Economic Viability of Developing Passive Recreational Opportunities in Puerto Rico: Insights for Sustainable Forest Management

Abstract

Forests offer a diverse range of ecosystem services for humans. For example, forests provide the space for passive recreational activities such as camping, hiking and bird watching. These activities are appealing to residents of all ages and in particular to those who cannot engage in more active recreational activities involving substantial expenditure of physical energy. The literature that examines the economic viability of developing passive recreational activities in forest ecosystems in developing countries is limited. Thus, using the Rio Hondo Community Forest in Puerto Rico as a case study, this study seeks to (1) estimate the benefits and costs of passive recreational opportunities, some of the nonmarket goods and services this forest offers, and (2) conduct a cost–benefit analysis to determine the economic viability of developing passive recreational activities. The results from this paper show that, in general, the residents prefer passive recreational activities demanding less effort. The development of such activities is economically viable. This study seeks to encourage inclusive forest management, ensuring that all groups of residents benefit from the forest, while simultaneously emphasizing sustainable practices.

1. Introduction

This study uses the Rio Hondo Community Forest (RHCF) in Puerto Rico as a case study. The RHCF is a secondary forest in Mayagüez that covers more than 65 acres with very high ecological value [1]. The forest hosts at least 42 different bird species, of which 10 are endemic [1]. Presumably, it was clear-cut at the beginning of the 20th century to plant sugar cane. Currently, only 5 percent of the original forest remains, and therefore one of the main goals is to preserve eight tree native species located within the forest border. Moreover, the forest can provide a wide range of recreational opportunities, including passive recreation (scenic view observation, low-level hiking, educational workshop participation, reading, meditation and camping). In this context, it is of key importance to develop sustainable management plans that protect the forest over time for future generations, are economically viable and allow residents’ enjoyment. To reconcile these objectives and guide authorities towards the decision-making process, the present study elicits the residents’ passive recreation preferences and conducts a cost–benefit analysis to assess the economic viability of most preferred activities (In Puerto Rico, there are a few studies on WTP for outdoor recreation in forest ecosystems.). The forest is protected by the US Forest Service, managed by the community, and supported by the Municipality of Mayagüez.

Ecosystem services or nonmarket goods and services are the benefits humans obtain directly or indirectly from ecosystems [2,3]. For example, humans benefit indirectly from carbon sequestration or from pollination services and obtain more direct benefits from timber and crops provided by forests. Less tangible direct benefits like recreational activities are also provided by forests [4]. According to Asim et al. (2015), active recreation involves both cognitive and physical effort, while passive recreation requires less physical effort [5]. Tavárez and Elbakidze (2019) provide examples of active activities (zip-lining or biking) and passive activities (meditation and cultural events) in forest ecosystems [6]. In this study, passive recreation is referred to as the activities that demand minimal physical effort from residents to enjoy nature.

Residents’ preferences and willingness to pay (WTP) for passive recreational opportunities provided by forests have been studied around the globe, including hiking, birdwatching, nature observation, driving within the forest and picnicking [6,7,8,9]. WTP for outdoor recreation depends on several aspects, such as the recreational attributes under evaluation, the sociodemographic characteristics of residents [6,7,8], the region [7] and residents’ environmental involvement [10,11]. Tavárez and Elbakidze (2019) used choice experiments to estimate residents’ WTP for recreational enhancements in the San Patricio Urban Forest of Puerto Rico, including bird watching, improved trails, a community garden and stage for activities [6]. The authors found that residents are willing to pay between USD 15 and USD 39 for recreational activities. Rivera-Acosta and Martínez-González (2020) used a contingent valuation method to estimate visitors’ WTP for recreational opportunities in the Finca Montaña forest of Puerto Rico and found that visitors are willing to pay USD 9.99 per visit for using the forest for outdoor recreation [12].

Moreover, other studies evaluate the aspects that enhance recreational experiences like the availability of information boards, waymarks and forest openings [13,14]. A better understanding of human preferences regarding recreation provided by forests is essential to design environmental policies that meet the needs. For example, passive recreation promotes wellbeing of vulnerable groups, such as elderly and disabled people, who generally cannot derive benefits from active recreational opportunities like zip-lining, mountain biking, high-level hiking, among others. However, the literature using the cost–benefit analysis of developing passive recreational opportunities in forest ecosystems is limited, especially in developing countries (Puerto Rico boasts one of the highest human development indexes (HDIs) in Latin America and the Caribbean region [15]. In 2012, Puerto Rico’s HDI stood at 0.865. However, when considering the inequality-adjusted human development index, this figure drops to 0.685. This makes Puerto Rico’s human development patterns similar to countries like El Salvador (0.675) and Bolivia (0.692).). Cost–benefit analysis is a useful tool to evaluate the economic viability of developing new projects or modifying existing programs [16,17]. The main goal of a cost–benefit analysis is to estimate the marginal benefits of a proposed project and compare them with the corresponding costs. The benefit–cost ratio presents the overall relationship between the relative costs and benefits of the proposed project; it helps illustrate how much money will be generated for each dollar spent on a project [16]. If the estimated net benefit surpasses the net costs, the project can be implemented, and policymakers can prioritize projects and existing resources.

