Exploring the influence of an extended theory of planned behaviour on preferences and willingness to pay for participatory natural resources management

Highlights

• We tested the effect of the theory of planned behaviour on preferences for local governance.

• We used a hybrid mixed logit model to analyse the data.

• We found a significant effect of TPB indicators on preferences.

• The hybrid model performed better than the baseline.

• Results suggested that TPB explains preference heterogeneity and should be considered for an effective destination management.

Abstract

In this paper we explore the role of the theory of planned behaviour in the context of preferences for a decentralized governance of natural resources. A choice experiment was carried out to elicit preferences of the tourists for alternative options of natural resource management in a case study in Italy and data were collected by means of personal interviews. Indicators to assess the planned behaviour of respondents were collected by means of Likert scales that were then included in the choice model. Differently from previous research on planned behaviour we use an extended version of the theory that includes moral norms and explicitly account for endogeneity of the indicators with a hybrid mixed logit model, in which a latent variable is used to explain both answers to the indicators and management choices. Results suggest a general preference of respondents for local governance and indicate that the latent variable has a significant effect on explaining preference heterogeneity and improve model fit. In addition, results suggest that the theory of planned behaviour is appropriate to model individuals' behavioural intention and can be used to tailor marketing activities aiming at increasing people's pro-environmental behaviours.

Introduction

Environmental values are difficult to evaluate in economic terms because of their non-market nature, therefore stated preferences surveys are frequently used to assess individuals' preferences. Stated preference surveys elicit willingness to pay (WTP) of good not traded in the market place, which is then used as a measure of the value people attach to that specific good (Hanley and Barbier, 2009). Choice Experiments (CE) are amongst the most flexible stated preferences approach, because they are based on the evaluation of single attributes of a good or service and allow addressing a very large number of situations (Hanley et al., 2002). In a CE setting the relevant attributes and attribute levels of a good or service are combined to create different alternatives and respondents are asked to choose their preferred in a sequence of choice tasks (Hensher et al., 2015).

It has been observed that using some behavioural theory to explain heterogeneity in preferences contributes to improve model estimates (Fischer and Hanley, 2007). In the environmental field, the theory on environmental value orientations towards wildlife help explaining different levels of WTP for wolves, lynx and salamanders (Grilli et al., 2018), while the new ecological paradigm is a predictor for preferences for endangered species (Aldy et al., 2012). Social norms were also found to have a significant effect on environmental preferences (Jones et al., 2010). In this paper we contribute to the behavioural literature of choices by studying the effect of the theory of planned behaviour (TPB) on environmental preferences. A novel aspect of our approach is that we assumed that planned behaviour is a latent and unobservable characteristics of the human behaviour, which can only be approximated by some stated measures such as Likert scales. The most common alternative approach is the direct inclusion of these variables in the utility function (Hess and Stathopoulos, 2013), which is problematic for at least two reasons, i.e. measurement errors and correlation with the error term (Hoyos et al., 2015), because Likert-type variables are only proxies that are collected with some degrees of inaccuracy. To solve this issue, we implemented a hybrid mixed logit model (HMXL). The HMXL is a simultaneous equation model in which a latent variable is used as a predictor for TPB indicators and for the choice model, so that TPB and CE are linked by this latent variable. Despite some increase in computational effort, the class of hybrid models is becoming popular because of the gain in explanatory power (Mariel et al., 2015).

The policy context of our analysis is the study of tourists' WTP for a participatory management of a tourism destination, i.e. the Monte Baldo Local Nature Park in the Province of Trento, which is located in the Italian Prealps. The traditional management and planning of territories has always been carried out by the Province of Trento, which is not efficient because it is not fully aware of the needs and opportunities of local communities and of the specific characteristics of natural resources. On the other hand, a co-management of local natural resources made together with local inhabitants may increase the quality of touristic offers and improve the recreational experience of visitors. A project to allow community-based management has been launched in several districts of the Province, including Monte Baldo. The project will create the so called “reserve network” for protected areas, aiming at involving local communities in management and balance ecological, social and economic aspects. Testing the theory of planned behaviour in this context is useful to understand the drivers of individuals' attitude towards natural resources management and the acceptability of conservation measures, so that potential sources of conflicts can be timely identified. In addition, the proposed model provides new insights on the sources of preference heterogeneity given by individuals' norms and beliefs.