0.041^***	0.022^*
		(0.009)	(0.009)	(0.010)
Cultural sites		0.082^***	0.076^***	0.036^***
		(0.012)	(0.013)	(0.009)
Travel guides		0.007^**	0.012^***	0.004^**
		(0.002)	(0.001)	(0.002)
Population (millions)		–0.001^***	–0.001^***	–0.000
		(0.000)	(0.000)	(0.000)
% Tourism		–0.018	–0.004	0.005
		(0.013)	(0.012)	(0.011)
Country pair connections
Embassy			2.400^***	2.351^***
			(0.166)	(0.163)
Region – same			1.934^***	1.939^***
			(0.146)	(0.145)
Colonial ties			1.088^***	1.108^***
			(0.136)	(0.133)
Type of country pairs
High out–degree centrality-high in–degree centrality				1.882^***
				(0.231)
High out–degree centrality-low in–degree centrality				0.513^**
				(0.191)
Low out–degree centrality-high in–degree centrality				1.140^***
				(0.208)
_cons	7.844^***	6.430^***	5.448^***	5.254^***
	(0.773)	(0.755)	(1.095)	(1.033)
Inflate
Arrivals	–48.670^***	–51.075^***	–49.510^***	–51.732^***
	(0.228)	(0.259)	(0.262)	(0.262)
_cons	26.460^***	27.549^***	26.655^***	27.796^***
	(0.097)	(0.098)	(0.099)	(0.099)
Lnalpha
_cons	1.565^***	1.406^***	0.950^***	0.904^***
	(0.022)	(0.038)	(0.045)	(0.043)
Alpha
_cons	4.781	4.080	2.584	2.470
	(0.106)	(0.155)	(0.117)	(0.105)
Wald chi²	581.300	1135.450	3111.310	4403.930
N	23773	13451	13106	13106
Robust Std. Err. in parentheses

p < 0.05; ** p < 0.01; *** p < 0.001.

Note: Negative binomial regressions are used because tourist flows involve count data with a significant amount of overdispersion. Low out–degree centrality flows (LODC) to Low in–degree centrality (LIDC) flows are the omitted category. The _cons coefficient is the estimate when all the variables in the model are evaluated at zero. The _alpha coefficient is the model’s dispersion parameter; if it is significantly greater than zero than a negative binomial model is a more appropriate model than a poisson. The lnalpha is the estimate of the log of the dispersion parameter, alpha.

Source: Data on travel flows come from UN World Tourism Organization, Yearbook of Tourism Statistics (2006), data on inequality measures are from the International Monetary Fund (GDP per capita), Human Development Indicators (United Nations Development Programme), and the World Bank’s World Development Indicators Index (Gini coefficient and percent manufacturing of GDP). Institutional links derive from the Mapping Globalization Project (empire and embassy ties), while cultural wealth measures derive from the United Nations World Heritage Center (World Heritage cultural sites) and four travel websites (Foders, Lonely Planet, Michelin and Fromers).

Holding everything else constant, embassy and colonial ties are both strongly and positively significant for DIT. These results show that history matters. Colonial ties beget familiarity with cultures and institutions, represent previously established pathways of travel, and encompass familial relationships that prompt travel between those locations. All of these factors suggest that familiarity and institutional connections facilitate travel. As I analyzed cross-sectional and not longitudinal data, the embassy data are harder to disentangle – do embassy relationships facilitate tourism or is it reverse causation: large amounts of travel prompt embassies ties? Although this question of causation is important to disentangle, what is observed is the very high association between this institutional connections and high DIT flows.²³

Additionally, in both origin and destination countries, both the existence of a country-specific travel guidebook and the presence of a World Heritage cultural site are positively associated with travel. This suggests that, overall, travelers tend to come from countries with cultural wealth, and they also tend to travel to destinations that contain cultural sites deemed worthy of global recognition from the UN and from the manufactures of travel guidebooks. Similar to the embassy data, the data from travel guidebooks are cross-sectional and I cannot infer causality – whether increases in the number of travel guidebooks prompts or follows travel; the selection process and consequences of possessing ‘cultural value’ should be further explored to understand the structural nature of cultural recognition. I find that when compared to LODC-LIDC, all other types – HODC-HIDC, HODC-LIDC, and LODC-HIDC – are positively associated with increases in travel; however, this tells us very little about whether the dynamics that occur within these types of flows are similar or dissimilar. It only suggests that a country within one of these flows is more highly associated with increases in travel.