A few studies conducted in both developed and developing countries show that, in general, conservation of forests is desirable, and that benefits exceed costs. This holds true for a study conducted in Eastern Finland, in one of the least developed rural communities bordering Russia [18]. In Norway, the benefits of preserving forest biodiversity and ecosystem services are substantially higher than the costs [19]. Moreover, a relatively recent study found that the benefits of planted forests is more than 100 times higher than the overall cost of the program when aggregating the public benefits of biodiversity at regional and national levels in New Zealand. For each NZ dollar spent, the public benefit is 149 NZ dollars. These findings reveal that NZ households support programs, especially if they sustain native species, such as the brown kiwi [20]. A study in China investigating the WTP and the costs and benefits of public green spaces for leisure activities, health and provision of ecosystem services is also in line with these results, but the cost–benefit ratio is smaller, 0.88 [21].

The results of the economic viability analysis of passive recreational activities in the RHCF intend to inform how resources can be efficiently allocated in ways reflecting the residents’ preferences and needs. In addition, the findings can assist by improving forest management towards more sustainable and equitable practices. Projects that study passive recreation address elements of sustainability and environmental justice [22,23]. Passive recreation tends to be more inclusive as it allows a broader range of visitors to benefit from forest ecosystems and who are encouraged to become more involved in management of resources, thus providing feedback to stakeholders on preferences, needs and sustainable practices [24]. In addition, recreation and physical activity in natural sceneries promote healthy lifestyles in communities [25], and recreation that seeks adequate use of natural spaces can contribute to sustainability (Sustainable recreational activities can contribute to Sustainable Development Goals (SDGs), in particular to Goal 15 of protecting, restoring and promoting sustainable use of terrestrial ecosystems, sustainably managing forests, combating desertification and halting and reversing land degradation and biodiversity loss. Recreation and physical activity in natural sceneries promote healthy lifestyles in communities and thus support Goal 3 of Good Health and Well-Being) [26].

This paper studies three key research questions: How much are the residents willing to pay for passive recreational activities? What are the key factors that affect the values placed? Is the implementation of passive recreational opportunities economically viable? Using the RHCF of Puerto Rico as a case study, the specific objectives are as follows:

(1)

To estimate the benefits and costs of passive recreational opportunities that the forest can provide;

(2)

To conduct a cost–benefit analysis to determine whether developing those recreational activities is economically viable.

This study uses choice experiment data from a questionnaire to estimate the economic benefits. Market prices, combined with quotes, are used for the estimation of costs. Multiple sensitivity analyses are conducted to check the robustness of the results of the cost–benefit analysis. The results suggest that developing passive recreational activities is economically viable, and that, among the evaluated activities, the residents are willing to pay more for passive recreational activities demanding less effort (e.g., educational workshops and camping). Future economic studies can use these methodologies and transfer benefit estimates from this research to evaluate whether it is worth allocating money to implement passive recreational opportunities in land with high ecological value. In general, community-based forest management (CBFM) is sustainable over a long term if it meets environmental, social and economic objectives [27]. In the context of this paper, if the benefits of passive recreational practices do not exceed the associated costs, it is hard to convince managers or officials to pursue such practices even if there are positive environmental and social outcomes. Thus, assessing economic viability is crucial to develop and maintain sustainable practices.

2. Study Area

Puerto Rico is the smaller island of the Greater Antilles and consists of a primary island and other minor islands, forming an archipelago situated in the Caribbean region, with a total population of 3.2 million [28]. The Commonwealth of Puerto Rico (1952) has been a US territory since 1898. Forest cover on the main island increased substantially after the 1940s mainly due to agricultural abandonment and land management. In 2018, forest cover represented about 52% of the land area [29].

Puerto Rico is vulnerable to natural hazards, including earthquakes and hurricanes. For example, in September 2017 Puerto Rico was severely affected by hurricanes Irma and Maria. In 2020, Puerto Rico was impacted by a set of frequent earthquakes. Unfortunately, after the disasters, forest management has not been a priority. For example, some forests remained closed for two years after hurricanes Irma and Maria [30,31].

The RHCF is in the municipality of Mayagüez, which can be found in the western portion of Puerto Rico. The forest has a total surface area of 27.5 hectares (~68 acres). Figure 1 shows a map of Puerto Rico in the Caribbean and the geographic location of the RHCF. This secondary forest naturally regrew after the 1970s. Recently, the US Forest Service provided funding for a management plan, aiming to ensure better practices [1]. For example, land use changes or significant land alterations are prohibited in the forest and the process to develop the management plan took into consideration the residents’ needs and priorities. Yet, it lacks the analysis of the economic viability of incorporating recreational opportunities. This paper investigates the residents’ preferences and the economic viability for developing passive recreational activities in this region.

