Cultural Distance and Chinese Outbound Tourism: Exploring the Moderating Effect of Geographical Distance

Abstract

In this study, the gravity model is applied via the dynamic generalised methods of moments estimation to assess the role of geographical distance in moderating the impact of culture on outbound tourism flows. The results show that cultural distance has a positive impact on Chinese outbound tourism flow. However, the effect of cultural distance on outbound tourism flow decreases as geographical distance increases since travellers to nearby destinations are more strongly influenced by cultural distance than travellers to more distant markets. These results therefore support the concept of the “diminishing effect of cultural distance” on the Chinese outbound tourism market. In addition, the results of this study serve as a basis for promoting the sustained contribution of Chinese outbound tourism to the development of destination management, which will help with the recovery of international tourism following the COVID-19 pandemic.

1. Introduction

Outbound tourism refers to international travel for the purpose of tourism and has stood at the core of tourism research in the past decades, with contributions on outbound tourism investigating travel flows out of a wide range of high-income countries such as Australia [1], South Korea [2], and the United Kingdom [3]. Following changes in the composition of incoming tourists in numerous destinations around the world, attention in the literature was recently diverted towards Chinese outbound tourism [4]. An important endeavour brought about by the democratisation of international travel hence resides in studying the factors determining flows from one specific country to another. Tourism flows between countries have commonly been studied from the perspective of macroeconomic indicators, including the gross domestic product (GDP), bilateral exchange rate, or consumer price index (CPI) of both outbound and inbound countries [5]. Beyond macroeconomic measures, another country-specific but non-economic variable was also brought into this debate to distil additional insights on factors underpinning cross-border tourism: culture [6,7,8].

Culture plays a vital role in determining how travellers select outbound destinations. Culture influences travellers’ preferences, intentions, and destination choices [7,9]. In the tourism literature, culture is often understood to reflect values, basic norms, beliefs, and ideas that develop over time [10]. Aside from the influence of national culture on the decision and decision process leading to travel decisions, another important way to leverage culture as a variable is through the use of cultural distance or, in other words, the quantified remoteness between the culture of an outbound traveller and that of the people inhabiting the potential travel destinations [11]. However, the effects of cultural distance (CD) on tourism decisions are complex. Cultural distance can be a double-edged sword in driving or dissuading tourists from patronising an international destination through cultural exploration or cultural conflict [8,12]. Therefore, how CD influences Chinese outbound tourism flows and its underlining impact mechanisms are worthy of further exploration.

Besides the central role of CD, the geographical distance (GD) between the origin country and destinations also influences travellers’ behaviours [13,14]. The GD tends to moderate the impact of CD on travellers’ destination choices according to its inconvenient (distinct cultural customs, language barriers, and different national systems) and beneficial factors (tourism attraction and novel travel experience) [15]. Specifically, compared with long-haul travel, short-haul travel entails less financial and emotional cost (e.g., travellers’ psychological uncertainty and worries on risky issues) [16]. Additionally, short-haul travellers are more willing to visit destinations they are familiar with or have previously visited to reduce their perception of ambiguity and uncertainty regarding the corresponding destination [16]. Therefore, the current study highlights the impact of CD and the moderating role of GD on Chinese outbound tourism flow.

The previous CD and tourism destination choice studies are characterised by three main limitations. First, owing to CD being relatively stable, and thus not easily changed in a short period of time [17], secondary panel data can enlarge the sample size and generalise model estimation at the national level compared with collecting primary data at the individual level. Second, the greater the GD between a home–destination pair, the more CD tends to be observed [18]. CD and GD are correlated, and sometimes even used as a proxy for one another [18]. However, there are differences between the two, which have not been explored [15]. The interaction relationship between GD and CD is therefore worthy of consideration. Third, as travellers are more willing to revisit destinations they are familiar with, tourism flow from the previous year is related to that of the current year [19], meaning that the data of outbound tourism flow tend to be dynamic and, therefore, preparing estimates using static models is inappropriate.

This study aims to investigate the effects of cultural distance and geographical distance on Chinese outbound tourism flow. In addition, an attempt is also made to evaluate the existence of an interaction between geographical distance and cultural distance. To this end, this study first investigates the effects of CD based on aggregate tourism flow at the country level. The perceived CD of individual travellers or CD scores of limited destinations cannot be unilaterally used for the assessment of tourism flows of tourist source countries [20]. Therefore, appropriate CD measurements are required when conducting macroanalyses of aggregate tourism flow data or microanalyses of travellers’ intentions at the individual level. Second, this study aims to validate the existence of a “diminishing effect of cultural distance” in travellers’ destination choices, which helps to shape patterns of the global tourism flow [21]. Knowing this interactive effect is essential for destination marketing organisations (DMOs) to formulate appropriate strategies to attract Chinese travellers and promote the sustained contribution of Chinese outbound tourism to the development of destination management. Last, this study aims to validate if the “persistence/reputation” impact could influence travellers’ decision-making process with respect to a particular destination in the context of Chinese tourism. The findings reveal that Chinese outbound travellers experience inertia of tourism destination selections, which means they are more likely to revisit destinations they have visited before [19].

