Investigating Change in the Willingness to Pay for a More Sustainable Tourist Destination in a World Heritage City

Abstract

The willingness to pay for sustainable tourism products and services has been widely discussed in the scientific literature. However, change in the willingness to pay over time has rarely been analysed. Such studies are important for understanding the impact that the increasing debate on sustainability, including environmental, economic, and social aspects, may have had on the willingness of tourists to pay more in order to ensure the sustainability of destinations. The aim of this article is to assess how the willingness of tourists to pay for sustainable tourism services has developed in the Spanish city of Cáceres, declared a World Heritage Site in 1986. Data from surveys conducted in 2012 and 2016 were used for this purpose. Logistic regression was applied to determine whether sociodemographic characteristics of tourists who visit the city influenced their willingness to pay in each year. A Chow test was applied to elucidate whether the differences between the years were statistically significant. The results obtained indicate that only the level of education determined willingness to pay, while origin, gender and age showed no effect. No significant change in the willingness to pay for sustainability was found among tourists in Cáceres between 2012 and 2016. These findings indicate that the willingness to pay for more sustainable tourism services did not increase during the period studied. We propose some measures such as a ‘municipal observatory of sustainable tourism’ in order to increase willingness to pay in this type of destination.

1. Introduction

The incorporation of the principles of sustainable development in the management of World Heritage cities is a very recent phenomenon, and its beginnings can be traced back to 2002 with the Budapest Conference and the Declaration published by the World Heritage Committee under the same name [1], at which time integration policies began to be adopted from the perspective of sustainable development according to the procedures of the World Heritage Convention [2].

The historical background of this perspective can be found in the International Charters on the conservation of heritage, historical towns, and urban areas published by the International Council on Monuments and Sites [3], which include the following: The Australia ICOMOS Charter for the Conservation of Places of Cultural Significance, known as “The Burra Charter” in 1979 [4]; the International Charter for the Conservation of Historical Towns and Urban Areas in 1987 [5]; the International Cultural Tourism Charter in 1999 [6], which emphasized the management of tourism at Places of Significant Heritage, as well as the so-called “Charter of Krakow 2000” regarding the principles of conservation and restoration of monumental heritage [7], and finally, the Xi’an Declaration on the Conservation of the Setting of Heritage Structures, Sites, and Areas in 2005 [8].

It is also worth highlighting the World Conferences sponsored by the United Nations, such as the Convention Concerning the Protection of World Cultural and Natural Heritage [9], which is a fundamental document in this process, as it jointly defends both natural and cultural heritage and introduces the recognition by member countries of the obligation to “identify, protect, preserve, rehabilitate and transmit to future generations the cultural and natural heritage located in their territory.” (Art. 4). This intergenerational solidarity was later invoked by the World Commission on Environment and Development in what is known as the “Brundtland Report”, which included the following in its own definition of sustainable development: “Sustainable development is development that meets the needs of the present generation without compromising the ability of future generations to satisfy their own needs” [10].

The European Union also has contributed to the debate on integrated urban sustainability, given that the so-called “Toledo Declaration”, signed in 2010 when Spain held the Council Presidency, addressed the comprehensive management of urban spaces from four perspectives: environmental, economic, social, and cultural. Moreover, it also incorporated the concept of “governance” [11] as the nexus among the four aspects [12], which also appeared in successive reports produced since then, such as the report by the Committee on Culture and Education of the European Commission, presented to the European Parliament in 2015: “Towards an Integrated Approach to Cultural Heritage for Europe”. Among the suggestions included in the report was an explicit reference to the vital economic relationship between cultural heritage and sustainable tourism, as well as to the need to take a global approach to the issue [13].