Relational, country-pair analysis

A relational country-pair analyses of globalized travel is interested in how relationships among DIT and the types of economic, institutional, and cultural inequalities differ depending on the context of the relationship; that is, it determines whether or not and how the identity (e.g. a ‘core’ or ‘peripheral’ country) of the origin and the destination matter. Results suggest that identities within country-pairs do matter. I find that different types of economic and cultural inequalities matter for different types of flows, but both types of institutional connections matter regardless.

Model 5 are results from HODC-HIDC flows, Model 6 is run on HODC-LIDC travel, Model 7 from LODC-HIDC, while Model 8 is based on LODC-LIDC flows. Supporting my first hypothesis, none of the inequality measures in the origin or the destination matter for globalized travel within country-pairs of similar structural position (HODC-HIDC/core to core; LODC-LIDC/peripheral-peripheral). Increases in travel between core country-pairs such as United States to Mexico, France to Canada, or Italy to China and between peripheral country-pairs including Korea to Japan,²⁴ Kenya to Uganda or New Zealand to Australia are not driven by economic inequalities but something else – institutions and culture.

Results suggest that cultural wealth in the form of both travel guidebooks and the presence of a UN World Heritage site matters for HODC-HIDC and LODC-LIDC flows, but the way it matters differs depending on country-pairs. For example, people from core countries in HODC-HIDC flows come from influential countries with cultural wealth vis-à-vis travel guidebooks, but not World Heritage sites, and select prestigious destinations with high cultural wealth in both measures. In contrast, people traveling within LODC-LIDC flows tend to originate in non-influential countries with high amounts of cultural wealth in the form of both travel guidebooks and World Heritage sites, and select non-prestigious destinations that contain cultural world heritage sites, but not travel guidebooks. In contrast, both types of institutions – colonial and embassy ties – matter in similar magnitude and significance for both types of flows (see Table 5).

Table 5.

Unstandardized coefficients from zero–inflated negative binomial regressions of 2006 tourist flows on relative inequalities, institutional links and cultural wealth by country–pair.