A gravity model is applied to a 15-year (2005–2019) panel dataset covering 43 Chinese popular outbound destinations. The generalised method of moments (GMM) technique is applied in the gravity model to address the dynamic model setting [19]. The findings of this study reveal that CD has a positive impact on Chinese outbound tourism flow. Nevertheless, CD has a diminishing effect on Chinese outbound tourism flow as the geographical distance grows.

This paper is structured as follows: Section 2 discusses the literature on the topic of CD, GD, and the role of the latter in moderating the former. Section 3 describes the methodology used in this paper, while Section 4 lays down the empirical results and findings. Section 5 comprises discussion, conclusion, limitations, and future research. Section 6 describes the implications.

2. Literature Review

2.1. Culture and Cultural Distance

Along with the lines of market transitions (from planned to the market-driven economy) and changes in social and cultural values [22], the Chinese international market has been vastly transformed. In addition, the liberalisation of visa procedures of the destinations and the relaxation of Chinese outbound tourism policies together with growing income and living standards result in a greater emphasis being placed on leisure, tourism, and intentions to explore different cultures [23]. Therefore, Chinese outbound travellers are more willing to visit destinations with larger cultural differences [15].

However, it is not easy to quantify and conceptualise the concept of culture, due to its complexity and intangibility [24]. CD refers to the degree of cultural differences between the home country and destinations, which is mainly characterised by language barriers, living habits, and social and cultural aspects [7]. In this regard, Hofstede’s definition [6] of cultural distance provides some guidance, including standard dimensions to measure cultural differences between two countries. The proposition of cultural distance overcomes the complexity and invisibility of culture and provides a quantitative approach to studying culture-related topics [25].

Cultural distance is a standard index used to measure the cultural difference between home and target destinations [26]. Kogut and Singh’s [25] work is widely used to calculate the cultural distance from Hofstede’s [6] four or six cultural dimensions. This index is measured by the variance-corrected differences with an arithmetic average between home country a and the target destination j. By using Hofstede’s [6] four or six cultural dimension scores, Kogut and Singh’s [25] cultural distance index can be calculated using the following formula:

$$C{D}_j={\sum }_{a=1}^n\{{\left(I_{aj}-I_{ai}\right)}^2/V_a\}/n$$

(1)

where CD_j is the cultural differences of destination j from the home country i. I_aj and I_ai stand for Hofstede’s cultural dimension a in destination j and home country i. Moreover, V_a is the variance of dimension a for the overall destinations in Hofstede’s list. In addition, a stands for the cultural distance dimension, and n refers to the total number of cultural dimensions (four or six cultural dimensions).

Hofstede’s [27] study validates that both four- and six-dimensional CD can be used to measure CD. This study employs four- and six-dimensional CD to test the robustness of Hofstede’s CD measurement in the context of tourism. Four-dimensional CD comprises the aspects of power distance index (PDI), individualism–collectivism (IDV), masculinity–femininity (MAS), and uncertainty avoidance index (UAI). Six-dimensional CD is extended upon the basis of CD4 with the addition of long-term orientation (LTO) and indulgence–restraint (IND) [27].

2.2. Geographical Distance and Distance Decay Theory

The geographical distance refers to the distance in kilometres between the major cities in the home country and destinations [28]. McKercher et al., [29] found that an increase in the geographical distance between two countries would yield a decline in the demand for goods or services. In addition to the limited number of studies on tourism flow, those that have focused on tourism demand lack empirical analysis of geographical distances. Cheung and Saha [30] and Liu et al., [17] both support the notion that culture similarity can positively influence tourism demand. However, whether this positive effect of cultural similarity exists in destinations with different geographical distance has not been discussed.

A paradox is subsequently identified when comparing differences in the behavioural intentions between short- and long-haul travellers, where the role of geographical distance cannot be neglected [26]. Based on distance decay theory, tourism studies have pinpointed that geographical distance has a crucial influence on tourism flow [14,29]. Except for the influence of geographical distance on traveller movements, geographical distance also has a filtering effect on travel motivation and tourism decision making [14]. Yan [31] tested the “filtering effect” of geographical distance on travellers visiting Hong Kong from closer and more distant countries and found that travellers from long-haul destinations were effectively excluded [31]. Hence, Chinese outbound travellers can be potentially filtered from visiting certain outbound destinations because of high travel expenses and potential high-risk problems [32].

2.3. Moderating Role of Geographical Distance on Cultural Distance

Lankhuizen and De Groot [33] used Hofstede’s [6] cultural distance index to investigate how culture influences trade. Through their work, they illustrated the negative effect of culture on bilateral trade. In the field of tourism studies, most existing research similarly support the negative influence of cultural distance on outbound tourism flows [7,34]. Yang et al., [34] supported the above proposition by observing, from their empirical results, that Chinese travellers are more likely to travel to destinations with similar cultures. Interestingly, the results of Yang et al., [34] were not consistent with existing studies that support the positive impact cultural differences may have on travellers’ destination intentions [32,35].