3. Materials and Methods

This section describes the procedures for data collection to estimate the results (Figure 2). The literature, input from stakeholders and an existing management plan were used to develop the choice experiment design. The questionnaire was tested in two focus group sessions held in the surrounding community in Mayagüez. Then, a pilot study with 30 participants was conducted to perform a preliminary evaluation of the survey. Finally, a refined version of the survey was administered among the residents around the RHCF, and the data collected were used for the estimation of recreational opportunities benefits. The benefits provided by recreational opportunities in the survey were subsequently compared with associated costs. The next subsections provide details of the conceptual structure shown in Figure 2.

3.1. Survey and Study Design

The survey instrument was divided into two main sections. The first applies to the choice experiment questions and the second was devoted to information about the residents’ sociodemographic characteristics, such as age, gender, education and income. Overall, the information obtained from the second section was used to build a profile of the sample surveyed and to better understand the responses to the choice experiments. At the end of the questionnaire, respondents were given an opportunity to provide comments about this study or to indicate any concerns about the overall management of the RHCF.

The questionnaire was distributed through in-person interviews by three interviewers who received training to reduce potential interviewer bias. Initially, we planned to administer the questionnaire to every other house in surrounding areas of the RHCF. However, during the data collection phase, we became aware of recent criminal activity near the forest. These illegal activities combined with the COVID-19 pandemic dampened the residents’ interest in participating in this study. Thus, we distributed questionnaires to all available households in the region. The data were collected in three phases: (1) Monday to Friday from 9:00 a.m. to 1:00 p.m., (2) Monday to Friday from 4:00 p.m. to 7:00 p.m. and (3) Saturday and Sunday from 10:00 a.m. to 3:00 p.m. This strategy allowed us to account for preference heterogeneity across the residents with different schedules during the week.

Two 2 h focus group meetings attended by 8–12 participants were held to test the questionnaire. Personnel of the board of directors of the RHCF and members of a local church helped with contact information to recruit focus group participants. To account for heterogeneous preferences among the residents, participants with different socioeconomic characteristics such as age, gender and education level were invited to the meetings. Focus groups were particularly useful for evaluating the survey length, vocabulary, clarity of questions, relevance of selected recreational attributes, maximum WTP and potential protest responses [6,32,33]. Focus group participants were allowed to make comments about the management or any other concern regarding the forest. The questionnaire was modified taking into consideration inputs from focus group participants.

We used a variety of control questions to evaluate responses to the choice experiments. Participants were asked to state reasons for supporting or not the hypothetical scenarios. We followed Bateman et al. (2002) and Tavárez and Elbakidze (2019) to identify protest responses [6,32]. Additionally, respondents were asked to state whether the questionnaire was difficult to understand, including the choice experiment exercise. To check the reliability of the results, a data analysis was conducted with all the responses and excluded both protest responses and surveys from respondents who stated that the questionnaire was very complicated. Respondents were also asked to rank in order of preference the recreational opportunities, which were compared with the choice experiment results.

3.2. Choice Experiment

The choice experiment is a commonly used method to obtain WTP estimates for multiple nonmarket goods and services [33,34], including recreational opportunities provided by forest ecosystems [6,8,14]. In choice experiments, respondents receive a predetermined list of tables (i.e., choice sets) composed of two or more hypothetical scenarios (i.e., alternatives) described by a set of attributes, and they are asked to select a preferred scenario. It is assumed that respondents will select the alternative that provides the best combination of attribute levels, i.e., combination of recreational activities, at given prices.

The selection of attributes and corresponding levels were based on the literature review [1,6], the input from the stakeholders and focus group consisting of the residents interested in outdoor recreation. We used the information from the stakeholders and the management plan of the RHCF to identify preliminary potential recreational activities that may be of interest to potential visitors. The selected recreational opportunities were validated in the focus group meetings. In addition to the recreational activities, a cost attribute was included to facilitate WTP estimates for each recreational activity, following the standard practice. The description of attributes and corresponding levels were provided to survey respondents at the beginning of the questionnaire (

Table 1. Description of attributes and corresponding levels.

Recreational Attributes	Attribute Description	Attribute Levels
Observation tower	Development of a wooden tower, approximately 30 feet high for forest observation and surrounding area.	Tower not available * Tower available
Educational workshops	Includes a variety of activities to promote residents’ interaction with the forest, such as crafts, community gardens, compost development, among other similar activities.	Workshops not available * Workshops available
Guided tour	Includes the services of a person with knowledge of the forest, such as its history, species, topography and climatology of the place. The tour guide also has knowledge about the sociodemographic information of the region.	Tour not available * Tour available
Camping	Area designated for camping within the forest. There will be two restrooms available to visitors. However, forest alterations to prepare these spaces will be minimal.	Camping not allowed * Camping allowed
Cost per visit	The amount of money you would pay per visit for each recreational improvement project option. The money paid must be considered as money spent that will not be used for other items.	USD 0 * USD 3 USD 6 USD 9 USD 12 USD 15

* The attribute level describes the status quo alternative.

