The role of tourism in European regions’ economic recovery: A spatial perspective

Abstract

The study searches for the presence of a tourism-led growth path after a negative shock. Both the presence of a tourism role in the recovery process and the existence of a spatial interdependence across regions are investigated. The purpose of the study is to look at insights, provided in particular by the experience of the 2007–2008 crisis in European regions, that can be useful to build a recovery after any future shocks, as for example the COVID-19 pandemic resulted to be. The findings show the presence of a strong spatial dependence in the recovery process and, furthermore, the positive potential that tourism activities exert on the ability to recover favouring positive spillovers across regions.

Keywords

economic recovery tourism spatial analysis European regions

Introduction

Tourism sector has recently gained greater attention in the literature debate because of its extraordinary expansion and the acknowledgment of its ever-increasing role in the world economy over the last decades. Indeed, with more than one billion tourists traveling every year, tourism has become a leading economic sector as demonstrated by its contribution to GDP which in 2019, before the COVID-19 pandemic, exceeded 10%. Moreover, with a share of 7% of the world’s total export it has become the third-largest export industry after fuels and chemicals. Before the pandemic led to a collapse in international travel and tourism, annual growth rates for tourism far exceeded those of world economies. This expansion was mainly due to the coincidence of different occurrences like the ever-increasing globalization process, with the emergence of developing countries representing new attractive markets both from a supply and demand point of view, and the diffusion of new travel and accommodation players and facilities (e.g., low-cost carriers, sharing economy, and booking platforms). These changes made travel and tourism easier and more affordable.

It is undeniable the contribution of tourism to income and employment opportunities (Lee and Chang, 2008), investments in new infrastructures, human capital and increasing competition (Blake et al., 2006), industrial development and improvements of national balance of payments (Balaguer and Cantavella-Jordá, 2002; Fahimi et al., 2018; Santamaria and Filis, 2019). This tourism-led growth hypothesis, was first proposed by Balaguer and Cantavella-Jorda (2002) who, by proving that the expansion of international tourism activities exerts economic growth, stated the presence of a theoretical and empirical link between inbound tourism and economic growth. This linkage results to be – despite sometimes non-linear and rather controversial (Cerisola and Panzera, 2024; Liu et al. 2024)-even more substantial in countries and regions where the cultural heritage and varieties of natural endowments draw huge flows of international visitors every year.

The emergency caused by the COVID-19 pandemic, however, led to a fall of the tourism sector and a sharp decline of its contribution to GDP to 5.3% in 2020 (WTTC, 2022). There is evidence that the share of tourism activities in GDP was revealed to be the single most important predictor of the growth shortfall in 2020, even when compared to other severe effects of the pandemic (Milesi Ferretti, 2021).

The last IMF World Economic Outlook (October 2023) stated that the global economy is recovering after the pandemic even if slowly and unevenly. Regarding the speed of the recovery, it appears to have been strongly conditioned by the cost-of-living crisis caused by the war between Russia and Ukraine and its effects on the energy and food markets. Whilst the presence of a global divergence in the recovery process can be traced back to different factors. One of them is certainly the strong demand for services that supported service-oriented economies, as for example important tourism destinations such as France and Spain, comparing to countries more specialized in manufacturing such as China and Germany. In particular, the IMF finds that “countries where tourism represents a high percentage of GDP have recorded faster recovery from the impacts of the pandemic in comparison to economies where tourism is not a significant sector” (IMF WEO, 2023), thus acknowledging that tourism is a key driver of the economic recovery.

In this line, the aim of our paper is to investigate the role played by the tourism sector in the economic recovery of European regions hit by a severe negative shock, by attempting to address both demand and supply-side effects of tourism-related factors. However, since the temporal proximity does not allow to have sufficient data to investigate regional economic paths emerged after the pandemic, we look at insights provided by the experience of the 2008 financial and economic crisis. The choice of 2008 financial crisis as a case study is derived by the need to identify a particularly severe event experienced by the European regions after the start of the monetary union. The length of the observation period is in fact long enough to take into account the effects of the policy strategies adopted by countries and regions to counter the crisis. Our main objective is to contribute to the tourism-led growth literature and unveil if the sector could contribute to economic recovery taking into account both the dynamic effects and the spatial interdependency in an integrated economic area. Our contribution fits inside the literature about the tourism-led growth nexus, which has recently experienced interesting developments linked to the complex nature of the relationship between tourist activities and economic expansion. The connection is often non-linear and bidirectional (Liu et al., 2024), it works better during early phases of economic growth (Saboori et al., 2023) and largely depends on the specialization of the territory (Sahni et al., 2023).

However, despite tourism activities are strictly connected to territorial characteristics and geographical proximity appears to be crucial to inter-regional tourism flows, standard empirical analyses do not properly account for tourism spatial dependence across neighbouring destinations. By contrast, different factors like for example productivity spillovers arising from labour migration, competition effects enhancing tourism enterprises use of innovation-based technologies and proficiencies, joint promotion undertaken by two or more destinations, negative events such as conflicts, political turmoil, diseases, or natural disasters, can favour the presence of spatial interdependencies across tourist destinations thus affecting their recovery process. The rational of using a spatial approach relies on the presence of spillovers and spatial effects that can either benefit or damage local development thus modifying the very destinations’ attractiveness (Yang and Wong, 2012). Despite the importance of this issue, studies detecting tourism spillovers remain scarce and, in most cases, they focus on effects arising mainly from non-tourism factors such as conflicts, terrorism or natural disasters (Drakos and Kutan, 2003; Gooroochurn and Hanley, 2005; Neumayer, 2004). The academic contribution of this paper would be to fill this gap by investigating whether the growth of proximate economies and changes in their growth influencing factors, and especially their tourism contexts, can have a significant role in regional economic recovery. The relevance of our contribution derives also from the geographical area of investigation: usually the existence of a tourism-led growth nexus is supposed to exist in developing countries, while almost neglected in advanced economies. To our knowledge, this is the first time that such an analysis is applied to European regions as a whole. The general intention is to add a piece to the mosaic of the complex field of investigation of the relationship between tourism and growth, in the awareness that it is still necessary to observe the phenomenon from a holistic perspective.

The empirical investigation is conducted through a standard dynamic panel data methodology as well as a dynamic spatial panel methodology. The first is the system GMM (Blundell and Bond, 1998; Blundell et al., 2000) suitable for a sample where the number of entities (153 European regions) is higher than time observations (11 years from 2009 to 2019). This methodology accounts for endogeneity and autocorrelation between dependent and explanatory variables and is suitable for estimates in which the coefficient of the lagged dependent variable might absorb the whole effect of explanatory variables. The second, instead, is the Spatial Durbin model that tests the tourism-led growth hypothesis by accounting for the presence of spatial interdependences in regions’ ability to react to shocks and in its main determinants.

The main finding of our investigation is the positive potential that tourism activities exert on the ability of a region to recover from an exogenous shock hitting its economy. Furthermore, by accounting for spatial dependence we demonstrate that local recovery is gained also through positive performances registered in proximate regions. Whether directed towards the region itself or towards neighbouring ones, larger tourist flows contribute to the recovery process and favour positive spillovers across regions. However, benefit of tourism expansion ends when flows, in the same region or in the neighbouring ones, exceed critical thresholds thus causing negative congestion effects. Finally, the path towards recovery seems to be held back by the presence of neighbouring regions specialized in traditional tourism activities, which suggests the existence of an ongoing competition across proximate destinations.

Our analysis indicates that inappropriate and excessive exploitation of tourist resources, as well as the potential conflict with resident population caused by excessive inbound visitors, may compromise the overall positive contribution that the tourism sector may exert on the economic recovery after a negative shock. This is another perspective from which to observe the non-linearity bidirectional causality between tourism and growth (Liu et al., 2024).

The policy implication that can be drawn by our findings is that a sustained growth in the tourism sector should be pursued whilst reconciling the ability to produce income and ensure a good quality of life. This combination, as well as fostering the economic resilience of the region, contributes also to strengthen the recovery in more proximate geographical areas.

