Preferences and Tourism Development under Uncertainty: An Empirical Study

Abstract

Using the Global Preferences Survey dataset, this paper examines the effects of six measures of preferences (altruism, negative reciprocity, patience, positive reciprocity, risk-taking, and trust) on the per capita international tourist arrivals and the per capita incoming tourist receipts. The data focus on 74 countries for the period from 1995 to 2019. The paper finds that citizens’ trust is positively related to tourism development, and its impact is statistically significant. This evidence indicates that a country with a higher level of trust in other nations’ people attracts more tourists and generates higher tourism receipts.

1. Introduction

In this paper, we examine the effects of a country’s preferences to attract international tourists. Our hypotheses suggest that indicators of preferences can be seen as an asset to promote tourism development. Cultural variables and heritage sites motivate international tourists to visit a country [1]. However, cultural variables are somehow more critical since international tourists want to feel an “emotional experience” rather than just visiting a heritage site [2]. Similarly, international tourists want to visit “happy countries” since they seek positive and joyful experiences during their visit [2].

This paper investigates whether preferences lead to a higher number of international tourists and increase tourism receipts. Mainly, we attempt to contribute to the existing empirical literature by investigating “non-economic” (preferences) drivers of international tourism development. Previous papers have examined the effects of “non-economic” indicators on international tourism development. For instance, [3] reviewed the literature on tourism demand until the mid-1990s and observed that most of the studies have focused on tourism’s economic determinants. The authors of [4] investigated the effects of the number of world heritage sites on tourism development over 2006–2009 in 66 countries. The authors find a positive impact on the number of world heritage on international tourist arrivals. This evidence is robust to consider the countries in different regions. The authors of [2] examine the effects of happiness on international tourist attraction in the cross-sectional data of 75 countries from 1999 to 2009. The authors observe that people prefer to go to “happier” countries and prefer to spend more in happier countries since the impact of happiness on tourist arrivals and tourism spending is positive. Several previous papers have focused on the determinants of the sustainability of tourism development (see, e.g., [5,6,7,8,9]). (see the literature review of [9] for the effects of psychological indicators on tourism development).

Our paper aims to contribute to the existing literature in several ways. First, we use the Global Preferences Survey dataset, which was recently introduced by [10,11]. To the best of the authors’ knowledge, this is also the first paper in the literature that uses the Global Preferences Survey dataset to investigate the effects of tourism development preferences. For this purpose, we use six indicators of preferences: altruism, negative reciprocity, patience, positive reciprocity, risk-taking, and trust. Second, to the best of the authors’ knowledge, this is the first paper to show the validity of the hypothesis that people want to visit countries with greater trust. Finally, we examine the effects of preferences on international tourism by considering a multi-country dataset. Specifically, we use the data from 74 countries with an average period from 1995 to 2019. We find that all preference measures are positively related to tourism development, but trust in tourism development is statistically significant. This evidence indicates that a country with a higher level of trust in other nations’ people attracts more tourists and generates higher tourism receipts.

The rest of the paper is organized as follows. Section 2 explains the data, the empirical model, and the estimation procedures. Section 3 discusses the empirical results, and Section 4 provides the conclusions.

2. Data, Model, Methodology

2.1. Data and Empirical Model

Data collected for this study came from four different sources, as shown in Table 1. A total number of eleven variables were considered: per capita international tourist arrivals, per capita incoming tourist receipts, economic freedom index, change in exchange rate, heritage, patience, risk-taking, positive reciprocity, negative reciprocity, altruism, and trust. The dataset covers the average period from 1995 to 2019, and the beginning date and the number of countries is due to the availability of the tourism data. The data frequency is annual, which is averaged over time, and the dataset includes 74 developing and developed countries (see Appendix A). The countries’ selection is related to the availability of the Global Preferences Survey (GPS) dataset. We utilized the data from the GPS, a globally representative dataset on risk and time preferences. The data were collected, as well as a rich set of covariates of more than 80,000 individuals. The samples were selected from 74 countries worldwide, representing 90 percent of both the world’s population and global income. Initially, the dataset was owned by Armin Falk, and the data are described in [10,11]. The GPS data include the representative population samples in many countries and ask about social and economic issues annually.