).

The choice experiment design included five attributes, four with two levels and one with five levels. The complete factorial design, which included all possible combinations of attribute levels, resulted in 80 possible alternatives (2⁴ × 5¹ = 80). Each choice set has two alternatives in addition to the status quo option where all attribute levels reveal the base scenario or current situation (Figure 3). A complete randomized design would include too many choice scenarios [34,35], increasing the burden on respondents and affecting data quality. Thus, an orthogonal fractional factorial design was employed using Sawtooth software (Lighthouse Studio 9.14.2). We used the orthogonal fractional factorial design due to its convenience and common use in choice experiments [36,37,38,39]. Although past studies have used 16 choice sets per respondent [40], or even 30 choice sets per respondent [41], we restricted the number of choice sets per respondent to 12 to avoid a potentially excessive burden and ensure high-quality data.

3.3. Theoretical Framework for Choice Experiments

The choice experiment method is grounded in Lancaster‘s consumer theory [42], which claims that respondents obtain utility from the goods’ attributes and not from the good itself. The econometrics are based on random utility theory [43]. The utility of individual n for good j can be decomposed into an observable and unobservable component ($U_{nj}=V_{nj}+{\epsilon }_{nj}$). It is assumed that participants would select alternative j over alternative i if the utility of alternative j is higher than the utility of alternative i ($U_{nj}>U_{ni}$). The probability of respondent n selecting alternative j over alternative i is as follows:

$$P_n\left(j\right)=(V_{nj}+{\epsilon }_{nj})>P_n\left(j\right)=(V_{ni}+{\epsilon }_{ni})\hspace{1em}\forall j \mathsf{ϵ} \mathrm{C}, j\ne i$$

(1)

Assuming the utility function is linear and additively separable [44], the indirect utility function can be expressed as follows:

$$V_{nj}={\alpha }_j+\beta X_j+\gamma Z_n+\delta C_j$$

(2)

where ${\alpha }_j$ is the constant term, or the alternative specific constant (ASC), that captures the preference for alternative j; β, γ and δ are coefficients; $X_j$ is the vector of all attributes; $Z_n$ is the vector of the interaction terms between the ASC and sociodemographic characteristics (SDCs) of respondents; and $C_j$ is the cost of alternative j. In this study, all recreational attributes are binary coded, and the cost attribute is continuous.

3.4. Estimation Models for Choice Experiments

We used three logit models for estimation of coefficients in the choice experiments [45]: a conditional logit model with main effects only, a conditional logit model that includes interaction effects and a random parameters logit model (RPLM) (We also employed a latent class analysis with two and three groups. However, the reported models provided a better model fit compared to the latent class models.). The conditional logit model with main effects only provides an initial assessment of preferences for choice experiment attributes. The conditional logit model with interaction effects allows determining whether individual profiles affect preferences and, subsequently, decision making in the choice experiment. The interaction effects are obtained by interacting the ASC, which takes the value of one if respondents selected a non-status quo alternative, with the SDCs of individuals. The RPLM is used to overcome some of the limitations of the conditional logit model, as explained in the next paragraphs.

The conditional logit model suffers some limitations, including the assumption of independence of irrelevant alternative (IIA) property and homogenous preferences, which may not be the case in empirical studies. We tested whether the estimates are independent of irrelevant alternatives following the test developed by Hausman and McFadden (1984) [46], which follows the chi-squared distribution with degrees of freedom equal to the number of coefficients estimated in the restricted model:

$${\mathsf{\chi }}^2={({\widehat{\beta }}_s-{\widehat{\beta }}_f)}^{\prime }{({\widehat{V}}_s-{\widehat{V}}_f)}^{-1}({\widehat{\beta }}_s-{\widehat{\beta }}_f)$$

(3)

where s denotes the estimators in the restricted subset after deleting one alternative, f is the estimator in the full set of alternatives, and ${\widehat{V}}_s$ and ${\widehat{V}}_f$ are the corresponding estimates of the covariance matrices. We carried out three tests by dropping one of the recreational project alternatives in each test. The results with five coefficients are reported in

Table 2. Test for independence of irrelevant alternative property.

Omitted Alternative	Chi-Square Value	Critical Value	Significant
Alternative 1	8.89	11.07	No
Alternative 2	17.59	11.07	Yes
Alternative 3	10.74	11.07	Yes

. The results show that the IIA property is violated when dropping alternatives two and three from the model but is not violated when dropping alternative one. To relax the IIA property and account for preference heterogeneity, an RPLM was employed. The RPLM is a widely used model in the choice experiment literature [6,14,34].