	(5)	(6)	(7)	(8)
	High out–degree centrality–high in–degree centrality	High out–degree centrality–low in–degree centrality	Low out–degree centrality–high in–degree centrality	Low out–degree centrality–low in–degree centrality
Origin characteristics
GDP ($1000s) per capita	0.000	0.000	–0.000^***	0.000
	(0.000)	(0.000)	(0.000)	(0.000)
HDI	–1.725	–1.916	6.191^***	0.916
HDI	(11.483)	(4.070)	(0.827)	(1.072)
Gini	–0.061	–0.030	–0.042^***	–0.019
Gini	(0.042)	(0.024)	(0.012)	(0.010)
% Manufacturing	0.003	0.002	0.016	0.023
% Manufacturing	(0.043)	(0.024)	(0.013)	(0.013)
Cultural sites	–0.003	0.019^*	0.048	0.050^*
Cultural sites	(0.012)	(0.008)	(0.033)	(0.020)
Travel guides	0.012^**	0.009^**	0.060^*	0.049^*
Travel guides	(0.005)	(0.003)	(0.026)	(0.023)
Population (millions)	–0.000	–0.001	0.001	–0.000
Population (millions)	(0.001)	(0.001)	(0.001)	(0.000)
% Tourism	–0.254^***	–0.180^***	–0.089^***	–0.061^***
% Tourism	(0.072)	(0.040)	(0.014)	(0.014)
Destination characteristics
GDP ($1000s) per capita	–0.000	0.000^**	0.000	0.000
GDP ($1000s) per capita	(0.000)	(0.000)	(0.000)	(0.000)
HDI	–0.731	–0.400	–9.327^**	0.294
HDI	(2.788)	(0.483)	(2.984)	(0.802)
Gini	0.018	–0.000	0.028	0.011
Gini	(0.038)	(0.007)	(0.037)	(0.008)
% Manufacturing	0.018	0.050^***	–0.112^***	0.017
% Manufacturing	(0.022)	(0.010)	(0.032)	(0.013)
Cultural sites	0.023^***	0.134^***	0.051^***	0.058^*
Cultural sites	(0.007)	(0.027)	(0.008)	(0.025)
Travel guides	0.011^**	0.150^***	0.010^**	0.024
Travel guides	(0.004)	(0.022)	(0.004)	(0.019)
Population (millions)	–0.000	–0.001	–0.000	–0.001
Population (millions)	(0.000)	(0.001)	(0.000)	(0.001)
% Tourism	–0.039	0.053^***	0.228^**	–0.011
% Tourism	(0.039)	(0.013)	(0.082)	(0.013)
Country pair connections
Embassy	2.285^***	1.604^***	1.646^***	2.372^***
Embassy	(0.413)	(0.093)	(0.237)	(0.215)
Region – same	2.364^***	1.384^***	2.190^***	2.080^***
Region – same	(0.251)	(0.284)	(0.217)	(0.193)
Colonial ties	1.661^***	1.099^***	0.992^***	1.253^***
Colonial ties	(0.416)	(0.201)	(0.157)	(0.173)
_cons	13.154	9.696^***	12.883^***	5.100^***
_cons	(7.350)	(2.407)	(3.343)	(1.157)
Inflate
Arrivals	–0.112^***	–14.269^***	–20.902^***	–51.253^***
Arrivals	(0.004)	(0.426)	(0.511)	(0.264)
_cons	22.840^***	23.760^***	23.729^***	27.520^***
_cons	(0.474)	(0.352)	(0.108)	(0.103)
lnalpha
_cons	0.114	0.412^***	0.634^***	0.997^***
_cons	(0.118)	(0.106)	(0.054)	(0.047)
alpha
_cons	1.121	1.509	1.886	2.710
_cons	(0.133)	(0.160)	(0.101)	(0.128)
Wald chi²	.	.	1755.650	1183.120
N	244	1466	1693	9703
Robust Std. Err. in parentheses

p < 0.05; ** p < 0.01; *** p < 0.001.

Both types of institutional connections also similarly matter for county-pairs that consist of structurally dissimilar positions: HODC-LIDC and LODC-HIDC. These institutional connections build enduring bridges within country-pairs regardless of their respective structural position. That these connections matter suggests that the presence of a bridge is important; however, the results do not say anything regarding the materials used (e.g. the content of these relationships), or how they were built (e.g. coercion or more friendly facilitation). For example, these data cannot speak to the qualitative differences between the relationship from Brazil to Canada (LODC-HIDC) or from the Netherlands to Egypt (HODC-LIDC), both of which have no colonial ties but do have institutional connections vis-à-vis embassies. Nor can it speak to the qualitative differences between travel within US and Philippines or Spain and Mexico country-pairs, both of which have embassy ties but also deep colonial histories. Knowledge of all three: that there is an institutional connection, how this connection was built, and the means with which it was, is important in order to holistically examine individual country-pair relationships. This article provides a first step in analyzing these ties at a macro-level, uncovering the magnitude and significance of both types of ties (see Table 5).

For structurally dissimilar country-pairs, cultural wealth is also positively and significantly important. For HODC-LIDC pairs, such as France to Egypt, which has six cultural UN World Heritage sites and five travel guides, or Switzerland to Australia, which has 15 different travel guides and two cultural UN World Heritage sites, both travel guidebooks and the presence of a World Heritage site are significant in the destination. They are also both positively significant in the origin, suggesting that travelers from places of high cultural wealth choose other places with relatively high cultural wealth as destinations. For LODC-HIDC pairs, such as Costa Rica (which has zero cultural sites but six travel guides) to Canada (which has six World Heritage sites and 32 travel guides), as well as from Norway (which has six World Heritage sites and two travel guides) to Greece (which has eight travel guides and 15 World Heritage sites) both travel guidebooks and World Heritage sites are important for the destination. However, for the origin, having a travel guidebook, but not a World Heritage site, is significant. These measures represent two different types of cultural wealth, and although both types of cultural wealth in the destinations are significant, for travelers from peripheral countries having a popular country-specific travel guidebook represents a more significant type of cultural wealth for their country compared to containing UN World Heritage sites.