We specified an econometric model for tourism demand to explain tourism development. The dependent variables for the model were per capita international tourist arrivals, measured by the number of persons who arrive in a given country as a tourist during the average period of 1995–2019, and per capita incoming tourist receipts (see, e.g., [12,13,14,15]). Our benchmark model is specified thus:

$$TA{\left(R\right)}_{it}={\beta }_0+{\beta }_1{X}_{it}+{\beta }_{2}PA{T}_{it}+{\beta }_{3}RI{P}_{it}+{\beta }_{4}PO{R}_{it}+{\beta }_{5}NE{R}_{it}+{\beta }_{6}AL{T}_{it}+{\beta }_{7}TR{U}_{it}+{\mu }_i+v_{it}$$

(1)

where the dependent variable TA(R) (proxies for tourism demand) is per capita international tourist arrivals and per capita incoming tourist receipts in country i at time t for which i = 1, 2, 3, … 74 and t = 1995, 1996, 1997, …, 2019. $X_{it}$ is the vector of controls, including the economic freedom index, the changes in the exchange rate, and the number of heritages. “PAT” is patience, “RIP” is risk preferences, “POR” and “NER” are positive and negative reciprocities, “ALT” is altruism, “TRU” is trust, respectively. ${\mu }_i$ is a country-specific effect while $v_{it}$ is a well-behaved error term. We test whether country-specific dummies should be included in Equation (1), and the outcome of the test suggests including them.

As a control regressor in Equation (1), the economic freedom index (EFI) captures the degree of economic freedom in the country i at time t on a scale from 0 to 100, with a higher score indicating high economic freedom. The inclusion of this variable in our model is in line with the argument that tourists are more likely to visit countries with high economic freedom (higher quality of institutions) than those with little or no economic freedom (see [16,17]). Data on this variable were taken from [18]. Heritage is also positively related to tourism development since these sites can attract more tourists [19]. The related data for the number of heritages are obtained from [20].

In the tourism demand literature, the exchange rate is reported to be a strong determining factor [21]. Changes in the country’s exchange rate imply not only changes in travel costs but a complete change in the tourists plan to travel to that country [22]. As a control regressor, therefore, we include changes in the exchange rate in Equation (1) to capture the cost of traveling and the intention to travel for tourism. We measure this variable by the nominal exchange rate of country i at time t relative to the U.S dollar, as also measured by [14]. We take data for this variable from the World Bank’s World Development Indicators [23].

Furthermore, Patience (PAT), as included in Equation (1), captures the willingness of tourists to give up something “today” for “tomorrow’s” benefit. The Risk Preferences (RIP) variable is included in Equation (1) to account for the tourists’ likelihood of taking a risk associated with tourism. As an intrinsic motivation to respond to behavior, reciprocity can either be positive or negative, with the former dealing with reward and the latter with punishment [19]. In Equation (1), we also consider the effect of both Positive Reciprocity (POR) and Negative Reciprocity (NER) on tourism development. Altruism (ALT) was measured using one qualitative and one quantitative item, both related to donations. Finally, the Trust (TRU) measure is based on one item, which asked respondents whether they assume that other people only have the best intentions. Apart from negative reciprocity and risk-taking, we expect all other preference measures to be positively related to tourism development (refer to [10,11] for the details of six indicators in the Global Preferences dataset).

2.2. Estimation Procedures

Consistent with the previous empirical literature (see, e.g., [2]), we applied the cross-country ordinary least squares (OLS), OLS estimators in estimating Equation (1), which provides relatively more long-term results than those of the correlation, as reported in Table 2. Estimating these results was conducted with robust standard errors to control the potential heteroscedasticity and autocorrelation throughout the estimated Models (1) to (7), as shown in Tables 3 and 4. We further applied the Breusch–Pagan Lagrange Multiplier (L.M.) test to ascertain whether the OLS estimators applied to our regression model fit the data under the null hypothesis that there is no cross-country effect. We rejected this hypothesis from the outcome of the test at the 5% level. Thus, we concluded that the estimations were made with the incorrect estimators, enabling us to include the dummy variable for cross-country effects.

Furthermore, in line with the Durbin–Wu–Hausman procedure, we tested for endogeneity to ensure no feedback effect among the variables. The various outcomes of the test suggest the absence of endogeneity. The outcome of the analysis indicated the rejection of this null hypothesis. Hence, we conclude that OLS cross-sectional estimations are relevant estimators. The empirical results obtained from these estimators’ application are presented in Tables 3 and 4, which we discuss in the following section.

3. Results

3.1. Preliminary Findings

We provide summary statistics of the variables in

Table 1. Descriptive statistics.