The application of sustainability criteria to World Heritage Cities faces a wide range of problems and conflicting interests that make it very difficult to achieve consensus on limits to their economic exploitation. For example, in the case of Italian tourist destinations, authors such as Cuccia and Guccio suggest not exceeding the accommodation capacity as well as not to focus on short-termism which has been used by local authorities as criteria to obtain the title of “World Heritage” [14]. Doing so can compromise long-term sustainable development, especially when one of the economic activities in cities with rich cultural heritage that benefits the most from such titles is tourism. This has also been documented in Spain, where tourism has been one of the main driving forces in transforming landscapes and their functionality, as attested by statistics from the World Tourism Organization [15].

Transformations caused by massive tourist activity have been reported since the 1970s by authors such as Turner and Ash [16], who saw this industry as a phenomenon similar to the European colonialism of the 19th century regarding its ability to irreparably destroy the ways of life of host societies. It should be noted at this point that such a position was dismissed by other authors as “emotional allegations” [17].

The experience of the last few decades seems to indicate the possibility of greatly reducing the negative impact of tourism on destinations and generating prosperity as a result, as long as environmental protection, culture, tourism, and urban policies are integrated, as these are the pillars of sustainable development [18]. Such integration must go hand in hand with the comprehensive management of tangible and intangible goods [19,20], as well as their authenticity as perceived by visitors, which has a decisive impact on tourists’ willingness to pay [21].

The importance of incorporating sustainability criteria into the touristic management of World Heritage Cities is increasingly being recognized by policy-makers and planners. There is also evidence that some tourists are willing to pay a certain amount extra for visiting sites that are perceived as sustainable, making sustainability an added value for destinations. For these trends to create a lasting, positive impact, a sufficient number of tourists have to be able and willing to reward efforts by destination managers to ensure sustainability. It is, therefore, important to investigate how tourist’s willingness to pay evolves over time. The purpose of this study is to test:

Hypothesis 1.

There is a significant change between 2012 and 2016 in the willingness to pay more for sustainable tourist services in the Spanish city of Cáceres.

Hypothesis 2.

There is influence of sociodemographic variables (origin, gender, age, and educational level) on willingness to pay more for sustainable tourism services.

Tourism sustainability has been present in the strategic policies of tourist destinations for at least the past 10 years. By now, its presence should have resulted in increased tourist awareness and become more lasting, as sustainability measures taken by destinations have become more widespread. For this reason, we want to check if the willingness to pay for tourism sustainability has changed in the city of Cáceres between the two years for which information is available in this regard (2012 and 2016).

2. Materials and Methods

The data in this study have been obtained from surveys that are collectively placed under the heading, “Survey on tourist perception of sustainability in a World Heritage city (Cáceres)”, which gathered opinions and/or knowledge about sustainable tourism that visitors to the city of Cáceres articulated during the months of March and April of 2012 and 2016, respectively. This survey is carried out every four years. Due to the COVID-19 pandemic it was not possible to conduct the survey in 2020, and the next round is planned for 2022. These survey periods coincide with the high season of tourism in the city, which includes Spring and Easter. To create the questions, the common questionnaire structure used in contingent valuation studies for estimating willingness to pay was consulted [22]. The surveys were carried out in the street by a team of interviewers who used a questionnaire written in Spanish and English, randomly selecting tourists in the Monumental City of Cáceres (as the most visited place), who gave oral responses to the questions on the form, which were then recorded by the interviewers.

The 2012 survey was taken by 486 people, and the 2016 survey by 474. According to data provided by the National Institute of Statistics regarding the accommodation occupancy survey, the city of Cáceres received 223,071 travellers in 2012 and 270,721 in 2016, including people who reside in Spain as well as those who live abroad. Using these two figures, the margin of error in carrying out the two surveys has been calculated at 5%, and the precision error stands at 4% [23].

The following socio-demographic characteristics were recorded for each respondent:

• Origin (Spanish or foreign)
• Gender
• Age
• Educational level

Among the survey questions, the following two were related to sustainability aspects:

Question 4: Do you take sustainability criteria into account when choosing a specific tourist destination or accommodation?