3.5. Cost–Benefit Analysis

The cost–benefit analysis is used to determine the economic viability of a project, such as new recreational activity-related projects. A project is said to be economically viable if the difference between the estimated benefits and estimated costs over the duration of a project is greater than zero [47]. The estimation of tangible costs like labor, materials and equipment is generally straightforward by using market prices and less costly than the estimation of the benefits of nonmarket goods and services. Quotes provided by private companies and market prices were used to obtain information on the cost of construction, materials and supplies, installation, maintenance, etc. The estimates of market size and benefits for nonmarket goods and services often require obtaining primary data through survey distribution, which is costly and time-consuming. For this study, the data on the benefits were based on the results from the choice experiments.

The benefit–cost ratio summarizes the overall relationship between the relative costs and benefits of a proposed project. The economic viability of a project can also be evaluated using benefit–cost ratios by assessing whether the ratio of estimated benefits over estimated costs is greater than one. However, the benefit–cost ratio also allows us to understand how much money is generated in the long term for each dollar invested in a project [16]. Thus, stakeholders and policymakers can prioritize the allocation of financial resources to recreational opportunities, conditional on the resulting benefit–cost ratios.

The cost–benefit analysis uses a discount rate to provide present values for all estimated future benefits and costs. The following equation was used to calculate present values [16]:

$$PV=FV/{\left(1+d\right)}^t$$

(4)

where PV is the present value, FV is the future value, d is the discount rate, and t is the year for which benefits (costs) are obtained (incurred). Unfortunately, the selection of the discount rate has influenced the results of the cost–benefit analysis and policy decisions in the past [47]. To overcome such limitation, we conducted a sensitivity analysis by allowing variations in the discount rate to verify the robustness of the results. In addition, we conducted a second sensitivity analysis by applying to the projected benefits the lower and upper bound of the 95% confidence interval of the monetary value that respondents were willing to pay for recreational activities [48]. In this context, the cost–benefit analysis supports policymakers and stakeholders in implementing new recreational activities aligned with the residents’ preferences and needs while also delivering new opportunities that support the sustainability of the RHCF.

4. Results and Discussion

4.1. Sociodemographic Information

A total of 168 residents completed the questionnaire from July to September 2022. However, eleven respondents did not fully complete the questionnaire. A total of 5652 observations were included in the regression models (157 respondents × 12 choice sets per respondent × 3 alternatives per choice set). Thirty-seven residents refused to participate in this study, resulting in a participation rate of 77%. Although this is a relatively high participation rate, it is lower than figures reported in Rivera-Acosta and González-Martínez (2020) and Tavárez and Elbakidze (2019) [6,12], who conducted similar studies in Puerto Rico and found participation rates above 90%. This result may be due to the COVID-19 pandemic. However, interviewers were informed about increased delinquency in the region, which probably also affected participation rate in this study.

We performed a power analysis following the procedure established by Bekker-Grob et al. (2015) and found that the minimum sample size needed for a statistical power of 80% at a 95% confidence interval was 245. Since the sample size in this study was 157, the resulting statistical power is 60%, suggesting that this study (when conducted repeatedly over time) is likely to produce a statistically significant result six times out of ten.

Table 3. Sociodemographic information of respondents.

Variables	Description	Mean (SD)	Median
Gender	Gender of respondent (1 = female, 0 = male)	0.55 (0.50)	1
Age	Age of respondent	60.31 (14.55)	64
Income	Total household income per month (1 = less than or equal to USD 500, 7 = more than USD 7000)	3.18 (1.66)	3
Education	Education of respondent (1 = none, 5 = graduate school)	3.36 (0.78)	3
Dependents	Number of households dependents	1.01 (1.24)	1

SD—standard deviation.

shows sociodemographic information of the surveyed respondents. Fifty-five percent of respondents are female, and the median age of respondents is 64. The average household income is USD 1001 to USD 2000 per month. The average education level among surveyed respondents is high school degree, with 32% of respondents completing a bachelor’s degree (not shown). The average number of dependents is 1.01. Overall, these figures are similar to the general population of Mayagüez. For example, the percentage of females in the region is 52%, median household income is USD 14,700 and percentage of the residents with a bachelor’s degree is 28%. However, the median age of the residents in this study is 64, which is higher than the median age of 50 in the Rio Hondo ward.

4.2. Choice Experiment Results

The sign and significance of coefficients are invariant across models (

Table 4. Regression results from choice experiments.

Variables	CLM	CLM with Interactions	RPLM
Observation tower	0.659 (0.070) ***	0.661 (0.070) ***	0.745 (0.109) ***
Educational workshop	0.989 (0.071) ***	0.989 (0.071) ***	1.112 (0.137) ***
Guided tour	0.677 (0.069) ***	0.679 (0.069) ***	0.776 (0.119) ***
Camping	0.777 (0.070) ***	0.778 (0.069) ***	0.867 (0.117) ***
Cost	−0.099 (0.009) ***	−0.099 (0.009) ***	−0.113 (0.016) ***
ASC	−0.222 (0.122) *	−0.383 (0.253)	−0.161 (0.230)
Income	-	0.182 (0.038) ***	-
Distance	-	−0.119 (0.058) **	-
Gender	-	−0.021 (0.116)	-
			Standard deviations
Observation tower	-	-	0.282 (0.489)
Educational workshop	-	-	0.737 (0.484) *
Guided tour	-	-	0.227 (0.413)
Camping	-	-	0.737 (0.438) *
Observations	6048	6048	6048
Respondents	157	157	157
AIC	3434	3412	3441

*** Significant at 0.01, ** significant at 0.05, * significant at 0.10. CLM—conditional logit model. RPLM—random parameters logit model. ASC—alternative specific constant. Standard deviations in parentheses.

