Does Distance Still Matter? Moderating Effects of Distance Measures on the Relationship Between Pandemic Severity and Bilateral Tourism Demand

Abstract

This study aims to investigate the moderating effects of various distance measures on the relationship between relative pandemic severity and bilateral tourism demand. After confirming its validity using actual hotel and air demand measures, we leveraged data from Google Destination Insights to understand daily bilateral tourism demand between 148 origin countries and 109 destination countries. Specifically, we estimated a series of fixed-effects panel data gravity models based on the year-over-year change in daily demand. Results show that a 10% increase in seven-day smoothed COVID-19 cases led to a 0.0658% decline in year-over-year demand change. The moderating distance measures include geographic, cultural, economic, social, and political distance. Results show that long-haul tourism demand was less affected by a destination’s pandemic severity relative to tourists’ place of origin. The moderating effect of national cultural dimensions indulgence versus constraints was also confirmed. Lastly, a discussion and implications for international destination marketing are provided.

Keywords

pandemic severity international tourism demand distance factors panel data gravity model Hofstede’s cultural dimensions

Introduction

The global outbreak of COVID-19 has brought devastating challenges to the tourism industry. As a contagious disease, COVID-19 transmission is facilitated by human movement. Therefore, in a collective effort to combat the pandemic, many countries closed their borders and imposed stringent restrictions on human mobility (Yang et al. 2021), posing a major hurdle to international tourism. Tourists have been reluctant to travel during the pandemic out of caution about potential health risks (Hang, Aroean, and Chen 2020). Their overall travel experiences may also be compromised, especially in destination countries where pandemic guidelines remain in place (Sigala 2020). These factors have led to plummeting international tourism demand. According to the UNWTO Tourism Dashboard (https://www.unwto.org/international-tourism-and-covid-19), international tourist arrivals declined by 73% in 2020, with an estimated loss of international tourism revenues totaling between 910 billion USD and 1.2 trillion USD.

Many empirical studies have investigated the economic impact of COVID-19 on the global tourism industry. Given a lack of updated tourism statistics reflecting the pandemic, most researchers have used tourism-related companies’ stock performance data to estimate the pandemic’s industry effects and regulatory policies (Kaczmarek et al. 2021; Sharma and Nicolau 2020). Scholars in applied econometrics have embraced various types of big data, such as air passenger searches and selection (Gallego and Font 2021), foot traffic data (Yang, Liu, and Chen 2020), and geotagged social media data (Huang et al. 2020), to estimate COVID-19’s economic impact on the tourism industry. Different from micro-econometric analysis, macro-economic modeling enables a better understanding of how different channels contribute to the pandemic’s effects and how such impacts diffuse across sectors. For example, Pham et al. (2021) applied tourism satellite account data to CGE modeling, a typical macro-economic model of tourism impact analysis, to evaluate COVID-19’s impact on Australian inbound tourism. Another macro-economic model, the dynamic stochastic general equilibrium model, was developed and calibrated by Yang, Zhang, and Chen (2020) to assess the pandemic as an external shock. A key advantage of macro-economic modeling is that it allows for multiple scenario analyses to simulate the effects of proposed policies intended to combat the pandemic and aid the broader industry.

Much of the literature has examined COVID-19’s impact based on a single country or a sample of destination countries without considering the dyadic nature of tourism demand (Morley, Rosselló, and Santana-Gallego 2014). Among various bilateral factors, distance has been deemed vital to international tourism demand. Studies have shown an array of distance measures (e.g., geographic, cultural, political, and economic) to be noteworthy. During the COVID-19 pandemic, individuals’ decision making may be guided by new norms in response to perceived uncertainty and danger; these norms may oppose those developed before the pandemic (Fernandez-Perez et al. 2021). Two questions thus arise: is distance still relevant in tourism demand modeling during the pandemic era? If so, then which distance measure is most suitable to explain bilateral tourism demand in this context? Meanwhile, although cultural factors have been long discussed in the tourism demand literature (Zhang, Li, and Wu 2019), few if any empirical studies have shed light on how such factors can shape international tourism demand amid COVID-19. Kaczmarek et al. (2021) found that tourism-associated companies from less individualistic countries suffered less than firms in more individualistic countries from the pandemic. However, no studies have yet discussed how origin-specific cultural traits may be more powerful than destination-specific ones in explaining bilateral demand.

In this study, we estimated a panel data gravity model on daily bilateral tourism demand across a large sample of countries from January 28, 2020 to January 11, 2021. Our work is intended to make several contributions to the literature. First, this study represents a pioneering effort to evaluate the effects of the COVID-19 pandemic on bilateral tourism demand based on dyadic demand data. Specifically, we estimated the impact of relative pandemic severity between destinations and origins based on a global dataset. By leveraging a global dyadic data with a large number of original and destination countries, we are expected to provide more generalizable insights. Second, we considered the moderating effects of various distance factors on the pandemic–tourism relationship to offer insight into the role of distance in shaping global bilateral tourism demand patterns in the COVID-19 era. In particular, we compared the effectiveness of different distance measures to demonstrate which distance factor represent the most relevant one. Last but not least, we examined how the effect of relative pandemic severity varied across origin countries with different cultural traits. By doing so, we scrutinized how cultural factors influence the ways tourists have interpreted and responded to the COVID-19 pandemic. Our results enrich the cross-cultural tourism literature by providing up-to-date findings on cross-cultural tourist demand amid COVID-19 and, more generally, revealing how cultural factors affect tourism demand in uncertain times.

Literature Review

Distance and Tourism Demand

The effect of geographic distance on tourism demand was well acknowledged in the early tourism literature based on the classic gravity model (Uysal and Crompton 1985). Distance goes beyond geographic space to measure physical accessibility, with a longer distance involving more travel impediments (Mckercher and Lew 2003). Long-distance travel is thus thought to be associated with a higher level of risk (Crotts 2004), which further inhibits demand. In tourism-related research, distance factors have been widely accepted as essential variables that may mold tourism demand (Ahn, Ekinci, and Li 2013; Balli, Balli, and Jean Louis 2016). Past studies have validated the influential roles of geographic and cultural distance on tourism demand (McKercher, Chan, and Lam 2008; Yang and Wong 2012). However, the roles of other distance measures, such as politics, administration, and economic aspects, have been largely neglected. To better incorporate different perspectives on distance factors, Ghemawat (2001) proposed the Culture, Administration, Geographic, and Economy (CAGE) framework, which broadly conceptualizes the role of distance in international strategies. According to Ghemawat (2001), these four dimensions cover most fundamental differences between countries. Researchers have adopted this framework to understand how bilateral differences can explain international trade and foreign direct investment (Tokas and Deb 2020). In the tourism context, the CAGE framework features bilateral attributes between origin countries and destinations that may affect tourists’ purchase behavior (Li and Katsumata 2020). Prior tourism studies also expanded the conceptual dimensions of distance to investigate the effects of various distance factors on tourism demand.

Geographic distance

As tourism inherently involves human mobility over space, geographic distance has been highlighted as critical to tourism demand. Geographic distance can be considered a proxy of the geographical impediment facing tourists (McKercher, Chan, and Lam 2008). This feature is a practical reflection of geography’s first law, which states that “everything is related to everything else, but near things are more related than distant things” (Tobler 1970). Similar to other rational consumers, tourists maximize their utility among destination alternatives based on various constraints such as cost. Considering the cost of accessing available resources, people prefer to use resources that are closer to them (McKercher, Chan, and Lam 2008). Distance decay theory lays a solid theoretical foundation for the effect of geographic distance, such that tourists’ demand is greater for destinations with greater spatial proximity and becomes smaller as destinations move farther from one’s place of origin. Geographic distance represents a financial and nonfinancial opportunity cost in tourism: travelers with limited time budgets are more likely to visit a destination entailing a shorter distance and less travel time. Relatedly, if travelers need to visit a long-haul destination, they will often stay longer to maximize their benefits (Park, Yang, and Wang 2019).

Distance decay patterns vary across research contexts. Bull (1991) suggested that tourism demand peaks near a place of origin and subsequently decreases exponentially as perceived costs related to travel distance and time increase. McKercher (1998) discovered that tourism demand fell only upon surpassing a certain distance threshold. Several scholars have recognized that the impact of geographic distance is not monotonic. Using a global sample, Mckercher and Lew (2003) found that the presence of an effective tourism exclusion zone significantly shapes tourism demand. Secondary peaks are likely to occur far from a place of origin, where the pull of unique attractions outweighs travel-related friction. Geographic distance can also complicate effective communication and hinder information flows (Borgatti and Cross 2003). Tourists from geographically distant countries may therefore have access to limited information when making decisions. Geographic distance also heavily shapes tourists’ perceived similarity, which is highly correlated with psychological distance (Liu et al. 2018). As psychological distance is one of the most prominent factors informing tourists’ behavior (Hateftabar 2021), the impact of geographic distance on tourism demand can be influenced by multiple psychological variables (Crompton 1979).