Question 11: Would you be willing to pay an increase in the price of a tourist service to enjoy a stay in a more sustainable destination?

2.1. Logistic Regression

This model was estimated using GRETL software 23, version 1.9.4 [24], a cross-platform software package for econometric analysis, written in the C programming language. It is free and open-source software. One of the econometric analyses that Gretl performs is the analysis of limited dependent variables, which includes the logit model that has been estimated in this paper.

This statistical technique has been used frequently in tourism research, in areas as diverse as market segmentation [25], identification of the determining factors of innovation in tourism [26], predictive determinants involved in tourist loyalty toward a destination [27] and describing the demand for ornithological tourism in Extremadura [28].

The statistical technique used for the present study was the binary logistic regression statistical analysis, as it tries to explain the behaviour of a qualitative characteristic (social phenomenon) based on a set of aspects that are also qualitative, using a non-linear probability model of dichotomous response. An explanation of this technique can be found in Agresti and Train [29,30].

Regarding the binary logistic regression model in question, the dependent variable is the random variable that represents the following question: “Would you be willing to pay a higher price for a tourist service in order to enjoy a stay in a more sustainable destination?” Being a dichotomous response question (Yes or No), it therefore justifies the choice of the model, which allows one to estimate the probability of an affirmative response to the question, and whether or not it is influenced by different factors (origin, gender, age, and educational level).

For this question, there were subsamples for 2012 and 2016, establishing a binary logistic regression model for each sample, as well as a joint model that contains both subsamples, which can be restricted (RM) or unrestricted (URM).

The restricted model RM contains both of the subsamples together and evaluates the independent variables for the dependent variable as a whole, assuming there is no intrinsic difference between the data from one subsample and another. On the other hand, the unrestricted model (URM) contains both of the subsamples together but duplicates the independent variables with a dichotomisation in which their effects are nullified in one of the years. For this case, the reference subsample is for 2012, so for the URM model, the duplicate independent variables obtain values of 0 for this year and 1 for 2016. In this way, the significance of the coefficients of the independent variables is sought, considering the difference between years.

To determine if there are differences between the two subsamples, a Chow test was carried out. The Chow test for contrast is performed to evaluate whether the parameters of a regression model are identical for all the samples analysed, and therefore it assesses whether the structure of relationships between variables remains constant. The null hypothesis to be tested is as follows:

$$H_0: \{\begin{array}{c}{\beta }_{0,1}={\beta }_{0,2}=\beta \\ {\beta }_{1,1}={\beta }_{1,2}=\beta \\ {\beta }_{2,1}={\beta }_{2,2}=\beta \\ .\\ .\\ .\\ {\beta }_{i,1}={\beta }_{j,2}=\beta \end{array}$$

where the alternative hypothesis is stated as follows:

$$H_1: \exists {\beta }_{0i}\ne {\beta }_{0j}, \exists {\beta }_i\ne {\beta }_j$$

We have proceeded to evaluate and interpret the estimations of the restricted model (composed of the total sample, years 2012 and 2016), and then the unrestricted model is presented, which includes the control variable D, having a value of 0 for the 2012 data and 1 for the 2016 data. Next, the existence of change is contrasted through the Chow test, and finally, in the event that there is change, both the restricted model for 2012 and that of 2016 are presented and interpreted.

In our case, given that the GRETL results offer the logarithm of the likelihood function (Log-likelihood), the contrast used has been the Log-likelihood ratio test between both of the models, which is provided by the following expression:

$$D=-2\left[\log \left({\mathrm{\Lambda }}_1\right)-\log \left({\mathrm{\Lambda }}_2\right)\right]$$

where $\log \left({\mathrm{\Lambda }}_1\right)$ is the logarithm of the likelihood function of the restricted model, and $\log \left({\mathrm{\Lambda }}_2\right)$ is the logarithm of the likelihood function of the unrestricted model.