The paper is organized as follows. The next section provides information on the theoretical and empirical literature on the subject. The third section refers to the empirical investigation and presents: the dataset and a descriptive analysis that highlights geographical details of regional economic recovery; the GMM and spatial panel methodologies as long as the estimation results; some robustness checks. Finally, the fourth section draws conclusions and derives some policy implications.

Tourism contribution to growth and economic recovery

Due to its acknowledged contribution to economic growth (since Balaguer and Cantavella-Jorda, 2002), the potential of tourism for the economic recovery has received considerable attention in the recent literature (Amore et al., 2018; Dogru and Bulut, 2018; Prayag et al., 2020). In particular, this potential showed up concretely after the pandemic crisis, when economies with large travel and tourism sectors demonstrated stronger economic resilience comparing to economies with smaller tourism sectors (IMF WEO, 2023). This led local governments and policymakers to consider tourism development as a crucial engine for the economic resilience not only after severe natural disasters, as usually considered in the previews reference literature, but also after general economic shocks. The evidence in 2023 international tourism reveals that by 2024 several countries will recover pre-pandemic levels, despite a context of uncertainty due to recent geopolitical tensions and conflicts.

The link between tourism and economic growth traces back to the seminal theoretical contribution by Butler on the evolution of tourist destinations, the so-called Tourism Area Life Cycle (TALC) approach (Butler, 1980). According to this view, tourism exerts positive effects on GDP through a direct impact of demand for tourism services as well as an increase in activities “related” to tourism. Since then, following different methodologies and extending analyses to various time periods, scholars have sought to verify whether and the extent to which tourism development contributes to the economic growth of a destination (Balaguer and Cantavella-Jorda, 2002; Brida et al., 2016a, 2016b; Dogru and Bulut, 2018; Gunduz and Hatemi, 2005; Kim et al., 2006; Merida and Golpe, 2016; Tugcu, 2014). The broad consensus on tourism’s driving role in boosting economic growth, favoured the birth of the “Tourism-Led Growth” hypothesis. The contribution of tourism is highlighted through various channels like for example higher levels of tourism income, employment, foreign exchange of production, government revenues due to multiplier effects and improvements in the balance of payments. However, on the other side there is also evidence that exploitation of tourism resources, due to an excessive and uncontrolled increase of arrivals, may cause conflicts between resident populations and the stakeholders of the tourism sector. Basing on the latter detrimental effect, recent studies turned their interest to aspects that have been overlooked so far, for example identifying critical thresholds for the “carrying capacity” in case of massive arrivals (Alkhathlan and Javid, 2013; Ehigiamusoe, 2020; Mirza and Kanwal, 2017) or potential controversial effects caused by congestion, environmental degradation and noise (Canale and De Siano, 2021; Capo et al., 2007; Cheer and Lew, 2017; De Siano and Canale, 2022). As such, tourism development has become an important issue for most governments.

The literature on the causal relationship between tourism development and economic growth allowed to formalize four different tourism-growth nexus hypotheses. The first is the tourism-led growth hypothesis, introduced by Balaguer and Cantavella-Jorda (2002), assuming that tourism development leads to economic growth (Adamou and Clerides, 2010; Brida et al., 2016a; Brida and Pulina, 2010; Cortes-Jimenez and Pulina, 2010; De and Kyaw, 2016; Husein and Kara, 2011; Rasool et al., 2021). According to this view, tourism development may rise beneficial spillovers on growth through different channels: foreign earnings to invest in the local economy, pushing tourism enterprises towards highly efficient provision of services due to higher competition in their market, positive scale economies and decreasing production costs, among others. The conservation hypothesis, on the contrary, points out a unidirectional causality from economic growth to tourism development. This assumption relies on the greater attractivity gained by destinations by means of economic policies aimed at enhancing public and private investments on infrastructures and, among others, on cultural and entertainment activities (Aslan, 2013; Oh, 2005; Payne and Mervar, 2010). The feedback hypothesis, instead, suggests a bilateral causality between tourism development and economic growth (Brida et al., 2015; Demiröz and Ongan, 2005; Lee and Chang, 2008; Tang and Ozturk, 2017; Tugcu, 2014) while the neutrality hypothesis denies the presence of a significant causality between tourism development and economic growth (Dogru and Bulut, 2018; Merida and Golpe, 2016; Tugcu, 2014). Recently several contributions made additional efforts to compose these different perspectives, considering the tourism-led growth nexus as a very complex phenomenon having peculiar features in each single destination. Examining the case of Thailand through a dynamic copula-based GJR-GARCH empirical model, Liu et al. (2024) find that tourism and expected macroeconomic conditions are non-linearly and bidirectional affected. Using a threshold dynamic regression model, Sahni et al. (2023) document an asymmetric effect of tourism on economic growth in African, Asian, and Latin American countries. Those with a low and medium level of growth receive higher benefits from tourism. The low and middle growth stage is found to be central also in Saboori et al. (2023) who – trough a quantile regression method - deserve attention to the feature of tourist market diversification, which in turn is said to be detrimental for countries belonging to the highest quantile of the GDP. The issue of the controversial effect of cultural tourism on economic development is examined in Cerisola and Panzera (2024). They find, through a structural equation estimation method, that the connection is mutually positive as it increases demand and at the same time enhances the identity of the territory attracting external private and public investments that reinforce growth and the identity itself. However, the path of tourism expansion might generate controversial effects when exceeding a certain threshold.

Regarding the role of tourism on the capacity to recover, previous studies focused only on destinations hit by natural disasters (Cellini and Cuccia, 2015; Cheng and Zhang, 2020; Dogru and Bulut, 2018; Orchiston and Higham, 2016; Ritchie, 2004). These studies highlighted the beneficial impact of tourism development on, above all, new capital investments, reconstruction of damaged infrastructures, transport accessibility and higher employment opportunities. The contribution to income levels and quality of life, on one side, and the presence of backward and forward links with other productive sectors (agriculture, manufacturing, fishing and finance (Lin et al., 2018), on the other, made tourism to play a central role in the overall economic recovery of destinations hit by environmental shock.

When extending the analysis to the aftermaths of a more general economic shock, empirical evidences are not conclusive. First of all, higher tourism specialization or overdependence on tourism industry can make destinations even more vulnerable (Adamou and Clerides, 2010; Parrilla et al., 2007), thus jeopardizing their capacity to shift on a sustainable long-run growth path. The micro-level analysis carried out by Prayag et al. (2020), instead, proves that since resilience is a “dynamic, multi-dimensional and multi-scale” process, the recovery of tourism organizations, being whippier, represents a valid example of favourable relationships between different types of resilience (psychological, employee and organizational resilience). Finally, the extent to which individuals, communities and tourism businesses are resilient, and how quickly they are able to adapt after a negative shock, can determine the duration and the trajectory of the recovery process itself (Hall et al., 2016, 2018).

An additional aspect that deserves attention when investigating the contribution of tourism to economic recovery, is the potential existence of spillovers across neighbouring destinations. These could arise because of various channels. Migration across proximate areas, for example, favours the exchange of knowledge and skills which contribute to increase the overall labour productivity (Gu et al., 2006; Yang and Wong, 2012). Geographical or productive proximity strengthens the competition among enterprises and encourages the adoption of new innovation-based technologies and higher proficiencies (Romão et al., 2017). Joint promotion undertaken by two or more neighbouring destinations can generate positive spatial interdependences (Gooroochurn and Hanley, 2005; Ouchen and Montargot, 2022). And, finally, negative shocks such as conflicts, political turmoil, diseases or natural disasters, by contrast, may exert negative impacts on the economic performances of proximate destination areas (Drakos and Kutan, 2003; Neumayer, 2004; Sönmez and Graefe, 1998).

In this vein, it becomes important to account for the presence of spatial interdependencies. However, out of fewer contributions detecting spatial interdependence in the tourism-led growth process, there are no spatial analyses dealing with this aspect when detecting the linkage between tourism and economic recovery.