), which can be used as a reliability check. The sign of the cost coefficient is negative, indicating that the probability of choosing an alternative decreases as the cost increases. This result is consistent with economic theory, which is also important for the reliability of the results. As expected, the sign of all recreational attributes is positive, suggesting that the recreational opportunities increase the probability of selecting an alternative. The standard deviations of the RPLM show heterogeneous preferences for educational workshops and camping facilities.

The interaction effect between the ASC and income is significant and positive, suggesting that households with higher incomes are more likely to support recreational opportunities in the RHCF. Higher household income implies higher disposable income that may be allocated to leisure-associated activities, such as recreation opportunities. The interaction effect between the ASC and distance is significant and negative, which indicates that the greater the distance from the RHCF, the lower the probability of supporting the recreational opportunities. This outcome is expected as the residents near the forest obtain more benefits than the residents living farther away. For example, the residents near the forest can receive benefits from scenic views, birdwatching and noise reduction. The interaction effect between the ASC and gender is not significant. Other SDCs of respondents like education and age were not included in the model due to a high degree of multicollinearity.

Marginal WTP for recreational activities can be estimated using parameters from conditional and random parameters logit models. Marginal WTP can be estimated using the negative ratio of the coefficient of the attribute of interest and the cost coefficient [34,35]. The residents are willing to pay USD 6.63, USD 9.93, USD 6.82 and USD 7.81 for the observation tower, educational workshops, guided tours and camping, respectively (

Table 5. Willingness to pay (USD) for recreational opportunities.

Recreational Activities	CLM	CLM with Interactions	RPLM
Observation tower	6.66 (4.93–8.41)	6.63 (4.91–8.36)	6.58 (4.84–8.33)
Educational workshop	9.99 (7.88–12.10)	9.93 (7.84–12.02)	9.83 (7.72–11.94)
Guided tour	6.84 (5.09–8.59)	6.82 (5.08–8.56)	6.86 (5.11–8.62)
Camping	7.86 (6.01–9.70)	7.81 (5.99–9.64)	7.67 (5.80–9.54)

Note: 95% confidence intervals (in parentheses) were calculated following the Krinsky–Robb method [49].

). Although the activities are not the same, our results are similar to WTP values for recreational opportunities in forest ecosystems reported in prior studies [8]. Our results are lower than WTP estimates for other recreational opportunities in the San Patricio Urban Forest of Puerto Rico reported by Tavárez and Elbakidze (2019), who found higher estimates (up to USD 39) using a one-time payment. However, our sample reported willingness to visit the site for recreational opportunities four times a year on average. This implies, for example, a WTP of USD 39.72/year for educational workshops and USD 31.24/year for camping. Thus, after correcting stated demand, our results are higher than the estimates reported by Tavárez and Elbakidze (2019).

Rivera-Acosta and González-Martínez (2020) found that the residents are willing to pay USD 10 for recreational activities in the Finca Montaña forest in Puerto Rico, and these estimates are higher than the values reported in this study. Nonetheless, they examined the residents’ WTP for a bundle of recreational opportunities, such as zip-lining, mountain biking, birdwatching, educational activities and hiking. In this study, the residents are willing to pay USD 16.69 for educational workshops and a guided tour. Our estimates are thus higher than those reported in their study.

4.3. Cost–Benefit Analysis

We identified multiple educational workshops that could be offered in the RHCF, including composting, community garden, aromatic candles and craft-related activities. The composting workshop can be offered in two or three hours. The two-hour workshop includes educational information through brochures and materials for the activity and has a cost of USD 5. The three-hour workshop includes theoretical information via a presentation and materials for the activity and costs USD 20. According to the choice experiment results, the residents are willing to pay USD 9.93 for workshops in the RHCF. Therefore, the two-hour workshop is economically viable, resulting in a benefit–cost ratio of 1.99. Conversely, the three-hour workshop is not viable, even with the WTP estimate in the upper end of the confidence interval (USD 12.02). For the community garden workshop, there are also two options. The first option costs USD 1 and includes theoretical information only. The second option includes theoretical information and materials for the activity with a cost of USD 35. Using the choice experiment results, the first workshop is viable; however, the second workshop is not viable even with the WTP estimate in the upper end of the confidence interval. Nonetheless, the first option may not be aligned with the residents’ expectations, as they were requested to express a WTP for hands-on workshops. The aromatic candle workshop presents various alternatives: all materials included (USD 22, 20 participants required); materials, an instructional manual and meals (USD 30); and all materials included (USD 20, 25 participants required). These alternatives are not economically viable as workshops. Craft activities cost USD 8, which includes simple materials like sticks, papers, scissors, crayons and watercolors. This option is economically viable and has a benefit–cost ratio of 1.24. However, if using the WTP estimate (USD 7.84) in the lower end of the confidence interval with a benefit–cost ratio of 0.98, this option is not economically viable.