Although I have found similar patterns in institutional connections and cultural wealth between similar (HODC-HIDC, LODC-LIDC) and dissimilar country-pairs (HODC-LIDC, LODC-HIDC), this is not the case when examining economic inequality measures. While the former has no economic inequality measure that is significant, the latter do. For influential core to non-prestigious peripheral pairings (HODC-LIDC), GDP per capita and percentage manufacturing are both positively significant for the destinations, but there are no significant economic inequality measures in the origin. Travelers from core countries choose peripheral countries that have higher overall wealth and higher industrial development than their peers. Their high economic and industrial development may compensate for their low degrees of influence, because these destinations presumably have similar amenities offered in their country of origin. However, for non-influential peripheral to prestigious core county-pairs, HDI and percentage manufacturing are both negatively significant for the destinations, while GDP per capita and the Gini coefficient are both negative and HDI is positive for the origin country. These flows are characterized by movement from peripheral countries with relatively high human development, low within-country inequality, but relatively low overall wealth to core destinations marked by low industrial and social (human) development than their peers. One interpretation suggests that the wealthy of the peripheral travel to core countries, and they have little concern for economic inequality.

Discussion

Conclusions

Travel is one facet of the much larger, complex phenomenon of globalization. Analyzing these flows contributes to ongoing discussions concerning structural and economic inequality, illustrates the enduring patterns of institutional connections and cultural wealth across the globe, and uncovers how globalization comprises many processes, each with differential effects depending on the identity of their respective ties.

I expand our general knowledge on the structure of globalization by incorporating travel flows. Globalized travel reflects global power structures and shows the limited reach of globalization. This article highlights the degree to which globalized travel is highly concentrated within and between countries with high in- and out-degree of centrality – those that are both highly influential and prestigious, and adds to the literature on globalized travel (e.g. Rimmer, 1998; Smith and Timberlake, 2001) by using UN World Tourism Organization data and focusing on countries rather than cities.

This article also expands work on how examining the content of one-to-one relationships matters for macro-level analyses (Bandelj, 2002, 2009). While much of network and globalization literature analyzes inequality vis-à-vis countries’ relative positions within a world system or by their individual resources, this article suggests that inequality between countries is structured by their relative position to one another. Additionally, rather than focusing on economic inequality, it brings together three broad traditions of research that examine globalization – international political economy, institutionalism, and cultural wealth – in order to demonstrate how globalization is a web of interconnected country-pairs that is defined by relational and country-specific attributes.

Running the analyses with country-pair types as variables to the dataset shows the significance of both types of institutional connections and for both types of cultural wealth measures in the origin and in the destination. By including multiple economic variables for both the origin and the destination it shows how those who travel globally are those from already relatively privileged social and economic countries. However, by running the analyses through datasets focused exclusively on each type of travel flow, I am able to uncover a more nuanced pattern of how relative economic and cultural inequality are connected to travel and how globalized travel is contextual, depending on the identity of the origin and destination. For example, although both types of institutional connections continue to remain important, by looking at the multiple cultural wealth variables by type of country-pairs I find that for certain destinations, one type (World Heritage sites) is important across flows, while the other (travel guidebooks) is not, but for origin countries, travel guidebooks are important across flows, while World Heritage sites are not. Furthermore, I find that similarly positioned countries do not have significant relationships with economic inequality measures in the origin or the destination, but economic inequality matters in HODC-LIDC and LODC-HIDC flows, and how it matters differs for each type. These more nuanced relationships between variables and type of travel flows can only be explored when the analyses is run by country-pair, essentially allowing an interaction between each variable and each type of flow. It illustrates that the effects of globalization are neither universal nor consistent but depend upon the identities of countries involved and their relationships with one another.