Cultural distance

Solomon et al. (2003) defined culture as the shared meanings, rituals, traditions, and norms that exist among members of society. In the tourism field, culture can be considered a country’s unique values, which may affect tourists’ destination choices and travel experiences (O’Leary and Deegan 2003). Cultural distance reflects the level of cultural dissimilarity between an origin and destination and can shape tourists’ destination perceptions (Liu et al. 2018). As suggested in the literature, cultural distance can both positively and negatively affect tourism demand. On one hand, cultural distance can bolster the appeal of tourism, especially among travelers enthusiastic about cultural novelty (Yang, Liu, and Li 2019). A high level of cultural distance reflects cultural diversity between countries, which can fulfill tourists’ desire for exotic experiences and satisfy visitors’ curiosity around pursuing different cultural activities (Kastenholz 2010). For example, based on tourist survey data, McKercher and Chow (2001) found that tourists from Western countries with a large cultural distance were more likely than Asian tourists to participate in local cultural activities in Hong Kong.

On the other hand, scholars have noted that cultural distance can harm tourism demand. First, cultural conflicts, cultural ethnocentrism, and cultural boycotts triggered by cultural dissimilarity can keep destinations from attracting visitors from other cultural backgrounds (Yang and Wong 2012). The higher the perceived risk of visiting more culturally distant destinations, the lower one’s travel intentions (Crotts 2004). Second, cultural distance can produce major barriers to communication and interaction. These issues can evoke psychological insecurity among tourists and deter visitation (Ye, Zhang, and Yuen 2013). For instance, Vietze (2012) found that visitors prefer to choose destinations with a background similar to their own culture (e.g., in terms of language or religion). Third, cultural boundaries may influence tourists’ attitudes and perceptions (Wong 2015); that is, a similar cultural background can enhance travelers’ satisfaction with tourism and commercial services (Park, Yang, and Wang 2019). Lastly, stronger cultural distance increases the monetary and nonmonetary costs necessary for tourists to ensure safety, reduce uncertainty, and improve the quality of their stay (Yang, Liu, and Li 2019).

Due to the intangible nature of culture, the measurement of cultural distance has attracted substantial scholarly attention. Most research has focused on uni-dimensional proxies, such as language similarity, religious similarity, or institutional and social norms, to assess countries’ cultural differences (Yang and Wong 2012). Some scholars have acknowledged the multidimensional nature of cultural distance and used multidimensional scores based on a national cultural framework. The four most popular are the World Values Survey framework (Inglehart 2004), which is based on public opinion data; Schwartz’s framework (Schwartz 1999), which reflects individuals’ expressions of their preferences; the Global Leadership and the Organizational Behavior Effectiveness framework, which is based on a survey of middle managers in three industries; and Hofstede’s framework, which is based on the mean scores of variables for each country. Hofstede, Hofstede, and Minkov (2005) national cultural dimension index is often adopted thanks to its simplicity, convenience, and intuitiveness. This index’s structure is highly stable; it incorporates common factors that can affect countries’ cultures worldwide. The index thus accurately depicts cultural differences between countries over time and has been frequently applied within the tourism literature (Reisinger and Crotts 2010).

Economic distance

Economic distance refers to the economic connection between two countries, with a stronger economic connection implying more frequent business interaction and business travel (Khan, Toh, and Chua 2005). Inbound tourists may also be affected by trade factors such as trade agreements (Saayman, Figini, and Cassella 2016). Empirical studies have shown that a closer economic connection enhances destinations’ image and minimizes information barriers to collectively promote bilateral tourist flows (Kumar, Prashar, and Jana 2019). Lastly, destinations may be more appealing to tourists from origin areas featuring stronger mutual economic interactions. Daily information exposure stemming from a tighter economic connection (e.g., as seen in advertisements and news reports) leads consumers in origin areas to be more likely to purchase products from economically connected countries (Khan, Toh, and Chua 2005)—including tourism and travel products.

Political distance

International tourism demand is particularly vulnerable to fast-changing bilateral political relations. Various political events have been found to significantly affect tourism demand, such as China’s tourism sanction on South Korea in 2017, China’s travel ban on Japan due to the territorial dispute in 2012, and long-lasting sanctions around U.S. citizens’ travel to Cuba. By contrast, a stable political connection between two countries can facilitate bilateral tourism demand. In the international trade literature, political distance reflects the political proximity between countries (Dajud 2013); it mirrors countries’ common political interests or shared political ideologies. This factor partially captures different countries’ prevailing bureaucratic, institutional, and political structures. Research suggests that political distance is thus important to consider (Dajud 2013). Inspired by prior research, political distance may influence other established tourism demand determinants (e.g., travel document trait process, interest similarity, and shared values) as well. Closer political connections often come with favorable entry policies for bilateral travel and a more positive destination image through tighter political linkages. Political proximity can be defined in multiple ways: the United Nations vote correlation can be used to measure similarity in countries’ political interests; data from the Polity IV Project can reveal political regime differences between countries; and bilateral visa restrictions and the number of embassies can contextualize bilateral political connectedness. In the tourism domain, Gil-Pareja, Llorca-Vivero, and Martínez-Serrano (2007) confirmed that the tourism-enhancing effect of embassies and consulates ranged between 15% and 30% in a gravity model of G-7 countries. Neumayer (2010) indicated that visa restriction lowered tourism demand by 52%–63% in a global bilateral travel dataset.

Social distance

Social distance reflects countries’ social disconnectedness (Jiang, Li, and Cutter 2021). In tourism, social distance partly refers to the perceived psychological distance between hosts and visitors and indicates a host community’s acceptance of and hospitality toward tourists (Thyne, Watkins, and Yoshida 2018). This aspect can substantially mold a destination’s image and tourism demand. A destination’s social image can also be easily distorted by social differences between origins and destinations; however, increasing destination familiarity can eliminate these biases (Tasci 2009). Studies on social distance have documented bilateral immigrant stocks as a major determinant of tourism demand due to the positive advertising effects from immigrants on their home countries (Balli, Balli, and Jean Louis 2016; Saayman, Figini, and Cassella 2016). Furthermore, immigrant stock attracts visiting friends and relative tourists in addition to inspiring business travel by establishing business connections with immigrants’ country of origin (Seetaram 2012).

Hypothesis Development

International tourism is vulnerable to various crises and disasters. A number of studies have explored how natural or human-made disasters influence tourism demand; focal events include earthquakes (Huang and Min 2002), financial crises (Goh and Law 2002), epidemics and pandemics (Pan, Wu, and Song 2012), and hurricanes (Woosnam and Kim 2014). Akin to other disasters, the COVID-19 pandemic has brought great uncertainty to tourists’ perceived travel safety and experiences (Pernecky 2020). Perceived risk of the turbulent macro-environment could also lead to a downturn in tourism supply, marked by employee furloughs or layoffs and business closures. Further, human mobility and tourist–host interaction are paramount in travel, yet many countries imposed travel bans and quarantine policies to slow the spread of COVID-19 (Gössling, Scott, and Hall 2021).

Traveling internationally involves a considerable level of travel risk and hence can elicit negative emotions such as anxiety and fear (Reisinger and Mavondo 2005). Once such negative emotions takes hold, many people exercise caution when making travel decisions through the demonstration effect (Seabra et al. 2013). During the current time of the COVID-19 pandemic, such risk-associated concerns and feelings of negative emotions (e.g., anxiety, sadness, fear) can trigger prevention behaviors and suppress individuals’ travel desire (Kim, Seo, and Choi 2021). The literature to date demonstrate that COVID-19 introduced striking operational challenges to the tourism industry (Sigala 2020), jeopardized tourism experiences (Sigala 2020) and caused significant increases in time and monetary costs to travel (Errett, Sauer, and Rutkow 2020). As a result, international tourism demand has fallen sharply (Yang et al. 2021).

While the COVID-19 pandemic is causing a global regression in international tourism demand (Yang et al. 2021), we argue that destinations that suffer more severely from the pandemic may face more distressing downfall in international travel demand, particularly from source markets where the pandemic is less severe. It is human nature for individuals to engage in self-referencing processes to understand unknown object and events (Debevec and Romeo 1992). When the external condition is unknown, individuals will rely on self-relevant information that is stored in memory as a reference line to interpret new, incoming information (Debevec and Romeo 1992). In the current context where the pandemic is causing considerable and vast changes to market conditions and industry practice (Errett, Sauer, and Rutkow 2020; Kim, Seo, and Choi 2021; Sigala 2020), tourists are likely to use their experiences in the place of origin as a reference line to interpret and evaluate the conditions in an international travel destination. For individuals who are based in a place of origin where the pandemic is less spread, traveling to destinations that suffer more severely from the COVID-19 pandemic means exposure to heightened levels of health and travel risks (Errett, Sauer, and Rutkow 2020). As such self-referenced evaluation processes heighten individuals’ perceived risk of traveling, tourists will avoid traveling internationally to destinations that suffer more severely from the COVID-19 pandemic (as compared with their place of origin). We thus propose that:

H1. A destination’s pandemic severity relative to a place of origin adversely affects international tourism demand.

The effect of the pandemic’s severity on bilateral tourism demand can be further moderated by distance for several reasons. First, according to the spatial interaction model, high destination attractiveness prompts tourists to visit distant destinations (Yang, Fik, and Zhang 2017). High-level attractiveness may not necessarily be compromised by the pandemic. Travel is also less substitutable in this case, causing the pandemic to have a lower impact on long-haul tourism demand. Long-haul travel is often accompanied by higher time, emotional, and financial investment as well; as such, canceling or postponing a trip could be prohibitively expensive.