This inconsistency can be explained by the distance decay theory, wherein short- and long-haul travellers differ significantly in their travel motivations, behaviours, and preferences [35]. This is because the geographical distance may effectively filter out travellers from visiting a more culturally distant destination according to time availability, financial budget, willingness, and language ability. The cost effect of cultural distance is determined by numerous factors inconveniencing outbound travellers, including distinct cultural customs, language barriers, and different national systems [32]. Moreover, the demand effect refers to the beneficial factors obtained from the cultural distance, including tourism attraction and novel travel experience [15]. Therefore, from both demand and cost aspects, the effect of geographical distance may interact with the effect of cultural distance in impacting on travellers’ international flows. In this regard, geographical distance provides a promising way to moderate cultural distance and Chinese outbound tourism issues. Since few studies have examined this interactive effect, we have attempted to address this research gap in our study to explain how geographical distance and cultural distance work together to influence travellers’ outbound destination choices.

3. Methodology

3.1. The Gravity Model

In this study, the gravity model is applied to investigate the impact of cultural distance and geographic distance influencing Chinese outbound tourism flow. The gravity model [36,37], which is extended from Newton’s law of gravitation force, is used to estimate flows of international trade. The gravity model allows the GD between two places to be weakened or strengthened by CD [15,38,39].

In this study, the definitions applied in this model are as follows: This study defines Y_ijt as Chinese outbound tourism flow between China i and destinations j in year t. GDP_it represents Chinese real GDP per capita (USD), while GDP_jt is real GDP per capita of outbound destinations (USD). Moreover, GD_ij represents the geographical distance between Shanghai (China) and the capital of destinations. The rationale is that Shanghai is one of the main contributors to outbound Chinese tourism, given that it is at the centre of the East China coastline [40]. The gravity model is

$$Y_{ijt}=f\left[\frac{GD{P}_{it}\times GD{P}_{jt}}{G{D}_{ij}}\right]$$

(2)

Moreover, a natural logarithm is used to transform both positive and non-dummy variables in order to lower the scale and heteroscedasticity effects. The regression equation is specified as follows, where µ_ij stands for the individual effect and ε_ijt is the random error term:

$$ln{Y}_{ijt}={\beta }_0+{\beta }_{1}lnGD{P}_{it}+{\beta }_{2}lnGD{P}_{jt}+{\beta }_{3}lnG{D}_{ij}+u_{ij}+{\mathsf{\epsilon }}_{ijt }$$

(3)

This study adopts the gravity model to explain Chinese outbound tourism flow, including variables to control for geographic, culture, economic conditions, special incidents, and the potential interaction term between CD and GD, shown as follows:

$$\begin{gathered} ln{Y}_{ijt}={\beta }_0+{\beta }_{1}lnGD{P}_{it}+{\beta }_{2}lnGD{P}_{jt}+{\beta }_{3}lnG{D}_{ij}+{\beta }_{4}lnC{D}_{ij}+{\beta }_5\left(lnG{D}_{ij}\right)\times \left(lnC{D}_{ij}\right)+{\beta }_{6}lnRE{R}_{ijt} \\ +{\beta }_7{t}_{2009}+{\beta }_{8}contig+u_{ij}+{\mathsf{\epsilon }}_{ijt } \end{gathered}$$

(4)

where Y_ijt is the total number of overnight tourists from home country i (China) to destination j. GDP_it and GDP_jt are GDP in country i and j, respectively. GDP in the home country can be interpreted as a proxy of Chinese travellers’ income level that influences outbound tourism decisions via travel budgets. GDP in destinations can be considered as the quality indicator, security, and health systems at the destination that influence travellers’ outbound tourism decisions via considerations of the destinations [41]. GD_ij is the geographical distance (in km) between home country i and destination j, which potentially influences the demand for tourism products or services [29]. CD_ij is the key research variable of interest in this paper; it refers to the cultural distance, which can quantitatively reflect the cultural differences between i and j [7]. Moreover, this study uses the interaction terms of GD_ij and CD_ij to explore their interactive effects on Chinese outbound tourism flow. The setting of the interaction term is based on the distance decay theory [35] that travellers’ destination selections tend to be influenced by cost (inconvenience factors caused by CD) and demand (beneficial factors caused by CD) effects. GD can be used to moderate CD and outbound tourism flow according to time availability, language ability, financial budget, and travel willingness to visit a nearby or a more culturally distant destination [15,32]. RER_ijt (real exchange rate between China i and destinations j, using RMB as the fixed currency) is a control variable in Equation (4) because it is a crucial factor influencing Chinese outbound tourism supply [42]. Considering the impacts of the financial crisis and spatial distance to tourism flow separately, two dummy variables are added in Equation (4), which are t₂₀₀₉ (the impact of the financial crisis in 2009; t₂₀₀₉ = 1, 0 in other years) and contig (the border relationship between the i and j, border = 1, no border = 0) [15].

3.2. Methodology

Ordinary least squares (OLS), the fixed effects model, and the random effects model can be used in estimating the gravity model. However, it is not possible to use the OLS method to solve unobserved heterogeneity problems, with its tendency to yield inconsistent and inefficient estimators (see

Table A1. Results of the OLS estimation.

	OLS
	lntour
lnGDPit	1.843 ***
	(0.143)
lnGDPjt	0.063 ***
	(0.023)
lnGDij	−1.550 ***
	(0.108)
lnCD6ij	2.368 **
	(1.185)
(lnGDij) × (lnCD6ij)	−0.269 **
	(0.135)
lnRERijt	−0.293 ***
	(0.023)
t2009	(0.217)
	(0.189)
contig	1.760 ***
	(0.344)
constant	2.345
	(1.493)
VIF	1.660
N	644

Standard errors in parentheses (** p < 0.05, *** p < 0.01). VIF value is calculated based on the entire independent variables except for the intersection term of (lnGD_ij) × (lnCD6_ij).