Among the studies related to the presence of spatial interdependence in the tourism-led growth path, Romão (2015) found positive spatial effects for European regions’ growth driven by tourism demand benefits; Yang and Fik (2014) referred to spatial spillovers and spatial heterogeneity to explain differences in tourism growth across a large sample of Chinese cities; inter-regional agglomeration effects in tourism are found by Majewska (2015) in Poland; and Kim et al. (2021), using microeconomic panel data on local authority districts in the UK, found significant spatial spillover effects on agglomeration economies and productivity in the tourism industry.

As far as we know, our study represents the first attempt to investigate the contribution of tourism on economic recovery not only by applying standard dynamic methodologies but also a spatial econometric perspective.

Empirical analysis

Data and descriptive analysis

The sample used in our investigation is given by a balanced panel of 153 European regions (NUTS2 level) belonging to 10 countries – Austria, Belgium, Denmark, France, Germany, Greece, Italy, Spain, Portugal and the Netherlands-observed over the period 2009–2019, for a total of 1683 observations¹.

The dependent variable is the Economic Recovery index, obtained as follows:

E R_{i t} = \frac{G D P p c_{i t}}{G D P p c_{i 0}}

with GDPpc_i0 equal to region’s i per capita GDP (in purchasing power) in 2008 (year of the peak before the financial crisis) and GDPpc_it indicating the same variable in year t (with t = 2009, …, 2019).

E R_{i t} > 1

indicates that the region at time t recovers after the slowdown caused by the 2007–2008 global financial crisis - the grater the index, the higher the GDP in year t in respect to the peak year before the crisis - while

E R_{i t} < 1

indicates that the GDP level at time t does not recover in respect to the peak year level, that is the absence of recovery.

The main explanatory variable is tourist arrivals (Arr_it) capturing the number of arrivals (in millions) in European region i at time t. It is given by the sum of foreign and domestic arrivals at tourist accommodation establishments and represents a standard measure to detect the tourism phenomenon inside a territory (Brida et al., 2016b; Ouchen and Montargot, 2022; Yang and Wong, 2012).

Figure 1 provides a first insight of the correlation between the annual averages of the recovery index and the arrivals (in millions) at regional level in Europe. The line that interpolates the data suggests that tourism arrivals and the economic recovery move together providing space for our investigation strategy.

Figure 1.

Economic recovery and tourist arrivals: panel mean (2009–2019). Source: own elaboration on Eurostat dataset.

The presence of a potential spatial linkage between economic recovery and tourism development across neighbouring regions is suggested by the distribution of the recovery index as shown in Figure 2 for the year 2019. The map shows a clear spatial clusterization of the economic recovery: regions in the Nord-Est stand out for their higher resilience; Central Europe and the Iberian peninsula show medium-high values; and Italian and Greek regions present the lowest values. The geographical distribution of the index hints at the existence of an underlying spatial dependence in the recovery process that deserves to be investigated.

Figure 2.

Economic recovery index in European regions in 2019. Source: own elaboration on Eurostat dataset.

Figure 3 shows the geographical distribution of tourism arrivals. Mean values of arrivals, beyond a spatial correlation across European regions, show a considerable variability within each country. In detail, the Iberian peninsula, with few exceptions, France, Nord of Italy and Austria present the highest values while Eastern Europe shows slightly lower values and Greece, South of Italy and North-West the lowest ones.

Figure 3.

Regional arrivals, mean values over the period 2009–2019. Source: own elaboration on Eurostat dataset.

The comparison between recovery (Figure 2) and tourism arrivals (Figure 3) suggests the presence of a positive impact of tourism in European regions.

In this study we investigate the role played by tourism by considering also other indicators detecting the demand side as well as the supply side of this sector. In detail, we consider: Location Quotient in accommodation (LQ^ACC), calculated as the ratio between the percentage of regional employment in Accommodation and Food activities and the same percentage at national level (Watson and Deller, 2022); Location Quotient in Art and Entertainment calculated as the percentage of regional employment over the national employment percentage in these activities (LQ^ENT); and Tourism Territorial Pressure (TTP), given by the ratio between thousands of presences (number of arrivals multiplied by the number of nights) and the population per km² (population density). LQ values greater than one indicate greater specialization and therefore a relative dependence of the region on the sector. TTP index, based on the additional pressure that tourism presences exert on the territory, enables to capture the environmental and economic dimensions of the impact as well as the attitude of the resident population towards possible overcrowding phenomena. TTP squared is considered for non-linear effects due to overtourism (Canale and De Siano, 2021; De Siano and Canale, 2022). The demand-side is accounted by Arrivals and Tourism Territorial Pressure (both at level and square), while the supply-side is detected through both the Location Quotients. To isolate these effects, we replicate the estimation of the model including separately tourism demand and tourism supply side components.

The empirical model includes also the usual controls suggested by the growth literature: Investment growth rate (INV^RATE) based on gross fixed capital formation; Activity rate (ACT^RATE) given by labour force aged 15–64 as percentage of the total population aged 15–64, to measure the general macroeconomic contribution of the labour input; Tertiary Education (EDU^TER) as the percentage of the whole population with a tertiary educational level, to account for the human capital contribution; and, finally, a dummy variable assuming the value of one for years from 2011 to 2019 (Dummy₂₀₁₁) (see Figure 1). The latter enables to account for the sovereign debt crisis that in many European countries was followed by structural adjustment programs which have inevitably affected both growth and the trend in aggregate demand.

The whole dataset has been retrieved from Regional Eurostat statistics at NUTS2 level. Table 1 lists the variable labels and summarizes the statistical features of our dataset. All variables are stationary according to Pesaran CADF test (Pesaran, 2004) which accounts for cross sectional dependence in series that may result from unobservable common shocks. The only exceptions are for TTP and TTP squared for which we use first differences in the model estimation. At the bottom of Table 1 we present the result of the variance inflation factor (VIF) test, implemented as a post-estimation test of a simple OLS regression in order to verify how much the variance of the estimated regression coefficients is increased because of collinearity. This test provides useful information about the eventual presence of “repeated” variables in the estimates that could occur because of the contemporaneous inclusion of variables – for example ARR and the TTP index – capturing similar information. The value of VIF is 1.17 far below the threshold value of 5, which allows us to keep all listed variables.

Table 1.

Variables and summary statistics.

Variables	Description	Obs	Mean	Std. Dev	Min	Max	CADF* test
ER	Economic recovery index	1683	1.041267	0.123768	0.697479	1.613636	−2.052***
LQ^ACC	% of regional employment in accommodation and food branches as ratio of national ones	1683	0.951596	0.229959	0.368501	2.747039	−2.515***
LQ^ENT	% of regional employment in entertainment branch as ratio of national ones	1683	1.000974	0.121792	0.684531	1.663457	−1.939***
ARR	Millions of arrivals in the region	1683	4,202,126	4,713,911	46169.7	4.05E + 07	−2.461***
TTP	Ratio between presences and population density	1683	4,182,798	9.15E + 07	21.15872	2.91E + 09	−0.931
ACT^RATE	Ratio between the labour force aged 15–64 and the total population aged 15–64	1683	0.722245	0.082991	0.034327	0.831354	−1.955***
INV^RATE	Change in gross fixed capital formation	1683	0.817304	10.96026	−49.8692	120.726	−2.532***
EDU^TER	Percentage of people with at least a tertiary education level	1683	27.62436	8.265668	11	58.4	−2.520***

*Null hypothesis: non-stationarity (Pesaran, 2004).

VIF (mean): 1.17.

Source: own elaboration on Eurostat data at NUTS2 level.