Construal level theory (CLT) may provide another explanation in the current discussion. Essentially, this theory explains how psychological distance affects an individual’s thoughts and behavior (Trope and Liberman 2010). Various distance factors determine tourists’ psychological distance between an origin and destination. Psychological distance further shapes the way travelers’ process destination and travel related information. More specifically, individuals focus on low-level, detailed, and specific contextual information when considering psychologically close (e.g., short-haul destinations) but high-level, abstract information to interpret objects or events in distant destinations (e.g., long-haul destinations) (Trope and Liberman 2010; Yang, Wu, and Yang 2018). For destinations that are psychologically close (e.g., short-haul destinations), tourists are thus likely to interpret pandemic severity based on a low construal level and in greater detail. For destinations that are psychologically distant (e.g., long-haul destinations), tourists will presumably process pandemic-related information with a high level of construal and in a more abstract manner.

Previous research in CLT documented that the difference in concrete vs. abstract thinking may be associated with promotional versus preventional focus (Lermer et al. 2015; Sagristano, Trope, and Liberman 2002). Abstract construals promote focus on desirability features of objects and events (i.e., attractiveness of traveling to a destination), whereas concrete construals promote focus on feasibility features (i.e., accessibility of traveling to a destination) (Lermer et al. 2015; Sagristano, Trope, and Liberman 2002). As concrete construal fosters detailed thinking, COVID-19 related health and travel risks may exacerbate in suppressing travel desire. Therefore, the distance between destination and origin may moderate the effect of pandemic severity on tourism demand. When considering destinations that are psychologically far away (e.g., long-haul destinations), travelers are more likely to adopt abstract thinking which drives them to focus on the overall attractiveness of the destination and thus underestimate the risk. On the other hand, when considering destinations that are psychologically close (e.g., short-haul destinations), travelers are more likely to shift their focus to concrete, detailed, accessibility related features. As such, risks concerns will loom larger to influence travel decision-making and hence magnify the effect of pandemic severity in suppressing travel desire. Put simply, pandemic severity should play a larger role in lowering tourism demand to short-haul destinations closer to a given origin area. With the increase of psychological distance, the effect of pandemic severity on lowering tourism demand will be weakened. Based on the literature and prior discussion, we propose the following hypotheses:

H2. Distance has a moderating effect, weakening the negative impact of pandemic severity on tourism demand:

H2a. Geographic distance has a moderating effect, weakening the negative impact of pandemic severity on tourism demand.

H2b. Cultural distance has a moderating effect, weakening the negative impact of pandemic severity on tourism demand.

H2c. Economic distance has a moderating effect, weakening the negative impact of pandemic severity on tourism demand.

H2d. Political distance has a moderating effect, weakening the negative impact of pandemic severity on tourism demand.

H2e. Social distance has a moderating effect, weakening the negative impact of pandemic severity on tourism demand.

Cultural traits dictate tourists’ behavior and destination selection (Correia, Kozak, and Ferradeira 2011). Under Hofstede’s national cultural framework, an origin country’s long-term orientation versus short-term orientation (LTOWVS) and indulgence versus restraint (IVS) are likely to moderate the pandemic’s impact on international tourism demand. LTOWVS measures the extent to which people plan for the long term rather than the short term. A society with a long-term orientation assigns more value to healthy states and cherishes life more (Gokmen, Baskici, and Ercil 2021). Individuals with a long-term orientation tend to make plans earlier for the future, suggesting that tourists with higher LTOWVS will be apt to avoid potential risks during the pandemic (e.g., threats to long-term health and well-being due to infection). The impact of pandemic severity on tourism demand would be enhanced in this case. IVS reflects the degree to which different societies allow for gratification of basic human desires related to relishing life and having fun. In higher-IVS societies, people are more likely to freely enjoy life and pursue instant gratification (Wang 2021). They also may be less interested in abiding by anti-pandemic regulations; thus, high-IVS countries will likely face more difficulty coordinating suitable anti-COVID-19 measures (Gokmen, Baskici, and Ercil 2021). The moderating effects of IVS will oppose those of LTOWVS. Both of these factors are closely related to social distancing (Wang 2021) and case fatality rates (Erman and Medeiros 2020), leading to the following hypotheses:

H3. Origin-specific cultural traits have moderating effects on pandemic severity relative to tourism demand:

H3a: The origin country’s LTOWVS has a moderating effect, strengthening the negative impact of pandemic severity on tourism demand.

H3b: The origin country’s IVS has a moderating effect, weakening the negative impact of pandemic severity on tourism demand.

Research Method

Empirical Model Specification

We applied a traditional gravity-type model to examine bilateral tourism demand between countries. In economics, gravity modeling has become very popular due to its success in explaining trade flows among countries (Kabir et al. 2017). Thus, it is a natural way to investigate the determinants of tourism flows by means of a gravity model considering that international tourism represents a form of international trade (Khadaroo and Seetanah 2008). From a theoretical perspective, gravity models focus on the spatial interaction between origins and destinations and have often been used to assess the effects of distance variables on tourism demand (Morley, Rosselló, and Santana-Gallego 2014). As a type of spatial interaction model, the gravity model provides several major advantages: (1) it covers explanatory variables more fully by integrating a destination’s tourism supply factor in source–destination pair analysis; (2) it involves origin–destination interaction, enabling more systematic analysis; and (3) it can account for time-related variables and geospatial factors simultaneously (Morley, Rosselló, and Santana-Gallego 2014). Those advantages thus ensure the predictive power of applying the gravity framework in our study. In this paper, we propose the following econometric model:

\ln y_{i j t} = δ * \ln p a n d e m i c_{i j t} + x_{i t}^{'} β + w_{j t}^{'} η + λ_{i j} + γ_{t} + ε_{i j t}

(1)

where i indicates the origin country, j indicates the destination country, and t indicates the date of observation. In this model, $y_{i j t}$ denotes year-over-year (YoY) tourism demand change, calculated as its current value divided by the value of the same day last year; $p a n d e m i c_{i j t}$ , as the major variable of interest, captures relative pandemic severity between a destination and origin country. In terms of control variables, $x_{i t}$ represents a set of origin-specific independent variables, while $w_{j t}$ represents a set of destination-specific independent variables. In addition, $λ_{i j}$ is the origin–destination-pair-specific effect for the pair origin i and destination j. This variable is more general than the specification of origin- and destination-specific effects in a panel data setting (Westerlund and Wilhelmsson 2011), as the effect captures all time-invariant effects of origin–destination pairs (e.g., various distance measures between an origin and destination country). Furthermore, $γ_{t}$ is the date-specific effect for date t, and $ε_{i j t}$ is the normal error term. We also introduced moderating variables $Z_{i j t}$ into the model:

\begin{array}{l} \ln y_{i j t} = δ * \ln p a n d e m i c_{i j t} + η * Z_{i j t} * \ln p a n d e m i c_{i j t} \\ + x_{i t}^{'} β + w_{j t}^{'} η + λ_{i j} + γ_{t} + ε_{i j t} \end{array}

(2)

Due to having a large number of origin–destination pairs, we used the full Gauss-Seidel algorithm to estimate the model. This approach is expected to generate results identical to fixed-effects estimators (Guimarães and Portugal 2010). All standard errors were estimated based on the clustered standard error of origin–destination pairs.

Variable Operationalization

We collected bilateral international tourism demand data from Google Destination Insights. This analysis tool packages Google search data to provide information on the frequency of people’s Google searches about flights, accommodations, and both in a given area. The index is query–share-based; the highest-volume day within a certain time range is normalized to 100. This parameter can capture how tourism demand evolves. In our model, the major variable of interest, lnpandemic, reflects the relative severity of the pandemic in a destination compared with the origin country. This variable was measured by the log of the seven-day smoothed number of confirmed COVID-19 cases per 1,000,000 population in a destination compared with the origin country. We collected associated data from the Johns Hopkins University dashboard and Europe CDC. A negative and significant estimate of lnpandemic lends support to H1.

As the dependent variable was collected based on the specific origin-destination pair, and any time-invariant destination-specific, origin-specific, and destination-origin-specific variables cannot be incorporated Therefore, some popular variables in tourism gravity modeling, such as origin/destination GDP and origin/destination population, became unavailable in this daily gravity model. The control variables were as follows:

• lno_departure/lnd_departure: the number of flights departing from major international airports in the origin/destination country relative to that number on the same day last year (in log). Data were collected from the International Civil Aviation Organization (ICAO). These variables reflect a reduction in air traffic for origins and destinations, a factor that can shape bilateral tourism demand (Ozer Balli, Balli, and Tsui 2019).

• lnd_room_supply: the number of hotel rooms available in the destination country relative to that number on the same day last year (in log). Data were obtained from STR, the most popular hotel data vendor across the world. This variable captures a decrease in destinations’ receiving capacity and tourism supply level (Yang et al. 2021).

• o_stringency_index/d_stringency_index: an index measuring the degree of stringency in an origin/destination country’s policy responses to the pandemic. Data were gathered from Oxford’s COVID-19 Government Response Tracker. These two variables cover the degree of government-imposed mandatory social distancing, which will greatly influence tourism performance (Chen et al. 2020).