). The result from the Hausman test (Prob > chi2 = 0.0000) supports the adoption of the fixed effects model instead of the random effects model. The fixed effects model is an appropriate method to deal with unobserved heterogeneity problems, but it cannot deal with time-invariant variables (e.g., cultural distance and geographical distance). In this regard, the least squares dummy variable (LSDV) fixed effects model with year fixed effects is adopted when time-invariant variables are included in the gravity model [9]. At the same time, the LSDV fixed effects model results do not allow the rejection of the null hypotheses of “no year fixed effects” at a 5% significance level, which means that approaches based on the use of a fixed effects model are not appropriate for estimation (See

Table A2. Results of the LSDV-fixed effect estimation.

	LSDV-FE
	lntour
lnGDPit	1.815 ***
	(0.236)
lnGDPjt	0.062 ***
	(0.023)
lnGDij	1.549 ***
	(0.108)
lnCD6ij	2.406 **
	(1.192)
(lnGDij) × (lnCD6ij)	−0.273 **
	(0.136)
lnRERijt	0.293 ***
	(0.023)
t2009	−0.365
	(0.225)
contig	1.778 ***
	(0.346)
2005.yr	0.000
	(.)
2006.yr	−0.022
	(0.244)
2007.yr	−0.12
	(0.233)
2008.yr	−0.229
	(0.228)
2009.yr	0.000
	(.)
2010.yr	−0.327
	(0.223)
2011.yr	−0.318
	(0.223)
2012.yr	−0.251
	(0.224)
2013.yr	−0.212
	(0.227)
2014.yr	−0.238
	(0.231)
2015.yr	−0.130
	(0.235)
2016.yr	−0.083
	(0.240)
2017.yr	−0.028
	(0.246)
2018.yr	−0.012
	(0.252)
2019.yr	0.000
	(.)
constant	2.722
	(2.195)
N	644.000

Standard errors in parentheses. ** p < 0.05, *** p < 0.01.

).

In a static framework of panel data, the possibility of dynamics and endogeneity cannot be satisfied. In the tourism context, Naude and Saayman [19] argued that the “persistence/reputation” impact could influence travellers’ decision-making process of a particular destination. Compared with a completely new and unfamiliar destination, travellers are more likely to return to destinations they have visited before and to recommend this place to their friends when they have had a good travel experience. To overcome the endogeneity problem, system-generalised methods of moments estimator (SYS-GMM) is an appropriate approach to estimate models with highly persistent dependent variables and allows the existence of time-invariant variables, shown as Equation (5):

$$\begin{gathered} ln{Y}_{ijt}={\beta }_0+{\beta }_{1}ln{Y}_{ijt-1}+{\beta }_{2}lnGD{P}_{it}+{\beta }_{3}lnGD{P}_{jt}+{\beta }_{4}lnG{D}_{ij}+{\beta }_{5}lnC{D}_{ij}+{\beta }_6\left(lnG{D}_{ij}\right)\times \left(lnC{D}_{ij}\right)+{\beta }_{7}lnRE{R}_{ijt}+ \\ {\beta }_8{t}_{2009}+{\beta }_{9}contig+u_{ij}+{\epsilon }_{ijt } \end{gathered}$$

(5)

The endogeneity problem is likely to yield inaccurate estimations by OLS when lnY_ijt−1 tends to be correlated with lnY_ijt. Instrumental variables are capable of overcoming this problem. This study uses the lagged term lnY_ijt−1 as the instrumental variable for its strong relationship with lnY_ijt, and it is uncorrelated with the error term of lnY_ijt−1. Two-step GMM is used in this study because it results in more reliable estimation, and it is therefore capable of providing robust estimates when the number of destinations exceeds the time series dimensions in the current study [43].

3.3. Research Design

The aims of this study are to (a) analyse the impact of cultural distance on Chinese outbound tourism flow and (b) identify and discuss the moderation effect of geographical distance in influencing the impact of cultural distance on Chinese outbound tourism flow. For these purposes, the gravity model is applied to investigate the underlying factors that influence Chinese outbound tourism flow [15,38,39]. The gravity model has the advantage of specifying a positive association between trade flows and the economic masses of two bilateral countries and a negative association between the trade flows and the associated geographical distance [44], enabling scholars to explore outbound tourism flows from both economic and non-economic perspectives [45].

This study adopts three groups of independent variables. The core factor, cultural distance (CD), is included in the first group. Previous studies reveal that CD has a significant effect on travellers’ destination selections [33,34,35]. However, whether CD has positive or negative effects is inconsistent, so it becomes crucial to examine how CD influences Chinese travellers’ overseas destination choices. The second group consists of standard gravity variables and economic variables: real GDP per capita, geographical distance (GD) between the home country and destinations in the pair, and real bilateral exchange rate (RER). GDP in home countries and destinations and GD between them are the main variables in the setting of gravity model [37]. RER is an effective control variable that can influence outbound tourism decisions [42]. The last group includes two dummy variables: one is the common border (contig) and the other is the year dummy (t2009-financial crisis), which are widely used in international tourism studies [15,46]. The details and data sources for the variables are summarised in

Table 1. Variable explanation and data source.