Methodology and estimation results

After checking the stationarity of the variables, the study employs two different approaches to investigate the influence of tourism on European regional economic recovery. The first, known as System GMM, is a dynamic panel model specification accounting for autocorrelation between dependent and explanatory variables. This technique is suitable in the case of a large N and small t, that is in the case of a number of observed individuals (153 European regions) much higher than the observations over time (11 years). It uses moment conditions, in which the lagged differences of the dependent variable are used as instruments in the level equation. Through the first differences this estimator removes the regional-specific effect and uses instruments to form moment conditions and eliminates the problem of heterogeneity across panel members. The system GMM is an extension of the Arellano–Bond (Arellano and Bond, 1991) estimator that accommodates large autoregressive parameters and a large ratio of the variance of the panel-level effect to the variance of idiosyncratic error (Blundell and Bond, 1998; Blundell et al., 2000), therefore providing robust results in case of equations to be estimated with lagged dependent variables. Since the estimator must be free of autocorrelations in the idiosyncratic errors, the Arellano-Bond test for first- and second-order autocorrelation in the first-differenced errors is performed after the estimation (tests results in Table 2).

Table 2.

Tourism and economic recovery in European regions (2009–2019): SYS-GMM and SDM Estimation results.

Dependent variable: index of economic recovery (ER)
Variables	SYS-GMM	Spatial durbin model
Variables	SYS-GMM	Main	WX	LR_Direct	LR_Indirect	LR_Total
ER_t-1	1.025*** (0.025)	0.750*** (0.0141)	−0.292*** (0.0283)	0.759*** (0.0138)	0.0987*** (0.0214)	0.857*** (0.0216)
LQ^ACC	0.008 (0.016)	0.0428*** (0.0112)	−0.0489* (0.0282)	0.0388*** (0.0121)	−0.0438 (0.0498)	−0.00498 (0.0563)
LQ^ENT	−0.007 (0.008)	−0.00383 (0.00548)	0.00142 (0.0128)	−0.00338 (0.00581)	−0.00130 (0.0209)	−0.00467 (0.0242)
ARR	8.72e − 09*** (2.36E − 09)	1.44e − 09* (8.55e − 10)	8.09e − 09*** (1.44e − 09)	2.58e − 09*** (8.49e − 10)	1.43e − 08*** (2.29e − 09)	1.69e − 08*** (2.62e − 09)
TTP	4.82e − 13 (9.71e − 12)	8.07e − 12*** (5.92e − 12)	−2.11e − 11 (7.71e − 12)	6.10e − 12** (6.45e − 12)	−2.81e − 11 (1.43e − 11)	−1.43e − 20 (9.43e − 21)
TTP²	−4.20e − 21 (6.10e − 21)	−1.22e − 21* (2.82e − 21)	7.21e − 21 (3.96e − 21)	−2.11e − 21* (3.12e − 21)	−1.22e − 20 (7.37e − 21)	−1.43e − 20 (9.43e − 21)
INV^RATE	0.0005*** (0.0001)	0.000204*** (5.21e − 05)	0.000325*** (8.93e − 05)	0.000261*** (5.20e − 05)	0.000687*** (0.000136)	0.000948*** (0.0001)
ACT^RATE	0.320*** (0.105)	0.230*** (0.0562)	−0.0416 (0.0987)	0.239*** (0.0558)	0.115 (0.152)	0.354** (0.171)
EDU^TER	−0.0001 (0. .0001)	5.01e − 05 (7.94e − 05)	9.98e − 06 (0.000158)	6.21e − 05 (8.88e − 05)	5.97e − 05 (0.000284)	0.000122 (0.0003)
Dummy₂₀₁₁	−0.015*** (0.002)	0.0203*** (0.00493)	−0.0248*** (0.00515)	0.0184*** (0.00458)	−0.0253*** (0.00488)	−0.00683** (0.00287)
Constant	−0.268*** (0.069)
rho		0.470*** (0.0260)
sigma2_e		0.000335*** (1.23e − 05)
Regional FE	NA	YES	YES	YES	YES	YES
Observations	1530	1530	1530	1530	1530	1530
R-squared		0.690	0.690	0.690	0.690	0.690
Number of regions	153	153	153	153	153	153
Post-estimation tests
AB test ord. 1	−4.174***
AB test ord. 2	0.821
SDM vs SAR
Chi2	175.80 (0.000)
Prob>chi2	175.80 (0.000)
SDM vs SEM
Chi2	150.17 (0.000)
Prob>chi2	150.17 (0.000)
SDM vs SAC
AIC (SDM)	−7654.036
AIC (SAC)	−7780.486

Notes. The dependent variable is the economic recovery index (ER). The first column shows the system GMM results and the following columns display the spatial durbin model estimates: regions’ explanatory variables (main), neighboring regions’ explanatory variables (WX) and long run direct, indirect and total effects of regions’ explanatory variables, respectively. In both models, the explanatory variables of interest are arrivals (ARR), tourism territorial pressure (TTP) and its square values (TTP²) and indicators of regional tourism specialization, local quotient in accommodation services (LQ^ACC) and local quotient in entertainment activities (LQ^ENT). TTP and TTP² are at first-differences. The baseline controls for economic growth included in the model are defined in the text and described in Table 1. Standard errors in parenthesis.

*p < .1; **p < .05; ***p < .01.

The estimation strategy defined by the system GMM technique is described by the following equation:

E R_{i t} = α_{1} E R_{i t - 1} + α_{2} L Q_{i t}^{A C C} + α_{3} L Q_{i t}^{E N T} + β_{1} A R R_{i t} + γ_{1} T T P_{i t} + γ_{2} T T P_{i t}^{2} + ζ X_{i t} + μ_{i} + ε_{i t}

(1)

with i = 1, …153 and t = 2009, …2019. X indicates the whole set of controls, μ_i the region-specific fixed effects and ε_it the error term. In running the estimates, the regional-specific fixed effects are removed because of the transformation in first differences.

The second is a dynamic spatial panel model which also captures the role of geographical interconnections as well as spillovers across proximate regions (Anselin, 2013; Elhorst, 2017; LeSage and Pace, 2009, 2014). Geographical proximity is captured by a binary contiguity matrix where weights are set equal to 1, if regions share a border, and 0 otherwise. As regions in the sample have a different number of neighbours, we row-standardize the matrix in order to obtain proportional weights for each contiguous region. Following LeSage and Pace (2009), we investigate the presence of spatial dependence by using the general Spatial Durbin model (SDM) specification, including both spatially lagged dependent and independent variables. Specific log-likelihood ratio tests and information criteria (tests results in Table 2) are then employed to test the suitability of the SDM versus alternative spatial models (SAR, SEM, SAC). The SDM specification, in particular, has the advantage to account for the whole complex set of spatial dependencies arising both as an endogenous interaction, driven by recovery process in neighbouring regions, and as an exogenous interaction, from changes in the explanatory variables in the region itself and in its neighbours. Regarding the effects of changes in internal independent variables, the SDM allows to separate the direct effects, on region’s own dependent variable, from the indirect effects. The latter, also defined as spillover effects, represent the influence back that region’s dependent variable receives when any change in its explanatory variable affects neighbours’ dependent variable (LeSage and Dominguez, 2012; LeSage and Pace, 2014).

The Spatial Durbin model is described by the following equation:

E R_{i t} = ρ W E R_{i t} + τ_{1} E R_{i t - 1} + α_{1} L Q_{i t}^{A C C} + β_{1} L Q_{i t}^{E N T} + γ_{1} A R R_{i t} + δ_{1} T T P_{i t} + φ_{1} T T P_{i t}^{2} + ζ_{1} X_{i t} + τ_{2} W E R_{i t - 1} + τ_{1} E R_{i t - 1} + α_{2} W L Q_{i t}^{A C C} + β_{2} W L Q_{i t}^{E N T} + γ_{2} W A R R_{i t} + δ_{2} W T T P_{i t} + φ_{2} W T T P_{i t}^{2} + ζ_{2} W X_{i t} + μ_{i} + ε_{i t}

(2)

where ρ is the spatial autoregressive parameter, W is a 153 × 153 spatial contiguity matrix, independent variables multiplied by W refer to values of these variables in neighbouring regions, μ_i are region-specific fixed effects and ε_it the error term.