Wilks [31]. shows that the D statistic follows an $x^2$ asymptotic distribution with $df2-df1$ degrees of freedom, where $df1$ and $df2$ denote the degrees of freedom of models (1) and (2), respectively. If the p-value associated with this D statistic is lower than the significance level, the presence of change can be assumed, and consequently, it is possible to conclude that there are significant differences in the fit of the binary logit model for the two surveys (2012 and 2016). This test of change, which compares the willingness to pay for tourism sustainability in the city of Cáceres in both 2012 and 2016, will allow us to accept Hypothesis H1 if the existence of change is confirmed, or reject it if the existence of change is not confirmed.

This test was used to check whether there was a structural change between the variables of the joint model (RM) and the models of the subsamples. In other words, it was used to verify whether the use of a joint regression was statistically justified, in comparison to the model separated by subsamples.

An explanation of this method can be found in Chow [32], Novales Cinca [33], Pérez Salvador, & García Salazar [34], and new applications for its use can be seen in Nielsen & Whitby [35].

2.2. Defining the Variables and Modelling

These variables have been generated as dichotomous binary variables whose categorisation is as follows:

Origin has two possible values:

$$Orig=\{\begin{array}{c}1 if they come from Spain \\ 0 if they are foreigners\end{array}$$

Likewise, gender has two possible values:

$$G=\{\begin{array}{c}1 if the person is a woman \\ 0 if the person is a man \end{array}$$

Age, which was divided into 4 groups, was transformed into 2 variables that have 2 possible values:

$$AG1=\{\begin{array}{c} 1 if they are between 18 and 40 years of age\\ 0 for the rest. \end{array}$$

$$AG2=\{\begin{array}{c}1 if they are between 41 and 65 years of age\\ 0 for the rest. \end{array}$$

Respondents who are older than 65 years old are the ones for whom both AG1 and AG2 have the value of 0, or in other words, AG1 = AG2 = 0.

Finally, educational level, which was previously divided into 4 categories, was transformed as well into 2 variables with 2 possible values:

$$EST1=\{\begin{array}{c}1 if they have not studied or have a primary school education\\ 0 for the rest \end{array}$$

$$EST2=\{\begin{array}{c}1 if they have a secondary school education\\ 0 for the rest \end{array}$$

The coding $EST1=EST2=0$ would be for those with university studies.

Continuing with the modelling, the equation of the logistic model is a non-linear exponential function, which can be presented as a linear function of the probability ratio of the logistic function through a logarithmic transformation, whereby:

$$P\left(Y_i=1\right)=\frac{\exp \left(z\right)}{1+\exp \left(z\right)}$$

Therefore:

$$z={\beta }_0+{\beta }_{1 }Ori{g}_i+{\beta }_{2 }{G}_i+{\beta }_{3 }AG{1}_i+{\beta }_{4 }AG{2}_i+{\beta }_{5 }EST{1}_i+{\beta }_{6 }EST{2}_i{e}_i$$

where $P\left(Y_{i }=1\right)$ represents the probability that tourist i is willing to pay a higher price for a more sustainable product or service.

3. Results

3.1. Restricted Model

The level of education (variables EST1 and EST2) had a statistically significant influence on the willingness to pay for a sustainable tourist destination (p < 0.0001), while the other explanatory variables (origin, gender, and age) had not (

Table 1. Estimation of the restricted model.

Explanatory Variable	Coefficient	p-Value
Const	0.950637	0.0054
Origin	−0.360892	0.1303
Gender	0.0774163	0.5893
AG1 (age 18–40)	−0.448488	0.1057
AG2 (age 41–65)	−0.205411	0.4392
EST1 (no studies or primary school education)	−1.79104	<0.0001
EST2 (secondary school education)	−0.728659	<0.0001
McFadden’s R-Squared	0.065376	Likelihood ratio test: Chi-square (6) = 79.1052 [0.0000]
Adjusted R-squared	0.053806
Log-Likelihood	−565.4484
Akaike Criterion	1144.897
Schwarz Criterion	1178.324
Hannan-Quinn Criterion	1157.683

Source: Prepared by the author based on calculations using GRETL.

