The Impact of Technology, Economic Development, Environmental Quality, Safety, and Exchange Rate on the Tourism Performance in European Countries

Abstract

The study investigates the contribution of technology (TECH), quantified by Internet penetration, in influencing tourism performance (TP) among the top ten touristic nations in Europe: France, Spain, Italy, Turkey, the United Kingdom, Germany, Greece, Austria, Portugal, and the Netherlands. Using panel data from 2000–2022, the study includes additional structural controls like environment quality, gross domestic production (GDP) per capita, exchange rate (ER), and safety index (SI). The Method of Moments Quantile Regression (MMQR) is employed to capture heterogeneous effects at different levels of TP, and Driscoll–Kraay standard error (DKSE) correction is employed to make the analysis robust against autocorrelation as well as cross-sectional dependence. Spectral–Granger causality tests are also conducted to check short- and long-run dynamics in the relationships. Empirical results are that TECH and SI are important in TP at all quantiles, but with stronger effects for lower-performing countries. Environmental quality (EQ) and GDP per capita (GDPPC) exert increasing impacts at upper quantiles, suggesting their importance in sustaining high-level tourism economies. ER effects are limited and primarily short-term. The findings highlight the need for integrated digital, environmental, and economic policies to achieve sustainable tourism development. The paper contributes to tourism research by providing a comprehensive, frequency-sensitive, and distributional analysis of macroeconomic determinants of tourism in highly developed European tourist destinations.

1. Introduction

Tourism has become one of the most vital and rapidly expanding industries within the modern global economy, being significantly responsible for earning GDP, jobs, and foreign exchange earnings [1]. The globe’s most popular region remains Europe, with countries such as France, Spain, Italy, and the United Kingdom consistently ranking as among the world’s most popular destinations [2]. In such economies, tourism is not only a source of exports of services but also a generator of regional development, cultural heritage, and infrastructure investment. With growing global competition and active tourist behavior (through digitally informed, experience-driven, and engagement-seeking tourists who actively use digital tools to plan, book, and review their travels), technological adoption has emerged as a key driver of TP, altering how destinations are marketed, tourists make decisions, and experiences are delivered [3,4].

The expansion in availability and use of Internet technologies (digital booking systems, review systems, AI guides, and mobile connectivity) has redefined the tourism sector. Such technologies have unlocked opportunities for smart tourism development, destination marketing, and personalized services. Empirical research shows that Internet penetration significantly improves tourism arrivals and expenditure, particularly in countries that involve digital platforms in their tourism policies [5,6]. Macro-level examinations of the role of TECH, as proxied by Internet penetration, in shaping TP in developed economies are scarce. Furthermore, most studies conflate tourism economies as homogeneous, ignoring the possibility that structural factors could have variable impacts across countries with diverse levels of tourism development [6,7]. Digital infrastructure refers to the foundational systems that support Internet-enabled tourism services, including broadband coverage, mobile network penetration, digital literacy, ICT infrastructure, and e-government tourism platforms. These systems enable smart tourism tools such as digital booking, virtual tours, real-time navigation, and personalized services. Improved digital infrastructure enhances accessibility, competitiveness, and experience quality, particularly for tech-savvy and international tourists.

This study fills these gaps by analyzing TECH ‘s impact on TP for the top ten most touristic European countries—France, Spain, Italy, Republic of Turkey, United Kingdom, Germany, Greece, Austria, Portugal, and the Netherlands. This paper makes several contributions. First, it offers a macroeconomic framework that considers TECH (Internet access), quality of environment, per capita GDP, ER, and SI as forces behind TP as measured by international tourist arrivals and tourism receipts. Second, it applies the MMQR to determine how such relationships vary at various quantiles of TP, which captures distributional heterogeneity not observed by mean models. Third, the study utilizes DKSE to be resistant to cross-sectional dependence and autocorrelation and follows up with the analysis using spectral–Granger causality tests to ascertain whether the causality is short-term or long-term in nature.

The questions for the study are as follows:

-

Is improved access to TECH (Internet) making a significant contribution to the performance of tourism in key European countries?

-

How do EQ, income levels, changes in ER, and SI affect TP, and are the effects different for high- and low-performance countries?

-

Are the causal linkages persistent over time, or do they differ at short- and long-term horizons?

While this study contributes new findings, it is not without limitations. Internet penetration, while widely used as a proxy, may or may not capture more qualitative aspects of digital tourism infrastructure (e.g., smart city features or user engagement). Additionally, environmental data such as the Environmental Performance Index (EPI) is gathered on a biennial basis, so its granularity is limited. It is applied to the ten best-performing European destinations, and although these countries host most European tourists, it is possible that the results may not apply to emerging economies or developing nations.

The rest of the paper is structured as follows; Section 2 provides an extensive literature review on determinants of TP, with a key focus on TECH and structure. Section 3 introduces the sources of data, definition of variables, and econometric methodologies. Section 4 describes and interprets the empirical findings of the MMQR, DKSE, and spectral–Granger causality models. Section 5 discusses the findings. Section 6 concludes with significant policy implications for enhancing TP based on digital, economic, and environmental strategies.

Although a growing body of literature has explored how factors such as GDP per capita, environmental quality, and digital adoption affect tourism outcomes, most existing studies rely on average-effect models that overlook differences across countries with varying levels of tourism performance (e.g., high-volume versus emerging markets). Additionally, while the impact of digitalization has often been examined using e-tourism proxies or social media metrics, the macroeconomic effect of Internet penetration—particularly in a comparative European context—remains insufficiently studied. This paper contributes to filling these gaps by (1) employing MMQR to capture distributional heterogeneity, and (2) using the spectral–Granger causality test to distinguish between short- and long-term causal dynamics. These methods allow us to provide a more nuanced understanding of how structural factors—including technology, safety, and environmental sustainability—affect tourism performance in the ten most visited European destinations.

2. Literature Review

The intersection of tourism and TECH has been gradually taking center stage in scholarly and policy discussions, particularly when considering the case of developed economies where digitalization is a dominant force in defining travel patterns, destination planning, and tourist satisfaction [3,4,5]. This literature review synthesizes contemporary literature across five main areas: the role of TECH in tourism, the impact of EQ, the impact of economic performance (per capita GDP), the impact of ER volatility, and the role of SI and political stability in the determinants of tourism flows.

2.1. Technology and Tourism

Innovations in the digital economy have revolutionized destination marketing and experience. Scholars point out that technological innovation in terms of access to the Internet facilitates tourism development through enhanced destination promotion, easier booking procedures, and real-time information exchange [3,5]. Smart tourism refers to the integration of advanced digital technologies (such as mobile applications, big data analytics, and interconnected service platforms) into tourism ecosystems to enhance visitor experiences, destination management, and sustainability. These innovations contribute significantly to competitiveness in advanced economies by enabling real-time services, personalized travel options, and data-driven planning, thereby strengthening destination appeal and operational efficiency [4]. Empirical research also verifies the role of digital connectivity in the positive role [8]. For instance, ref. [6] states that Internet penetration plays an impactful role in Europe’s foreign tourist arrivals by making destinations convenient and accessible to tourists. The same was verified by ref. [7], who established that digital tourism platforms (e.g., TripAdvisor or Airbnb) impact tourist decision making and demand.

The Internet has had a significant impact on the tourism industry, both on suppliers and consumers [9]. Evidence suggests that Internet access and mobile penetration have positive effects on global tourism, with Internet presence having a 20.8% effect. Information TECH use has an effect on 86.1% of global tourism [10]. Internet TECH has become crucial to the tourism sector, where a requirement exists to come up with models that are able to quantify its impact [11]. The Internet has revolutionized the organization of the tourism industry, restructuring both its operational and strategic processes. It has intensified competition and improved the bargaining power of consumers and suppliers, while reducing dependence on traditional intermediaries such as tour operators and travel agents [12,13]. Access methods through mobile applications (enabling real-time bookings, personalized travel planning, navigation, and instant feedback) have become essential in delivering flexible, user-friendly tourism experiences. In parallel, the rise of peer-to-peer service platforms like Airbnb and Uber has significantly reshaped the delivery and consumption of tourism services by fostering decentralized, cost-effective, and experience-driven alternatives to conventional offerings. These platforms have contributed to greater market diversity and consumer empowerment, demanding adaptive strategies from traditional stakeholders [9].

2.2. Environmental Quality and Tourism

Environmental sustainability has emerged as both a competitive advantage and a necessity for modern tourist destinations. A nation’s environmental sustainability is typically measured by the EPI, and studies have demonstrated that it serves as a key facilitator of tourism flows [14,15]. More and more tourists prefer destinations with clean air, preserved landscapes, and green infrastructure, such as eco-friendly public transport systems, green-certified hotels, and sustainable waste and water management facilities. International visitors particularly favor European destinations that implement favorable sustainability policies, such as carbon offset programs, Blue Flag coastal certifications, and protected area management strategies [16]. However, successful destinations often face the risk of overtourism, which may lead to environmental degradation and compromise their long-term attractiveness [17].

Evidence indicates an intricate relationship between EQ and tourist performance. Tourist development tends to have a negative effect on environmental performance, although this may decrease after some time [18,19]. The quality of institutions is significant, and it could work to reduce environmental degradation from tourism while promoting tourism development and economic growth [20]. EQ indicators, such as air and water quality, have a positive correlation with inbound tourism demand and duration of stay [21]. At the local level, environmental management has a U-shaped correlation with performance by tourism businesses. The initial application may negatively affect performance due to institutional conflict, but after environmentally sound values are institutionalized, performance improves, particularly with foreign inbound tourists [22]. These findings highlight the imperative of balancing tourism development with environmental protection and identify institutional quality as key in addressing this balance.

2.3. Economic Performance (GDPPC) and Tourism

Tourism demand is strongly associated with economic well-being. Countries with higher GDPPC tend to invest more in infrastructure, public services, safety, and promotional campaigns, which collectively enhance their tourism appeal [1,23]. Additionally, wealthier populations are more likely to engage in outbound travel, while wealthier countries also tend to offer superior tourism products and experiences. Empirical evidence supports a bidirectional causality between tourism and GDP, especially in developed and OECD economies, where increases in tourist arrivals and receipts can stimulate economic growth, and conversely, economic growth reinforces tourism expansion [24,25,26,27]. These findings align with the notion that international tourism behaves as a luxury good with high income elasticity [28].

2.4. Exchange Rate and Tourism

The ER influences the relative price of travel. A depreciation of currency will reduce the cost of a destination for foreign visitors, and appreciation will reduce its competitiveness [29,30]. However, the magnitude of this effect is debated. Ref. [23] argues that it is greater in price-sensitive tourism markets, while in Western Europe (where travel is culturally, and by default, safely and infrastructure-led), the ER rate effect will be weaker or lagged. Ref. [31] also found that ER volatility may deter tourists due to price uncertainty, especially in countries that are strongly reliant on international inbound tourism outside the Eurozone.

Studies of the impact of ER on TP have mixed outcomes. Whereas some evidence presents a considerable negative effect of ER on European inbound tourism [32], others present evidence that the ER has little effect on international demand for tourism in all countries [33]. Introduction of a single currency, such as the Euro, has been observed to have a moderate positive effect on tourist flows, increasing them by a rate of around 12%. Further, less volatile ER regimes are observed to have a stimulating effect on tourist flows [34]. Interestingly, two-way causality between the ER and tourism development has been observed, wherein increases in tourist numbers and tourism receipts have been found to appreciate the ER in some cases [35]. These findings show how the relationship between TP and ER is interdependent and complex in different contexts [36].

2.5. Safety and Tourism

Safety and security are fundamental considerations in tourists’ destination choices and are widely recognized as critical components of tourism competitiveness. Destinations perceived as safe attract more visitors, while those experiencing political instability, terrorism, or high crime rates often face significant declines in tourist arrivals [37,38,39]. Even perceived SI can strongly influence travel decisions and reduce tourism flows [40].

Destinations with high security ratings are generally more resilient to external shocks and are preferred by risk-averse travelers. In the post-COVID era, SI concerns have expanded to include public health infrastructure and emergency response readiness, emphasizing the broader scope of SI in tourism planning. Empirical studies provide further insights into the multidimensional role of SI [41,42]. A global study found that while security risks negatively affect tourist numbers, they can have spillover effects on employment and leisure-related expenditures [43]. In the ex-Yugoslav countries, a strong correlation between improved security conditions and the revival of tourism sectors has been documented [44].

In the context of policy, the Travel and Tourism Development Index (TTDI) of the World Economic Forum underscores SI as a core pillar of tourism development strategy [45]. However, gaps remain. For instance, a study in Albania revealed that SI protocols were poorly understood among tourism stakeholders, pointing to the need for improved awareness and implementation at the local level [46]. These examples collectively demonstrate that SI plays both a preventive and enabling role in tourism success. Integrating SI into national and regional tourism strategies is essential for enhancing competitiveness and resilience, especially in regions vulnerable to crises or undergoing socio-political transitions.

2.6. Gaps in the Literature

While extensive research has investigated individual determinants of tourism, relatively few studies apply quantile-based econometric approaches that capture heterogeneous effects across different levels of TP [19,44,47]. Additionally, although TECH is frequently analyzed through e-tourism or social media proxies [3,48], its long-term macroeconomic influence, particularly via Internet penetration, remains underexplored in comparative analyses of European tourism economies [6]. This study fills such gaps using the MMQR and DKSE frameworks to analyze average and distributional effects in the top ten European performers, while integrating environmental and SI considerations.

3. Data and Methodology

3.1. Data

A panel dataset balanced between the years 2000 to 2022 for the top ten tourism destinations in Europe, i.e., France, Spain, Italy, Turkey, United Kingdom, Germany, Greece, Austria, Portugal, and the Netherlands, is employed in this analysis. The nations are selected because they are consistently ranked among the world’s most popular destinations for international visitor arrivals and receipts by the United Nations World Tourism Organization (UNWTO).

Details of the variables are presented in

Table 1. Variables’ explanations.

	Variable	Abbreviation	Explanation	Sources
Dependent Variable	Tourism Performance	TP	Measured by a composite indicator of international tourist arrivals and tourism receipts from tourism (USD).	[1,49]
Independent Variable	Technology	TECH	Measured by the proportion of the population with access to the Internet, reflecting the extent of digital connectivity to which tourists and service providers have access.	[3,48,49]
Control Variables	Environmental Quality	EQ	Assessed by the Environmental Performance Index (EPI), a two-year composite indicator covering air, water, biodiversity, and climate policy.	[14,50]
	GDP per Capita	GDPPC	Reflects the level of economic development, affecting infrastructure, tourism potential, and investment in public services, in constant 2015 USD.	[23,49]
	Exchange Rate	ER	Nominal exchange rate of local currency against USD, included to determine the relative cost of traveling to each country.	[29,49]
	Safety Index	SI	A measure of political stability, crime rate, and law enforcement effectiveness, indicating the perceived safety of the destination.	[37,51]

.

3.2. Model Specification

To empirically examine the impact of TECH and other structural factors on TP, we utilize the following panel data regression model:

TPit = α0 + α1 Techit + α2 EQit + α3 GDPPCit + α4 ERit + α5 SIit

where TPit is the tourism performance (international tourist arrivals and tourism receipts) in country i at time t. TECHit presents Internet access (% of population) as a proxy for the TECH. The greater rate of Internet access indicates the digital capacity of a country. A higher rate of Internet access is expected to improve destination visibility, enable smart tourism tools, and maximize tourist experience, and thus improve TP. EQit is EQ, as proxied by EPI. Environmentally clean and sustainable destinations are more attractive, especially to high-spending and environment-conscious tourists. A positive effect is anticipated.

GDPPCit presents the GDPPC, which reflects economic development. Wealthier countries will have better tourism infrastructure, quality of services, and capacity for promotion. Therefore, GDPPC would improve the performance of tourism. ERit is ER that measures cost-effectiveness for foreign visitors. Fluctuations in ER affect the relative expense of visiting a country. Depreciation will tend to raise tourism, while appreciation discourages it; thus, the anticipated value of α4 will be negative. Finally, SIit is the safety index, measuring public security and political stability. Public security and political stability are important determinants of decision for tourists. Levels of the SI should raise tourism performance.

3.3. Methodology

To examine the dynamic and heterogeneous impact of TECH and other structural variables on TP, a multi-step econometric strategy is employed.

3.3.1. Cross-Sectional Dependence and Heterogeneity Tests

Due to the possibility of interdependence between European nations, Pesaran’s cross-sectional dependence test [52], the Frees test, and Fisher’s test are utilized to identify contemporaneous correlation across units. Parameter heterogeneity is tested using the Pesaran and Yamagata slope homogeneity test [53] in order to confirm parameter heterogeneity. These tests are important in ensuring correct model selection.

3.3.2. Panel Unit Root and Cointegration Tests

Stationarity of the data is tested through CIPS (cross-sectionally augmented IPS) and CADF (cross-sectionally augmented Dickey–Fuller) unit root tests, which are suitable in the context of cross-sectional dependence. Upon identifying the order of integration, the Westerlund cointegration test is used to identify whether there exists a long-run equilibrium relationship between variables [54].

3.3.3. Method of Moments Quantile Regression (MMQR)

The primary method used here is the MMQR, which allows for the estimation of the effects of covariates at various quantiles of the dependent variable [44]. The approach captures distributional heterogeneity since it recognizes that the effect of explanatory variables may vary for low-, medium-, and high-performing tourism countries. MMQR is resistant to outliers and does not consider homogeneous effects for the panel.

The MMQR accounts for heterogeneity across countries by estimating the effects of independent variables on different quantiles of TP. This approach captures structural differences among countries with varying tourism capacities and digital maturity, enabling us to observe how determinants such as TECH or GDP affect low-, middle-, and high-performing destinations differently. Additionally, slope heterogeneity tests [40] further validate that the relationship between tourism drivers and performance is not uniform across all cross-sectional units (countries), justifying a flexible modeling strategy.

3.3.4. Driscoll–Kraay Standard Errors (DKSE)

To validate average effects and test for robustness, the fixed-effects DKSE model is employed [55]. It addresses heteroskedasticity, autocorrelation, and cross-sectional dependence, which suit macro-panel data with interdependent units, such as European economies [56].

3.3.5. Spectral–Granger Causality Test

To capture variation across time horizons, the spectral–Granger causality test is applied [57]. This method decomposes causal relationships into frequency bands, allowing us to distinguish between short-run dynamics (e.g., fluctuations due to temporary ER shocks) and long-run structural linkages (e.g., persistent effects of digital infrastructure or GDPPC). These time-sensitive insights are important for formulating both tactical and strategic tourism policies [58,59].

The Figure 1 outlines the sequence of data selection, variable definition, pre-estimation diagnostics, econometric modeling (MMQR, DKSE), and causal testing (spectral–Granger).

3.4. Data Processing and Software

The panel dataset was compiled using Microsoft Excel, where raw data from the World Development Indicators, Environmental Performance Index, and World Governance Indicators were cleaned, merged, and transformed. All econometric analyses were conducted using Stata 17. Variables were tested for stationarity, heterogeneity, and cross-sectional dependence using the xtcips, xtcd, and slopeh commands. MMQR was implemented using the user-written command mmqreg, and DKSE values were estimated using xtscc. The spectral–Granger causality analysis followed the approach proposed by ref. [43], implemented via a custom Stata routine. Missing values were handled via listwise deletion, and no imputation was performed. All variables were expressed in logarithmic form to reduce heteroskedasticity and ensure comparability of coefficients.

4. Empirical Findings

To validate the use of panel data econometrics and ensure robust model estimates, tests for cross-sectional dependence (CSD) are conducted. These tests are necessary because the presence of cross-sectional dependence in panel data implies that shocks or structural forces affecting one country will also influence other countries (which is particularly relevant to integrated European tourist markets).

The Pesaran test is for average pairwise correlation in the residuals across cross-sectional units. According to the results in

Table 2. Cross-sectional dependency tests.

Test	Statistic	p-Value
Pesaran	46.42 ***	0.00
Fisher	187.741 ***	0.00
Frees	15.45 ***	0.00

Note: (***) indicates that the estimated parameters are significant at the 1% significance level.

, the highly significant test statistic (46.42) with a p-value of 0.00 offers strong evidence of cross-sectional dependence among countries in the panel. This result indicates that the tourism dynamics in one country (e.g., France) are likely to be affected by or impact other countries (e.g., Spain, Italy), which implies the need to account for such dependencies in model selection. Also, the Fisher-type test, based on the concatenation of p-values of separate unit root tests (or independence tests), also supports cross-sectional dependence in the form of a large and significant test statistic (187.741, p-value = 0.00). This conforms with the assumption that shocks (a European-wide digitalization trend or macroeconomic fluctuation, for instance) may affect TP in the sample nations at the same time. Finally, the Frees test estimates cross-sectional dependence as the sum of squared rank correlations. The Frees statistic (15.45) is significant and again rejects the null hypothesis of cross-sectional independence at the 1% level.

To determine the stationarity properties of variables within the panel data model, we perform two second-generation unit root tests: the cross-sectionally augmented IPS (CIPS) test and the cross-sectionally augmented Dickey–Fuller (CADF) test. The tests are appropriate where there is cross-sectional dependence, as supported by the results in Table 2.

The results of the tests in

Table 3. Unit root test.

Variable	CIPS	CADF
TP	−2.80 ***	−4.06 ***
EQ	−3.57 ***	−6.87 ***
TECH	−4.08 ***	−8.73 ***
GDPPC	−3.57 ***	−6.88 ***
ER	−1.98	−1.06
SI	−3.65 ***	−7.15 ***
∆ER	−2.36 **	−1.68 *
Critical values	α = 10%: −2.12
	α = 5%: −2.25
	α = 1%: −2.51

Note: (***), (**), and (*) indicate that the estimated parameters are significant at the 1%, 5%, and 10% significance level, respectively.

indicate that TP, EQ, TECH, GDPPC, and SI are stationary at that level, since their test statistics in terms of absolute values are more than the critical value at a 1% significance level (−2.51) in both CIPS and CADF tests. This verifies that these variables are of integrated order zero [I(0)], since they do not contain unit roots and are stationary in levels. In another hand, the ER cannot reject the null hypothesis of unit root in both tests because its values (−1.98 in CIPS and −1.06 in CADF) lie outside the 10% critical value. This indicates that ER is level non-stationary, i.e., I(1). The first difference in ER (ΔER) is stationary at the 5% level in the CIPS test and at the 10% level in the CADF test. This confirms ER to be stationary after the first difference test, confirming its I(1) nature.

Table 4. Heterogeneity slope test.

Δ	p-Value	Δ Adj	p-Value
−3.70 ***	0.00	−6.75 ***	0.00

Note: (***) indicates that the estimated parameters are significant at the 1% significance level.

shows the results of the slope homogeneity test, which is frequently used to determine whether the slope coefficients of cross-sectional units (countries, in the current research) are heterogeneous or homogeneous. Two of these tests for the given statistics are the Δ test and the adjusted Δ (Δ Adj) test, both of which were proposed by Pesaran and Yamagata [40]. The Δ value of −3.70 is significant at the 1% level (p = 0.00), giving very strong evidence of non-slope homogeneity under the null hypothesis. This means that the relationship between the independent variables (e.g., GDPPC, TECH) and the dependent variable (TP) differs from one country to another. The adjusted Δ statistic, which controls for small sample bias and cross-sectional dependence, further supports this result with a larger and more significant test value (−6.75, p = 0.00). This powerfully establishes that the slope coefficients are not homogeneous across the ten European countries in the panel.

Table 5. Cointegration tests.

Test	Statistic	p-Value
Westerlund	22.89 ***	0.00
Modified Dickey–Fuller	4.32 **	0.00
Dickey–Fuller	2.07 ***	0.01
Augmented Dickey–Fuller	3.36 ***	0.00
Unadjusted modified Dickey–Fuller	−5.41 ***	0.00
Unadjusted Dickey–Fuller	−10.74 ***	0.00

Note: (***) and (**) indicate that the estimated parameters are significant at the 1% and 5% significance level, respectively.

provides the results of a number of panel cointegration tests to examine if there exists any long-run equilibrium relationship between TP and its determinants (TECH, GDPPC, ER, SI, and EQ) in the selected European countries.

Westerlund test accommodates cross-sectional dependence as well as heterogeneity. The highly significant test statistic (22.89, p = 0.00) refutes the null hypothesis of no cointegration and thus establishes that there is a long-term relationship among the variables in the panel. The residual-based panel unit root tests of the estimated long-term equation (modified DF, DF, ADF, and their counterparts without adjustment) also present robust evidence of cointegration. All the test statistics are significant at the 5% or 1% levels. The negative signs of unadjusted tests (e.g., −10.74 and −5.41) confirm the stationarity of residuals, a pre-condition for cointegration. The combined evidence from the It’s okay.Westerlund as well as residual-based tests overwhelmingly supports the presence of panel cointegration between the concerned variables. It also indicates that despite short-run volatility, TP and its determinants move towards a long-term equilibrium over time among the top ten tourist destinations in Europe.

Table 6. MMQR model (dependent variable: TP).

Quantile	EQ	TECH	GDPPC	ER	SI	Intercept
5	0.01	0.01 ***	0.99 ***	0.00	0.35 ***	3.08 ***
25	0.14 *	0.04 ***	1.14 ***	−0.02	0.36 ***	2.52 ***
50	0.21 ***	0.03 ***	1.22 ***	−0.03	0.37 ***	2.18 ***
75	0.35 ***	0.02 *	1.39 ***	−0.05	0.38 ***	1.53 **
95	0.51 ***	0.01	1.58 ***	−0.08	0.39 ***	0.77

Note: (***), (**) and (*) indicate that the estimated parameters are significant at the 1%, 5%, and 10% significance level, respectively.

shows findings from the MMQR method, which allows us to estimate heterogeneous effects across different quantiles of the outcome variable (TP). This is particularly helpful where the impact of predictors varies across different distribution of the outcome variable and captures structural asymmetries between high- and low-performing nations.

The TECH effect (as caused by Internet penetration) is positive and significant at lower to median quantiles (5th to 75th), i.e., increasing access to the Internet improves TP, especially in low- to mid-performing economies. The coefficient declines across the quantiles and becomes insignificant at the 95th percentile, indicating diminishing returns to TECH in already high-performing economies. Also, EQ is positive and increasing over quantiles. This implies that EQ is increasingly important to high-performing tourism countries, perhaps as tourists become more and more environmentally aware in their decisions.

GDPPC is always positive and highly significant across all quantiles, with coefficients increasing from 0.99 to 1.58. This is a measure that shows economic prosperity increases TP and that its effect increases in the case of higher-performing countries. However, the effect of ER is negative, though not statistically significant at smaller quantiles, and becomes negative and borderline significant at larger quantiles. This suggests that currency appreciation (i.e., stronger local currency) can reduce international tourist arrivals on the basis of lower price competitiveness, especially in higher-performing tourist economies. Additionally, the SI has a significantly positive and overall positive impact on all quantiles. As one would anticipate, SI is an important determinant of tourism regardless of the current performance level of a nation.

The DKSE estimations provide robust average findings of the determinants of TP, confirming the central roles of TECH, EQ, income levels, and SI. The pooled panel regression estimates with DKSE are presented in

Table 7. DKSE (Driscoll–Kraay standard errors) method (dependent variable: TP).

	Coefficient	Std. Err.	t	p > t
EQ	0.25 **	0.12	2.14	0.04
TECH	0.03 ***	0.01	4.17	0.00
GDPPC	1.27 ***	0.33	3.87	0.00
ER	−0.03	0.03	−1.37	0.18
SI	0.37 ***	0.06	6.36	0.00
Intercept	2.00 ***	0.74	2.70	0.00

Note: (***) and (**) indicate that the estimated parameters are significant at the 1% and 5% significance level, respectively.

, which adjust for heteroskedasticity, autocorrelation, and cross-sectional dependence—issues already diagnosed in earlier tests (Table 2, Table 3 and Table 4). The model estimates the average long-term impact of independent variables on TP across the ten European countries.

The coefficient of TECH is highly positive (0.03) at the 1% level. This confirms that expanded Internet access positively contributes to TP, consistent with the hypothesis that TECH enhances destination accessibility, marketing scope, and tourists’ experience. Moreover, EQ is statistically significant and has a positive effect (0.25 at 5% level), which indicates that better environmental conditions attract more foreign tourists, validating the importance of sustainability in tourism development.

GDPPC with a coefficient of 1.27 (p < 1%) shows that economic prosperity is strongly related to good performance in tourism. Wealthier countries possess better infrastructure, services, and advertising capacity, thus being more appealing to visitors. The coefficient of ER is statistically insignificant (p = 0.18) but negative (−0.03). Although higher ER can reduce a country’s affordability to visitors, this effect is not observed to be strong in the pooled model. Finally, SI exerts a positive and significant effect (0.37, p < 0.01) on tourist performance. This emphasizes the important role of political stability and public safety as pulls for international tourists.

In order to explore further the dynamic interactions between TP and its determinants, a spectral–Granger causality test was used (

Table 8. Spectral–Granger causality test results.

Frequency	TECH → TP	EQ → TP	GDPPC → TP	ER → TP	SI → TP
0.01	0.32	0.39	0.34	0.05	0.21
0.2	0.27	0.09	0.47	0.06	0.30
0.4	0.49	0.23	0.33	0.25	0.30
0.6	0.33	0.21	0.60	0.29	0.15
0.8	0.13	0.27	0.26	0.01	0.16
1.00	0.45	0.29	0.30	0.04	0.10

and Figure 2). This test enables us to determine whether the independent variables (TECH, EQ, GDPPC, ER, SI) drive TP in various frequency domains, thus separating short-term (high-frequency) from long-term (low-frequency) causal relationships. Values of TECH causality are relatively high for low and medium frequencies, with peaks around 0.45. This indicates that Internet availability exerts long-lasting causal effects on TP that are strong, which is consistent with MMQR and DKSE results.

EQ causality increases at lower frequencies, indicating the long-term impact of environmental sustainability on tourism attractiveness. This corroborates the argument that clean air, biodiversity, and environmental legislation have a cumulative influence in the long run. The causality test of GDPPC shows the maximum and strongest causality values (approximately 0.6), particularly at low frequencies, which further verifies that economic prosperity significantly and persistently affects TP for all horizons. Values of ER causality are less strong and are largely at higher frequencies; this signifies that the variability of the ER affects tourist performance in the short run, perhaps due to temporary changes in relative travel affordability or prices. Uniform but relatively weak causality at all frequencies for SI dominates, with stronger influences at lower frequencies, reinforcing once again the importance of long-term stability and public safety in sustaining global tourist flows.

In general, the spectral–Granger causality test results supplement the earlier panel model results and add temporal sensitivity by placing emphasis on the frequency domain effect of each determinant; long-term tourism policy must address TECH infrastructure, environmental protection, and economic growth. Short-term policy measures (e.g., ER management or improvement in security concerns) could buffer near-term impacts but have little effect in the long run.

5. Discussion

This study employed two complementary econometric techniques (MMQR and DKSE regression) to analyze the impact of TECH, EQ, income, ER, and security on TP in ten key European tourist destinations.

Both methods consistently exhibit a positive and statistically significant effect of Internet penetration on TP. This confirms earlier claims by [3,4], which emphasized the importance of digital infrastructure and smart tourism tools in advanced economies (as discussed in Section 2.1). The MMQR specification also indicates that TECH is more of a driving factor in lower-performing countries (e.g., 5th and 25th quantiles), with diminishing returns in high-performing ones (not significant at the 95th quantile). This implies that the expansion of digital infrastructure is vital for emerging tourism destinations, where it significantly enhances international visibility and business efficiency. The DKSE model also supports this result with a substantial, robust mean effect (β = 0.03, p < 0.01), indicating that TECH is one of the main drivers of tourism growth in the sampled countries. These findings align with ref. [3], who highlight the role played by digitalization in enabling smart tourism and competitiveness of destinations. This corroborates ref. [48], who describe how e-tourism websites expand market coverage. This substantiates that TECH is a key tourism driver and supports digital transformation; however, returns diminish in saturated markets [60].

MMQR results show that EQ has a greater impact, with higher TP (aligning with [14,61], who stressed that environmental sustainability is not only a competitive advantage but also a necessity for long-term tourism viability), being non-significant at lower quantiles and highly positive at the 95th quantile (β = 0.51, p < 0.01). This suggests that high-performing tourist destinations rely increasingly on environmental sustainability to maintain their appeal. The DKSE model also corroborates a significant average effect (β = 0.25, p = 0.04), reaffirming the fact that tourists, particularly from high-income source markets, perceive environmental cleanliness and sustainability as crucial determinants in destination choice. These results are consistent with ref. [61], in which the authors explain that environmental concerns are at the heart of maintaining tourism in over-touristed destinations. Furthermore, ref. [19] empirically confirmed that there is a strong relationship between environmental indicators and tourism. This aligns with ref. [2]’s research on environmentally conscious tourists wanting to visit destinations with strong environmental policy. The impacts of the ER are weak or nondescript. Unlike the results of [62], who concluded that the ER is critical in price-sensitive Asian tourist markets, European tourism appears less susceptible to currency change, perhaps due to its multi-faceted attractiveness (culture, business, and heritage).

Both models identify GDPPC as a strong and stable predictor of TP, as documented in the literature by [1,23]. The increasing marginal effect across quantiles (MMQR) supports the feedback loop between economic development and tourism identified by ref. [27]. The MMQR model reveals the effect of GDPPC on increasing tourism success—from β = 0.99 at the 5th quantile to β = 1.58 at the 95th—while the DKSE model supports a significant average effect (β = 1.27, p < 0.01). This would imply that wealthier countries can invest more in tourism infrastructure, promotion, and services, and that this is mirrored by their superior performance. The relationship between tourism and national income is strongly established. For example, ref. [1] confirms that GDP is one of the primary determinants of TP, especially through its effect on infrastructure and destinations’ image. Moreover, ref. [63] found GDP to be a main determinant of tourist demand in Europe.

Although traditional theories predict strong ER effects [29], our findings suggest that European tourism markets are less price-sensitive. This echoes the mixed results reported by [33,34]. This would imply that price competitiveness may be a bigger issue in advanced tourist economies but a smaller issue in developing economies, where other determinants like SI and accessibility may override it. More recent studies such as ref. [23] also present mixed results, particularly in European markets where travel decisions are less price-elastic. In another hand, political stability and SI are not something that can be bargained for in tourist development, validating ref. [37]’s claims about risk perception’s role in destination choice. Their study reiterates GDP and SI as global pillars of tourism growth, which is consistent with ref. [23], but GDP’s effect becomes more significant in wealthier nations.

Both models confirm a positive, strong, and highly significant effect of SI on TP. MMQR results detect that this effect is marginally stronger from the 5th to the 95th quantiles (β = 0.35 to 0.39), while the DKSE model detects a highly significant mean coefficient (β = 0.37, p < 0.01). The uniformly strong role of safety across all quantiles is consistent with the findings of [37,40], who argued that real and perceived SI are central to tourist decision making. This finding corroborates the reality that political stability and perceived safety are universal needs for international tourist appeal. Ref. [38] provides empirical evidence that political violence and SI deficits significantly reduce tourist arrivals.

Cumulatively, both methods (MMQR and DKSE) affirm that TECH, SI, GDPPC, and EQ are key pillars of sustainable TP. The ER’s effects, though theoretically significant, are less robust (highlighting the complexity of price sensitivity in European tourist markets). Generally, the DKSE model treats the effect linearly, but MMQR establishes that the impact of TECH diminishes in developed markets (e.g., Spain and France). This nuance verifies ref. [3], which claims that digital saturation yields diminishing marginal returns in advanced economies. Unlike linear models, MMQR shows that EQ is the most important factor for high-performing destinations, consistent with ref. [64].

6. Conclusions and Policy Recommendations

In this research, the long-term and distributional effects of major economic and structural drivers— TECH (proxied by Internet accessibility), EQ, GDPPC, ER, and security—on the performance of tourism in the leading ten touristic nations in Europe (France, Spain, Italy, Turkey, the United Kingdom, Germany, Greece, Austria, Portugal, and the Netherlands) were examined. With robust panel data techniques, including MMQR, DKSE, and the spectral–Granger causality technique, we found ubiquitous and consistent evidence that digital infrastructure, environmental sustainability, national income levels, and public safety all contribute significantly to TP, but their impacts differ by country and by timescale.

Our MMQR findings reveal that TECH most profoundly impacts low- and middle-performing tourist economies, providing evidence for the hypothesis that digital transformation can act as a compensatory mechanism for less competitive destinations through enhanced accessibility, visibility, and service efficiency. However, in high-performing countries, the marginal returns of TECH dwindle, highlighting the need for innovation and not just connectivity. EQ shows a progressively positive impact at higher quantiles of tourism, reminding us that sustainability is not just an ethical requirement but also a strategic benefit for developed tourist markets. GDPPC is the most robust determinant in all models, highlighting the importance of economic wealth for preserving tourism infrastructure, human resources, and advertising efforts. Even though the ER is revealed to have a negative but statistically insignificant effect in mean models, its short-term effect is more evident in the frequency domain, so it can be inferred that price competitiveness can affect tourist selection in the short term, even if less so in the long term. Finally, the persistent and significant effect of SI when using all methods confirms its key role in attracting foreign visitors.

These findings have several significant policy implications for the growth of European tourism. High priority must be given to continued investment in digital infrastructure, and particularly in countries and regions lagging behind in Internet penetration and smart tourism capacity. Enlarging coverage with cheap, high-speed Internet not only adds to the tourist experience but also enables more effective digital marketing and coordination among tourism value chains. At the same time, policy frameworks must elevate environmental sustainability as a prime pillar of tourism planning. This involves strengthening green regulations, promoting green certification of tourism businesses, developing eco-tourism infrastructure, and incorporating environmental measures in destination marketing activities. In addition, tourism-supportive macroeconomic measures such as transport infrastructure investment, training of human resources, and support for small- and medium-sized tourism enterprises should foster continued economic growth, which can generate both employment and exchange revenues.

While ER policies are not easily manageable over the short term, governments and central banks should monitor currency fluctuations that can affect destination competitiveness. Where appropriate, tourism ministries can complement monetary policy with flexible pricing policies and targeted subsidies to offset short-term affordability shocks. Furthermore, the necessity of SI means closer coordination among tourism authorities, law enforcement, and public health institutions will be required. Ensuring SI for residents and tourists alike (either by preventing crime, preparing for disasters, or controlling health risks) is imperative in order to maintain destinations’ attractiveness and resilience.

Lastly, the transition to a more competitive, sustainable, and inclusive European tourism sector will depend not on single-point policy actions but on a comprehensive strategy that simultaneously promotes digitalization, sustainability, economic resilience, security, and price flexibility. All of these measures must be supported by evidence-based policy making and continuous monitoring to make tourism a resilient force for development and international co-operation in the post-pandemic world economy.

Limitations and Directions for Future Research

Despite the comprehensive design of this study, several limitations should be acknowledged. First, the proxy used for digital infrastructure—namely, Internet penetration—captures general connectivity but may not reflect more advanced components such as mobile booking platforms, digital literacy, or smart tourism systems. Second, the environmental quality variable (EPI) is based on biennial data, which may smooth out year-to-year environmental changes. Third, the analysis is limited to the top ten European tourism destinations, which may restrict the generalizability of findings to emerging or lower-performing tourism economies.

Future research could address these limitations by incorporating more detailed indicators of digital engagement (e.g., platform usage metrics and AI-enabled services), extending the sample to developing regions, or applying dynamic versions of MMQR and time-varying causality methods. Additionally, further investigation into the post-COVID resilience strategies adopted by destinations could yield valuable policy insights for crisis-era tourism planning.