Several moderating variables were adopted to empirically test H2 and H3:

• lndist_geo: geographic distance (in kilometers) between the capital cities of an origin and destination country. Data were obtained from the CEPII Gravity database and were used in their natural logarithm form. A positive and significant estimate of lndist_geo lends support to H2a.

• lndist_cul: Cultural distance was operationalized in accordance with Hofstede’s framework. Following Kogut and Singh’s (1988) formula, we combined Hofstede’s six cultural dimensions—individualism, uncertainty avoidance, power distance, masculinity, long-term orientation, and indulgence—into the following composite index: $C D_{o d} = \sqrt{\sum_{i = 1}^{6} {{(I_{i o} - I_{i d})}^{2} / V_{i}}}$ . We used the logarithm of $C D_{o d}$ in our model. A positive and significant estimate of lndist_cul lends support to H2b.

• lndist_eco: We measured economic distance as the inverse of the average percentage of bilateral trade flows between an origin and destination over total international trade to reflect countries’ economic distance (in log). Data were constructed from the CEPII Gravity database. A positive and significant estimate of lndist_eco lends support to H2c.

• lndist_pol: Political distance was estimated in terms of the ideological distance between countries as evidenced by their United Nations General Assembly votes, indicating the comparability of actions between countries over time. We adopted data from Bailey, Strezhnev, and Voeten (2017), estimating countries’ “ideal points” by analyzing voting behavior in the United Nations General Assembly. This variable is often taken as a proxy of the ideological distance between countries (Schneider and Tobin 2020). A positive and significant estimate of lndist_pol lends support to H2d.

• lndist_soc: We calculated social distance as the inverse of the average percentage of bilateral immigrants’ stocks between an origin and destination over their total immigration (in log). Data were collected from the International Migrant Stock dataset of the United Nations. A positive and significant estimate of lndist_soc lends support to H2e.

• o_lto: The long-term orientation versus short-term orientation index of origin countries was taken from Hofstede’s national cultural dimension index. A negative and significant estimate of o_lto lends support to H3a.

• o_ivr: Countries’ indulgence versus restraint index was again taken from Hofstede’s national cultural dimension index. A positive and significant estimate of o_ivr lends support to H3b.

Data Validity

Before taking Google’s Destination Insights data as a proxy of bilateral tourism demand, we checked the data’s validity. The data were aggregated based on Google searches, and past literature confirmed the Google search data represent an accurate approximation of actual tourism flows (Siliverstovs and Wochner 2018). Destination Insights also provide aggregate destination data, which we compared to actual performance data to confirm validity. First, we compared YoY daily destination aggregate accommodation demand data (lnd_DI_accommodation) with YoY daily destination room demand data provided by STR (lnd_room_demand). The top panel of Figure 1 shows the accompanying scatterplot reflecting a positive relationship. We next regressed lnd_room_demand on lnd_DI_accommodation in a traditional fixed-effects panel data model; the within-R² value reached 0.4267 with an overall R² value of 0.3010. Second, we evaluated the correlation between YoY daily destination aggregate air demand data (lnd_DI_air) and YoY daily flight departures from major international airports in the destination country provided by ICAO (lnd_air_departure). The bottom panel of Figure 1 indicates a clear positive relationship between the two variables. Estimation results from a fixed-effects panel data model revealed a within-R² value of 0.4298 and an overall R² value of 0.4070. These results demonstrate that the Google Destination Insights data serve as a reasonable proxy of international tourism demand. Lastly, we checked the data based on alternative bilateral tourism demand data, namely daily geotagged Twitter data (Li et al. 2021). This data removed non-human tweets and considered single- and cross-day human movement across countries. Of note, users’ home location was not considered in this dataset. The Pearson correlation coefficient between these two variables was 0.2216 after logarithmic transformation.

Figure 1.

Scatterplots of data validation.

Data Description

Our final dataset contained 613,889 observations from 148 origin countries and 109 destination countries. Figure 2 maps the frequency of countries in the dataset as either destinations or points of origin. As shown, the data covered a large number of observations from/to Europe and North America. Table 1 presents the descriptive statistics of variables in the econometric model. We gathered 613,889 observations across all demand data from Google Destination Insights (with 4,681 origin–destination pairs), 416,787 observations based on accommodation demand (with 4,057 origin–destination pairs), and 549,665 observations based on air demand (with 3,640 origin–destination pairs). All data spanned January 28, 2020–January 11, 2021. As the data cover the year-over-year change, the data in 2019 were also considered in many variables. We focused on unbalanced panel data, and the number of observations per origin–destination pair ranged from 2 to 311. We also checked variance inflation factors (VIFs), and VIF measures for all variables were far below the suggested cutoff value of 10 (Dormann et al. 2013), suggesting the absence of multi-collearity issues.

Figure 2.

Map of countries in the dataset.

Table 1.

Descriptive Statistics of Variables in the Econometric Model.

Variable	All			Accommodation			Air
Variable	Obs.	Mean	Std. Dev.	Obs.	Mean	Std. Dev.	Obs.	Mean	Std. Dev.
lnDI_all	613,889	−0.835	0.961
lnDI_accommodation				416,787	−0.979	0.972
lnDI_air							549,665	−0.900	0.737
lnpandemic	613,889	−0.249	2.694	416,787	−0.204	2.712	549,665	−0.294	2.618
lno_departure	613,889	−1.754	0.986	416,787	−1.746	0.974	549,665	−1.748	0.992
lnd_departure	613,889	−1.768	0.979	416,787	−1.743	0.990	549,665	−1.764	0.976
lnd_room_supply	613,889	−0.062	0.098	416,787	−0.065	0.102	549,665	−0.061	0.098
o_stringence_index	613,889	60.418	18.351	416,787	60.515	18.410	549,665	60.503	18.279
d_stringence_index	613,889	60.132	18.381	416,787	60.090	18.163	549,665	60.098	18.394
lndist_geo	613,889	8.122	1.023	416,787	8.111	1.063	549,665	8.088	1.004
lndist_cul	492,800	1.147	0.467	354,562	1.141	0.473	444,925	1.146	0.465
lndist_eco	561,542	11.890	2.605	389,236	11.417	2.617	504,261	11.704	2.483
lndist_pol	613,889	−0.551	1.337	416,787	−0.541	1.342	549,665	−0.558	1.329
lndist_soc	445,288	12.348	3.111	319,714	11.918	3.049	410,350	12.204	3.043
o_lto	587,425	48.334	21.747	401,312	48.243	21.655	528,815	48.571	21.472
o_ivr	580,756	48.332	20.038	397,918	49.755	19.176	522,350	48.623	19.987
Country pairs	4,681			4,057			3,640
Origin	148			148			143
Destination	109			109			109

Empirical Results

Results Based on All Demand

Table 2 lists the estimation results of econometric models based on all demand data from Google Destination Insights. Model 1 estimated the sample without any interaction terms with an adjusted R² of 0.476. In the model, lnpandemic was estimated to be −0.00658, which was statistically significant at the .01 level. Specifically, a 10% increase in confirmed COVID-19 cases (seven-day smoothed) in a destination relative to a point of origin led to a 0.0658% decline in the YoY change of bilateral tourism demand as measured by Google Destination Insights data. H1 was thus supported. We then re-estimated these data using destination-specific confirmed cases (seven-day smoothed). The model’s goodness-of-fit indices showed that the model containing a relative confirmed case measure (Model 1) the model with destination-specific confirmed cases measure (results available upon requests). In this model, the signs of the estimates of control variables were largely consistent with our expectations: lno_departure and lnd_departure were each estimated to be positive and significant, suggesting that airport volume was positively related to international tourism demand during the COVID-19 era. In particular, the estimate of lnd_departure was larger, highlighting the role of destination airports in shaping bilateral tourism demand compared with origin airports. Both o_stringence_index and d_stringence_index were estimated to be negative and significant; in other words, a stricter stringency policy from origin and destination governments appeared to temper bilateral demand. The estimates of these two variables were of similar magnitude, showing that the effect sizes of origin and destination governments’ policies were equally important. Lastly, lnd_room_supply was estimated to be statistically insignificant.

Table 2.

Estimation Results Based on All Demand.