We used available information from locally guided tours in agritourism, ecotourism and visits to historical sites. All tours included information about the site and a walk through the site. The price of the guided tours ranged from USD 8 to USD 12 per person for up to 4 h. Although we identified all-day tours for USD 75 and USD 95 in national parks, including meals and transportation, we excluded those tours from our analysis as they do not fit within the framework of this study. The choice experiment results and cost estimates suggest that guided tours are not economically viable. But if using the WTP estimate in the upper end of the confidence interval (USD 8.56), the USD 8 tour option turns out to be economically viable.

The camping area and observation tower are long-term projects and require further analysis. The respondents reported that, on average, they are willing to visit the RHCF four times a year to enjoy the recreational activities offered. We used this information to project the total benefits provided by recreational opportunities. The residents are willing to pay USD 26.52 and USD 31.24 a year for the observation tower and camping facilities, respectively. At a 6% discount rate, this represents a total benefit of USD 106 and USD 125 per resident over a period of five years for the observation tower and camping, respectively. The estimated total population in the region is 3640 residents. However, eighty-five percent of the residents are adults (US Census, 2021) and only seventy-four percent of respondents were willing to pay for the recreational opportunities. By extrapolating this figure, we can argue that at least 2290 adult residents are interested and willing to pay for the activities. Therefore, the benefits for the region’s population at a 6% discount rate and over a period of five years are USD 210,398 and USD 247,845 for the observation tower and camping facilities, respectively. It is worth noting that these estimates represent a lower bound, as residents from other regions can visit the forest for passive recreation purposes.

We used prices from local companies to estimate the cost of constructing camping facilities and the observation tower in the RHCF. For the camping facilities, we mainly explored the construction of two public restrooms near the area designated for camping. Cost varied depending on the size and the inclusion of a septic tank. The cost without a septic tank is USD 20,000 per restroom, including labor salaries. The cost of construction per restroom facility, including septic tank and labor is USD 30,000. Two restrooms with one septic tank cost USD 50,000. Commercial bills for utilities (water and electricity) are estimated at USD 700/month. Salaries for maintenance staff are estimated at USD 34 per day (4 h workday at USD 8.50/h), which corresponds to the minimum wage in Puerto Rico. Associated cleaning costs are estimated at USD 1200/year. Maintenance costs related to wood repair and varnish of the camping facilities are estimated at USD 2000 every 2 years. As for building the observation tower, prices vary depending on how sophisticated the tower is. We selected a relatively basic design that fulfills the objective of facilitating an appreciation of scenic beauty. The construction cost of the 30-foot wooden observation tower is estimated at USD 80,000; the maintenance costs associated with wood repair and varnish are estimated at USD 3500 every 2 years.

The opportunity cost represents the loss of potential gains from other alternatives when one alternative is adopted. Currently, land use changes are prohibited in the RHCF. Therefore, the opportunity cost of alternative land uses is nearly zero. However, there is an opportunity cost associated with the money spent on recreational opportunities. For example, the revenue lost from interest can be calculated as the opportunity cost of each recreational activity developed. We used 4% as the interest rate, which corresponds to the interest paid by credit unions. In addition to the opportunity cost, we also included an insurance fee for the duration of the project. Insurance is particularly important in Puerto Rico, a region highly vulnerable to natural disasters. A local insurance company provided estimates of an insurance policy that covers damages caused by hurricanes and earthquakes. Although the construction of the camping facilities is less costly than the observation tower, labor, utilities and maintenance costs make the observation tower less costly in the long term, even though the observation tower has a higher opportunity cost and insurance fee associated with the policy (

Table 6. Benefits and costs (USD) of the recreational opportunities at a 6% discount rate.

Recreational Activities	Year 0	Year 1	Year 2	Year 3	Year 4	Total
Benefits
Observation tower	0	57,282	54,040	50,981	48,095	210,398
Camping	0	67,477	63,658	60,054	56,655	247,845
Costs
Observation tower	80,000	3396	6715	3023	6372	99,506
Camping	50,000	21,745	24,830	19,353	24,634	140,563

). The results indicate that developing both recreational opportunities is economically viable. The benefit–cost ratio is 2.11 and 1.76 for the observation tower and the camping facilities, respectively, suggesting that the observation tower provides greater benefits per dollar.

