Questioning the Role of Tourism as an Engine for Resilience: The Role of Accessibility and Economic Performance

Abstract

During the last decades, tourism activities were promoted by local and regional policy-makers as a universal solution for rural areas’ suffering from massive depopulation, technological delay, or economical struggles. A large debate flourished in the literature on whether and to what extent tourism could play a role in supporting rural localities. Some valid evidence was brought by researchers backing the cure-all role of tourism, as well as by those who criticized the limited, or even negative, impact of tourism on rural areas. However, following the economic crisis of 2008, the attention switched to a newer and more relevant topic: Does tourism increase the resilience performance of rural areas? Our paper tries to answer this question by focusing on both economic and demographic resilience, which are the most sensitive sectors during an important shock. Following a detailed territorial breakdown according to a twofold typology (spatial accessibility and number of employees), correlations were used to determine the impact of tourism activities on building resilience for each type of rural territory. The results indicated a positive effect of tourism activities upon economic and demographic resilience performance in highly accessible rural areas, while in peripheral areas, the impact was insignificant. This paper provides new insights into the various roles that tourism plays in rural areas and offers suggestions for local policy-makers.

1. Introduction

Tourism represents one of the most dynamic economic branches of the last four decades, and its role in local and regional economies has constantly grown, reaching over 1.4 billion arrivals worldwide and generating over 10% of global GDP before the outbreak of the COVID-19 pandemic [1]. While it was initially considered a predominantly urban activity, both in terms of origin and destination of tourist fluxes [2], it spread rapidly in rural areas, to the point of becoming a central point in rural development strategies [3,4]. The diffusion of tourism activities was followed by an increase of interest in the long-term effects they had on destinations. With respect to rural tourism, a considerable amount of literature has analyzed the role of tourism in supporting the social and economic development of rural areas and found rather beneficial effects of tourism activities [5,6,7,8,9,10].

Overall, the positive effect of tourism on rural areas is dominated by two major segments. First, there is the positive economic impact provided by job creation. Tourism was found to be a constant job generator in rural areas [11], consequently inducing economic diversification [12,13], increasing the average revenues [11], revitalizing rural localities with weakening agricultural activities [7], resuscitating rural localities in peripheral regions [14,15,16], and producing an overall economic growth [17,18,19]. The ease of creating new jobs is due to the lack of high standards in comparison with other economic sectors, which makes tourism very attractive, especially for the younger population [20], as well as a cheaper reconversion of human resources from other sectors (such as agriculture, low industry, or craftsmanship) [21]. The economic impact is by far the most scrutinized, with considerable positive evidence from studies in Europe, America, and, more recently, Asia [22,23].

Second, the most acknowledged social improvement produced by tourism in rural areas is the reduction of outgoing migration, with positive examples in both Northern [20,21] and Southern [20,24] peripheral areas being recorded in the literature. Identified either through a demographic stabilization in declining localities [20] or backed by a clear reduction of the outgoing migration [21,25], the main impact of tourism upon rural communities tends to be demographical. The reduction of outgoing migration was accompanied by several social improvements, such as amelioration of rural female employment [11], improvement of residents’ perceived well-being [11,26,27], increase of overall quality of life for the residents [28,29], and higher motivation of residents in preserving and promoting natural and cultural heritage [30,31,32].

Unsurprisingly, local and regional stakeholders, as well as decisional actors, considered tourism as the main engine for the regeneration of peripheral rural localities, thus assigning considerable resources for carrying out tourism-centric strategies [3,4]. However, this enthusiasm was reduced recently by the criticism and skepticism of some scholars who signaled the poor connection between tourism and sustainable development [13,24,33,34,35].