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8
	All	All	All	All	All	All	All	All
lnpandemic	−0.00658*** (0.002)	−0.0318* (0.019)	−0.0153*** (0.006)	−0.0275*** (0.010)	−0.00577*** (0.002)	−0.0137 (0.010)	0.000258 (0.005)	−0.0159*** (0.005)
lno_departure	0.0909*** (0.006)	0.0911*** (0.006)	0.100*** (0.006)	0.0945*** (0.006)	0.0907*** (0.006)	0.103*** (0.007)	0.0931*** (0.006)	0.0931*** (0.006)
lnd_departure	0.144*** (0.006)	0.144*** (0.006)	0.146*** (0.006)	0.146*** (0.006)	0.145*** (0.006)	0.145*** (0.007)	0.144*** (0.006)	0.146*** (0.006)
lnd_room_supply	−0.0577 (0.057)	−0.0575 (0.057)	0.0794 (0.062)	−0.0358 (0.061)	−0.0587 (0.057)	−0.0246 (0.061)	−0.0295 (0.058)	−0.0133 (0.058)
o_stringence_index	−0.00235*** (0.000)	−0.00232*** (0.000)	−0.00195*** (0.000)	−0.00233*** (0.000)	−0.00234*** (0.000)	−0.00295*** (0.000)	−0.00196*** (0.000)	−0.00195*** (0.000)
d_stringence_index	−0.00231*** (0.000)	−0.00234*** (0.000)	−0.00189*** (0.000)	−0.00226*** (0.000)	−0.00233*** (0.000)	−0.00282*** (0.000)	−0.00233*** (0.000)	−0.00219*** (0.000)
lndist_geolnpandemic*		0.00301 (0.002)
lndist_cullnpandemic*			0.00850** (0.004)
lndist_ecolnpandemic*				0.00169** (0.001)
lndist_pollnpandemic*					0.00238* (0.001)
lndist_soclnpandemic*						0.000346 (0.001)
o_ltolnpandemic*							−0.000127 (0.000)
o_ivrlnpandemic*								0.000201** (0.000)
constant	−0.145*** (0.025)	−0.144*** (0.025)	−0.195*** (0.027)	−0.155*** (0.027)	−0.144*** (0.025)	−0.120*** (0.030)	−0.178*** (0.026)	−0.172*** (0.026)
Year-month effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country-pair effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N (observations)	613,889	613,889	492,800	561,542	613,889	445,288	587,425	580,756
N (country pairs)	4,681	4,681	3,231	4,047	4,681	2,724	4,226	4,195
R ²	0.480	0.480	0.480	0.486	0.480	0.488	0.479	0.471
Adj. R²	0.476	0.476	0.476	0.482	0.476	0.485	0.475	0.467
AIC	1,292,455.6	1,292,429.3	995,540.2	1,157,370.8	1,292,424.5	877,808.4	1,214,894.8	1,199,955.3
BIC	1,292,523.6	1,292,508.6	995,618.0	1,157,449.4	1,292,503.8	877,885.5	1,214,973.8	1,200,034.2

Note: Country-pair-based clustered standard errors are presented in parentheses.

***

Indicates significance at the .01 level. **Indicates significance at the .05 level. *Indicates significance at the .10 level.

Models 2–8 in Table 2 depict the estimation results of models with interaction terms to examine the moderators of lnpandemic. Models 2–6 included the interaction terms of lnpandemic with different distance measures. Because data were unavailable for some distance measures, different models contained different numbers of observations, ranging from 445,288 (Model 6) to 613,889 (Model 2). For those five models, a slightly higher R² and adjusted R² of Models 4 and 6 indicating a stronger explanatory power when we consider economic or social distance measures as moderating variables in our estimation. Across the five models, the estimated coefficient of the interaction term was positive, consistent with our postulations from H2a to H2e. However, only the estimates of lndist_cul* lnpandemic (Model 3), lndist_eco* lnpandemic (Model 4), and lndist_pol* lnpandemic (Model 5) were estimated to be statistically significant, confirming the moderating effects of cultural distance, economic distance, and political distance on the pandemic–tourism relationship. H2b, H2c, and H2d were therefore supported by the empirical results. We further calibrated the turning point when the effect of relative pandemic severity became statistically insignificant using the Delta method with the main effect and interaction terms (Greene 2007). The marginal effect of lnpandemic became insignificant when the cultural distance between two countries exceeded (roughly) the 60th percentile when the economic distance was larger than (roughly) the 75th percentile and when the political distance was larger than (roughly) the 75th percentile.

Models 7 and 8 in Table 2 estimated the moderating effects of origin countries’ two national cultural dimensions of the origin country. Whereas o_lto* lnpandemic was estimated to be statistically insignificant in Model 7, o_ivr* lnpandemic was estimated to be positive and significant in Model 8. The outcome supported H3b, indicating that relative pandemic severity was less influential on tourism demand from culturally more indulgent countries of origin early in the COVID-19 pandemic.

Results From Accommodation and Air Demand

Table 3 displays the estimation results for the models of accommodation demand (Models 9–16) and air demand (Models 17–24). A higher value of R² and adjusted R² in all estimations indicates that our independent variables help to explain a larger proportion of variance for those two demand measures. In Model 9, the negative coefficient of lnpandemic was statistically insignificant, suggesting that YoY accommodation demand change was not sensitive to the relative level of pandemic severity in a destination. As indicated in Models 10 and 12, we identified the respective moderating effects of geographic distance (Model 10) and economic distance (Model 12) on the relationship between relative pandemic severity and bilateral tourism demand. Lastly, the negative and significant coefficient of o_lto*lnpandemic in Model 15 implies that tourists from more long-term-oriented cultures were more affected by pandemic severity. The results for air demand models were highly similar to those including all forms of demand; the correlation coefficient between all demand change and air demand change was 0.9962. In Model 17, the negative impact of relative pandemic severity was statistically significant. This effect was respectively moderated by cultural distance, economic distance, and political distance between origin and destination countries as shown in Models 19–21. Model 23 revealed that an origin country’s cultural dimension of long-term versus short-term orientation also moderated the pandemic–tourism relationship.

Table 3.

Estimation Results Based on Accommodation and Air Demand.

	Model 9	Model 10	Model 11	Model 12	Model 13	Model 14	Model 15	Model 16
	Accommodation	Accommodation	Accommodation	Accommodation	Accommodation	Accommodation	Accommodation	Accommodation
lnpandemic	−0.00176 (0.003)	−0.0499** (0.024)	0.00479 (0.007)	−0.0469*** (0.012)	−0.00255 (0.003)	−0.0135 (0.012)	0.0220*** (0.006)	0.000604 (0.009)
lndist_geolnpandemic*		0.00571** (0.003)
lndist_cullnpandemic*			−0.00575 (0.005)
lndist_ecolnpandemic*				0.00398*** (0.001)
lndist_pollnpandemic*					−0.00250 (0.002)
lndist_soclnpandemic*						0.000410 (0.001)
o_ltolnpandemic*							−0.000474*** (0.000)
o_ivrlnpandemic*								−0.0000529 (0.000)
Year-month effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country-pair effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	416,787	416,787	354,562	389,236	416,787	319,714	401,312	397,918
adj. R²	0.524	0.524	0.520	0.525	0.524	0.516	0.524	0.517
	Model 17	Model 18	Model 19	Model 20	Model 21	Model 22	Model 23	Model 24
	Air	Air	Air	Air	Air	Air	Air	Air
lnpandemic	−0.00769*** (0.002)	−0.0211 (0.016)	−0.0150*** (0.006)	−0.0275*** (0.009)	−0.00645*** (0.002)	−0.0173** (0.009)	0.00167 (0.005)	−0.0145*** (0.004)
lndist_geolnpandemic*		0.00161 (0.002)
lndist_cullnpandemic*			0.00706* (0.004)
lndist_ecolnpandemic*				0.00162** (0.001)
lndist_pollnpandemic*					0.00368*** (0.001)
lndist_soclnpandemic*						0.000635 (0.001)
o_ltolnpandemic*							−0.000188** (0.000)
o_ivrlnpandemic*								0.000130 (0.000)
Year-month effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Country pair effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	549,665	549,665	444,925	504,261	549,665	410,350	528,815	522,350
Adj. R²	0.587	0.587	0.584	0.591	0.588	0.581	0.586	0.577

Note: Country-pair-based clustered standard errors are presented in parentheses. Estimated coefficients of control variables are not provided for brevity.

***

Indicates significance at the .01 level. **Indicates significance at the .05 level. *Indicates significance at the .10 level.

Robustness Check

We conducted several major robustness checks on the results involving all demand. Estimation results for the robustness checks appear in the online Supplemental Materials. First, considering the lagged effect of pandemic severity (Zhang, Qian, and Hu 2021), we introduced a seven-day lag of lnpandemic, L7.lnpandemic, as the major variable of interest and re-ran all models in Table 2. In the model without interaction terms, L7.lnpandemic was estimated to be −0.00834, which was statistically significant at the .01 level. This effect size was larger than its counterpart in Model 1, revealing a more pronounced lagged effect of relative pandemic severity than the instant impact on bilateral tourism demand. Moreover, all five interaction terms with distance measures were estimated to be positive; among them, three were statistically significant. Also, o_ivr*L7. lnpandemic was estimated to be positive and significant.

Second, in addition to the origin–destination-country-pair-specific effects and date-specific effects, we considered origin–destination-country-pair-specific time trends (Bun and Klaassen 2007) to re-estimate all models in Table 2. In the model without interaction terms, lnpandemic was estimated to be −0.00402, statistically significant at the .05 level. This coefficient was smaller in magnitude than its counterpart in Model 1. All five interaction terms with distance measures were estimated to be positive, four of which were statistically significant. Moreover, o_lto* lnpandemic was estimated to be negative and significant while o_ivr* lnpandemic was estimated to be positive and significant.