Category	Variable	Definition	Data Source
Dependent Variable	Yijt	Number of Chinese outbound overseas travellers per destination	Euromonitor Database
Independent Variable	GDPit	Chinese real GDP per capita (US dollars)	CEPII-Chelem Database: http://www.cepii.fr/CEPII/en/bdd_modele/bdd.asp (accessed on 1 June 2020)
	GDPjt	Real GDP per capita of outbound destinations (US dollars)
	GDij	Geographical distance between Shanghai and capitals of destinations (km)	Distance calculator: http://zh.thetimenow.com (accessed on 1 June 2020)
	CDij	Cultural distance between home country and destinations	Hofstede official website: http://www.geerthofstede.com (accessed on 1 June 2020)
	RERijt	Real exchange rate between China and destinations	Word Bank Database: https://data.worldbank.org.cn (accessed on 1 June 2020)
	ERijt	Nominal exchange rate between China and destinations	Word Bank Database: https://data.worldbank.org.cn (accessed on 1 June 2020)
	CPIit CPIjt	Consumer Price Index for China Consumer Price Index for destinations	Word Bank Database: https://data.worldbank.org.cn (accessed on 1 June 2020)
Dummy	t2009	$\mathrm{Dummy} (\mathrm{financial} \mathrm{crisis} \mathrm{impact}; t_{2009}$$=1, t_{else}$ = 0)
	contig	Dummy (border = 1, else = 0)

.

RER is calculated using the following Formula (6). ER_it is the nominal exchange rate of RMB per USD while ER_jt is the nominal exchange rate of the destination currency per USD.

The nominal exchange rate of the destination currency per RMB is ER_ijt = ER_jt/ER_it. CPI_it and CPI_jt are the consumer price index for China and destinations, respectively. The real exchange rate RER_ijt can be defined as Equation (6).

$$RE{R}_{ijt}=\frac{E{R}_{jt}\times CP{I}_{it}}{E{R}_{it}\times CP{I}_{jt}}$$

(6)

Although Beijing is the capital of China, Shanghai ranks first as the place of departure in terms of outbound tourist flow. Considering the availability of cultural distance indices from the official Hofstede website, this research comprises a final selection of 43 destinations covering a varied range of geographical distances of Chinese outbound tourism, from some nearby destinations (e.g., Taiwan, Hong Kong, etc.) to some more distant destinations (e.g., US, South Africa, etc.) (see

Table 2. Chinese outbound destination list.

Continent	Destinations
Asia (10)	Taiwan, Hong Kong, South Korea, Japan, Philippines, Vietnam, Bangladesh, Thailand, Malaysia, Singapore
Europe (17)	Italy, France, Germany, UK, Poland, Ireland, Austria, Belgium, Denmark, Netherlands, Czech Rep., Norway, Portugal, Switzerland, Spain, Hungary, Russia
Oceania (2)	Australia, New Zealand
America (8)	US, Canada, Argentina, Brazil, Ecuador, Columbia, Peru, Chile
The Middle East (2)	UAE, Israel
Africa (4)	South Africa, Egypt, Morocco, Kenya

). The panel data employed in this study range from 2005 to 2019 (15 years) because the tourism data after 2020 tends to be influenced by COVID-19. The generalised method of moments (GMM) technique is applied in the gravity model because Chinese outbound travellers may have strong intentions to revisit destinations they are familiar with [15,19,47].

4. Results

Table 3. Descriptive statistics of the variables.

Variable	Obs	Mean	Std. Dev.	Min	Max
lnYijt	645	5.234	1.927	0.262	9.899
lnGDPit	645	8.662	0.333	8.040	9.138
lnGDPjt	645	9.819	1.231	6.442	11.583
lnGDij	645	8.839	0.804	6.532	9.886
lnCD6ij	645	0.863	0.623	−1.145	1.943
lnCD4ij	645	0.677	0.857	−1.915	1.943
lnRERijt	645	0.101	2.456	−2.726	7.946
lnERijt	645	0.103	2.470	−2.723	8.156
t2009	645	0.067	0.250	0	1
contig	645	0.068	0.252	0	1

shows the descriptive statistics for all variables. The results reveal that the bilateral data are time-sensitive (e.g., Chinese outbound tourism flow, nominal and real exchange rate), and GDP in destinations show relatively high volatility because of transnational differences. In comparison, “year fixed effects” variables (e.g., cultural distance, geographical distance) and Chinese real GDP are relatively stable. The unit root test (IPS with demean unit root test is used to mitigate the cross-sectionally dependence impact) is also used, and the result shows that the temporal variables are stationary at the 5% significance level (see

Table 4. Results of the Unit root test (for time series variables).

Test Method	lnYijt	lnGDPjt	lnRERijt
IPS with demean unit root test	−1.896 ** (0.029)	−2.513 *** (0.006)	−2.654 *** (0.004)

p-value in parentheses. ** p < 0.05, *** p < 0.01.

). Moreover, the variables in this model specification are free of multicollinearity problems, because the VIF value is from 1.03 to 2.29. AR (2) and the Sargan test result in

Table 5. Results of the SYS-GMM estimation.