The skepticism was due to studies reporting that tourism could induce a few negative socio-economic effects as well. Sometimes, the positive economic outcomes of tourism are reported to be only in the short term [24] and to manifest only in more developed economies, while in poorer rural areas, tourism could trigger an unbalanced income distribution [34,36], leading, in the long term, to social conflicts between hosts and guests [35,37]. A study focused on Eastern Europe found that the tourism-induced effects in rural areas were often overestimated [38], supporting the recent idea that tourism introduction in rural areas does not automatically lead towards success [13]. Some authors suggested that tourism had a positive—however, very limited—impact, which was mainly dependent on the local economic development level and accessibility index [11,39], while Eusébio, Kastenholz, and Breda [40] suggested straightforwardly that there was no significant economic contribution. It is reasonable to affirm that tourism development induces both opportunities and challenges for rural destinations, which calls for a more thorough analysis of its impact on rural communities.

Therefore, it is no surprise that tourism studies are shifting their focus from tourism-induced development, considered as rather controversial nowadays, to notions more strongly related to communities, sustainability, and well-being. During the last decades, a new concept, based on a notion that gained popularity in social sciences, joined the research and policy discourse: Resilience. In terms of regional development, resilience is seen as the capacity of a system to bounce back following a shock by either returning to its original state, adapting to a different state, or creating a new socio-economic equilibrium state [41,42,43]. The concept of resilience offers new and challenging perspectives on local and regional development, becoming a major objective in many tourism strategies due to its capacity to build a bridge between community needs, competitiveness, and sustainable tourism development [44,45]. In tourism studies, resilience is considered to be either an important part of sustainable development [46] or a better alternative—more community-focused—to the “traditional” paradigm of sustainable development [47]. In any case, the resilience concept provides a solid tool to investigate the impact of tourism on communities [48].

The first link between tourism and resilience appeared during the 1990s [49]. However, it should be mentioned that, until the 2000s, tourism resilience was analyzed almost exclusively in relation to the financial and environmental shocks, with only a limited emphasis on social components. However, the dangers and reactions triggered by overtourism in popular destinations in Europe and all over the world [35] engaged the scientific discourses with more relevant subjects for local communities, and put tourism-induced resilience into the frame of sustainability [46]. Given tourism’s high vulnerability towards financial, military, or socio-economic shocks [50,51,52,53], which makes it susceptible to quality and structural changes, the study of economic and societal behavior during and after a shock becomes a necessity. Resilience-based analysis provides a useful approach to understanding how a specific shock impacts the destinations, the economic development, and the local communities [54,55]. It also demonstrates a significant applicability for rural areas due to their higher inherent vulnerabilities towards shocks [56]. Resilience analysis, alongside system dynamics modeling, provides a wider and more coherent vision by taking into account uncertainty, non-linearity, and unexpected changes [27,57]. Moreover, for community-based tourism, it provides a useful guide for decision-makers in supporting sustainable development [58], given that the rapidly changing socio-economic environment calls for resilient communities in order to preserve their subsistence [59].

As a result, the most recent research on tourism risks and disaster management specifically calls for a better understanding of tourism resilience (and the factors that influence it), as well as tourism-induced resilience [60,61]. In the light of the most recent and most impactful shock to affect the tourism industry (COVID-19), the resilience of tourism destinations is brought again into discussion in various forms [61,62]. However, despite the attention given to the dynamic and unpredictable nature of tourism activities [33,63,64], only a limited amount of literature has focused on tourism-induced resilience. It is widely accepted that tourism increases the regional resilience capacity through economic diversification [12,13]; however, this usually applies only for the first stages of tourism implementation. When tourism monopolizes the destination’s economy, it reduces the resilience capacity by inducing mono-specialization and creating a huge dependency on a sole and highly vulnerable activity [50,51].

Our paper addresses the particular issue of tourism-induced resilience by identifying the main impact that tourism activities had on enhancing resilience performance in rural destinations. To this end, we took into consideration the relation between tourist arrivals and economic resilience (measured through the dynamics of the number of employees), as well as demographic resilience (measured through the demographic fluctuations).