Limitations

A major task for further work in macro-examinations of global phenomena is creating methodologies that address both the over-determined nature of these systems (since they are determined by a relatively small number of country-pairs or countries), and the idiosyncratic nature of each nation-state’s history, geography, economic systems, and their position in the global economy. Although this research is important for furthering the research on travel and global inequality at the macro-level, it does not produce models that capture both the over-determinism and idiosyncrasies within this system. While micro-, country- and region-specific research within the travel and globalization literatures addresses these complexities more completely, they need to more fully account for the characteristics of each flow, and how they differ depending on a given country’s position in the global economy. Integrated research of macro- and micro-phenomenon are necessary to further disentangle these complex relationships. Additionally, this study lacks longitudinal data, which are crucial for predicting and analyzing the ways these relationships change over time. This article provides only the first step by describing the global structure of travel vis-à-vis country-pairs and its connection to country-pair attributes.

Future research

Although this article addresses some concerns about the global structure of the travel industry, it also raises avenues for future research, including an examination of how country-pair relationships and their attributes change over time. For instance, has travel become more or less centralized in particular country-pairs over time? What accounts for these changes or lack thereof? Other research, such as investigating barriers to travel, such as visa or passport requirements, is essential for more thoroughly understanding the movement of globalized travel. Future research should also address the reasons for fluctuations in the effects of income inequality by country-pair. One way to do so is to include different measurements of income inequality, including purchasing power parity (PPP) and/or consumer price index (CPI). Although country-specific research addresses some of these questions, a macro-look into the global structure is needed. Future research should also seek to better understand what distinguishes different forms of travel. For example, is travel from Mexico to the United States significantly different from that of Germany to France, given that both types of flows are from countries with high out-degree centrality to countries with high in-degree centrality? More creative ways to refine and expand our knowledge of this increasingly important industry is necessary as it has consequences for individual nation-states’ GDP, employment rates, trade, and national images, as well as the global flows of people and goods. Additionally, since article examines only existing associations between country-pairs and their attributes, an important next step would be an analysis of would-be ties between non-connected country-pairs in order to be able to analyze causality or the determination of ties and travel flows.

Footnotes

Acknowledgements

I thank Miguel Centeno for his mentorship throughout this project, and Martin Ruef for his methodological insights from conception to completion. I also thank John Boli, Jason Beckfield, Christopher Chase-Dunn, Joseph Cohen, and Fredrick Wherry for their feedback on very early versions of the article, and the participants in the 2009–2010 Empirical Seminar at Princeton University for constructive criticism as I was beginning this research. Additionally, I thank David A. Smith and the anonymous reviewers for their generous feedback throughout the review process.

Funding

This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. DGE-0646086.

Notes

References

Baggio

(2007) The web graph of a tourism system. Physica A 379: 727–734.

Bandelj

(2002) Embedded economies: Social relations as determinants of foreign direct investment in core and Eastern Europe. Social Forces 81(2): 411–444.

Bandelj

(2008) From Communists to Foreign Capitalists: The Social Foundations of Foreign Direct Investment in Postsocialist Europe. Princeton, NJ: Princeton University Press.

Bandelj

(2009) The global economy as instituted process. American Sociological Review 74(1): 128–149.

Barabasi

A-L

(2003) Linked: How Everything is Connected to Everything Else and What it Means for Business, Science and Everyday Life. Cambridge: Plume.

Beckfield

(2003) Inequality in the world polity: The structure of international organization. American Sociological Review 68(3): 401–424.

Beckfield

(2010) The social structure of the world polity. American Journal of Sociology 115(4): 1018–1068.

Bestor

(2000) How sushi went global. Foreign Policy 121: 54–63.

Bhagwati

(2004) In Defense of Globalization. Oxford: Oxford University Press.

10.

Biggart

Guillen

(1999) Developing difference: Social organization and the rise of the auto industries of South Korea, Taiwan, Spain and Argentina. American Sociological Review 64(5): 722.

11.

Borgatti

Everett

Freeman

(2002) Ucinet for Windows: Software for Social Network Analysis. Cambridge, MA: Analytic Technologies.

12.

Borocz

(1996) Leisure Migration: A Sociological Study on Tourism. New York: Pergamon Press.