Variables	Definition	Data Source	Mean	Std. Dev.	Min.	Max.	Obs.
Per Capita International Tourist Arrivals	Logarithmic Form	[23]	−2.092	1.577	−6.470	0.930	74
Per Capita Incoming Tourist Receipts	Logarithmic Form	[23]	4.657	1.739	−0.596	7.638	74
Economic Freedom	Index from 0 to 10	[18]	6.799	0.795	4.240	8.330	74
Exchange Rate	Logarithmic Form	[23]	6.428	37.42	−4.430	321.7	74
Heritage	Number	[20]	8.657	9.636	0.000	40.63	74
Patience	Level of Index	[10,11]	0.004	0.370	−0.612	1.071	74
Risk-Taking	Level of Index	[10,11]	0.009	0.305	−0.792	0.970	74
Positive Reciprocity	Level of Index	[10,11]	−0.034	0.343	−1.038	0.570	74
Negative Reciprocity	Level of Index	[10,11]	0.008	0.277	−0.489	0.738	74
Altruism	Level of Index	[10,11]	−0.037	0.346	−0.939	0.906	74
Trust	Level of Index	[10,11]	−0.030	0.276	−0.706	0.609	74

. The mean value for our key explanatory variable (trust) is −0.030, with minimum and maximum values of −0.706 (in Uganda) and 0.609 (in Egypt), respectively. This evidence provides insight into tourists’ perception of whether people in the sampled countries have only the best intentions. Positive and negative reciprocities have a mean score of −0.034 and 0.008, with maximums of 0.570 (in Egypt) and 0.738 (in Cote d’Ivoire) and minimum values of −1.038 (in Mexico) and −0.489 (in Guatemala), respectively. Altruism has a mean score of −0.037, with a maximum of 0.906 (in Bangladesh) and a minimum value of −0.936 (in the Czech Republic). Tourists’ willingness to wait (i.e., patience) and their willingness to take risk (i.e., risk-taking) have a mean score of 0.004 and 0009 and a minimum and maximum values of −0.612 (in Nicaragua), 1.071 (in Sweden), −0.792 (in Portugal), and 0.970 (in South Africa), respectively. These results indicate that tourists are less likely to wait for future benefits and are less likely to take risk.

Furthermore, on average, the economic freedom index score for the 74 countries is 6.799, with 4.240 and 8.330 being the respective minimum and maximum values. This value indicates a considerable amount of economic freedom in the sampled countries that may be strong enough to attract tourists. Changes in exchange rate seem to be the most volatile among all the explanatory variables, with a standard deviation of 37.42 (or 37.42%), indicating “obstruction or uncertainty” to tourists.

Using the correlation matrix, we also examined the nature and strength of relationships among the variables, as presented in

Table 2. Correlation matrix.

Regressors	Tourist Arrivals	Tourist Receipts	Economic Freedom	Exchange Rate	Heritage	Patience	Risk-Taking	Positive Reciprocity	Negative Reciprocity	Altruism	Trust
Tourist Arrivals	1.000	–	–	–	–	–	–	–	–	–	–
Tourist Receipts	0.906	1.000	–	–	–	–	–	–	–	–	–
Economic Freedom	0.575	0.723	1.000	–	–	–	–	–	–	–	–
Exchange Rate	−0.012	−0.180	−0.431	1.000	–	–	–	–	–	–	–
Heritage	0.214	0.284	0.211	−0.056	1.000	–	–	–	–	–	–
Patience	0.348	0.459	0.482	−0.110	0.371	1.000	–	–	–	–	–
Risk Taking	−0.087	−0.154	−0.173	0.225	−0.154	0.243	1.000	–	–	–	–
Positive Reciprocity	0.056	0.113	0.044	−0.044	0.070	0.011	−0.258	1.000	–	–	–
Negative Reciprocity	−0.029	−0.069	−0.029	−0.074	0.139	0.273	0.188	−0.164	1.000	–	–
Altruism	−0.199	−0.163	−0.072	0.015	0.103	−0.014	−0.014	0.707	−0.138	1.000	–
Trust	0.219	0.329	0.237	−0.211	0.108	0.222	−0.076	0.369	0.144	0.278	1.000

. Where two or more variables are highly correlated (i.e., $cov\left(X_i,X_j\right)\ne 0$), any estimator applied to such data will be weak and such correlation is most likely to increase the probability of type 2 error [24].

From the correlation results, we observed that the economic freedom index was the highest correlated variable with international tourism arrival (receipt), with a correlation coefficient of 0.575. Altruism seems to be negatively related to tourism development, contrary to our expectation. However, as we expected, exchange rate changes, risk-taking, and negative reciprocity negatively correlate with the dependent variables (international tourism arrival (receipt)). In contrast, all other regressors are positively correlated with the dependent variables. We also observed that the correlation among all the regressors is weak, except between positive reciprocity and altruism. However, since most of the regressors are weakly correlated, we conclude that multicollinearity is not an issue in our estimations.