Table 2 shows estimation results. The first column presents estimates obtained with the system GMM which, being a dynamic estimator, reduces the observations for each geographical unit by one (total number of observations reduces from 1683 to 1530). The first thing that is worth noting is the significance and the high value of the lagged dependent variable (1.025***). This result – as described in the methodology section – assigns more validity to the coefficient of the other explanatory variables. Scrolling down the column it is possible to observe that the coefficients of the variables representing regional tourism specialization are not significant, therefore neither the presence of a specialization in accommodation activities nor a specialization in entertainment activities are able to foster economic recovery. The coefficient describing the effect of arrivals is very significant and equal to 8.72e − 09***. The index of congestion is not significant both when considering it at level and at its squared value, despite the signs are consistent with the presence of a non-linear effect (De Siano and Canale 2022). This is maybe due to the width of the territories considered. The variables usually included in the standard growth models-that is activity rate and investment rate-describe a positive connection with our recovery index (INV^RATE = 0.0005*** and ACT^RATE = 0.320***). Nothing can be said about the variable EDU^TER maybe due to the long time necessary for human capital to affect growth. Furthermore, as predicted by Figure 1, the year dummy variable is significant and negative, registering the adverse shock of the foreign debt crisis occurred in 2011. Finally, the AB test rejects the null-hypothesis of the presence of autocorrelation of second order. Therefore, our results support the conclusion that tourists’ flows foster recovery, despite tourism specialization appears to be a non-determinant factor.

The remaining columns of Table 2 present the results of the spatial model. Test statistics confirm the SDM as the most appropriate specification in detecting spatial interdependences of economic recovery across European regions. The estimate for spatial autoregression coefficient, which represents the spatial lag of the dependent variable, reveals that regions are positively influenced by neighbours’ recovery (0.470***). This implies that the recovery of proximate economies favors the increase of regional GDP and therefore the possibility for a region to close the gap with the peak level registered before the financial crisis.

As stated by LeSage and Pace (2009), in presence of spatial dependence there is no sense to look at isolated effects of internal changes in explanatory variables, as their total effect maybe conditioned by spillover effects. In detail, we find that arrivals exert the expected positive direct effects on regional economic recovery (2.58e − 09***). These effects are strongly reinforced by the presence of positive spillovers from neighbouring regions (1.43e − 08***). As long as tourism flows do not turn into mass arrivals, an increase of tourism pressure on the territory appears to have a propulsive effect on the recovery process (6.10e − 12**). However, once a critical threshold is exceeded, tourism pressure exerts a direct negative effect on regional resilience (coefficient for TTP² is −2.11e − 21*). Regarding the specialization in tourism activities, accommodation and food services show only a direct positive effect on regional GDP (0.0388***).

Among growth enhancing factors, activity and investment rates show a positive contribution confirmed by the sign of their direct effects on regional economic recovery (0.24*** and 0.0002***, respectively), while education is confirmed to have no impact on the recovery process.

The spatial Durbin approach enables to account also for potential effects driven by changes in neighbouring tourism contexts. In this regard, recovery appears to be positively affected by an increase in arrivals in neighbouring regions. Indeed, we find a positive and significant coefficient for tourism arrivals in proximate regions (8.09e − 09***). By contrast, regional recovery is harmed by a higher specialization in accommodation and food services activities in nearby regions (−0.0489*), suggesting the presence of a competition across proximate destinations aimed at attracting tourists’ flows and potential demand source. Finally, among the nearby growth enhancing factors, economic recovery appears to be positively affected by increasing investment rates.

At last, attempting to isolate tourism demand-side and supply-side effects on economic recovery we estimated, following the same econometric approaches, two alternative empirical models. The first including just the arrivals and the congestion effects to detect the demand side components, the second considering the variables accounting for specialization in the sector which observe the phenomenon from tourism firms’ supply-side. Both models include also the variables accounting for the general economic system dynamics. Results, presented in the appendix, substantiate the presence of tourism contribution to economic recovery, by means of the positive effect of arrivals, even when the two aspects of the market are considered separately. In summary, Table A1. “The contribution of tourism to economic recovery in European Regions: demand-side effects” tells us that no relevant difference is present in the significance and the value of the estimated coefficients. The only difference arises with the effects of tourism territorial pressure when detecting the presence of spatial dependence: regional capacity to recover after the financial crisis is harmed by TTP rising in neighbouring regions (−1.87e − 11***) and by regions’ themselves TTP spillovers (for both TTP at level and square). Similar consideration can be made when observing Table A2. “The contribution of tourism to economic recovery in European Regions: supply effects” where the coefficients of the variables accounting for tourism specialization remain not significant in the system GMM approach. By contrast, accounting for spatial dependence, specialization in accommodation services keeps the negative impact when increasing in neighbouring regions and a positive direct effect when registered within the region.

Robustness checks

With the aim of supporting the results presented above, as robustness check we further estimate a model where we substitute the number of arrivals in the region with the arrivals weighted by the resident population (ARR/POP). This change provides additional information about the contribution of tourism demand dynamics on the recovery path of each single region by considering the dimension of each economic territory. Furthermore, we introduced, in substitution of the dummy variable for the sovereign debt crisis, the year fixed effects. This second change allows a more comprehensive check on the eventual presence of a shock separately in each year of the sample period. Results, shown in Table 3, are very similar to those of the principal model presented in Table 2. When looking at the first column, presenting results of the system-GMM model, the coefficient of our main explanatory variable of interest (ARR/POP) is again significant and positive (0.005**), supporting the validity of our interpretative hypothesis about the role of tourism in economic recovery. The coefficients of variables accounting for the specialization in tourism sector are still not significant, while those referring to a general model of growth – excluding tertiary education - are significant and with the expected signs (INV^RATE = 0.0003***, ACT^RATE = 0.292***). The regional specific fixed effects are removed because of the system GMM empirical strategy using first differences as instruments of the level equation.

Table 3.

Tourism and economic recovery in European regions: robustness check.

Dependent variable: index of economic recovery (ER)
Variables	System GMM	Spatial durbin model
Variables	System GMM	Main	Wx	LR_Direct	LR_Indirect	LR_Total
ER_t-1	0.993*** (.024)	0.742*** (0.0140)	−0.257*** (0.0289)	0.747*** (0.0136)	0.0665*** (0.0224)	0.814*** (0.0212)
LQ^ACC	−0.000 (0.013)	0.0230** (0.0116)	−0.0698** (0.0276)	0.0158 (0.0122)	−0.0897** (0.0425)	−0.0739 (0.0489)
LQ^ENT	−0.005 (0.008)	−0.00382 (0.00537)	−0.00126 (0.0126)	−0.00353 (0.00566)	−0.00356 (0.0197)	−0.00709 (0.0229)
ARR/POP	0.005** (.003)	0.00506*** (0.00109)	0.0116*** (0.00345)	0.00657*** (0.00107)	0.0201*** (0.00484)	0.0266*** (0.00512)
TTP	1.06e − 12 (7.89e − 12)	6.87e − 12 (0.238)	−1.97e − 11 *** (0.009)	5.03e − 12 (0.402)	−2.52e − 11** (0.023)	−2.02e − 11 (0.175)
TTP²	−2.50e − 21 (4.25e − 21)	−6.65e − 22 (0.810)	−4.99e − 21 (0.200)	−1.12e − 21 (0.697)	−7.87e − 21 (0.182)	−8.99e − 21 (0.243)
INV^RATE	0.0003*** (0.0001)	0.000142*** (5.25e − 05)	0.000227** (9.14e − 05)	0.000174*** (5.63e − 05)	0.000417*** (0.000128)	0.000590*** (0.000151)
ACT^RATE	0.292*** (0.074)	0.208*** (0.0567)	−0.0108 (0.0951)	0.214*** (0.0536)	0.106 (0.136)	0.321** (0.147)
EDU^TER	−0.0001 (0. .0001)	7.43e − 05 (7.80e − 05)	−5.91e − 05 (0.000155)	7.89e − 05 (8.29e − 05)	−3.60e − 05 (0.000245)	4.29e − 05 (0.000292)
Constant	−0.206*** (0. .052)
Rho		0.407*** (0.0283)
sigma2_e		0.000321*** (1.18e − 05)
Regional FE	N.A.	YES
Year FE	YES	YES
Observations	1530	1530	1530	1530	1530	1530
R-squared		0.796	0.796	0.796	0.796	0.796
Number of regions	153	153	153	153	153	153
Post-estimation tests
AB test ord. 1	−3.951***
AB test ord. 2	0.726
SDM vs SAR
Chi2	127.39*** (0.000)
Prob > chi2	127.39*** (0.000)
SDM vs SEM
Chi2	80.39*** (0.000)
Prob > chi2	80.39*** (0.000)
SDM vs SAC
AIC (SDM)	−7838.259
AIC (SAC)	−7838.259