13.

Bourdieu

(1984 [1979]) Distinction: A Social Critique of the Judgement of Taste. Cambridge, MA: Harvard University Press.

14.

Britton

(1996) Tourism, dependency and development: A mode of analysis. In: Apostolopoulos

Leivadi

Yiannakis

(eds) The Sociology of Tourism: Theoretical and Empirical Investigations. London: Routledge, pp. 155–172.

15.

Brown

Hall

(2008) Tourism and development in the Global South: The issues. Third World Quarterly 29(5): 839–849.

16.

Centeno

Cohen

(2010) Global Capitalism: A Sociological Perspective. Cambridge: Polity.

17.

Centeno

Bandelj

Wherry

(2011) Political economy of global cultural wealth. In: Bandelj

Wherry

(eds) The Cultural Wealth of Nations. Stanford, CA: Stanford University Press, pp. 23–46.

18.

Chan

Lim

McAleer

(2005) Modelling multivariate international tourism demand and volatility. Tourism Management 26(3): 459–471.

19.

Chase-Dunn

Grimes

(1995) World–systems analysis. Annual Review of Sociology 21: 387–417.

20.

Clancy

(2001) Mexican tourism: Export growth and structural change since 1970. Latin American Research Review 36(1): 128–150.

21.

Clark

Beckfield

(2009) A new trichotomous measure of world–system position using the international trade network. International Journal of Comparative Society 50(1): 5–38.

22.

Colloredo–Mansfeld

(2002) An ethnography of neoliberalism. Current Anthropology 43(1): 113–137.

23.

Cornelissen

(2005) The Global Tourism System: Governance, Development and Lessons from South Africa. Burlington, VT: Ashgate.

24.

Crick

(1996) Representations of international tourism in the social sciences: Sun, sex, sights, savings, and servility. In: Apostolopoulos

Leivadi

Yiannakis

(eds) The Sociology of Tourism: Theoretical and Empirical Investigations. London: Routledge, pp. 15–51.

25.

Derudder

Witlox

(2005) An appraisal of the use of airline data in assessing the world city network: A research note on data. Urban Studies 42(13): 2371–2388.

26.

Derudder

Devriendt

Witlox

(2007) Flying where you don’t want to go: An empirical analysis of hubs in the global airline network. Tijdschrift voor Economische en Sociale Geografie 98(3): 307–324.

27.

Derudder

Witlox

Taylor

(2007) U.S. cities in the world city network: Comparing their positions using global origins and destinations of airline passengers. Urban Geography 28(1): 74–91.

28.

Deudney

Ikenberry

(2009) The myth of the autocratic revival: Why liberal democracy will prevail. Foreign Affairs 88(1): 77–93.

29.

Dollar

(2005) Globalization, poverty and inequality since 1980. The World Bank Research Observer 20(2): 145–175.

30.

Dos Santos

(1970) The structure of dependence. American Economic Review 60(2): 231–236.

31.

Dreby

(2010) Divided by Borders: Mexican Migrants and Their Children. Berkeley: University of California Press.

32.

Duina

(2011) Institutions and the Economy. Cambridge: Polity Press.

33.

Erkus-Ozturk

(2009) The role of cluster types and firm size in designing the level of network relations: The experience of the Antalya tourism region. Tourism Management 30: 589–597.

34.

Erkus-Ozturk

Eraydin

(2011) Factors of global connectivity in Antalya’s tourism. Annals of Tourism Research 38(4): 1300–1321.

35.

Frank

(1973) The development of underdevelopment. In: The Political Economy of Development and Underdevelopment. New York: Random House (originally published in Monthly Review, September 1966).

36.

Friedman

(2006) The World is Flat: A Brief History of the Twenty-first Century. New York: Farrar, Straus and Giroux.

37.

Gaggio

(2011) Selling beauty: Tuscany’s rural landscape since 1945. In: Bandelj

Wherry

(eds) The Cultural Wealth of Nations. Stanford, CA: Stanford University Press, pp. 90–113.

38.

Garin-Munoz

(2004) Madrid as a tourist destination: Analysis and modelization of inbound tourism. International Journal of Tourism Research 6: 289–302.