). The Chi-square likelihood ratio test of the previous model equal to 79.1052.

The model shows that the only variables that influence the willingness to pay for a more sustainable touristic destination in a world heritage city are EST1 and EST2. Moreover, their negative symbol means that the higher the level of education of the tourists, the greater is their willingness to pay.

3.2. Unrestricted Model

In the unrestricted model (

Table 2. Estimation of the unrestricted model.

Explanatory Variable	Coefficient	p-Value
Const	0.990296	0.0039
Origin	−0.568913	0.0775
Gender	0.0431634	0.8390
AG1 (age 18–40)	0.128126	0.7150
AG2 (age 41–65)	−0.100905	0.7549
EST1 (no studies or primary school education)	−1.78195	<0.0001
EST2 (secondary school education)	−0.916671	<0.0001
OrigD	0.350106	0.3665
GD	0.00846479	0.9764
AG1D	−1.01254	0.0126
AG2D	−0.183068	0.6312
EST1D	0.0580550	0.9044
EST2D	0.296009	0.3454
McFadden’s R-Squared	0.074322	Likelihood ratio test: Chi-square (12) = 89.9295 [0.0000]
Adjusted R-squared	0.052834
Log-Likelihood	−560.0363
Akaike Criterion	1146.073
Schwarz Criterion	1208.152
Hannan-Quinn Criterion	1169.818

Source: Prepared by the author based on calculations using GRETL.

), the influence of the educational level variable is confirmed, although the variable AG1D is also significant, in that the youngest age group in 2016 (18–40 years) shows a lower willingness to pay than the same age group in 2012.

The variables of this unrestricted model have been obtained by multiplying the original explanatory variables by the control variable D as follows:

$$OrigD=Orig\ast D$$

$$GD=G\ast D$$

$$AG1D=AG1\ast D$$

$$AG2D=AG2\ast D$$

$$EST1D=EST1\ast D$$

$$EST2D=EST2\ast D$$

The previous model presents a Chi-square likelihood ratio contrast equal to 89.929. The p-value of the statistic is 0.000. Therefore, at a significance level of 5%, it can be concluded that there is sufficient evidence to reject the null hypothesis and affirm that the model is statistically significant.

3.3. Chow Test

The statistic D = 10.8242 with n = 6 degrees of freedom has shown a p-value = 0.0940, and therefore, at a significance level of 5% there is not enough evidence to reject the null hypothesis and affirm that there has been a change between the regressions executed for the logit regression models of the subsamples and the model as a whole. This means that there is no change in terms of difference between WTP of tourists in 2012 and 2016.

4. Discussion

The comprehensive territorial vision regarding patrimonial conservation in World Heritage tourist destinations requires the economic evaluation of cultural heritage as a measure of progress toward sustainable development in these spaces, which involves multiple obstacles that impede its accomplishment, both economic and sociocultural [36]. It is necessary to take into account one of the most overlooked factors of research in the field of sustainability until merely a decade ago—the tourists themselves [37,38]. Knowledge of the economic dimension of historical heritage and its corresponding impact on the spaces where an activity takes place, its positive and negative externalities, is a prerequisite for any action that takes place in these spaces [39]. The difficulty in collecting relevant market data on cultural goods has led to the use of stated preference methodologies, such as the contingent valuation method (CVM) and choice model (CM) [40]. Contingent valuation is the most widely used method for appraising cultural assets, and the only one, according to some authors [41], as it allows empirical economic measurements not previously considered by economists. The surveys are easy to carry out and economical, as long as the asset to be appraised is explained correctly and the sample is taken from a well-defined target audience.

The argument against this method involves two considerations: the willingness to pay (WTP) is limited by the socio-economic status of the respondents, and the preferences considered are those of the current generation [42,43], with some authors calling for an approach based more on behaviours rather than attitudes [44], given its failure to predict specific behaviours, although it should be noted that it can be corrected through the principle of aggregation, which shows a much higher predictive capability based on general attitudes and personality traits [45].