Third, we considered the asymmetric effect of pandemic situation. As relative pandemic severity was considered, its effect size may differ based on the absolute pandemic severity level. In other words, although the relative level of pandemic severity is the same, high absolute destination severity (vs. high absolute origin severity) can impart a different effect from low absolute destination severity (vs. low absolute origin severity). Therefore, we added an interaction term of lnpandemic and absolute, where absolute = 1 if the absolute destination severity (seven-day smoothed number of confirmed COVID-19 cases per 1,000,000 population) is higher than the median value in the sample, and absolute = 0; otherwise. Our results show that the impact of relative pandemic severity is negative and significant (with a coefficient of −0.019) when the absolute pandemic severity is of high level in the destination. However, the impact is statistically insignificant when the absolute level of pandemic severity is low.

Fourth, as massive travel restrictions were imposed starting from March 2020, we re-estimated the model using the data starting from March 1, 2020. In the model, lnpandemic was estimated to be −0.00547. Other results are largely similar to the results from Table 2. Lastly, we used lndeaths, the log of the seven-day-smoothed number of COVID-19 deaths in a destination relative to that number in an origin country, to evaluate pandemic severity (Lee and Chen 2020). Based on the model estimation results using this alternative variable of interest, the coefficient was estimated to be −0.00976, which was statistically significant at the .01 level. The magnitude of the coefficient was larger than that of lnpandemic in Model 1, which was −0.00658. Only one interaction term with the distance measure was estimated to be positive and significant: lndist_cul*lndeaths. In addition, o_lto*lndeaths was estimated to be negative and significant while o_ivr*lndeaths was estimated to be insignificant.

Conclusions

We collected Google Destination Insights data from 148 origin countries to 109 destination countries to represent bilateral tourism demand. After confirming this metric’s validity as a proxy of demand, we estimated a series of panel data gravity models of daily bilateral tourism demand. Findings indicated that a 10% increase in confirmed COVID-19 cases (seven-day smoothed) in a destination relative to a point of origin resulted in a 0.0658% decline in the YoY change of bilateral tourism demand. This negative effect was proved to be significantly moderated by certain origin-to-destination distance measures, including cultural distance, economic distance, and political distance. Furthermore, the results show that the effect of pandemic severity was moderated by the origin country’s cultural traits, such as the indulgence versus restraint cultural dimension. More specifically, the negative impact of relative pandemic severity was less pronounced between countries distant from each other and from countries with high indulgence tendencies. The results also revealed discrepancies between the models for accommodation demand and air demand. Lastly, we confirmed our major empirical results via multiple robustness checks.

Overall, the results suggest that a destination’s pandemic severity relative to a point of origin is more effective than destination-specific pandemic severity in capturing the impact of COVID-19 in international tourism demand modeling. This outcome points to the merits of applying bilateral data for tourism demand modeling. Moreover, a global sample can provide more convincing and reliable estimates to guide policymakers’ and industry professionals’ decisions. Our findings further demonstrated that the effect of COVID-19 deaths was quite similar to that of confirmed COVID-19 cases, indicating that both measures are equally reflective of the pandemic’s impact on global tourism demand.

Our empirical results also highlighted distance and cultural factors as moderators of the pandemic–tourism relationship. Based on the spatial interaction model and CLT, we confirmed that distant destinations are less vulnerable to pandemic severity. These findings offer more credence to the distance effect with respect to tourism demand. The results can also be generalized to other global disasters. Our study represents a breakthrough attempt to explore and compare distance measures in bilateral tourism demand modeling. Although past research unveiled the roles of different distance measures separately (Yang and Wong 2012; Zhang, Seo, and Lee 2013), none has systematically evaluated a large set of distance measures to capture travel constraints from distinct perspectives. Our results illuminate a pronounced moderating effect of cultural distance, among various distance measures, on the pandemic–tourism relationship. This outcome underscores the importance of cultural distance in shaping tourists’ psychological distance between origin and destination countries during the COVID-19 era. Compared with other distance measures (e.g., geographic, political, economic, and social), cultural distance plays a key role in explaining tourism demand during crises such as the COVID-19 pandemic.

Findings regarding the moderating effects of origin countries’ cultural traits also enhance the cross-cultural tourism literature by clarifying cross-cultural tourist behavior in times of uncertainty. To our knowledge, this study is among the first to investigate the relationship between origin countries’ cultural traits and the COVID-19 pandemic in tourism. Our work thus serves as a springboard for future research, such as studies exploring novel ways to minimize crises’ adverse effects on the tourism industry. We also emphasized the importance of considering national cultures to investigate tourist behavior in the face of unexpected external shocks. By integrating cultural dimension theory with tourists’ decision-making process, we identified the roles of origin countries’ specific cultural traits in molding tourists’ behavior during a pandemic. Consistent with research from other business disciplines, our findings indicate that consumer behavior is largely influenced by a country’s cultural dimensions (Pantano et al. 2021). These results further imply that policymakers should combine technical measures with cultural elements to develop and enact effective policies (Gokmen, Baskici, and Ercil 2021).

Our results also offer several implications for destination marketing organizations, especially those relying heavily on the international market. First, international tourism destinations should closely monitor the pandemic situation in major source markets; relative pandemic severity can function as a key indicator to evaluate the potential of tourism marketing. Second, for countries hit hard by the pandemic, marketing organizations can focus on major source markets with more short-term-oriented and indulgent cultural traits. Tourism demand in these markets should be relatively less influenced by pandemic severity. Last but not least, as the short-haul tourism market will remain a core travel segment during and after the pandemic, such markets should be given more information on safety measures to alleviate pandemic-related concerns. In addition, tourism products that are associated with lower COVID-19 risk should be developed and promoted, such as outdoor activities.

Some limitations may temper the generalizability of our results. First and foremost, we proxied international tourism demand using online Google search data. These data may partially obscure measurement error, especially in countries in which Google’s search engine holds a low market share. Second, our data only covered part of the pandemic cycle; daily records of COVID-19 cases continued to be updated in the meantime. Estimates may also change throughout the pandemic, and conclusions may vary based on data from the recovery stage. Third, we tested our research hypotheses based on aggregate tourism demand; further micro-level insights (i.e., tourist-level data) could be intriguing. Therefore, we call for future studies to collect data from different stages of the pandemic and via surveys and experiments with individual tourists. Relevant findings could shed further light on COVID-19’s impact on tourism demand and how distance and cultural factors moderate this effect.

Supplemental Material

sj-docx-1-jtr-10.1177_00472875221077978 – Supplemental material for Does Distance Still Matter? Moderating Effects of Distance Measures on the Relationship Between Pandemic Severity and Bilateral Tourism Demand

Supplemental material, sj-docx-1-jtr-10.1177_00472875221077978 for Does Distance Still Matter? Moderating Effects of Distance Measures on the Relationship Between Pandemic Severity and Bilateral Tourism Demand by Yang Yang, Linjia Zhang, Laurie Wu and Zhenlong Li in Journal of Travel Research

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is supported by XJTLU Research Enhancement Funding, under grant No. REF-21-01-001.

ORCID iDs

Yang Yang

Linjia Zhang

Supplemental Material

Supplemental material for this article is available online.

Author Biographies

Yang Yang, PhD, is an Associate Professor at the Department of Tourism and Hospitality Management, Temple University. Email: yangy@temple.edu. His research interests include tourism and hospitality analytic.

Linjia Zhang, PhD, is an Assistant Professor at International Business School Suzhou, Xi’an Jiaotong- Liverpool University, China. Email: linjia.Zhang@xjtlu.edu.cn. His current research projects include nonparametric efficiency analysis, tourism performance index, and cluster life cycle.

Laurie Wu, PhD, is an Associate Professor at the Department of Tourism and Hospitality Management, Temple University. Email: laurie.wu@temple.edu. Her research focuses on service marketing and consumer behavior.

Zhenlong Li, PhD, is an Associate Professor of Geography at the University of South Carolina. E-mail: zhenlong@sc.edu. His research focuses on GIScience, geospatial big data, social media analytics, human mobility, and GeoAI.

References

Ahn

Taehong

Ekinci

Yuksel

Gang

. 2013. “Self-Congruence, Functional Congruence, and Destination Choice.” Journal of Business Research 66 (6): 719–23.

Bailey

Michael A.

Strezhnev

Anton

Voeten

Erik

. 2017. “Estimating Dynamic State Preferences From United Nations Voting Data.” Journal of Conflict Resolution 61 (2): 430–56.

Balli

Faruk

Balli

Hatice O.

Jean Louis

Rosmy

. 2016. “The Impacts of Immigrants and Institutions on Bilateral Tourism Flows.” Tourism Management 52: 221–9.

Borgatti

Stephen P.

Cross

Rob

. 2003. “A Relational View of Information Seeking and Learning in Social Networks.” Management Science 49 (4): 432–45.

Bull

1991. The Economics of Travel and Tourism. Melbourne: Pitman.

Bun

Maurice J. G.

Klaassen

Franc J. G. M.

2007. “The Euro Effect on Trade is Not as Large as Commonly Thought.” Oxford Bulletin of Economics and Statistics 69 (4): 473–96.

Chen

Ming-Hsiang

Demir

Ender

García-Gómez

Conrado Diego

Zaremba

Adam

. 2020. “The Impact of Policy Responses to COVID-19 on U.S. Travel and Leisure Companies.” Annals of Tourism Research Empirical Insights 1 (1): 100003.