	GMM1	GMM2	GMM3	GMM4	GMM5
	lntour	lntour	lntour	lntour	lntour
lnYijt−1	0.862 ***	0.856 ***	0.867 ***	0.871 ***	0.856 ***
	(0.011)	(0.012)	(0.010)	(0.012)	(0.011)
lnGDPit	0.242 ***	0.251 ***	0.254 ***	0.253 ***	0.243 ***
	(0.016)	(0.017)	(0.022)	(0.024)	(0.017)
lnGDPjt	0.120 ***	0.106 ***	0.032 *	0.01	0.109 ***
	(0.022)	(0.020)	(0.018)	(0.023)	(0.020)
lnGDij	−0.173 ***	−0.083	−0.044	−0.004	−0.046
	(0.035)	(0.058)	(0.051)	(0.048)	(0.055)
lnCD6ij	0.421 ***	2.455 ***			2.916 ***
	(0.043)	(0.793)			(0.752)
(lnGDij) × (lnCD6ij)		−0.233 ***			−0.290 ***
		(0.090)			(0.085)
lnCD4ij			0.410 ***	1.487 ***
			(0.035)	(0.493)
(lnGDij) × (lnCD4ij)				−0.121 **
				(0.056)
lnRERijt	0.008	0.013	0.012	0.010
	(0.008)	(0.011)	(0.013)	(0.013)
lnERijt					0.025 **
					(−0.011)
t2009	−0.153 ***	−0.154 ***	−0.150 ***	−0.154 ***	−0.151 ***
	(0.006)	(0.008)	(0.006)	(0.008)	(0.007)
contig	0.241 ***	0.253 ***	0.223 ***	0.263 ***	0.244 ***
	(0.028)	(0.032)	(0.011)	(0.014)	(0.031)
constant	−1.287 ***	−1.910 ***	−1.595 ***	−1.729 ***	−2.149 ***
	(0.392)	(0.397)	(0.363)	(0.352)	(0.383)
Sargan (p-value)	0.3942	0.3573	0.3483	0.3492	0.3619
AR(1) (p-value)	0.0004	0.0004	0.0004	0.0003	0.0004
AR(2) (p-value)	0.8286	0.8311	0.8310	0.8309	0.8345
N	602	602	602	602	602

Note: lnY_ijt is the lagged term of lnY_ijt. AR (1) and AR (2) are the test for first-order and second-order serial correlations, respectively. Standard errors in parentheses (* p < 0.1, ** p < 0.05, *** p < 0.01).

demonstrate that the model satisfies the null hypotheses of SYS-GMM estimation (“no second-order autocorrelation” and “effective instrumental variables”).

The overall regression results (see Table 5) revealed that the coefficients of innate explanatory variables in the gravity model are in line with expectations. In particular, the number of outbound travellers from different destinations is directly proportional to the economic scale (GDP) of China and the destination but negatively influenced by the geographical distance between China and the destination. This confirms the applicability of the tourism gravity model. In terms of economic factors, the exchange rate is positively correlated with Chinese outbound tourism flow. For the dummy factors, “Border with China” has a positive impact on the number of Chinese travellers to outbound destinations, whereas the 2009 financial crisis has a significant negative impact on Chinese outbound tourism to various destinations. In addition, the lagged term of dependent variables has a significant positive impact on Chinese outbound tourism flow. This indicates that Chinese outbound travellers experience strong inertia to a certain destination, which is consistent with findings in the existing tourism literature [15,47]. In this sense, the preferences and habits of Chinese outbound travellers have a crucial impact on the selection of outbound destinations.

In order to test the robustness of the dynamic model setting, this study adopted both four- and six-dimensional cultural distance measurements developed by Kogut and Singh [25]. For both cultural distance indices, lnCD has a significant positive impact on Chinese outbound tourism flow in GMM estimation (1) and (3) in Table 5. When adding the interaction term, cultural distance still has a significant and positive influence, while the coefficients of the interaction term of (lnCD) × (lnGD) are negative in GMM estimation (2) and (4). A simple slope diagram was established to verify the interaction effect between cultural distance and geographical distance on outbound tourism flow [48]. The distribution of geographical distance is shown in Figure 1 with a lower quartile and upper quartile to represent the low geographical distance level (Low GD) and high geographical distance level (High GD). Figure 2 shows the relationship between CD and outbound tourism flow at the low GD versus the high GD by taking GMM (2) in Table 5 as an example [49].

The slope of low GD is steeper than that of high GD, indicating that there is an interaction effect between GD and CD. The difference in slopes between the two implies that the effect of CD on outbound tourism flow at the low GD is greater than that at the high GD. When Chinese travellers explore places with high cultural differences (high CD), they tend to choose nearby destinations (low GD), while others that visit places with low cultural differences (low CD) prefer to visit distant destinations (high GD). In other words, short-haul Chinese travellers (low GD) are more willing to explore the cultural differentiated characteristics of tourism destinations (high CD), while long-haul Chinese travellers (high GD) are less willing to perceive cultural differentiated characteristics (low CD).