39.

Goldstone

(2001) Making the World Safe for Tourism. New Haven, CT: Yale University Press.

40.

Greenbaum

Hultquist

(2006) The economic impact of terrorist incidents on the Italian hospitality industry. Urban Affairs Review 42(1): 113–130.

41.

Gullette

(2007) Migration and tourism development in Huatulco, Oaxaca. Current Anthropology 48(4): 603–610.

42.

Hanneman

Riddle

(2005) Introduction to Social Network Methods (free introductory textbook on social network analysis). Available at: http://faculty.ucr.edu/~hanneman/nettext/C10_Coreity.html#Degree (accessed 1 November 2012).

43.

Hubbell

(2008) Rethinking dependency theory: The case of Dominica, the rascal state. Journal of Third World Studies 25(1): 95–109.

44.

Huntington

(1996) The Clash of Civilizations and the Remaking of World Order. New York: Simon & Schuster.

45.

Inglehart

Welzel

(2009) How development leads to democracy. Foreign Affairs 88(2): 33–48.

46.

Ireland

(1993) Gender and class relations in tourism employment. Annals of Tourism Research 20(4): 666–684.

47.

Keeling

(1995) Transport and the world city paradigm. In: Knox

Taylor

(eds) World Cities in a World-System. Cambridge: Cambridge University Press, pp. 115–131.

48.

Kim

Chen

M-H

Jang

(2006) Tourism expansion and economic development: The case of Taiwan. Tourism Management 27(5): 925–933.

49.

Korzeniewicz

Moran

(2009) Unveiling Inequality: A World–historical Perspective. New York: Russell Sage Foundation.

50.

Kowalski

(2011) When cultural capitalization became global practice: The 1972 World Heritage Convention. In: Bandelj

Wherry

(eds) The Cultural Wealth of Nations. Stanford, CA: Stanford University Press, pp. 73–89.

51.

Lanfant

M-F

(1980) Introduction: Tourism in the process of internationalization. International Social Science Journal 32(1): 14–43.

52.

Louch

Hargittai

Centeno

(1999) Phone calls and fax machines: The limits of globalization. The Washington Quarterly 22(2): 83–100.

53.

Mahutga

(2006) The persistence of structural inequality? A network analysis of international trade, 1965–2000. Social Forces 84(4): 1863–1889.

54.

Miguens

JIL

Mendes

JFF

(2008) Travel and tourism: Into a complex network. Physica A 387: 2963–2971.

55.

Milne

Ateljevic

(2001) Tourism, economic development and the global-local nexus: Theory embracing complexity. Tourism Geographies 3(4): 369–393.

56.

Neal

(2010) Refining the air traffic approach to city networks. Urban Studies 47: 2195–2215.

57.

Neal

(2012) Creative employment and jet set cities: Disentangling causal effects. Urban Studies 49(12): 2693–2709.

58.

Nemeth

Smith

(1985) International trade and world-system structure: A multiple network analysis. Review (Fernand Braudel Center) 8(4): 517–560.

59.

Pierson

Skocpol

(2002) Historical institutionalism in contemporary political science. In: Katznelson

Milner

(eds) Political Science: The State of the Discipline. New York: WW Norton & Company Inc., pp. 693–721.

60.

Prebisch

(1950) The Economic Development of Latin America and Its Principal Problems. New York: United Nations for ECLA.

61.

Prebisch

(1959) Commercial policy in the underdeveloped countries. American Economic Review 49(2): 251–273.

62.

Przeworski

Limongi

(1997) Modernization: Theories and facts. World Politics 49(2): 155–183.

63.

Raab

Milward

(2003) Dark networks as problems. Journal of Public Administration Research and Theory 13(4): 413–439.

64.

Rimmer

(1998) Transport and telecommunications among world cities. In: Lo

F-C

Yeung

(eds) Globalization and the World of Large Cities. Tokyo: United Nations University Press, pp. 433–470.

65.

Rivera

(2008) Managing ‘spoiled’ national identity: War, tourism and memory in Croatia. American Sociological Review 73: 613–634.

66.