Notes. The dependent variable is the economic recovery index (ER). The first column shows the GMM results and the following columns display the Spatial Durbin model estimates: regions’ explanatory variables (main), neighboring regions’ explanatory variables (Wx) and direct, indirect and total effects of regions’ explanatory variables, respectively. In both models, the explanatory variables of interest are Arrivals (ARR), Tourism Territorial Pressure (TTP) and its square values (TTP2) and indicators of regional tourism specialization, Local Quotient in Accommodation services (LQ^ACC) and Local Quotient in Entertainment activities (LQ^ENT). The baseline controls for economic growth included in the model are defined in the text and described in Table 1. Standard errors in parenthesis.

*p < .1; **p < .05; ***p < .01.

Regarding the spatial specification of the modified model, the first evidence is the presence of a strong positive spatial dependence across regions’ capacity to recover, which confirms previous results shown in Table 2. Then, it is noteworthy the outcome associated to the variable of interest ARR/POP that, while maintaining the positive expected sign, presents a considerably larger size, both when arrivals are registered in the region itself and in neighbouring regions. By weighting arrivals on the population of the region, results highlight even more the positive contribution of tourism to the regional economic recovery. An increase of Tourism territorial pressure (TTP), instead, appears to be always detrimental, when registered in the region itself (due to spillover effects) and when occurring in proximate regions. Regarding, growth contributing factors, capital (INV^RATE) and labour (ACT^RATE) accumulations remain statistically significant and do not show significant changes either in the sign or in the size of their estimates.

Conclusions

The role of tourism within the economic activity is gaining an ever-increasing attention. This is true not only in developing countries, but also in advanced economies where it is considered at the centre stage of debates about its contribution to growth and recovery. This is also the case of countries belonging to Europe which, constrained by supranational rules in managing policy interventions, are in search of instruments to sustain economic recovery. In this context, tourism seems to cover a central role, as it is a dynamic sector able to rapidly react after a systemic negative shock.

As matter of fact, our investigation reveals the positive potential that tourism activities exert on the ability of a region to recover from a negative exogenous shock hitting its economy. The panel dynamic estimation strategy accounting for endogeneity reveals that the main factor influencing recovery is tourists’ arrivals, hence revealing a fast and important connection between tourism and its contribution to growth. This result, when limiting the analysis to the System-GMM model, is independent from tourism specialization but rather goes together with the standard components of growth (investments and activity rates). It is the increase of tourism arrivals that, contributing positively to aggregate demand, enhance the recovery process.

Similar and more articulated findings are revealed when spatial dependence based on geographical proximity is considered for European regions. Tourists’ arrivals do promote GDP growth, thus contributing to enhance the capacity to recover after a negative shock, but, as the congestion increases, tourism turns to be detrimental for the economic recovery. Moreover, the direct effect of regional tourism flows is strongly reinforced by the presence of spillover effects from neighbouring regions. A positive contribution is provided also by the internal specialization in the accommodation and food services activities (0.0388***) while increasing specialization in proximate regions appears to be detrimental.

Results do not change significantly if we consider separately the indicators of tourism demand and supply side, both for the system-GMM and the spatial model specifications.

What this analysis highlights is that - besides the increase of the tourism context-a consistent boost to close the gap after the crisis is driven by the recovery of proximate economies. However, regions benefit from increasing tourist flows to neighbouring regions but suffer when the latter are characterized by higher tourism territorial pressure or higher specialization in traditional tourism activities such as hospitality and catering businesses.

Our analysis supports the conclusion that the tourism sector plays a central role for economic growth, due especially to the effect of increasing demand. However, it suggests also that an inappropriate and excessive exploitation of touristic resources might produce conflict with inhabitants. The ever-increasing visitors flows thus compromises the overall positive contribution that tourism may exert on the economic recovery.

The policy implication is that, since tourist flows have a fast dynamic, focusing on promoting the territory and a convincing image of the destination, even in times of crisis, provides rapid recovery tools, once the turbulence has passed. Sustained growth in the tourism sector should be pursued whilst reconciling the ability to produce income and ensure a good quality of life and, finally, dampening competition between proximate regions. Policy makers should implement also measures to accompany increasing demand with a production structure capable of responding to the growing demands of visitors. Quality of services, transport efficiency and integrated management of the tourist supply in each destination are necessary elements to consolidate the sector’s contribution to GDP and limit its volatility. As a matter of fact, as tourism demand rapidly contributes to recovery, it can also exert a negative pressure on GDP in times of turbulence of regional, national, and international economies. Therefore, our study suggests non only that the tourism sector should be protected from exogenous negative shocks as its rapid decline could amplify the effects on regional GDP, but also that an excessive share of the tourism sector inside national economy, in presence of instability, makes the production structure fragile.

The main limitation of our study is that GDP could provide a too wide perspective to examine the effects on the territory. Therefore, especially regarding the pressure on the territory, future investigation should refer to narrower areas – such as cities or historical sites – to fully account for the contribution of tourism to GDP.

Footnotes

Declaration of conflicting interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Rita De Siano

Note

Author biographies

Rita De Siano, Associate professor in Economic Policy. Research interests: Regional economic resilience, regional economy and territorial dualism, tourism and economic growth, role of women in growth and development processes.

Rosaria Rita Canale, full professor in Economic Policy. Research interests: macroeconomic policies in the European Union, territorial dualism, tourism and economic growth.

Appendix

Table A1.

The contribution of tourism to economic recovery in European Regions: demand-side effects

Dependent variable: Index of Economic Recovery (ER)
		Spatial Durbin model
Variables	System GMM	Main	WX	LR_Direct	LR_Indirect	LR_Total
ER_t-1	1.025***	0.754***	-0.292***	0.762***	0.0935***	0.855***
	(0. 026)	(0.0141)	(0.0284)	(0.0137)	(0.0218)	(0.0213)
ARR	8.73e-09***	1.53e-09*	8.35e-09***	2.69e-09***	1.47e-08***	1.74e-08***
	(2.36e-09)	(8.58e-10)	(1.45e-09)	(8.46e-10)	(2.24e-09)	(2.56e-09)
TTP	3.65e-13	9.20e-12	-1.87e-11***	7.98e-12	-2.26e-11*	-1.46e-11
	(9.72e-12)	(0)	(0)	(0)	(0)	(0)
TTP²	-4.36e-21	-1.18e-21	-6.27e-21	-1.92e-21	-1.11e-20*	-1.30e-20
	(6.12e-21)	(0)	(0)	(0)	(0)	(0)
INV^RATE	0.0005***	0.000208***	0.000338***	0.000266***	0.000700***	0.000966***
	(0.0001)	(5.24e-05)	(8.96e-05)	(5.32e-05)	(0.000140)	(0.000164)
ACT^RATE	0.323***	0.233***	-0.0392	0.245***	0.115	0.360**
	(0.107)	(0.0564)	(0.0986)	(0.0557)	(0.153)	(0.168)
EDU^TER	0.0001	5.19e-05	1.47e-05	5.56e-05	4.83e-05	0.000104
	(0.0001)	(7.97e-05)	(0.000158)	(9.25e-05)	(0.000263)	(0.000322)
Dummy₂₀₁₁	-0.015***	0.0217***	-0.0263***	0.0194***	-0.0263***	-0.00688**
	(0.002)	(0.00495)	(0.00517)	(0.00442)	(0.00502)	(0.00272)
Constant	0.0001***
	(0.0001)
rho		0.464***
		(0.0261)
sigma2_e		0.000340***
		(1.25e-05)
Regional FE	N.A.	YES
Observations	1530	1530	1530	1530	1530	1530
R-squared		0.646	0.646	0.646	0.646	0.646
Number of regions	153	153	153	153	153	153
Post-Estimation Tests
AB test ord. 1	-4.1827***
AB test ord. 2	0.8166
SDM vs SAR
Chi2				174.67***
Prob>chi2				(0.000)
SDM vs SEM
Chi2				151.86***
Prob>chi2				(0.000)
SDM vs SAC
AIC (SAM)				-7770.423
AIC (SAC)				-7636.636

Notes: The dependent variable is the Economic Recovery index (ER). The first column shows the System-GMM results and the following columns display the Spatial Durbin model estimates: regions’ explanatory variables (main), neighboring regions’ explanatory variables (WX) and long run direct, indirect and total effects of regions’ explanatory variables, respectively. In both models, the explanatory variables of interest are demand-side tourism components: Arrivals (ARR), Tourism Territorial Pressure (TTP) and its square values (TTP2). The baseline controls for economic growth included in the model are defined in the text and described in Table 1. Standard errors in parenthesis. * p < 0.1; ** p < 0.05; *** p < 0.01.

Table A2.

The contribution of tourism to economic recovery in European Regions: supply-side effects

Dependent variable: Index of Economic Recovery (ER)
		Spantertial Durbin model
Variables	System GMM	Main	WX	LR_Direct	LR_Indirect	LR_Total
ER_t-1	1.067***	0.752***	-0.295***	0.765***	0.149***	0.914***
	(0. 017)	(0.0142)	(0.0280)	(0.0136)	(0.0230)	(0.0215)
LQ^ACC	80.0233	0.0441***	-0.0552*	0.0391***	-0.0534	-0.0143
	(0.0163)	(0.0114)	(0.0284)	(0.0125)	(0.0498)	(0.0570)
LQ^ENT	-0.004	-0.00613	-0.00366	-0.00647	-0.0104	-0.0169
	(0.008)	(0.00553)	(0.0129)	(0.00605)	(0.0229)	(0.0266)
INV^RATE	0.0006***	0.000246***	0.000460***	0.000333***	0.00102***	0.00136***
	(0.0001)	(5.22e-05)	(8.53e-05)	(5.19e-05)	(0.000135)	(0.000154)
ACT^RATE	0.209***	0.248***	0.0798	0.281***	0.372**	0.653***
	(0.065)	(0.0568)	(0.0979)	(0.0557)	(0.167)	(0.185)
EDU^TER	0.0001	3.30e-05	-3.97e-05	3.33e-05	-4.95e-05	-1.62e-05
	(0.0001)	(8.01e-05)	(0.000159)	(9.28e-05)	(0.000296)	(0.000362)
Dummy₂₀₁₁	-0.0124***	0.0208***	-0.0232***	0.0189***	-0.0222***	-0.00336
	(0.002)	(0.00496)	(0.00519)	(0.00477)	(0.00520)	(0.00299)
Constant	-0.213***
	(0.048)
rho		0.507***
		(0.0248)
sigma2_e		0.000343***
		(1.27e-05)
Regional FE	N.A.	YES
Observations	1530	1530	1530	1530	1530	1530
R-squared		0.600	0.600	0.600	0.600	0.600
Number of regions	153	153	153	153	153	153
Post-Estimation Tests
AB test ord. 1	-4.256 ***
AB test ord. 2	0.853
SDM vs SAR
Chi2				202.46***
Prob>chi2				(0.000)
SDM vs SEM
Chi2				127.92***
Prob>chi2				(0.000)
SDM vs SAC
AIC (SDM)				-7727.258
AIC (SAC)				-7640.954

Notes: The dependent variable is the Economic Recovery index (ER). The first column shows the System-GMM results and the following columns display the Spatial Durbin model estimates: regions’ explanatory variables (main), neighboring regions’ explanatory variables (WX) and long run direct, indirect and total effects of regions’ explanatory variables, respectively. In both models, the explanatory variables of interest are supply-side tourism components: Local Quotient in Accommodation services (LQ^ACC) and Local Quotient in Entertainment activities (LQ^ENT). The baseline controls for economic growth included in the model are defined in the text and described in Table 1. Standard errors in parenthesis. * p < 0.1; ** p < 0.05; *** p < 0.01.

References

Adamou

Clerides

(2010) Prospects and limits of tourism-led growth: the international evidence. Review of Economic Analysis 3: 287–303.

Alkhathlan

Javid

(2013) Energy consumption, carbon emissions and economic growth in Saudi Arabia: an aggregate and disaggregate analysis. Energy Policy 62: 1525–1532.

Amore

Prayag

Hall

(2018) Conceptualizing destination resilience from a multilevel perspective. Tourism Review International 22(3–4): 235–250.

Anselin

(2013) Spatial Econometrics: Methods and Models. Berlin, Germany: Springer Science Business Media.

Arellano

Bond

(1991) Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies 58(2): 277–297. DOI: 10.2307/2297968.

Aslan

(2013) Tourism development and economic growth in the Mediterranean countries: evidence from panel granger causality tests. Current Issues in Tourism 17(4): 363–372.

Balaguer

Cantavella-Jordá

(2002) Tourism as a long-run economic growth factor: the Spanish case. Applied Economics 34(7): 877–884.

Blake

Sinclair

Soria

JAC

(2006) Tourism productivity: evidence from the United Kingdom. Annals of Tourism Research 33(4): 1099–1120.

Blundell

Bond

(1998) Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics 87: 115–143. DOI: 10.1016/S0304-4076(98)00009-8.

10.

Blundell

Bond

Windmeijer

(2000) Estimation in dynamic panel data models: improving on the performance of the standard GMM estimator. In: Baltagi

(ed) Nonstationary Panels, Cointegrating Panels and Dynamic Panels. New York, NY: Elsevier, 53–92.

11.

Brida

Pulina

(2010) A Literature Review on the Tourism-Led-Growth Hypothesis. Sassari, Sardinia: Centre for North South Economic Research, University of Cagliari and. Working Paper CRENoS 2010/17.

12.

Brida

Lanzilotta

Pereyra

, et al. (2015) A nonlinear approach to the tourism-led growth hypothesis: the case of the MERCOSUR. Current Issues in Tourism 18(7): 647–666.

13.

Brida

Cortes-Jimenez

Pulina

(2016a) Has the tourism-led growth hypothesis been validated? A literature review. Current Issues in Tourism 19(5): 394–430.

14.

Brida

Lanzilotta

Pizzolon

(2016b) Dynamic relationship between tourism and economic growth in MERCOSUR countries: a nonlinear approach based on asymmetric time series models. Economics Bulletin 36(2): 879–894.

15.

Butler

(1980) The concept of a tourist area cycle of evolution: implications for management of resources. Canadian Geographies/Géographies Canadiennes 24(1): 5–12. DOI: 10.1111/j.1541-0064.1980.tb00970.x.

16.

Canale

De Siano

(2021) Territorial pressure and tourism contribution to GDP: the case of Italian regions. International Journal of Tourism Research 23(5): 891–900.

17.

Capo

Font

Nadal

(2007) Dutch disease in tourism economies: evidence from the balearics and the canary islands. Journal of Sustainable Tourism 15(6): 615–627.

18.

Cellini

Cuccia

(2015) The economic resilience of tourism industry in Italy: what the ‘great recession’ data show. Tourism Management Perspectives 16: 346–356.

19.

Cerisola

Panzera

(2024) Heritage tourism and local prosperity: an empirical investigation of their controversial relationship. Tourism Economics 0(0). DOI: 10.1177/13548166241234099.

20.

Cheer

Lew

(eds) (2017) Tourism, Resilience and Sustainability: Adapting to Social, Political and Economic Change. London, UK: Routledge.

21.

Cheng

Zhang

(2020) Is tourism development a catalyst of economic recovery following natural disaster? An analysis of economic resilience and spatial variability. Current Issues in Tourism 23(20): 2602–2623.

22.

Cortes-Jimenez

Pulina

(2010) Inbound tourism and long-run economic growth. Current Issues in Tourism 13(1): 61–74.

23.

Kyaw

(2016) Tourism development and growth. Annals of Tourism Research 60: 23–26.

24.

De Siano

Canale

(2022) Controversial effects of tourism on economic growth: a spatial analysis on Italian provincial data. Land Use Policy 117: 106081. DOI: 10.1016/j.landusepol.2022.106081.

25.

Demiröz

Ongan

(2005) The contribution of tourism to the long-run development: a space-time analysis on European regions. The Region 4(3): 1–17.

26.

Dogru

Bulut

(2018) Is tourism an engine for economic recovery? Theory and empirical evidence. Tourism Management 67: 425–434.

27.

Drakos

Kutan

(2003) Regional effects of terrorism on tourism in three mediterranean countries. Journal of Conflict Resolution 47(5): 621–641.

28.

Ehigiamusoe

(2020) Tourism, growth and environment: analysis of non-linear and moderating effects. Journal of Sustainable Tourism 28(8): 1174–1192.

29.

Elhorst

(2017) Spatial panel data analysis. Encyclopedia of GIS 2: 2050–2058.

30.

Fahimi

Akadiri

Seraj

, et al. (2018) Testing the role of tourism and human capital development in economic growth. A panel causality study of micro states. Tourism Management Perspectives 28: 62–70.

31.

Gooroochurn

Hanley

(2005) Spillover effects in long-haul visitors between two regions. Regional Studies 39(6): 727–738.

32.

Kavanaugh

, et al. (2006) Human resources management in China’s hotel industry. China Tourism Research 2(3): 209–245.

33.

Gunduz

Hatemi

(2005) Is the tourism-led growth hypothesis valid for Turkey? Applied Economics Letters 12(8): 499–504.

34.

Hall

Malinen

Vosslamber

, et al. (2016) Business and Post-disaster Management: Business, Organisational and Consumer Resilience and the Christchurch Earthquakes. London, UK: Routledge.

35.

Hall

Prayag

Amore

(2018) Tourism and Resilience: Individual, Organisational and Destination Perspectives. Bristol, UK: Channel View Publications.

36.

Husein

Kara

(2011) Research note: re-examining the tourism-led growth hypothesis for Turkey. Tourism Economics 17(4): 917–924.

37.

IMF (2023) World Economic Outlook: Navigating Global Divergences. Washington, DC: IMF.

38.

Kim

Chen

Jang

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

39.

Kim

Williams

Park

, et al. (2021) Spatial spillovers of agglomeration economies and productivity in the tourism industry: the case of the UK. Tourism Management 82: 104201.

40.

Lee

Chang

(2008) Tourism development and economic growth: a closer look at panels. Tourism Management 29(1): 180–192.

41.

LeSage

Dominguez

(2012) The importance of modeling spatial spillovers in public choice analysis. Public Choice 150(3–4): 525–545.

42.

LeSage

Pace

(2009) Introduction to Spatial Econometrics. New York, NY: CRC Press.

43.

LeSage

Pace

(2014) The biggest myth in spatial econometrics. Econometrics 2(4): 217–249.

44.

Lin

VSS

Yang

(2018) Where can tourism-led growth and economy-driven tourism growth occur? Journal of Travel Research 58(5): 760–773.

45.

Liu

Ramos

Yang

, et al. (2024) Analysing the dynamic co-movement between tourism and expected economic growth considering extreme events. Tourism Economics 30(1): 3–26. DOI: 10.1177/13548166221129441.

46.

Majewska

(2015) Inter-regional agglomeration effects in tourism in Poland. Tourism Geographies 17(3): 408–436.

47.

Mérida

Golpe

(2016) Tourism‐led growth revisited for Spain: causality, business cycles and structural breaks. International Journal of Tourism Research 18(1): 39–51.

48.

Milesi-Ferretti

(2021) The COVID-19 travel shock hit tourism-dependent economies hard. Hutchins Center Working Papers.

49.

Mirza

Kanwal

(2017) Energy consumption, carbon emissions and economic growth in Pakistan: dynamic causality analysis. Renewable and Sustainable Energy Reviews 72: 1233–1240.

50.

Neumayer

(2004) The impact of political violence on tourism: dynamic cross-national estimation. Journal of Conflict Resolution 48(2): 259–281.

51.

(2005) The contribution of tourism development to economic growth in the Korean economy. Tourism Management 26(1): 39–44.

52.

Orchiston

Higham

JES

(2016) Knowledge management and tourism recovery (de) marketing: The Christchurch earthquakes 2010–2011. Current Issues in Tourism 19(1): 64–84.

53.

Ouchen

Montargot

(2022) Non-spatial and spatial econometric analysis of tourism demand in a panel of countries around the world. Spatial Economic Analysis 17(2): 262–283.

54.

Parrilla

Font

Nadal

(2007) Tourism and long-term growth a Spanish perspective. Annals of Tourism Research 34(3): 709–726.

55.

Payne

Mervar

(2010) Research note: the tourism growth nexus in Croatia. Tourism Economics 16(4): 1089–1094. DOI: 10.5367/te.2010.0014.

56.

Pesaran

(2004) General diagnostic tests for cross section dependence in panels. CWPE: 0435.

57.

Prayag

Spector

Orchiston

, et al. (2020) Psychological resilience, organizational resilience and life satisfaction in tourism firms: insights from the Canterbury earthquakes. Current Issues in Tourism 23(10): 1216–1233.

58.

Rasool

Maqbool

Tarique

(2021) The relationship between tourism and economic growth among BRICS countries: a panel cointegration analysis. Future Business Journal 7(1): 1–11.

59.

Ritchie

(2004) Chaos, crises and disasters: a strategic approach to crisis management in the tourism industry. Tourism Management 25(6): 669–683.

60.

Romão

(2015) Culture or nature: a space-time analysis on the determinants of tourism demand in European regions. Discussion Papers Spatial and Organisational Dynamics. CIEO- Research Centre for Spatial and Organizational Dynamics, University of Algarve, 3.

61.

Romão

Guerreiro

JPSM

Rodrigues

(2017) Territory and sustainable tourism development: a space-time analysis on European regions. The Region 4(3): 1–17.

62.

Saboori

Ghaderi

Soleymani

(2023) A revised perspective on tourism-economic growth nexus, exploring tourism market diversification. Tourism Economics 29(7): 1812–1835. DOI: 10.1177/13548166221134183.

63.

Sahni

Nsiah

Fayissa

(2023) Tourism specialization, growth stage, and economic growth. Tourism Economics 29(6): 1533–1550. DOI: 10.1177/13548166221111311.

64.

Santamaria

Filis

(2019) Tourism demand and economic growth in Spain: new insights based on the yield curve. Tourism Management 75: 447–459.

65.

Sönmez

Graefe

(1998) Influence of terrorism risk on foreign tourism decisions. Annals of Tourism Research 25(1): 112–144.

66.

Tang

Ozturk

(2017) Is tourism a catalyst of growth in Egypt? Evidence from Granger non-causality and the generalised variance decomposition analysis. Anatolia 28(2): 173–181.

67.

Tugcu

(2014) Tourism and economic growth nexus revisited: a panel causality analysis for the case of the Mediterranean region. Tourism Management 42: 207–212.

68.

Watson

Deller

(2022) Tourism and economic resilience. Tourism Economics 28(5): 1193–1215.

69.

World Travel and Tourism Council (WTTC) (2022) Travel and Tourism Economic Impact Global Trends. London, UK: World Travel and Tourism Council. Oxfordeconomics.com, dowloadable at. https://wttc.org/research/economic-impact

70.

Yang

Fik

(2014) Spatial effects in regional tourism growth. Annals of Tourism Research 46: 144–162.

71.

Yang

Wong

(2012) A spatial econometric approach to model spillover effects in tourism flows. Journal of Travel Research 51(6): 768–778.