5. Discussion, Conclusions, Limitations, and Future Research

5.1. Discussion

Firstly, the results of the SYS-GMM estimation indicate that there is a significant positive relationship between cultural distance (in terms of CD4 and CD6) and the number of Chinese travellers to that destination, where CD4 stands for four-dimensional CD, including power distance index (PDI), individualism–collectivism (IDV), masculinity–femininity (MAS), and uncertainty avoidance index (UAI); CD6 is extended upon the basis of CD4 with the addition of long-term orientation (LTO) and indulgence–restraint (IND), which validate the robustness of Hofstede’s cultural distance measurement in investigating outbound tourism flow [27]. Use of CD6 reveals the greater impact on tourism flow owing to travellers’ preferences for future-oriented cultural tourism market (LTO), and the degree of social indulgence in desire and pleasure (IND) are highly correlated with traveller satisfaction [50].

Secondly, the real exchange rate is positively correlated with Chinese outbound tourism flow in GMM regression (1)–(4). Moreover, the nominal exchange rate has a significant positive influence on Chinese outbound tourism flow in GMM regression (5). A possible explanation is that the nominal exchange rate could indicate the home country’s international competitiveness compared with the destination exchange market. Travellers usually exchange foreign currencies from banks or use credit cards because they are likely to consume goods or services in particular destinations [47]. Therefore, the nominal exchange rate could be a more appropriate control variable in international tourism flow studies [51]. This study uses RMB as a fixed currency. In this sense, it is indicated that appreciation of RMB could promote Chinese outbound tourism flow. This is because appreciation of RMB would reduce travellers’ tourism costs and further lead to an enhancement of the Chinese outbound tourism supply.

Finally, there is a negative relationship between geographical distance and the number of Chinese travellers to that destination. These results are consistent with the conclusion of McKercher’s [14] study that tourism flow declines with the growth of geographical distance between source and outbound markets. However, the geographical distance displays a significant negative influence in GMM regression (1) while presenting a negative impact on Chinese outbound tourism flow in GMM regression (2)–(4). This implies that the influence of geographical distance may not be a simple linear impact but tend to interact with other factors. According to McKercher [14]), the influence of geographical distance on demand is not a deterministic factor by itself. This study reveals that there is a significant negative interactive effect of geographic distance and cultural distance on the number of Chinese travellers to that destination. This implies that geographical distance has a significant effect under the joint interaction of cultural distance, which echoes the distance decay theory of increases in cost brought about by greater geographical distances [13,29]. Therefore, the impacts of cultural distance on outbound tourism flow can be moderated by geographical distance. The travel decisions of travellers visiting long-distance destinations are less likely influenced by cultural differences than those of travellers visiting short-haul destinations. Thus, this study helps to explain the “diminishing effect of cultural distance” between cultural distance and Chinese outbound tourism flow [21]. Further explanations will be provided in the following section.

5.2. Conclusions, Limitations, and Future Research

This study aims to determine the effects of cultural distance (and geographical distance on Chinese outbound tourism flow. In addition, this study also attempts to evaluate the existence of the interactive effect of geographical distance and cultural distance (CD) on the outbound tourism flow. The findings of the current study reveal that CD positively influences Chinese travellers’ outbound tourism choices. However, Chinese travellers are more likely to explore greater cultural differences in short-haul destinations, while they are more willing to visit places with fewer cultural distances in more distant destinations.

This study contributes to international tourism research by validating the existence of the “diminishing effect of cultural distance” in travellers’ destination choices by employing 15-year (2005–2019) dynamic panel data with dynamic generalised methods of moments (GMM) estimation. According to the results of the gravity model, CD may have a more substantial impact on outbound tourism flows in nearby tourism destinations. In contrast, a weaker influence of CD on the outbound tourism flows may be found in more distant destinations.

This study has a few limitations that could limit the generalisability of the findings. First, the cultural distance of only 43 destinations was tested. Future research should include other countries. Secondly, Hofstede’s culture dimensions oversimplify the cultural diversity between countries since they do not allow for variation over time or reflect personal diversity. In this sense, future studies should attempt to investigate the cultural distance at an individual, rather than aggregated, level through the collection of primary data. Finally, it is suggested that more trend factors (e.g., migration and globalisation) that can co-vary with CD be included in future research.

6. Implications

6.1. Theoretical Implications

Firstly, this study contributes to international travel research by demonstrating that CD positively influences Chinese travel behaviours. These results support the “benefit of foreignness” theory that travellers are strongly motivated to visit more culturally distant destinations [35]. Existent studies on investigating developed countries’ outbound markets usually reveal a negative impact of CD on travellers’ destination selections [7,10]. However, Chinese travellers’ demand for experiencing the local culture of destinations remains strong even though the interest in local culture from travellers of other countries is decreasing. Chinese travellers’ growth in purchasing power has encouraged them to sample cultural activities and consume cultural products [52].

Secondly, the results of this study indicate that Chinese travellers tend to participate in short-distance travel. The results also support the distance decay theory that tourism flow declines as the GD between two places increases. However, compared with time consumption and language barriers, saving transportation costs is not the main reason Chinese outbound travellers visit short-haul destinations [53]. As mentioned in the Chinese outbound tourism expenditure report [54], Chinese travellers pay increased attention to experiencing local cultures and customs when travelling abroad. Except for the standard products such as “air ticket + hotel”, the demand for personalisation and quality in tourism is prominent. From this point of view, embedding cultural factors to improve the quality of the personalised travel experience is particularly important, especially in attracting Chinese travellers to visit short-haul destinations. Therefore, a more in-depth understanding is needed to support short-distance travel. This study highlights a research gap in studies investigating the factors that explain Chinese travellers visiting short-haul destinations.

Thirdly, an interesting finding is made in that Chinese travellers are willing to explore greater cultural differences in short-haul destinations, for instance, Japan and Singapore. The underlying reason may be the homogenisation of attractions in such destinations. For example, when travellers’ primary travel purpose is to visit Disneyland, they can go to Hong Kong or Japan instead of the United States. The homogenisation of attractions supports the argument that short-distance tourism can provide the experience when there are greater cultural differences [55]. Moreover, Chinese travellers are more likely to visit places with fewer cultural differences in long-haul destinations. A possible explanation is that long-haul Chinese travellers have fewer preferences in perceiving cultural difference characteristics. For example, when Chinese travellers arrive at some long-haul destinations, such as the UK or the US, they will generally visit Chinatown for its restaurants and entertainment sectors [56]. However, cultural differences are usually greater in more distant destinations than in nearby destinations. Therefore, there are contradictions in this tourism mode, although these contradictions lead to interaction effects between CD and GD. The effect of interaction, which may exist in Chinese outbound travellers, has not previously been investigated in the existing literature. The contradiction theory can be explained by travellers’ desire to gain cultural experience but without the intention of visiting more distant destinations. A possible explanation is that China’s main holidays are at most seven days [57], so the remaining time for exploring the local culture is insufficient when deducting the flight time. At the same time, Asian destinations can increasingly provide cultural experiences that Chinese travellers are interested in. Therefore, determining Chinese travellers’ preferences in response to perception of cultural differences in nearby destinations is a potential research topic for researchers to investigate further, especially since Chinese travellers may be reluctant to participate in long-distance travel after COVID-19.

6.2. Managerial Implications

Firstly, knowing the way to promote the sustainability of Chinese tourism is crucial for the worldwide tourism market [58]. From this point of view, this study provides valuable insights into China. As the world’s largest developing country in terms of population, China has also become the world’s largest source of outbound travellers [59]. Therefore, tourism practitioners must provide products that meet travellers’ consumption habits and psychological needs by understanding Chinese travellers’ choices. This study reveals that CD positively impacts Chinese outbound travellers’ destination choices. Compared with other developed countries and regions, China’s outbound tourism market is not entirely mature, which means many Chinese citizens are still first-time outbound travellers [60]. Therefore, to “experience unprecedented travel” by combining the local cultural factors from attractions, local food, entertainment, and shopping activities is a strong requirement for Chinese outbound travellers. This study provides a reference for researchers to further investigate the cultural factors influencing Chinese outbound travellers in selecting their destinations after the COVID-19 pandemic [61]. Accordingly, the findings will help the recovery of international tourism since Chinese outbound travellers are the major source of international tourism.

Secondly, although outbound tourism development has undergone more than ten years of changes, the results of this study still support McKercher and du Cros’s [35] findings that travellers from more culturally distant places are more easily motivated to visit a particular destination for cultural reasons. This implies that Chinese outbound travellers are willing to perceive the characteristics of culture in tourism destinations. A possible reason is that potential Chinese travellers’ interest in experiencing foreign cultures and interacting with locals has not changed. The culturally different characteristics of destinations still attract Chinese outbound travellers to travel internationally, which echoes Brown’s [62] viewpoint that cultural immersion is the primary motivation for travellers. Therefore, the results of this research suggest that DMOs are needed to promote the cultural immersion experience by re-orientating tourism services and products to attract Chinese travellers. However, CD between China and destinations cannot vary over time. Thus, it is more effective to promote the perceived CD of Chinese travellers [63]. For example, DMOs can develop the local cultural competence by offering customised tourism services and specific tourism routines with cultural attractions and heritage to cater to distinct travellers’ desires and preferences. Moreover, specific and accurate advertising, marketing strategies, and safety protection policies are needed to stimulate tourism flows and reduce travellers’ psychological worries about safety issues, especially in the post-COVID-19 period [64].

Finally, this study highlights the interaction effect between GD and CD on Chinese outbound travellers’ choice of destinations, which indicate that the positive influence of CD on Chinese outbound tourism flow for nearby destinations is more substantial. For example, of the five Universal Studios, the three in Asia attract more Chinese travellers than the ones in America [65]. Moreover, Asian cultures, including Korean Hallyu and Malaysian and Indonesian Nyonya, represent a natural attraction for Asian travellers due to their greater relatedness to China [66]. Therefore, this kind of cultural attractiveness is strengthened. At the same time, Chinese travellers can explore European characteristics in Asian destinations. For example, travellers visiting Macao usually have a strong impression of Portuguese culture [67]. Therefore, DMOs could actively strengthen the image of a destination in potential tourism source markets to improve its visibility. For instance, virtual reality (VR) positively improves travellers’ attitudes toward a particular destination. Furthermore, theme parks, local cultural characteristics, and areas with an exotic touch have greatly attracted Chinese travellers. The above tourism marketing strategies could motivate Chinese outbound travellers to visit overseas destinations. At the same time, the images and popularity of the destinations are effectively developed.