In several studies, the willingness to pay more for eco-labelled products has been linked to pro-environmental values and attitudes with regard to the sociodemographic characteristics of the respondents [46,47]. There is also a higher probability of accepting a price increase if respondents agree with the purpose of the payment [48]. The results of our study coincide with their conclusions: a sociodemographic variable influences WTP, in this case, the educational level. Other factors involved in the willingness to pay more include the influence of sustainability on the expectations of tourists, the experiences offered by the destinations, and the desire to visit these sites again [49]. In this sense, our studies should include additional variables in the future—beyond the socioeconomic ones—that affect the issues related to this approach.

The fact that the willingness to pay more for a stay in a more sustainable destination in the city of Cáceres did not increase during the study period may be due to several factors, but as mentioned in Section 3, the willingness to pay is higher when respondents agree with the purpose of the price increase.

It is important to explain what specific measures are being carried out in the field of sustainability, both by the administrative authorities, who have the option of promoting the destination as sustainable, and by private establishments, which can add value to their services if they align the increase in their prices with the objectives of sustainability, communicating this information to their potential clients.

Taking into account the above, theoretical models of heterogeneous costs and benefits between governments, communities, and individuals with the purpose of improving policies in complex socio-ecological systems (SES) make sense [50].

5. Conclusions

As there was no change between the logistic regressions carried out with regard to the question asked, there were not differences in the results obtained between the 2012 and 2016 surveys when analysing whether tourists would be willing to pay a higher price for a tourist service to enjoy a stay at a more sustainable destination. There was not any evidence for a change in WTP between 2012 and 2016, and only level of education of respondents had an effect on their WTP in each year. Thus, Hypothesis 1 (H1) has not been confirmed but Hypothesis 2 has been (H2).

In order to coherently carry out a sustainability policy that will be endorsed by a higher socio-economic bracket, it is necessary to have analysis data on tourists that takes into account their preferences and expectations about a World Heritage city in relation to their valuation of sustainability, since failing to do so might result in the implementation of measures by administrative authorities and/or business owners that may not be well-received by tourists, either due to a lack of information, or because the measures are not perceived as worthy of a higher price at the time of their implementation.

In any case, it is necessary to clearly inform tourists that these price increases are going to result in an improvement in the tourism management of the municipality, in terms of sustainability that could translate into improvements: better waste treatment, less overcrowding of the destination, greater economic compensation to the resident population, etc.

For this reason, we believe that the establishment of a Municipal Tourism Sustainability Observatory, which would bring together and coordinate the needs of administrative authorities, tourism companies, and visitors for the purpose of carrying out advertising and awareness campaigns based on a rigorous analysis of demand, could be an essential element in highlighting the importance of the effort made by all of the actors involved toward sustainability.

Among the limitations of this study, the time frame bears mentioning, because even though the study aimed to include the year 2020, the third wave of surveys had to be suspended due to the COVID-19 pandemic and the lockdowns imposed. Moreover, one must consider that it would be appropriate to extend the study to other World Heritage cities with characteristics similar to those of Cáceres in order to obtain a comprehensive view of demand patterns in these areas, which could be a powerful decision-making tool for public administrators as well as private companies, who would be able to learn more about the behaviour of tourists and how to reinforce such behaviour, or not, depending on each case.

It would be necessary to carry out a homogenization in the questions of the surveys, especially with the concept of sustainability, to avoid different interpretations of it by the tourists surveyed.

In any event, the aforementioned limitations will be addressed when tourist activity returns to our World Heritage Cities. This will broaden current research with new surveys in 2022 or 2023 to discern whether or not the COVID-19 pandemic has had an impact on tourists’ awareness of sustainability, and on their greater or lesser willingness to pay more for sustainable tourism products and services.