For the next part, the paper will be structured as follows: First, the method used for the statistical approach and the index construction will be presented; then, a section comprising the main results of our research will follow, and, finally, the last part will be entirely dedicated to offering explanations for the results as well as to discussing the implications for policy-makers and academics.

2. Materials and Methods

In order to identify the resilience performance, we had to analyze the recovery behavior of destinations in relation with a shock. We selected the economic crisis of 2008 as the most recent, wide-spread, and statistically measurable shock. The area of study is represented by Romanian rural localities. We used the smallest statistical administrative units (LAU-2) in order to have a clearer image of tourism-induced resilience at the local level.

The choice of Romanian rural areas was motivated by two main reasons. First, the impact of the economic crisis upon tourist arrivals is easily identifiable in Romanian rural areas, as it represented the only cause for the reduction of tourism flows between 2008 and 2010, therefore avoiding shock overlaps. Second, as in many other Eastern European countries, rural tourism received increased recognition in Romania as a sustainable development tool for rural economies and communities [6], with national strategies encouraging its spread. Therefore, the choice of Romanian rural areas for testing the resilience performance of rural destinations offers the advantage of easily isolating the aftershock recovery and putting it in relation with the demographic and economic behavior of the locality.

Given that tourism impacts are highly dependent on the local economic development and accessibility index [11,39,65,66], rural LAU-2s of Romania areas were classified according to three criteria: (i) Volume of tourism flows (significant vs. non-significant/non-existent), (ii) economic performance, and (iii) spatial accessibility to urban areas. Economic performance and spatial accessibility were computed as composite indexes.

• Volume of tourism flows (significant vs. non-significant/non-existent): By rural tourism destination, we understand a destination in a rural area receiving tourists during each year of the study period (2008–2014). This choice was made in order to demarcate the destinations where tourism is a constant economic activity, and not an ephemeral one.
• Economic performance: Economic performance was calculated as the average of the z-scores for the number of firms per 1000 inhabitants and the turnover per inhabitant. LAU-2s displaying average z-scores above 0 were classified as being “Performant”, whereas the rest were classified as being “Under-performant”.
• Spatial accessibility to urban areas: Spatial accessibilities were computed as road travel times from each rural LAU-2 barycenter to the closest second-tier (generally above 200,000 inhabitants) and third-tier (above 25,000 inhabitants) urban centers, respectively. Travel times were calculated as follows:

Acc = TTR3 + TTR3*TTR2,

(1)

where Acc = accessibility level, TTR3 = travel time to the closest third-tier urban center, and TTR2 = travel time to the closest second-tier urban center. Finally, accessibility levels were converted into z-scores: LAU-2s displaying negative values were classified as “Central”, whereas the rest fall into the “Peripheral” category.

As a consequence, each of the criteria splits the rural destinations into two classes: Destination vs. Non-Destination; Performant vs. Under-performant; Central vs. Peripheral. A total of eight classes resulted by crossing the six categories resulting from the above-mentioned classifications (see Figure 1 and

Table 1. Average number of tourist arrivals and Defert index for all types of rural localities.

	Average no of Arrivals (2008–2014)	Average Population (2008–2014)	Defert Index * (%)
Central and Performant—0	76,663	1,807,626	4.2
Central and Performant—1	522,798	604,402	86.5
Peripheral and Performant—0	20,763	345,940	6.0
Peripheral and Performant—1	176,991	166,035	106.6
Central and Under-performant—0	46,623	3,856,133	1.2
Central and Under-performant—1	94,991	237,620	40.0
Peripheral and Under-performant—0	12,253	2,506,200	0.5
Peripheral and Under-performant—1	97,313	190,959	51.0

* Defert index = [Average no. of tourist arrivals]*100/[Average annual population]; 0 = Non-significant tourism destinations and non-touristic areas (not included in the analysis); 1 = Tourism destinations.

).