3.2. Cross-Sectional Data Estimations

Preference measures of tourism are our primary concern in this work, and thus, we will concentrate our discussion on them. In

Table 3. Benchmark regression: per capita international tourist arrivals.

Regressors	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Constant Term	−11.26 *** (1.844)	−11.43 *** (1.602)	−11.44 *** (1.585)	−11.44 *** (1.560)	−11.26 *** (1.560)	−11.16 *** (1.573)	−11.16 *** (1.573)
Economic Freedom	1.320 *** (0.258)	1.344 *** (0.227)	1.346 *** (0.225)	1.334 *** (0.222)	1.311 *** (0.223)	1.309 *** (0.223)	1.309 *** (0.223)
Exchange Rate	0.011 *** (0.002)	0.012 *** (0.002)	0.012 *** (0.002)	0.011 *** (0.002)	0.011 *** (0.002)	0.012 *** (0.002)	0.012 *** (0.002)
Heritage	0.012 (0.017)	0.013 (0.014)	0.013 (0.015)	0.015 (0.015)	0.017 (0.015)	0.012 (0.015)	0.012 (0.015)
Patience	0.123 (0.422).	–	–	–	–	–	0.154 (0.411)
Risk-Taking	–	−0.116 (0.450)	–	–	–	–	−0.133 (0.439)
Positive Reciprocity	–	–	0.152 (0.390)	–	–	–	0.157 (0.403)
Negative Reciprocity	–	–	–	−0.235 (0.496)	–	–	−0.247 (0.522)
Altruism	–	–	–	–	0.762 * (0.455)	–	0.705 (0.501)
Trust	–	–	–	–	–	0.670 ** (0.298)	0.703 *** (0.275)
Number of Countries	74	74	74	74	74	74	74
R–Squared (Within)	0.406	0.406	0.407	0.407	0.433	0.435	0.462

Notes: The dependent variable is the per capita international tourist arrivals. The robust standard errors are in (). *** p < 0.01, ** p < 0.05, * p < 0.010.

, we attempted to examine the individual effect of each of these measures on tourism development (i.e., tourists’ arrivals and tourism receipts), and regression results show that only two of these measures (altruism and trust) are statistically significant at 10% and 5% levels of significance, respectively. The economic freedom index and exchange rate changes proved to be statistically significant at a 1% significant level in all estimations. In Models (5) and (6), altruism and trust also report statistically significant values at 10% and 5% levels, respectively. However, when we include all preferences, only the coefficient of trust is statistically significant; its impact on per capita tourist arrivals is positive.

In

Table 4. Benchmark regression: per capita incoming tourist receipts.

Regressors	(1)	(2)	(3)	(4)	(5)	(6)	(7)
Constant Term	−6.288 *** (2.059)	−6.889 *** (1.727)	−6.896 *** (1.714)	−6.951 *** (1.750)	−6.776 *** (1.693)	−6.447 *** (1.654)	−6.543 *** (1.842)
Economic Freedom	1.578 *** (0.289)	1.662 *** (0.244)	1.665 *** (0.242)	1.671 *** (0.248)	1.637 *** (0.240)	1.602 *** (0.234)	1.705 *** (0.241)
Exchange Rate	0.006 *** (0.002)	0.007 *** (0.002)	0.007 *** (0.002)	0.007 *** (0.002)	0.007 *** (0.002)	0.008 *** (0.002)	0.009 *** (0.002)
Heritage	0.019 (0.013)	0.022 * (0.011)	0.022 * (0.011)	0.022 ** (0.011)	0.026 ** (0.012)	0.021 * (0.012)	0.023 * (0.013)
Patience	0.408 (0.431)	–	–	–	–	–	0.452 (0.405)
Risk-Taking	–	−0.230 (0.430)	–	–	–	–	−0.261 (0.402)
Positive Reciprocity	–	–	0.398 (0.323)	–	–	–	0.405 (0.317)
Negative Reciprocity	–	–	–	−0.257 (0.454)	–	–	−0.273 (0.481)
Altruism	–	–	–	–	−0.638 (0.489)	–	−0.689 (0.502)
Trust	–	–	–	–	–	1.135 ** (0.483)	1.241 *** (0.473)
Number of Countries	74	74	74	74	74	74	74
R–Squared (Within)	0.566	0.562	0.567	0.563	0.577	0.591	0.625

Notes: The dependent variable is the per capita incoming tourist receipts. The robust standard errors are in (). *** p < 0.01, ** p < 0.05, * p < 0.010.

, trust only proves to be the significant determiner in tourism receipts. The outcome for trust reveals that tourism destinations whose people have a comparatively higher level of trust in other people tend to attract more tourists and generate higher tourism receipts.

3.3. Discussion of the Findings

Through the application of structural equation modelling (SEM), [25] also finds that among other variables, interpersonal trust is a significant determinant of tourism institutions. However, this finding seems to relate more to tourism institutions than to tourism receipts. Similar discussions are also accentuated in various papers (see e.g., [26,27,28,29]). On the other hand, [30] describes altruism as socially acceptable activities, including philanthropy and charity. Within the purview of social exchange theory and altruistic surplus phenomenon, it is argued that some altruistic behaviors such as helping strangers (e.g., tourists) tend to make individuals vulnerable to risky social environments that are capable of attracting criminals. The authors of [28] even articulate that being a “good Samaritan” to strangers is likely to “leave one vulnerable to a crime committed by the person being helped, such as a scam or a theft of some sort.”

Contrarily, and as it relates to tourism, however, it is believed that selfless behaviors of country’s residents that can benefit tourists are more likely to promote not just one-way tourism but repeat tourism as well (see [31,32]). Within this argument, our findings indicate that altruism or altruistic behaviors have a significant impact on tourists’ arrival (see Table 3) but are insignificant for tourism receipt (see Table 4). In essence, the results in Table 3 indicate that an improvement in the residents’ level of altruistic behaviors (i.e., being more helpful, such as giving charity or donations) will, on average, increase the number of tourist arrivals, and this is statistically significant at the 10% level. Similar findings and discussions can be found in [33].

Although risk-taking and negative reciprocity have both reported a negative effect on both tourism arrival and tourism receipts, in both cases, neither of these variables is statistically significant. This evidence implies that tourists may be willing to retaliate if they have been treated unjustly, even if it will come at a cost or punish someone for unfair behavior towards themselves or a third party [19]. Note that such actions have no significant effect on their arrival at a tourism destination. This evidence is probably due to confidence that tourists may have in political and economic freedom and the rule of law, peace, and stability of a country. The outcome is also not surprising because, for the most substantial part of the countries included in our sample, there is the pervasiveness of the rule of law and a significant level of peace and order. These results are similar to [34], whose results show that perceived risk influenced attitudes toward visiting Australia.

In some tourism destinations, it is often possible to observe that some community members participate in some positive behaviors, such as helping strangers for the benefit of their community and their self-development [33]. These behaviors are argued to have some potential effect on tourism demand. However, our results presented in Table 3 and Table 4 seem to contradict this discussion.

Like the negative reciprocity, positive reciprocity also reports statistically insignificant values for tourist arrivals and tourism receipts. These results indicate that although the tourists may accept help from strangers and return such help with some positive responses/favors, such behavior does not influence tourism development. It is not unlikely that the tourists may be willing to behave positively with a tourism destination’s residents [35]. Nevertheless, such behavior does not matter to their arrival as tourists. A similar outcome is also observed for the coefficient of patience, indicating that it matters to neither tourism arrival nor tourism receipts. We could not find any prior research that supports or contradicts this outcome.

Two control variables (economic freedom and exchange rate changes) are statistically significant in estimating tourism arrival and receipts. At the same time, heritage is not a significant determiner of tourism arrival but is notable for tourism receipts.

These results indicate the importance of assessing tourists’ preferences when designing policy implications for tourism development sustainability. Particularly, trust can affect tourists’ travel decisions in more sustainable directions.

4. Conclusions

In this paper, we used the Global Preferences Survey dataset to examine the effects of preferences on tourism development. We focused on the indicators of altruism, negative reciprocity, patience, positive reciprocity, risk-taking, and trust to examine their effects on the per capita international tourist arrivals and the per capita incoming tourist receipts. We focus on the cross-sectional data from 74 countries for the period from 1995 to 2019. It is found that all preference measures are mixed related to tourism development, but there are positive effects of trust on indicators of tourism development, and they are statistically significant.

The evidence in this paper implies that a country whose people place a higher level of trust in other nations’ people attracts more tourists and provides higher tourism receipts. In terms of this evidence, there are some implications for marketing purposes. For instance, tourism firms and travel agencies can benefit from destinations’ trust characteristics, alongside classical selling issues. Therefore, it can be a good idea to use the GPS data as potential determinants of other tourism indicators. However, the results are limited for 74 countries due to the availability of the data. Future papers can focus on the purposes of visits, which can be affected by psychological and cultural indicators by using the cross-country datasets covering the COVID-19 pandemic era.