Correia

Antónia

Kozak

Metin

Ferradeira

João

. 2011. “Impact of Culture on Tourist Decision-Making Styles.” International Journal of Tourism Research 13 (5): 433–46.

Crompton

John L.

1979. “Motivations for Pleasure Vacation.” Annals of Tourism Research 6 (4): 408–24.

10.

Crotts

John C.

2004. “The Effect of Cultural Distance on Overseas Travel Behaviors.” Journal of Travel Research 43 (1): 83–8.

11.

Dajud

Camilo Umana

. 2013. “Political Proximity and International Trade.” Economics and Politics 25 (3): 283–312.

12.

Debevec

Kathleen

Romeo

Jean B.

1992. “Self-Referent Processing in Perceptions of Verbal and Visual Commercial Information.” Journal of Consumer Psychology 1 (1): 83–102.

13.

Dormann

Carsten F.

Elith

Jane

Bacher

Sven

Buchmann

Carsten

Carl

Gudrun

Carré

Gabriel

Marquéz

Jaime R. García

Gruber

Bernd

Lafourcade

Bruno

Leitão

Pedro J.

, et al. 2013. “Collinearity: A Review of Methods to Deal With It and a Simulation Study Evaluating Their Performance.” Ecography 36 (1): 27–46.

14.

Erman

Aysegul

Medeiros

Mike

. 2020. “Exploring the Impact of Cultural Variability on COVID-19-Related Mortality: A Meta-Analytic Approach.” Research Square. doi:10.21203/rs.3.rs-39104/v1.

15.

Errett

Nicole A.

Sauer

Lauren M.

Rutkow

Lainie

. 2020. “An integrative review of the limited evidence on international travel bans as an emerging infectious disease disaster control measure.” Journal of Emergency Management 18 (1):7–14.

16.

Fernandez-Perez

Adrian

Gilbert

Aaron

Indriawan

Ivan

Nguyen

Nhut H.

2021. “COVID-19 Pandemic and Stock Market Response: A Culture Effect.” Journal of Behavioral and Experimental Finance 29: 100454.

17.

Gallego

Inmaculada

Font

Xavier

. 2021. “Changes in Air Passenger Demand as a Result of the COVID-19 Crisis: Using Big Data to Inform Tourism Policy.” Journal of Sustainable Tourism 29: 1470–89.

18.

Ghemawat

2001. “Distance Still Matters: The Hard Reality of Global Expansion.” Harvard Business Review 79 (8): 137–47.

19.

Gil-Pareja

Salvador

Llorca-Vivero

Rafael

Gallego

Inmaculada

Martínez-Serrano

José Antonio

. 2007. “The Impact of Embassies and Consulates on Tourism.” Tourism Management 28 (2): 355–60.

20.

Goh

Carey

Law

Rob

. 2002. “Modeling and Forecasting Tourism Demand for Arrivals With Stochastic Nonstationary Seasonality and Intervention.” Tourism Management 23 (5): 499–510.

21.

Gokmen

Yunus

Baskici

Cigdem

Ercil

Yavuz

. 2021. “The Impact of National Culture on the Increase of COVID-19: A Cross-Country Analysis of European Countries.” International Journal of Intercultural Relations 81: 1–8.

22.

Gössling

Stefan

Scott

Daniel

Hall

C. Michael

. 2021. “Pandemics, Tourism and Global Change: A Rapid Assessment of COVID-19.” Journal of Sustainable Tourism 29 (1): 1–20.

23.

Greene

William H.

2007. Econometric Analysis. 6th ed. Upper Saddle River, NJ: Prentice Hall.

24.

Guimarães

Paulo

Portugal

Pedro

. 2010. “A Simple Feasible Procedure to Fit Models With High-Dimensional Fixed Effects.” The Stata Journal 10 (4): 628–49.

25.

Hang

Haiming

Aroean

Lukman

Chen

Zhifeng

. 2020. “Building Emotional Attachment During COVID-19.” Annals of Tourism Research 83: 103006.

26.

Hateftabar

Fahimeh

. 2021. “The Impact of Psychological Distance on Tourists’ Length of Stay: Survival Analysis.” Journal of Hospitality and Tourism Management 46: 1–11.

27.

Hofstede

Geert

Hofstede

Gert Jan

Minkov

Michael

. 2005. Cultures and Organizations: Software of the Mind. Vol. 2. New York, NY: McGraw-Hill.

28.

Huang

Jen-Hung

Min

Jennifer C. H.

2002. “Earthquake Devastation and Recovery in Tourism: The Taiwan Case.” Tourism Management 23 (2): 145–54.

29.

Huang

Xiao

Zhenlong

Jiang

Yuqin

Xiaoming

Porter

Dwayne

. 2020. “Twitter Reveals Human Mobility Dynamics During the COVID-19 Pandemic.” PLoS One 15 (11): e0241957.

30.

Inglehart

Ronald

. 2004. Human Beliefs and Values: A Cross-Cultural Sourcebook Based on the 1999-2002 Values Surveys. México: Siglo XXI.

31.

Jiang

Yuqin

Zhenlong

Cutter

Susan L.

2021. “Social Distance Integrated Gravity Model for Evacuation Destination Choice.” International Journal of Digital Earth. Published online April 20. doi:10.1080/17538947.2021.1915396.

32.

Kabir

Mahfuz

Ruhul

Salim

Rashid

Al-Mawali N.

2017. “The gravity model and trade flows: Recent developments in econometric modeling and empirical evidence.” Economic Analysis and Policy 56: 60–71.

33.

Kaczmarek

Tomasz

Perez

Katarzyna

Demir

Ender

Zaremba

Adam

. 2021. “How to Survive a Pandemic: The Corporate Resiliency of Travel and Leisure Companies to the COVID-19 Outbreak.” Tourism Management 84: 104281.

34.

Kastenholz

Elisabeth

. 2010. “‘Cultural Proximity’ as a Determinant of Destination Image.” Journal of Vacation Marketing 16 (4): 313–22.

35.

Khadaroo

Jameel

Seetanah

Boopen

. 2008. “The role of transport infrastructure in international tourism development: A gravity model approach”. Tourism Management 29(5): 831–40.

36.

Khan

Habibullah

Toh

Rex S.

Chua

Lyndon

. 2005. “Tourism and Trade: Cointegration and Granger Causality Tests.” Journal of Travel Research 44 (2): 171–6.

37.

Kim

Ellen Eun Kyoo

Seo

Kwanglim

Choi

Youngjoon

. 2021. “Compensatory Travel Post COVID-19: Cognitive and Emotional Effects of Risk Perception.” Journal of Travel Research. Published online October 6. doi:10.1177/00472875211048930.

38.

Kogut

Bruce

Singh

Harbir

. 1988. “The Effect of National Culture on the Choice of Entry Mode.” Journal of International Business Studies 19 (3): 411–32.

39.

Kumar

Mukesh

Prashar

Sanjeev

Jana

R. K.

2019. “Does International Tourism Spur International Trade and Output? Evidence From Wavelet Analysis.” Tourism Economics 25 (1): 22–33.

40.

Lee

Chien-Chiang

Chen

Mei-Ping

. 2020. “The Impact of COVID-19 on the Travel and Leisure Industry Returns: Some International Evidence.” Tourism Economics. Published online November 11. doi:10.1177/1354816620971981.

41.

Lermer

Eva

Streicher

Bernhard

Sachs

Rainer

Raue

Martina

Frey

Dieter

. 2015. “The Effect of Construal Level on Risk-Taking.” European Journal of Social Psychology 45 (1): 99–109.

42.

Katsumata

Sotaro

. 2020. “The Impact of Multidimensional Country Distances on Consumption of Specialty Products: A Case Study of Inbound Tourists to Japan.” Journal of Vacation Marketing 26 (1): 18–32.

43.

Zhenlong

Huang

Xiao

Tao

Ning

Huan

Xinyue

Huang

Binghu

Xiaoming

. 2021. “ODT FLOW: Extracting, Analyzing, and Sharing Multi-Source Multi-Scale Human Mobility.” PLoS One 16 (8): e0255259.

44.

Liu

Hongbo

Xiang (Robert)

Cárdenas

David A.

Yang

. 2018. “Perceived Cultural Distance and International Destination Choice: The Role of Destination Familiarity, Geographic Distance, and Cultural Motivation.” Journal of Destination Marketing & Management 9: 300–9.

45.

McKercher

Bob

. 1998. “The Effect of Distance Decay on Visitor Mix at Coastal Destinations.” Pacific Tourism Review 2 (3–4): 215–23.

46.

McKercher

Bob

Chan

Andrew

Lam

Celia

. 2008. “The Impact of Distance on International Tourist Movements.” Journal of Travel Research 47 (2): 208–24.

47.

McKercher

Bob

Chow

So-Ming Billie

. 2001. “Cultural Distance and Participation in Cultural Tourism.” Pacific Tourism Review 5 (1–2): 23–32.

48.

Mckercher

Bob

Lew

Alan A.

2003. “Distance Decay and the Impact of Effective Tourism Exclusion Zones on International Travel Flows.” Journal of Travel Research 42 (2): 159–65.

49.

Morley

Clive

Rosselló

Jaume

Santana-Gallego

Maria

. 2014. “Gravity Models for Tourism Demand: Theory and Use.” Annals of Tourism Research 48: 1–10.