Two main variables were employed in our study: “Change in the number of employees at the LAU-2 level” as a proxy for rural economic resilience and “population change at the LAU-2 level” as a proxy for rural demographic resilience. Both variables were computed for two distinctive time frames: 2008–2010 to capture the economic/demographic resistance following the 2008 shock, and 2010–2014 to capture the economic demographic recoverability. “Number of employees” reflects “all persons with an individual labor contract/agreement for a definite or indefinite duration (including seasonal workers, the manager, or the administrator) for all sectors of activity” [67].

It is important to keep in mind that the geographical distribution of the employees was registered at the LAU-2 level by taking into account the locality in which the employees carry out their activities. As a consequence, they constitute a highly relevant proxy for the local economic performance of the LAU-2 they are attributed to. “Population change” was computed using the permanent resident population, based on the person’s address in the LAU-2 “where he/she declares to have the main dwelling, printed as such on their identity card and registered by the administrative bodies of the state” [68]. Though not as reliable as the census or the resident population, it is the only and largest detailed population database for local administrative units on a year-by-year basis.

In order to achieve our aim, we employed descriptive statistics and correlations. All analyses and data processing were carried out in SPSS v. 22 (IBM Corp. (2013); IBM SPSS Statistics for Windows, Version 22.0, Armonk, NY: IBM Corp). The first part of the results stems from descriptive statistics, where tourism destinations were compared to non-tourism destinations on all the indexes of interest (employees and population). This allowed us to zoom further on tourism destinations only by analyzing the link between tourist arrivals and employees/population for the years during (2008–2010) and after crisis (2010–2014), detailed by accessibility and economic performance.

In order to limit the variations in tourist arrivals, we decided to employ a simple three-year moving average [(X_t-1 + X_t + X_t+1)/3]. For example, tourist arrivals for 2008 were computed as the unweighted mean of arrivals from 2007, 2008, and 2009. The choice for data normalization was motivated by the dynamic and unpredictable nature of tourism destinations [33,63,64]. Moving averages allow the smoothing of short-term fluctuations, highlight long-term trends, and are frequently used in economics and marketing [69].

3. Results

Our analysis was carried out for the resistance period (2008–2010), the main period of shock manifestation when most of the tourism destinations recorded a significant drop in tourist arrivals, as well as for the recovery period (2010–2014), when the destinations displayed significant growth in tourist arrivals. While the economic crisis manifested for a longer period (until 2012 or even 2016 in some European regions), the tourism manifestation of the crisis manifested only until 2010. This is not surprising, as previous studies showed that tourist arrivals in European countries displayed higher levels of resilience compared to other economic sectors following the 2008 financial crisis [70,71,72].

For our first analyses, we compared rural tourism destinations (as described above) with rural LAU-2s that shared the same economic performance and accessibility typology that displayed tourist arrivals only sporadically or never, the aim being to observe the differences in numbers of employees and population during the resistance and recovery periods.

The first results (Figure 2, left) suggest no major difference between tourism destinations and non-destinations in terms of labor market resistance to and recovery from economic crisis. The trajectories mainly depend on the economic profile and geographic accessibility of the rural community and less on the existence (or not) of significant tourism activities. Thus, irrespective of the existence or not of significant tourism activities, Central and Performant localities display low resistance but very high recoverability rates, whilst the rest of the rural areas are generally faced with low resistance and low recoverability alike. It is surprising and somehow counter-intuitive that Peripheral rural localities, both Performant and Under-performant, display better recovery rates if they are not tourism destinations. It seems that tourism activities negatively affect the economic resilience of rural areas. On the other hand, the right-hand side of Figure 2 suggests that local communities (LAU-2s) that are tourism destinations report better demographic trajectories (population change) compared to non-significant/non-touristic local communities, and this holds for each of the four types of territories studied. Moreover, while the differences are only minor for the Peripheral areas, they seem very strong for the Central ones.

However, these results do not fully explain the relationship between tourist arrivals and economic or demographic resilience; therefore, a more in-depth analysis was performed specifically for the tourism destinations. The results—which, from this point onwards, will reference only tourism destinations—indicate, as seen in

Table 2. Pearson correlations between tourist arrivals and number of employees, as well as population, according to the accessibility-based typology.

	Type of Territory	Number of Employees		Population
		2008–2010	2010–2014	2008–2010	2010–2014
Tourist arrivals	Accessibility-based typology
	Central	0.32 **	0.26 **	0.19 **	0.18 **
	Peripheral	0.02	−0.01	−0.16 **	−0.13 **

** p < 0.00; * p < 0.05.

,

Table 3. Pearson correlations between tourist arrivals and number of employees, as well as population, according to the economic-performance-based typology.

	Type of Territory	Number of Employees		Population
		2008–2010	2010–2014	2008–2010	2010–2014
Tourist arrivals	Economic-performance-based typology
	Performant	0.26 **	0.22 **	0.12 *	0.15 **
	Under performant	0.13 *	0.11 **	0.01	0.03

** p < 0.00; * p < 0.05.

and

Table 4. Correlations between tourist arrivals and number of employees, as well as population, according to the detailed typology.

	Type of Territory	Number of Employees		Population
		2008–2010	2010–2014	2008–2010	2010–2014
Tourist arrivals	Detailed typology
	Performant and Central	0.30 **	0.22 **	0.19 **	0.17 **
	Performant and Peripheral	−0.07	−0.10	−0.24 **	−0.20 **
	Under-performant and Central	0.17 *	0.18 **	0.02	0.07
	Under-performant and Peripheral	0.03	0.04	−0.05	−0.02

** p < 0.00; * p < 0.05.

, that there are significant links between tourist arrivals and employees, as well as population. Table 2 reports the Pearson correlations identified when analyzing tourism destinations by accessibility. It can be noticed that tourist arrivals correlate positively with the number of employees during the resistance and the recovery periods, suggesting that the higher number of tourist arrivals, the higher the number of employees, but also that the drop in tourists during the resistance period caused a drop in employees as well. However, this is valid only for Central tourism destinations, and not for Peripheral ones. Regarding the correlations between tourist arrivals and population, they are significant for both Central and Peripheral rural localities, but are smaller. Nonetheless, it is interesting to notice that the correlations are positive for Central destinations while being negative for Peripheral destinations. These results suggest that, in Central areas, the high values of tourist arrivals go hand in hand with high values of population, while in Peripheral areas, the higher the arrivals, the lower the population values.

Table 3 reports the Pearson correlations identified when analyzing tourism destinations by economic performance. It can be noticed that tourist arrivals correlate positively with the number of employees during the resistance and recovery period for both Performant and Under-performant destinations, while the correlations between tourist arrivals and population are significant only for Performant destinations, suggesting a relation between tourism and demographic resilience; however, this is only when the rural destinations are also performant from an economic point of view. It is interesting to notice that, for the destinations that are economically performant, the link between the tourist arrivals and employees is stronger than for the destinations that are economically under-performant.

Table 4 analyzes the link between tourist arrivals and the employees, as well as the population, for the four crossed typologies of tourism destinations. Two major connections are revealed by this table: (a) It can be noticed that for Central destinations, the correlation between tourist arrivals and employees remains significant during both resistance and recovery periods, regardless of their economic performance, while for Peripheral destinations, the same correlations fail to be significant, regardless of their economic performance. (b) A similar pattern emerges when analyzing the correlation between tourist numbers and population; however, this time, Performant destinations, regardless of their accessibility, manage to display significant values (although Performant and Peripheral destinations display negative relations, which will be tackled in the Discussions section), while both types of Under-performant destinations show no significant relation.

4. Discussion

Our paper analyzed the relation between tourist arrivals and economic resilience (the number of employees), as well as demographic resilience (fluctuations in total population), in order to isolate the impact that tourism activities had on enhancing resilience performance in rural destinations. The results open the field for discussions and explanations regarding the relation between tourism and resilience.

4.1. Tourism-Induced Economic Resilience

Statistical analysis suggests one main result concerning the relationship between tourist arrivals and local labor markets. This result implies that tourism’s impact on the number of employees varies, especially according to the territorial (“Peripheral”/”Central”) context. Whether the analysis was carried out only according to the accessibility-based typology or according to a detailed typology, statistically significant relationships at different levels of intensity were found for Central destinations. However, despite differences in correlation intensity, one may notice that correlations are positive in all territorial contexts displaying statistically significant values, thus indicating that the dynamics of tourist arrivals and the change in the local labor market tend to follow similar trajectories, no matter the period. The positive relationship suggests that an increasing/decreasing number of tourist arrivals significantly contributes to increasing/decreasing the size of the local labor market. It can also suggest that a dynamic local labor market might trigger increasing tourism business, and, therefore, arrivals; however, given that tourism managed to display quicker recoveries than any other economic sector [71,72], it is safe to assume that the relation usually goes from tourism towards other sectors. Alternatively, maybe there is a more complex relationship between the two, mediated by a third party, which requires further investigation in order to determine the causality direction.

Complementarily, no relation was found between tourist arrivals and the size of the local labor market in peripheral areas of Romania. Such territorial variations of tourism’s impact go against a prevailing perception stating that tourism contributes to reducing territorial inequalities [11,14,70]. It is worth mentioning that while significant relationships between tourist arrivals and economic resilience manifest for both Performant and Under-performant areas, the relationship is always stronger for the Performant areas. This could be an argument for the conditionality of tourism’s positive impact; this conditionality is given by pre-existent services, as well as a diversified job market, in order to manage to use tourism as leverage for economic resilience.

A secondary and negative relationship seems to emerge from these results as well. While tourist arrivals proved be a significant determinant of resilience for rural communities, especially following a major economic crisis and with a visible effect through the relations existent during the recovery period, they also position themselves as a source of vulnerability, given the significant relation during the resistance period. The drop in tourist numbers during the crisis was connected with the drop in employees in the same rural destinations. Our paper supports findings from previous studies pointing to a more complex impact of tourism on territorial development imbalances [73], and this might be particularly true during and following periods of crisis. The higher correlation coefficients during the resistance period compared to the recovery period suggest that tourism also has a destabilizing effect on the local labor market during the first stages of an economic crisis. Tourism creates a vulnerability, which manifests rather quickly. However, tourism also triggers the response to this vulnerability by creating and inducing higher resilience performance during the recovery period. It is still worth mentioning the tendency of tourism-induced resilience to manifest only in more developed areas, an issue already mentioned in literature [34,36].

4.2. Tourism and Local Demographic Trends

For rural Romanian localities, outgoing migration represented a growing issue during the last decades, as most of them recorded drops in the total population numbers. Some localities situated in peripheral regions registered extremely high values in population decline between 1992 and 2018, with a record of 61.2% in Gogosu, Dolj county (in the Under-performant and Peripheral class in our analysis) [68]. Tourism was seen by local and regional actors as a method of demographic stabilization for the depopulated areas, following positive reports from similar areas [24].

Our study found that statistical relations between tourist arrivals and demographics are more diverse compared to the case previously discussed. This is mainly due to the fact that both positive and negative significant correlations were found between number of arrivals and LAU-2s’ demographic size. The results should be analyzed from a twofold perspective: First, the accessibility-based typology shows that in Central rural destinations, the arrivals were positively correlated with demographics for both periods of resistance and recovery. This result could be indicative of the fact that the growth in tourist numbers goes hand in hand with the increase in population (the same applies for negative dynamics following a shock as well). However, the results indicated that in Peripheral destinations, these relations, while still significant, were reversed. Second, the economic-performance-based typology indicated an existent and significant relation only for Performant rural destinations. Running the same set of analyses for a more detailed typology of localities (Performant and Central, Performant and Peripheral, Under-performant and Central, Under-performant and Peripheral) leads to stronger and more detailed correlations.

As counterintuitive as it may seem, such results are, in fact, expectable. Tourism arrivals and population size go hand in hand in dynamic areas (Central and Performant), which attract jobs, in-going migration, and also tourist arrivals. On the contrary, Peripheral and Performant areas report a negative correlation, which may be explained through two different approaches: (a) The increasing number of tourists is leading to a depopulation of the peripheral areas, which are economically performant (the development of tourism activities can produce shifts in local populations and labor force demand [24]), and (b) rapidly depopulating rural destinations registered a growth in tourist arrivals (phenomenon widely recorded in the 1990s in peripheral areas [14,16]). It is difficult to assert with a high degree of certainty which of the two scenarios is causing the negative correlations due to the complex nature of tourism activities, and even a mix of both explanations is highly possible. In order to have a clearer image of the relationships between tourists and population in the Peripheral and Performant category, a case study approach is required, which amply overcomes the aim of this study.

With more acute economic problems and out-migration comes a higher relative role of tourism in community development (those areas are still Performant; hence, we must assume a policy-based reaction). The fact that correlations prove to be stronger when more detailed territorial typologies are taken into account points out that the territorial contextualization of the tourism–demography nexus should be taken more seriously into account. At least in some cases, it may lead to more nuanced results. The negative performance of peripheral areas could be linked to the lack of important entrepreneurial and tourism skills [16,74], or the lack of good administrative skills from the local authorities [16]. These results support the strategy of tapping into and creating broad networks in rural areas in order to connect communities from destinations and ensuring, in time, local benefits from tourism development.

5. Conclusions

The overall aim of this study was to identify the role of tourism activities on resilience enhancement in rural areas. The case of Romanian rural LAU-2 localities was presented and analyzed using Pearson correlations and a twofold territorial typology. While previous studies focused on the various direct economic effects induced by tourism activities, our research represents one of the few approaches on this topic, especially for Central and Eastern Europe. The study identified relationships previously unknown between tourism and the behavior of rural localities during the resistance or recovery period following a shock.

First, we found that rural tourism destinations managed to display more positive demographic evolutions than rural non-tourism localities, and that the evolutions were better when the localities were economically performant and central. However, the relationship fails to manifest when tourism and non-tourism localities were compared regarding economic behavior. Second, when the analysis was carried out only for tourism destinations, significant evidence of tourism-induced resilience performance was observed, but rather for central and economically performant destinations.

The paper brings new empirical evidence regarding the complex role that tourism can have in shaping the behavior of rural localities, especially in times of crisis or shortly after. Such empirical evidence has theoretical implications for the development of the “destination resilience” construct, mainly by showing that the role of tourism as resilience enhancer can be activated only when the economic context and the geographical accessibility allow it. While tourism is acknowledged to be resilient [71,72], it cannot generate resilience performance by itself, which can be later transferred towards the entire economic system of rural destinations. However, further research is needed to see whether tourism triggers the economic recovery or vice-versa.

The current study has several limitations. First, although the demographic microdata employed were most the most detailed and reliable available data, they capture neither temporary migration nor temporary residence changes. Further research could target in-depth investigations aimed at examining local demographic changes on a case study basis. Second, changes in employment numbers have been examined only globally, as a sum of all sectors of activity. Upcoming studies could question local employment changes induced by tourism by looking at spillover effects on each economic branch separately.

For local and regional policy-makers, the results should provide food for thought in order to build multisectorial economic strategies, connecting tourism activities with existing local activities, such as agriculture or craftmanship. These results have practical implications on building destination tourism strategies by showing that tourism-induced resilience is context dependent. As such, strategies for peripheral destinations should plan for resilience outcomes that are different from the ones that are conceived for centrally located destinations.