Rogerson

Visser

(2006) International tourist flows and urban tourism in South Africa. Urban Forum 17(2): 199–213.

67.

Scheper-Hughes

(2006) Biopiracy and the global quest for human organs. NACLA Report on the Americas 39(5): 14–22.

68.

Shih

H-Y

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

69.

Singer

(1949) Economic progress in underdeveloped countries. Social Research 16: 1–11.

70.

Smith

Timberlake

(2001) World city networks and hierarchies, 1977–1997: An empirical analysis of global air travel links. American Behavioral Scientist 44: 1656–1678.

71.

Smith

White

(1992) Structure and dynamics of the global economy: Network analysis of international trade 1965–1980. Social Forces 70(4): 857–893.

72.

Smith

(2005) Mexican New York: Transnational Lives of New Immigrants. Berkeley: University of California Press.

73.

Snyder

Kick

(1979) Structural position in the world system and economic growth, 1955–1970: A multiple-network analysis of transnational interactions. American Journal of Sociology 84(5): 1096–1126.

74.

Steinmetz

(2008) The colonial state as a social field: Ethnographic capital and native policy in the German overseas empire before 1914. American Sociological Review 73(4): 589–612.

75.

Stiglitz

(2002) Globalization and Its Discontents. New York: WW Norton & Company.

76.

Taylor

Derudder

Witlox

(2007) Comparing airline passenger destinations with global service connectivities: A worldwide empirical study of 214 cities. Urban Studies 28(3): 232–248.

77.

Taylor

Derudder

Gago Garcia

(2009) From North-South to ‘Global’ South? An investigation of a changing ‘South’ using airline flows between cities, 1970–2005. Geography Compass 3(2): 836–855.

78.

Teixeira

Fernandes

(2012) Tourism time series forecast – Different ANN architectures with time index input. Procedia Technology 5: 445–454.

79.

Thelen

(1999) Historical institutionalism in comparative politics. Annual Review of Political Science 2: 369–404.

80.

United Nations World Tourism Organization (1995) Technical Manual: Concepts, Definitions and Classifications for Tourism Statistics (1). Madrid: World Tourism Organization.

81.

United Nations World Tourism Organization (2006) Tourism Factbook. Available at: http://www.wtoelibrary.org/content/v486k6/?v=search (accessed 9 October 2009).

82.

United Nations World Tourism Organization (2008) Tourism Market Trends 2006 – World Overview & Tourism Topics. Madrid: World Tourism Organization.

83.

Van Nuffel

Saey

Derudder

. (2010) Measuring hierarchical differentiation: Connectivity and dominance in the European urban network. Transportation Planning and Technology 33(4): 343–366.

84.

Wallerstein

(1990) Culture as the ideological battleground of the modern world-system. Theory, Culture and Society 7: 31–55.

85.

Wallerstein

(2004) World-Systems Analysis: An Introduction. Durham, NC: Duke University Press.

86.

Wasserman

Faust

(2009) Social Network Analysis: Methods and Applications. Cambridge: Cambridge University Press.

87.

Watson

(2006) Transnationalism, localization and fast foods in East Asia. In: Watson

(ed.) Golden Arches East: McDonald’s in East Asia, 2nd edn. Stanford, CA: Stanford University Press.

88.

Wherry

(2007) Trading impressions: Evidence from Costa Rica. ANNALS American Academy of Political and Social Science 610: 217–231.

89.

World Economic Forum (2009) Travel & Tourism Competitiveness Report. Available at: http://www.weforum.org/en/initiatives/gcp/TravelandTourismReport/index.htm.

90.

Yahya

(2003) Tourism flows between India and Singapore. The International Journal of Tourism Research 5: 347–367.

91.

Zelizer

(2005) The Purchase of Intimacy. Princeton, NJ: Princeton University Press.

92.

Zelizer

(2009) Intimacy in economic organizations. In: Bandelj

(ed.) Economic Sociology of Work. Bradford: Emerald, pp. 23–56.

93.

Zelizer

(2010) How I became a relational economic sociologist and what does that mean? Paper presented at the University of California at Davis for a Conference on Relational Work.