50.

Neumayer

Eric

. 2010. “Visa Restrictions and Bilateral Travel.” The Professional Geographer 62 (2): 171–81.

51.

O’Leary

Sinéad

Deegan

Jim

. 2003. “People, Pace, Place: Qualitative and Quantitative Images of Ireland as a Tourism Destination in France.” Journal of Vacation Marketing 9 (3): 213–26.

52.

Ozer Balli

Hatice

Balli

Faruk

Tsui

Wai Hong Kan

. 2019. “International Tourism Demand, Number of Airline Seats and Trade Triangle: Evidence From New Zealand Partners.” Tourism Economics 25 (1): 132–44.

53.

Pan

Bing

Doris Chenguang

Song

Haiyan

. 2012. “Forecasting Hotel Room Demand Using Search Engine Data.” Journal of Hospitality and Tourism Technology 3: 196–210.

54.

Pantano

Eleonora

Priporas

Constantinos-Vasilios

Devereux

Luke

Pizzi

Gabriele

. 2021. “Tweets to Escape: Intercultural Differences in Consumer Expectations and Risk Behavior During the COVID-19 Lockdown in Three European Countries.” Journal of Business Research 130: 59–69.

55.

Park

Sangwon

Yang

Wang

Mingshu

. 2019. “Travel Distance and Hotel Service Satisfaction: An Inverted U-Shaped Relationship.” International Journal of Hospitality Management 76: 261–70.

56.

Pernecky

Tomas

. 2020. “Critical Tourism Scholars: Brokers of Hope.” Tourism Geographies 22 (3): 657–66.

57.

Pham

Tien Duc

Dwyer

Larry

Jen-Je

Ngo

Tramy

. 2021. “COVID-19 Impacts of Inbound Tourism on Australian Economy.” Annals of Tourism Research 88: 103179.

58.

Reisinger

Yvette

Crotts

John C.

2010. “Applying Hofstede’s National Culture Measures in Tourism Research: Illuminating Issues of Divergence and Convergence.” Journal of Travel Research 49 (2): 153–64.

59.

Reisinger

Yvette

Mavondo

Felix

. 2005. “Travel Anxiety and Intentions to Travel Internationally: Implications of Travel Risk Perception.” Journal of Travel Research 43 (3): 212–25.

60.

Saayman

Andrea

Figini

Paolo

Cassella

Silvio

. 2016. “The Influence of Formal Trade Agreements and Informal Economic Cooperation on International Tourism Flows.” Tourism Economics 22 (6): 1274–300.

61.

Sagristano

Michael D.

Trope

Yaacov

Liberman

Nira

. 2002. “Time-Dependent Gambling: Odds Now, Money Later.” Journal of Experimental Psychology General 131 (3): 364–76.

62.

Schneider

Christina J.

Tobin

Jennifer L.

2020. “The Political Economy of Bilateral Bailouts.” International Organization 74 (1): 1–29.

63.

Schwartz

Shalom H.

1999. “A Theory of Cultural Values and Some Implications for Work.” Applied Psychology 48 (1): 23–47.

64.

Seabra

Cláudia

Dolnicar

Sara

Abrantes

José Luís

Kastenholz

Elisabeth

. 2013. “Heterogeneity in Risk and Safety Perceptions of International Tourists.” Tourism Management 36: 502–10.

65.

Seetaram

Neelu

. 2012. “Immigration and International Inbound Tourism: Empirical Evidence From Australia.” Tourism Management 33 (6): 1535–43.

66.

Sharma

Abhinav

Nicolau

Juan Luis

. 2020. “An Open Market Valuation of the Effects of COVID-19 on the Travel and Tourism Industry.” Annals of Tourism Research 83: 102990.

67.

Sigala

Marianna

. 2020. “Tourism and COVID-19: Impacts and Implications for Advancing and Resetting Industry and Research.” Journal of Business Research 117: 312–21.

68.

Siliverstovs

Boriss

Wochner

Daniel S.

2018. “Google Trends and Reality: Do the Proportions Match?” Journal of Economic Behavior & Organization 145: 1–23.

69.

Solomon

Sheldon

Greenberg

Jeff

Schimel

Jeff

Arndt

Jamie

Pyszczynski

Tom

. 2003. “Human Awareness of Mortality and the Evolution of Culture.” In The Psychological Foundations of Culture, edited by Schaller

Mark

Crandall

Christian S.

, 24–49. New York, NY: Psychology Press.

70.

Tasci

Asli D. A.

2009. “Social Distance: The Missing Link in the Loop of Movies, Destination Image, and Tourist Behavior?” Journal of Travel Research 47 (4): 494–507.

71.

Thyne

Maree

Watkins

Leah

Yoshida

Mari

. 2018. “Resident Perceptions of Tourism: The Role of Social Distance.” International Journal of Tourism Research 20 (2): 256–66.

72.

Tobler

Waldo R.

1970. “A Computer Movie Simulating Urban Growth in the Detroit Region.” Economic Geography 46: 234–40.

73.

Tokas

Kalpana

Deb

Arnab Kumar

. 2020. “CAGE Distance Framework and Bilateral Trade Flows: Case of India.” Management Research Review 43 (10): 1157–81.

74.

Trope

Yaacov

Liberman

Nira

. 2010. “Construal-Level Theory of Psychological Distance.” Psychological Review 117 (2): 440–63.

75.

Uysal

Muzaffer

Crompton

John L.

1985. “An Overview of Approaches Used to Forecast Tourism Demand.” Journal of Travel Research 23 (4): 7–15.

76.

Vietze

Christoph

. 2012. “Cultural Effects on Inbound Tourism Into the USA: A Gravity Approach.” Tourism Economics 18 (1): 121–38.

77.

Wang

Yan

. 2021. “Government Policies, National Culture and Social Distancing During the first Wave of the COVID-19 Pandemic: International Evidence.” Safety Science 135: 105138.

78.

Westerlund

Joakim

Wilhelmsson

Fredrik

. 2011. “Estimating the Gravity Model Without Gravity Using Panel Data.” Applied Economics 43 (6): 641–9.

79.

Wong

IpKin Anthony

. 2015. “A Multimethod Multilevel Study of Heritage Transmission: The Role of Culture on Tourist Interest and Authenticity.” Journal of Travel Research 54 (5): 672–85.

80.

Woosnam

Kyle M.

Kim

Hyun

. 2014. “Hurricane Impacts on Southeastern United States Coastal National Park Visitation.” Tourism Geographies 16 (3): 364–81.

81.

Yang

Altschuler

Benjamin

Liang

Zhengkang

Xiang (Robert)

. 2021. “Monitoring the Global COVID-19 Impact on Tourism: The COVID19tourism Index.” Annals of Tourism Research 90: 103120. doi:10.1016/j.annals.2020.103120.

82.

Yang

Fik

Timothy J.

Zhang

Hong-lei

. 2017. “Designing a Tourism Spillover Index Based on Multidestination Travel: A Two-Stage Distance-Based Modeling Approach.” Journal of Travel Research 56 (3): 317–33.

83.

Yang

Liu

Hongbo

Chen

Xiang

. 2020. “COVID-19 and Restaurant Demand: Early Effects of the Pandemic and Stay-at-Home Orders.” International Journal of Contemporary Hospitality Management 32 (12): 3809–34.

84.

Yang

Liu

Hongbo

Xiang (Robert)

. 2019. “The World is Flatter? Examining the Relationship Between Cultural Distance and International Tourist Flows.” Journal of Travel Research 58 (2): 224–40.

85.

Yang

Wong

Kevin K. F.

2012. “The Influence of Cultural Distance on China Inbound Tourism Flows: A Panel Data Gravity Model Approach.” Asian Geographer 29 (1): 21–37.

86.

Yang

Laurie

Yang

Wan

. 2018. “Does Time Dull the Pain? The Impact of Temporal Contiguity on Review Extremity in the Hotel Context.” International Journal of Hospitality Management 75: 119–30.

87.

Yang

Zhang

Hongru

Chen

Xiang

. 2020. “Coronavirus Pandemic and Tourism: Dynamic Stochastic General Equilibrium Modeling of Infectious Disease Outbreak.” Annals of Tourism Research 83: 102913.

88.

Ben Haobin

Zhang

Hanqin Qiu

Yuen

Peter P.

2013. “Cultural Conflicts or Cultural Cushion?” Annals of Tourism Research 43: 321–49.

89.

Zhang

Cheng

Qian

Li-Xian

Jian-Qiang

. 2021. “COVID-19 Pandemic With Human Mobility Across Countries.” Journal of the Operations Research Society of China 9: 229–44. doi:10.1007/s40305-020-00317-6.

90.

Zhang

Jun

Seo

Sangyun

Lee

Hoonyoung

. 2013. “The Impact of Psychological Distance on Chinese Customers When Selecting an International Healthcare Service Country.” Tourism Management 35: 32–40.

91.

Zhang

Yingsha

Xiang (Robert)

Tong (Stoltzer)

. 2019. “The Impacts of Cultural Values on Bilateral International Tourist Flows: A Panel Data Gravity Model.” Current Issues in Tourism 22 (8): 967–81.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB