In Search of a Social Optimum in Tourism Flows: The Role of Cultural Heritage in Italian Municipalities

Abstract

Tourism represents a contributor to local economic conditions as well as to fund raising for heritage preservation. However, concerns related to excessive flows of tourists have been raised from many parts. In fact, overtourism can lead to social, cultural, environmental, and economic costs. This is especially true in cultural destinations. Therefore, touristic destinations both benefit from tourism-related advantages and are subject to the associated drawbacks. This work aims at investigating the sustainability of tourism flows in Italian municipalities from the residents’ perspective by jointly analyzing localization benefits and costs for residents -among which tourism is both a cost and a benefit. Particularly, the work uses a neoclassical economic approach to look for a social optimum in tourism flows, defined as the level that would guarantee the optimal allocation of tourists among Italian municipalities. This enables us to identify municipalities encountering overtourism or undertourism, according to our model. The research is deepened by considering local specificities in terms of heritage endowment and cultural touristic vocation.

Introduction

It is quite well established that tourism represents a contributor to local economic conditions (see, among others, Balaguer & Cantavella-Jordá, 2002; Brida & Pulina, 2010; Faber & Gaubert, 2019; Hohl & Tisdell, 1995; Kadiyali & Kosová, 2013; Liu & Song, 2017). Particularly, in cultural destinations tourism is also a resource to support cultural activities and material heritage maintenance and preservation.

However, concerns related to excessive flows of tourists have been raised from many parts. In fact, over-tourism might generate social, cultural, environmental, and economic costs, mainly affecting and damaging local communities and therefore becoming unsustainable. This may be especially the case in cultural destinations, where, among other inconveniences, overtourism could damage cultural heritage (see, among others, Adie et al., 2020; De Luca et al., 2020; Neuts & Nijkamp, 2012; Popp, 2012; Rasoolimanesh et al., 2019). In this respect, the overcrowded situation of cities like Venice has sadly become extremely famous and has also been the object of some relevant literature (Russo, 2002).

Therefore, tourism represents both a benefit and a cost in economic terms for destinations.

Regarding this issue, the current debate has developed around the well-known (but difficult to define) concepts of overtourism, carrying capacity, and sustainability of tourism flows (see, for instance, Peeters et al., 2018).

The present work enters this discussion adopting an economic reasoning and proposing a novel methodology for identifying an equilibrium level in tourism flows. Specifically, we look for the optimal distribution of tourists across Italian municipalities that guarantees the balance between economic benefits and costs. We do so by equating residents’ localization benefits and residents’ localization costs associated with one additional tourist—with tourism being considered both as a localization benefit and cost—within a neo-classical framework. This eventually allows us to identify the social optimum for the whole system. Specifically, the term social implies that the optimum refers to the entire system of municipalities, that is, it represents the optimal allocation of tourists among Italian municipalities from a resident perspective. We distinguish the areas according to their (positive or negative) distance from such an optimal level of tourism flows, therefore identifying overtourism or undertourism situations, and we investigate if and how this is linked to the presence of cultural heritage.

Thus, this paper contributes to the existing knowledge in mainly two ways: firstly, by providing a theoretical and conceptual framework that allows one to look at tourism as simultaneously an economic localization benefit and cost; secondly, by applying such an approach to a granular territorial level, that is, Italian municipalities.

To do so, the work is structured as follows: the next section revises the existing literature on tourism, overtourism and carrying capacity with a particular focus on heritage destinations (Section 2); subsequently, the theoretical framework to define the social optimum in tourism flows is explained (Section 3) and the empirical application is presented (Section 4). Finally, the results are displayed and discussed (Section 5) and the conclusions are proposed (Section 6).

Tourism and Local Development: The Risk of Overtourism and the Role of Cultural Heritage

Tourism has been widely described as an engine for growth, for instance in terms of employment creation (Hohl & Tisdell, 1995; Kadiyali & Kosová, 2013; Liu & Song, 2017), generation of positive economies of scale and scope (Brida & Pulina, 2010), and export of non-tradable services, implying tourists flows rather than goods flows (Faber & Gaubert, 2019). In addition, tourists’ consumption generates inflows of foreign currency, contributing to the balance of payment through foreign exchange earnings (Seetanah, 2011) and higher production levels are found out when efficiency is obtained through competition between local firms and analogous activities in different tourist destinations (Dritsakis, 2004).

An interesting part of the relevant literature is founded on the tourism-led growth hypothesis, which was developed and analyzed by Balaguer and Cantavella-Jordá (2002). As outlined by Brida and Pulina (2010), this idea stems directly from the export-led growth hypothesis, adapted to tourism. According to this theory, exports drive economic growth by enhancing the efficiency of resource allocation and increasing production volumes. Additionally, exports lead to a rise in investment levels.

Overall, as highlighted by Nunkoo et al. (2020), who conducted a meta-analysis on more than 100 studies dedicated at exploring the tourism-led growth hypothesis, there is a general support for a causality nexus between tourism and economic development (Balaguer & Cantavella-Jordá, 2002; Cortés-Jiménez, 2008; Liu & Song, 2017), although the results are sensitive to various factors. In this respect, tourism may be considered as a benefit for the residents who decide to localize in a given place.

However, some important negative effects associated with tourism have been stressed, too. In fact, there may be significant social, cultural, environmental, and economic tourism-related costs, mainly affecting and damaging local communities (Adie et al., 2020; Biagi et al., 2012; Richards, 2017; Riganti & Nijkamp, 2008). ¹ Therefore, tourism is also a cost for the residents who decide to localize in a particular area.

Balancing these aspects is not easy, and in fact when the drawbacks overcome the advantages situations of so-called overtourism may arise, that is, “… situation(s) in which the impact of tourism, at certain times, and in certain locations, exceeds physical, ecological, social, economic, psychological, and/or political capacity threshold” (Peeters et al., 2018, p. 22). This can ultimately be a consequence of tourism strategies focused on increasing visitor numbers, as currently and commonly pursued worldwide (ibidem).

Even though overtourism is a relatively new term in the public and academic debate (Capocchi et al., 2019; Celata & Romano, 2022), the phenomenon is not a new one, as problematic forms of tourism crowding and their effects on local communities and environment have been studied for decades. Yet, there is much evidence that the character of tourism in many locations is changing rapidly and becoming an actual problem. However, the notion is accompanied by an inevitable degree of indeterminacy (Koens et al., 2018; McCool & Lime, 2001; UNWTO, 2018), which of course carries with it some evident measurement issues. In fact, there are no widely accepted methodologies for the quantification of thresholds and therefore of overtourism (Buitrago & Yñiguez, 2021).

By its very nature, the overtourism phenomenon is associated with tourist numbers, the type and time frame of their visit, and a destination’s carrying capacity (Peeters et al., 2018). Originally identified in association with environmental issues (Wagar, 1964, 1974), this is defined in the tourism domain by the (UNWTO, 1981) as the “maximum number of people that may visit a tourist destination at the same time, without causing destruction of the physical, economic and socio-cultural environment and an unacceptable decrease in the quality of visitors’ satisfaction” (see also the interesting contributions by O’Reilly, 1986; Saveriades, 2000).

In this respect, it is stimulating to note how sometimes the perspective considered is the one of tourists, especially in the first approaches to carrying capacity, while some other times the residents’ point of view prevails. In fact, carrying capacity can be seen as either the maximum number of visitors that can be accommodated keeping the quality of their experience unchanged, or as the maximum number of visitors that can be accommodated by a given destination under conditions of maximum stress (Canestrelli & Costa, 1991).

In any case, the concept remains subjective and idiosyncratic, as it might reflect the interest of some specific actors (e.g., real estate investors) at the expense of others (e.g., residents). Besides, with equal tourism flows, different territories might be characterized by different carrying capacities, depending on local specificities (Buitrago & Yñiguez, 2021; Gowreesunkar & Vo Thanh, 2020). Overall, carrying capacity is multidimensional as environmental, economic, psychological and perceptual factors need to be considered (Cocossis, 2002; Diedrich & García-Buades, 2009).

Finally, when specifically associated with heritage cities, carrying capacity may be defined as “the number of visitors an art city can absorb without hindrance of the other social and economic functions it performs” (Glasson, 1994). Focusing in particular on heritage destinations, numerous studies in scientific literature have emphasized the growing significance of cultural heritage in tourism. The UNWTO has acknowledged this trend by recognizing the concept of “cultural tourism” and defining it as “a type of tourism activity in which the visitor’s essential motivation is to learn, discover, experience and consume the tangible and intangible cultural attractions/products in a tourism destination” (UN Tourism General Assembly (2017).

However, it is quite easy to see how culture and tourism are characterized by a love-hate relationship (Russo, 2002). In fact, on the one hand, cultural heritage is a crucial and widespread tourism resource globally, making heritage tourism one of today’s most prominent forms of travel (Timothy, 2014). In this respect, numerous studies emphasize the growing significance of cultural heritage in tourism. Nevertheless, on the other hand, several socio-economic costs have been associated specifically with cultural tourism, also underlining lack of community involvement in planning and in the related development processes. In their famous paper van der Borg et al. (1996) explain indeed that heritage cities draw many visitors, bringing both benefits and costs. When costs outweigh benefits, tourism development becomes unsustainable, necessitating interventions. Analyzing six heritage cities in Europe, the authors find out that tourism threatens the vitality of local economies, the integrity of heritage, and residents’ quality of life. Therefore, urgent measures are needed to control and guide visitors’ flows. Other works highlight specifically how the mismatch between cultural (e.g., history, memory) and economic forces (e.g., tourism, trade) in a historic city can lead to the erosion of place identity (Boussaa, 2021) and how heritage tourism may result in disproportionate site exploitation, folklorization, museumification, or extensive (excessive?) construction of tourism venues (e.g., Gravari-Barbas, 2024).

The above-mentioned considerations suggest that while tourism serves as a significant economic resource for destinations, bringing economic benefits, it also comes with several drawbacks that result in socio-economic costs for places. The balance between these two faces of tourism is something that the literature is trying to analyze in order to find the sustainable level of tourist flows. In this sense, tourism destinations look for novel methods to try to assess and measure carrying capacity before overtourism occurs (Peeters et al., 2018).

An important strand of literature deals with this by analyzing tourism’s impact on local residents’ lives and perceptions (Sharpley, 2014). Relying on primary data, this literature measures social carrying capacity in terms of the perceived impacts of tourism on residents’ lives, with positive perceptions suggesting the limit has not been reached yet (e.g., Saveriades, 2000) or, in other terms, the benefits still exceed the costs. Expanding on this, Tokarchuck et al. (2017) and Tokarchuk et al. (2021) examine the tourism effects on well-being and life satisfaction, identifying a threshold where negative impacts outweigh benefits (i.e., an overtourism condition). Their findings show an inverted U-shaped relationship between tourism intensity and life satisfaction, with social carrying capacity corresponding to the parabola’s vertex.

While they focus on local residents’ perceptions and well-being related to tourism, we propose a different approach to determine the optimal level of tourism flows. Our analysis is grounded in a neoclassical economic reasoning, according to which tourism represents both a localization benefit - an economic advantage that attracts residents—and a localization cost—a deterrent due to its negative economic effects. We look at both positive and negative relevant elements (see also Cerisola & Panzera, 2024), identifying residents’ localization benefits and costs and looking for a social optimum in tourism flows, as explained in the next section. This approach allows us to apply a unique methodology to the whole system of Italian municipalities—without focusing on a single case study—, therefore identifying the optimal allocation of tourists among destinations.

The Theoretical Framework: Definition of the Social Optimum in Tourism Flows

In order to conceptually identify the social optimum in tourism flows, we resort to a neo-classical theoretical-conceptual framework consistent with what has been done within the optimal city size theory, on which we draw to build our reasoning (Alonso, 1971; Richardson, 1972, 1973). More specifically, the optimal city size theory aimed at determining the equilibrium in urban size and it was rooted in turn in the neoclassical stream of location choice models such as those of Von Thünen, Alonso, and Fujita. The idea behind the model is that the location choices of individuals are driven by utility maximization, achieved when marginal location costs equal marginal location benefits. Therefore, the equilibrium in urban size is determined by the balance between localization benefits and localization costs. In Camagni et al. (2013), this depends on a range of tangible and intangible factors besides the city size, which itself represents both a benefit and a cost. In their work, the optimal city size is therefore determined by solving the equation that balances marginal benefits and marginal costs of location for an additional inhabitant.

In this study, similarly to the optimal city size theory, we start with the idea that individuals balance benefits and costs in their location choices. We then substitute urban size with tourism flows. In fact, tourism can be considered both a benefit and a cost in location decisions. Following the reasoning put forward in Section 2 and proposed by the literature, tourism flows generate both positive and negative economic externalities for city residents, making tourism both a benefit, and a cost for local inhabitants. We therefore maximize residents’ net benefits by equating marginal benefits (which include tourism flows) and marginal costs (which also include tourism flows). Finally, we solve this equation with respect to tourism to establish the equilibrium level of tourism flows.

In more detail, in our approach the social optimum in tourism flows is reached when the localization benefits acquired by the residents through one additional tourist equal the associated localization costs. Such an ideal level is the “social” optimum since it implies the best situation for the system as a whole (aggregate level). This condition should be assumed as a target by a government interested in the economic sustainability of tourism flows from a socio-economic point of view. In this respect, it is interesting to note that van der Borg et al. (1996) highlight the need to forsake laissez faire (see also Agyeiwaah, 2020).

Since tourism is obviously not the only benefit or cost of location for residents, we resorted to the urban economics literature (see also Camagni et al., 2013) to identify other common residents’ localization benefits (B) and residents’ localization costs (C) as functions of different elements, as follows:

• (1a) B = f(tourism, cultural heritage, agglomeration economies, amenities, diversity, human capital)

• (2a) C = f(tourism, congestion, malaise, environmental damage, housing prices)

Equation (1a) devises residents’ localization benefits as a function of:

tourism, which is a localization benefit for the residents through job creation and increased (tax) revenues. In addition, many tourism products can also be enjoyed by residents -i.e., festivals, restaurants, natural and cultural attractions, outdoor recreation opportunities (Boley et al., 2018);

cultural heritage, which is a localization benefit in terms of the related value chain (ESPON, 2019), of access to beauty and inspiration (Cerisola, 2019a, 2019b), and of well-being (Perucca, 2019);

agglomeration economies represent the well-known advantages attainable through a concentrated localization (Camagni, 1993; Hoover, 1937; Isard, 1956);

amenities are a localization benefit, in the form of intangible advantages, such as accessibility to high-quality public services -schools, hospitals; to a variety of recreational services -theaters, cinemas; to high education services – universities; to cultural capital – museum and historical monuments (Carlino & Saiz, 2019; Cheshire & Magrini, 2006; Rappaport, 2007);

diversity is a localization benefit in terms of creativity (Cohen & Levinthal, 1990; Jacobs, 1961; Landry, 2008);

human capital is a localization benefit because it provides a more sophisticated, innovative, and stimulating environment (Camagni et al., 2013).

Equation (2a), instead, describes residents’ localization costs as a function of:

tourism, which is a localization cost for the residents since it generates congestion in terms of crowding, traffic and parking problems (Jacobsen et al., 2019), incentivizes petty crime (e.g., Biagi et al., 2012; Biagi & Detotto, 2014; King et al., 1993; Montolio & Planells-Struse, 2016), and increases the cost of living (Celata & Romano, 2022). In addition, it can produce friction between tourists and residents, and changes in residents’ way of life (Diaz-Parra & Jover, 2021);

congestion is a localization cost through disadvantages such as traffic and difficulties in the management of waste and other public services (Camagni et al., 1986);

malaise is a localization cost intended in terms of social conflict (e.g., Biagi & Detotto, 2014; Boakye, 2010; Ryan, 1993);

environmental damage is a localization cost in terms of pollution for example, (Apergis & Payne, 2012; Holden, 2009; Kousis, 2000);

housing prices is a localization cost in terms of «cost of the city» (Jeanty et al., 2010).

Particularly important is that, as explained above, tourism is at the same time a localization benefit and a localization cost for the residents. Our conceptual and empirical reasoning takes advantage of this specific feature, as described in the following.

It is worth noticing that the variables other than tourism included in our model do not exhibit a specific relationship with tourism itself—this is intentional. These variables were deliberately chosen because they represent either localization benefits or costs for residents rather than tourists. In fact, these variables represent location advantages or disadvantages for urban inhabitants, and they are not meant to be indicators of overtourism or undertourism.

Equations (1b) and (2b) make explicit residents’ localization benefits and costs in a way that can be used to identify the shapes of the benefits and costs curves through their first and second derivatives with respect to tourism, according to a neo-classical framework. Equations (1c) and (1d), in fact, show how the benefits curve is expected to be concave, while equations (2c) and (2d) reveal a convex costs curve.

• (1b) B = tourism^ζ cultural heritage ^νagglomeration economies^ηamenities^Φdiversity^μhuman capital^π

• (2b) C = tourism^αcongestion^ιmalaise^βenvironmental damage^δrent^γ

• Marginal benefits

• (1c) δB/δ tourism = ζ tourism^ζ−1 cultural heritage ^νagglomeration economies^ηamenities^Φdiversity^μhuman capital^π > 0

• (1d) δ”B/δ”tourism = ζ (ζ−1) tourism^ζ−2 cultural heritage ^νagglomeration economies^ηamenities^Φdiversity^μhuman capital^π < 0

• Marginal costs

• (2c) δC/δ tourism = α tourism^α−1congestion^ιmalaise^βenvironmental damage^δhousing prices^γ > 0

• (2d) δ”C/δ”tourism = α(α−1)tourism^α−2congestion^ιmalaise^βenvironmental damage^δhousing prices^γ >0

To find the optimal level of tourism flows, we maximize residents’ net benefits by equating marginal benefits and marginal costs, that is, residents’ localization benefits, and residents’ localization costs of one additional tourist. ² Then, we solve the equation with respect to tourism (the mathematical derivation is reported in Appendix 1). The same mathematical derivation will lead to the equation to be econometrically estimated, as explained in the next section.

The theoretical framework is represented in Figure 1, which highlights the point where the marginal localization benefits (MLB) of one additional tourist equal the corresponding marginal localization costs (MLC) for the residents (panel a). A point where MLB exceed MLC represents a situation where the arrival of one additional tourist generates higher benefits than costs to residents (panel b), and therefore the increase in touristic flows is still advantageous for the inhabitants. Instead, in a situation where MLC are higher than MLB, an increase in the tourism flows would create economic damage to residents (panel c), as the additional costs implied by the arrival of one more tourist would exceed the associated benefits. The optimal situation for residents would be the point where MLB equal MLC (i.e., where net benefits are maximized). This situation is not automatically reached via market forces, since it mostly depends on tourists’ individual choices. Rather, reaching the social optimum would theoretically require the existence of a forward-looking decision-maker able to plan touristic flows for the best of the whole economic system. According to this theoretical framework we define overtourism as a situation in which MLC exceed MLB, and undertourism a situation in which MLB exceed MLC, as indicated in the following figure (Figure 1).

Figure 1.

Social optimum in touristic flows for the whole economic system. Panel (a) shows the social optimum; panel (b) represents a situation of undertourism in which marginal benefits exceeds marginal costs; panel (c) shows an overtourism situation in which marginal costs overcome marginal benefits

In the next section, the paper moves from the theoretical framework presented above to the econometric application aimed at applying the conceptual idea to the case of Italian municipalities.

The Empirical Identification of the Social Optimum in Tourism Flows

The Econometric Model

We apply the theoretical and conceptual framework put forward before to Italian municipalities. Italy is, in fact, an interesting case study, since it hosts an extraordinary wealth of cultural heritage, and it is a renowned touristic destination. As remarked by Bernini and Galli (2023, p. 2): “… nearly 5,000 museums, galleries, collections, archeological sites, and monuments open to the public in 2018. With more such institutions than any other country in the world, Italy has 58 sites on the UNESCO World Heritage List, and there is at least one cultural structure within every 50 km squared of the country.”

Particularly, based on the idea proposed in the previous section, the mathematical derivation of the equalization between marginal localization benefits and marginal localization costs and the subsequent solution with respect to tourism (see Appendix 1) leads directly to the equation to be estimated through regression analysis. More specifically, we estimate the following specification:

$$\begin{array}{c}\multicolumn{1}{c}{\log \mathrm{tourism}=\mathrm{\alpha }+{\mathrm{\beta }}_1\log \mathrm{CH}+{\mathrm{\beta }}_2\log \mathrm{pop}}\\ \multicolumn{1}{c}{+{\mathrm{\beta }}_3\log \mathrm{social}\mathrm{expenditure}+{\mathrm{\beta }}_4\log \mathrm{accessibility}}\\ \multicolumn{1}{c}{+{\mathrm{\beta }}_5\log \mathrm{parks}+{\mathrm{\beta }}_6\log \mathrm{diversity}}\\ \multicolumn{1}{c}{+{\mathrm{\beta }}_7\mathrm{loghuman}\mathrm{capital}+{\mathrm{\beta }}_8\log \mathrm{pop}\mathrm{density}}\\ \multicolumn{1}{c}{+{\mathrm{\beta }}_9\log \mathrm{crime}+{\mathrm{\beta }}_{10}\log \mathrm{pollution}}\\ \multicolumn{1}{c}{+{\mathrm{\beta }}_{11}\mathrm{loghousing}\mathrm{prices}+\epsilon }\end{array}$$

(3)

where the units of analysis are Italian municipalities, and the dependent variable (log tourism) is the logarithmic transformation of per capita arrivals in 2019. log CH is the logarithm of the number of physical elements of cultural heritage per square kilometer, while log pop (population) is our measure of agglomeration economies. log social expenditure (expenditure on interventions and social services per capita), log accessibility (a standardized highway accessibility index), and log parks (parks subject to ministerial architectural and archeological restrictions per square kilometer) are our proxies for amenities. log diversity represents creativity and is measured through sectoral fragmentation, log human capital (specialization in high-tech sectors) catches the sophistication of the environment, log pop density (inhabitants per square kilometer) represents congestion, log crime (voluntary, and attempted murders per capita) catches social malaise, while log pollution (incidence of cars with emission standards lower than Euro 4 class on total cars) measures the environmental damage. Finally, log housing prices is the corrected (see next sub-section) highest real estate market value per square kilometer. All the explanatory variables are measured for the year 2014.

Since the units of analysis are Italian municipalities, that is, entities that clearly spatially depend on each other in several aspects, such as, more evidently than others, public services, and logistics, we estimate a Spatial Durbin Model (SDM). This “produces unbiased coefficient estimates also if the true data-generation process is a spatial lag or a spatial error model” (Elhorst, 2010, p. 10). Our estimates are based on a row standardized inverse distance matrix.

Subsequently, we estimate the predicted values, which represent the optimal level of tourism flows we look for, that is, the point in which the net benefits related to tourism flows are maximized through equalizing marginal localization benefits and marginal localization costs. ³

The next subsection describes the data, their sources, and the computation of the indicators.

The Data

The dataset used to empirically test the conceptual framework is primarily sourced from ISTAT (Italian Statistical Office) and includes 7,898 Italian municipalities. The dependent variable (tourism flows) refers to 2019, while the regressors to 2014. 2019 has been selected to avoid including COVID-19 pandemic effects in the analysis, while 2014 has been chosen since it was the oldest year for which all the included regressors were available. The final number of municipalities included in the analysis is the result of an accurate matching between the list of municipalities in effect in 2014 and the one in effect in 2019. In fact, between 2014 and 2019, some Italian municipalities experienced changes, including mergers, incorporations, and divisions, which led to modifications in their ISTAT codification.

The next part of this section describes the variables selected in the attempt of assessing as accurately as possible the localization benefits and costs for residents.

Tourism flows are measured as the number of tourists per inhabitant. ISTAT provides the exact number of tourists for a number of municipalities (which represents about 98% of the total tourist arrivals). An aggregate number of tourists is provided by province for the other municipalities. We apportioned this aggregate number to the other municipalities using the share of employed people in sector I (accommodation and food service sector) which presents a 93% fully statistically significant correlation with the available number of tourist arrivals.

Among the benefits, cultural heritage is measured retrieving data from the Italian Ministry of Culture (MIBACT), specifically from the Risk Map of Italian material cultural heritage, which identifies heritage assets deserving protection through the legal recognition of cultural interest. We measured material cultural heritage using the number of assets classified as archeological and/or architectural. The number of cultural heritage elements has been further weighted by area to obtain an indicator of heritage exposure. As suggested by Cerisola (2019a), this serves as a measure of the local environment’s cultural intensity—essentially, how likely individuals are to encounter a tangible piece of cultural heritage in their surroundings. These data refer to 2004. Since we are primarily focused on tangible, immovable heritage, which is typically quite old, it is reasonable to assume that the distribution of these sites has remained fairly stable over time.

In addition to tourism and cultural heritage, other traditional localization benefits have been measured exploiting the following variables:

resident population represents our measure for agglomeration economies;

per capita social expenditures, number of historical parks per square km, and logistic accessibility measure different forms of amenities, that is, respectively, public services, recreational services/natural amenities, accessibility;

the complement of the Herfindahl-Hirschman Index computed on employment in 17 sectors measures the municipal diversity;

the specialization in high-tech sectors is used as a proxy for human capital.

Alongside, in addition to tourism, other traditional localization costs have been assessed using the following variables:

population per square km is used as a congestion measure;

the number of murders per 100,000 inhabitants has been selected as a proxy for social costs and, therefore, malaise;

the incidence of polluting cars represents our measure for environmental damage;

real estate highest prices per square meter in the municipality (housing prices) is a localization cost in terms of “cost of the city.”

Table 1 provides the list of the selected variables, together with the corresponding description, source, and reference year. The benefit or cost measured by each variable is also specified.

Table 1.

Variables’ Description.

Variable name	Computation	Class of variable	Data source	Year
Tourism	Number of arrivals per capita	Benefit and cost	ISTAT—Atlante Statistico dei Comuni	2019
Cultural heritage	Assets with ministerial architectural and archeological restrictions per kilometer	Benefit	Risk Map, Italian Ministry of Culture	2004
Agglomeration economies	Total population on 1st January	Benefit	ISTAT—Atlante Statistico dei Comuni	2014
Social expenditures	Expenditure on interventions and social services per capita	Benefit	ISTAT	2014
Natural amenities	Number of historical parks per square kilometer	Benefit	Risk Map, Italian Ministry of Culture	2004
Accessibility	Standardized index of accessibility based on the access to the highway network	Benefit	ISTAT	2022 4
Diversity	1-HHI, based on employment in 17 (1-digit level) sectors	Benefit	ISTAT	2014
Human capital	Share of employment in high-tech sectors	Benefit	ISTAT	2014
Congestion	Population per square kilometer	Cost	ISTAT	2014
Malaise	Voluntary committed and attempted murders per capita	Cost	Italian Ministry of Interior	2014
Environmental damage	Incidence of cars with emission standards lower than Euro 4 class on total cars	Cost	ISTAT	2014
Housing prices	(Corrected) Highest real estate market value regardless of the property category	Cost	OMI—Osservatorio del Mercato Immobiliare	2014

It is worth mentioning that the selection of these variables has been guided by two main factors: first, as previously mentioned, identifying the best indicators to measure the specified location benefits and costs; second, choosing variables not directly linked to tourism flows. In fact, equalizing marginal localization benefits and marginal localization costs, the subsequent solution with respect to tourism leads directly to equation (3) in which localization benefits and costs appear on the right-hand side and tourism flows on the left-hand side. However, some location benefits and costs might be influenced by tourism. Therefore, to avoid reverse causality issues we included variables as much as possible not correlated with tourism flows.

More specifically, since natural amenities can be made available/enjoyable to attract tourists, we resorted to historical (old) green areas (protected through the legal recognition of cultural interest) to avoid the risk of reverse causality with tourism. In addition, for measuring malaise, voluntary homicides and attempted murders per capita were considered in the attempt to exclude other types of crime that might be directly influenced (generated) by tourism (e.g., robbery). A similar reasoning was carried out for the environmental damage: since it can also be caused by tourism, we resorted to an indicator (incidence of polluting cars) clearly associated with residents rather than tourists.

Particular attention deserves the correction we operated on rent. Since tourists quite evidently affect it (Boussaa, 2021; Valente et al., 2022), rent was regressed on tourism flows and the residuals of this analysis are what we include as our measure of housing prices in the model described by Equation (3). The same approach was used, among others, by Caragliu and Nijkamp (2012) and by Dellisanti (2023) In our case, this correction changes the regression coefficient from positive into negative and reassures us on the accuracy of the procedure.

The following section (Section 5) displays and discusses the results.

Results and Discussion

After running a Spatial Durbin Model (whose outcome is displayed in Appendix 2) we estimated the predicted values of tourism flows which represent the optimal level we looked for, taking into account the localization benefits and costs for residents. To identify how much the actual tourism flows deviate from the optimal level in each municipality, we calculated the Mean Absolute Percentage Error (MAPE), that is, in our case, the difference between actual, and predicted tourism flows weighted by the actual flows as follows:

$$\mathrm{MAPE}=\frac{\mathrm{Actual}\mathrm{tourism}\mathrm{flows}-\mathrm{Predicted}\mathrm{tourism}\mathrm{flows}}{\mathrm{Actual}\mathrm{tourism}\mathrm{flows}}$$

(4)

We exploited the sign and the magnitude of the obtained values to determine whether each municipality experiences overtourism, undertourism, or balance level in tourism flows. Specifically, if the value of MAPE is greater than zero, the actual number of tourists per capita exceeds the optimal level, resulting in an overtourism condition. Conversely, if MAPE takes on a negative value, it suggests that the actual number of tourists per capita is below the optimal level, indicating a state of undertourism. Ideally, values equal to zero correspond to the balance level. We extended this condition to the 5% of municipalities whose MAPE values are closest to zero. Higher positive values indicate more severe overtourism issues, while higher negative values suggest a greater potential to accommodate additional tourists. The following figure (Figure 2) represents MAPE distribution between the 10th and 90th percentile. Values lower than the 10th percentile or greater than the 90th percentile are not displayed on this figure for readability reasons; however, it is important to anecdotally mention some of the municipalities which fall in these percentiles. Municipalities such as Trento, Bolzano and Aosta fall below the 10th percentile. Notably, these are quite big cities that are located in provinces that are primarily characterized by a strong mountain-oriented vocation, featuring numerous well-known, often smaller, destinations that instead experience overtourism issues. A different example of municipality falling below the 10th percentile is Scandicci, a town located in the metropolitan area of Florence, whose appeal is overshadowed by the far more famous art city. At the opposite end, expectedly, Venice ranks above the 90th percentile, confirming its severe overtourism condition. Additionally, smaller yet well-known seaside destinations fall within this range, including Tremiti Islands in Apulia; Portofino in Liguria or Lignano Sabbiadoro in Veneto, but also mountain destinations like Andalo in Trentino Alto Adige. In general, all the renowned heritage destinations in Italy suffer from overtourism issues. However, the municipalities more affected by overtourism seem to be those characterized by a natural touristic vocation. This is in line with what it is reported by Peeters et al. (2018) through a case studies’ analysis applied to Europe they suggest that the most vulnerable destinations are coastal, islands and rural heritage sites. This is related to the usual small size of these touristic destinations in terms of inhabitants. The touristic pressure exerted by the tourists is higher in these places in relative terms.

Figure 2.

MAPE distribution between the 10th and the 90th percentile.

To systematize more clearly the reasoning on our results and to explore in greater detail the role of cultural heritage, the classification of municipalities provided by the Italian Statistical Office (ISTAT) according to their touristic vocation has been considered. More in detail, ISTAT assigns a specific category to each municipality according to two criteria: the touristic vocation based on geographic (e.g., closeness to the seaside; height) and anthropic (great urban agglomerations) criteria; and touristic density (see for more details ISTAT, 2022). ⁵ Since we are mainly interested in heritage destinations, we aggregated ISTAT categories and obtained one main group of municipalities, that is, big cities and culture, history, art, and landscape destinations. The municipalities belonging to this group have been divided into overtourism or undertourism (destinations in balance have not been considered) and represented in red or green, respectively, in Figure 3. As the map shows quite clearly, heritage municipalities in overtourism and in undertourism are in fact localized very near one to the other, and this is especially true in central Italy. This leads to think about the suitability of a more balanced distribution of touristic flows among the different destinations.

Figure 3.

Overtourism, undertourism, and heritage in Italian municipalities.

In order to deepen the reasoning in this respect, the ISTAT category labeled no specific touristic vocation has been considered. Specifically, the municipalities belonging to this category and with touristic potential (i.e., the ones in undertourism) have been investigated in terms of their heritage endowment. In fact, many of them are endowed with a significant amount of cultural heritage (i.e., their material cultural heritage per square kilometer is greater than the median value). These municipalities show a significant value in terms of their heritage endowment and are characterized by undertourism. They are represented in blue in Figure 3, which shows that they are often located close to big cities and culture, history, art, and landscape destinations in overtourism with, therefore, great potential for valorization of less well-known destinations and touristic redistribution.

Therefore, a more balanced distribution of tourists could be achieved through enhanced cooperation and networks among destinations via, for instance, promotion of different attractions, better transportation options, and tours. As highlighted in Russell et al. (2022) the importance of social media in influencing traveling decisions of many individuals could be exploited by (less renowned) destination marketers to both tailor marketing messages as well as to address and monitor the destination image, potentially trying to attract visitors willing to distinguish themselves from mass tourism and usual must-see destinations. Examples of virtuous cooperation based on culture among destinations can be found in the Italian Museum Routes initiative. ⁶

It is important to note that our definition of overtourism and undertourism is based on the concept of maximizing net benefits for residents, which reflects a perspective rooted in economic reasoning. This approach focuses on assessing tourism’s impacts, emphasizing the balance between the positive and negative economic effects of tourism on a given destination. However, overtourism can also be analyzed from other perspectives, such as social, environmental, or cultural, which may lead to different conclusions. In this sense, it could be interesting to match the overtourism condition found through the methodology presented in this paper with residents’ perception.

Additionally, the effects of overtourism are not uniform within the same municipality; some neighborhoods may experience significant touristic pressure and negative consequences, while others remain largely unaffected. A precise analysis of such spatial variability is possible through a case study analysis only. Furthermore, the issue related to day-trippers—for which it is difficult to retrieve data—exacerbate the congestion of specific areas within municipalities. Nevertheless, one key strategy is to convince visitors to spread themselves around parts of the city beyond its historic core, similarly to what we suggest for different municipalities. ⁷

Conclusions

This work presented particularly interesting theoretical implications since it addressed the study of touristic flows, originally looking at their role of both benefit and cost for the residents. It did so through an innovative re-conceptualization of a neo-classical theoretical framework, empirically applied to a finely geographically disaggregated sample (Italian municipalities), with a specific focus on the role of cultural heritage.

Overall, the paper documented the existence of overtourism in many Italian destinations, but also the touristic potential of several other less acclaimed municipalities. In terms of practical implications, it appeared particularly important to develop strategies aimed at managing tourism in a sustainable way, especially through a more balanced distribution of the flows. Such objective could be pursued through suitable policies, which would of course require coordination among local administrations and that could be effectively based on the valorization of cultural heritage, especially in places that are not yet esteemed in this respect.

In general, traditional forms of mass tourism are no longer adequate for sustainable economic development (Timothy, 2014). When reasoning on sustainable tourism, the strategies that are available to manage excess visitor demand can be either supply-side measures (enlarging the use potential of the city) or demand-side measures (limiting the use of the city for tourism purposes). In addition, the interventions put in place can be either “hard” (affecting quantities in terms of number of tourists) or soft (affecting their behavior). In this paper, we suggest a more equilibrated distribution of tourism flows among municipalities. Clearly, in practice this should be managed through specific agreements (deals) among nearby-located populated agglomerations, implying the abandonment of laissez faire in tourism development, in favor of the adoption of an explicit tourism management policy (see van der Borg et al., 1996).

Some practical cases on this line of reasoning can be found in Italy. For example, the Umbria region museum system (Montella, 2020) has got, among its objectives, the valorization and connection of different (small) museums over the territory, in integration with the touristic and productive activities of the region.

Even more evident, in this regard, the role of the Uffizi Diffusi ⁸ project, which aims at taking its art to the people with the idea of delocalizing (and thus de-congestioning) the famous Florence museum and bringing several of its artworks outside the capital city. The object is to regenerate the region’s lesser-known towns and to encourage a more sustainable tourism. Therefore, an operation intended to celebrate the cultural and landscape heritage of Tuscany, for both tourists and locals.

An important aspect the present work did not focus on is the “quality” of tourism, that is, an absolutely intangible (but still significant) feature that would need additional research to be properly considered, especially within related policies. In this respect, some interesting strategies have been put forward in Italy in recent times in terms of slow tourism (e.g., Calzati & de Salvo, 2018; Klarin et al., 2023; Moira et al., 2017). Other admitted limitations of this study are linked to the absence of data on day-trippers, who could play a significant role in exacerbating congestion issues, as well as the unavailability of data on tourism seasonality.

In all, this paper opened the way to more conceptualized research on touristic flows and proposed a related empirical application that took cultural heritage in particular account. Further studies could be devoted to deepening the reasoning on the here proposed redistribution of flows, based on geographical and socio-economic networks among populated areas as well as matching this analysis with residents’ perception related to tourism.