The selection of the discount rate influenced the results of the cost–benefit analysis and policy decisions in the past [47]. We conducted a sensitivity analysis by allowing variations in the discount rate to account for uncertainties in future cash flow. We also conducted a second sensitivity analysis by applying to the projected benefits the lower and upper bound of the 95% confidence interval of the monetary value that respondents were willing to pay for recreational activities [48]. Both recreational opportunities are economically viable under all scenarios (

Table 7. Benefit–cost ratios over a five-year period at different discount rates.

Recreational Activities	3% Discount Rate	6% Discount Rate	9% Discount Rate
	Mean WTP
Observation tower	2.25	2.11	1.99
Camping	1.86	1.76	1.68
	Lower bound of WTP confidence interval
Observation tower	1.66	1.57	1.48
Camping	1.42	1.35	1.28
	Upper bound of WTP confidence interval
Observation tower	2.83	2.67	2.51
Camping	2.29	2.18	1.07

). The observation tower provides the highest benefit–cost ratio, indicating the investment in this activity is the most effective.

4.4. Discussion

Results that comply with validity tests for reliability are crucial for stated preference-based studies. In this study, we found that the cost coefficient in the choice experiments is negative, and the income coefficient is positive; both findings are consistent with the economic theory. Traditionally, WTP studies use these two findings as indicators of validity in stated preference-based studies.

This study finds that developing all recreational opportunities is economically viable over the entire duration of the project, yet some caveats should be discussed. First, the costs associated with the first year of the long-term projects (observation tower and camping facilities) are higher than the corresponding benefits. In fact, profits would not be observed until 3 years after the implementation of the projects. This result suggests that some initial incentives are needed to implement the recreational activities in the RHCF. Second, recreational opportunities costs were calculated following specific details. However, the cost of constructing the observation tower and camping facilities, for example, vary depending on the type of construction and specifications.

The board of directors of the RHCF pointed out that the observation tower is very attractive to the residents and that it is probably one of the best projects to develop in the forest, which is consistent with the surveyed residents’ preferences, based on the results from this study. However, the focus group participants prioritized other recreational activities. This finding indicates that relying on a single source for survey design or distribution may be problematic in some contexts. A reason for this outcome could be that, although minimal, the construction of the observation tower requires human intervention in forested lands, whereas the other recreational opportunities, including camping facilities, are expected to be provided in nonforested lands (grass). It may be possible that the focus group participants discounted the environmental degradation, i.e., reduction in ecosystem service provision, in the valuation exercise.

The current management plan of RHCF shows that promotion of visits is desirable, as well as the creation and enhancement of recreational activities [1]. At this time, the administration does not allow any infrastructure construction on site; however, the results from this study indicate that current and potential users would prefer the alternatives, such as building an observation tower and camping facilities. Incorporating preferences from an evidence-based study, as this research, will benefit future management plans. Successful conservation of the RHCF requires both protection of ecological assets and community engagement to sustain efforts and co-management. Therefore, plans should represent common recreational interests and address the needs.

The results from this study show that the residents are willing to pay more for educational workshops and camping opportunities (Table 5), which are the two activities that require less physical effort. Passive recreational opportunities allow a broader range of visitors to benefit from forest ecosystems and, therefore, can encourage multiple groups of residents to become more involved in forest management. The inclusion among groups of residents provides feedback to stakeholders and policymakers, which contributes to forest sustainability [24]. Achieving sustainable management is more plausible if multiple groups of residents receive benefits from an action, such as passive recreational development. In contrast, the lack of contribution of local communities may undermine continuation of forest management plans [50].

It is worth noting that crime activities may affect visitation rates for recreational activities. Therefore, benefit estimates for the region’s population, especially those related to long-term projects, need to be re-estimated in ways that represent other scenarios that may affect the demand for recreational activities. Future studies must consider these issues if crime continues to rise.

5. Concluding Remarks

Forests provide multiple benefits to society in the form of ecosystem services, including space for recreational opportunities. This study uses a cost–benefit analysis to determine the economic viability of implementing an observation tower, educational workshops, guided tours and space for camping with restroom facilities in the RHCF of Puerto Rico. The results show that recreational opportunities in the RHCF are economically viable. The benefit–cost ratios suggest that the observation tower and camping facilities are the two recreational opportunities with the highest benefit per dollar, and thus it is more feasible to implement these activities and that they remain over time.

This study contributes to the literature in two different ways. First, most published studies on WTP for recreational opportunities focus on estimating the potential benefits or income generated from the activities, but most of them do not examine the economic viability of implementing the projects and case studies in developing countries are few. Without an economic assessment of proposed activities, it is hard to convince authorities to implement or continue them, even if they bring positive environmental and social outcomes in sustainable management plans. Second, the most common outdoor recreational activities studied in forest ecosystems include hiking, mountain biking, picnicking and scenic beauty observation. We examined different passive recreational activities of interest to the residents aligned with local regulations.

Future programs aiming at implementing the passive recreational activities examined in this study need a wider discussion with the residents. It is advised to organize meetings to discuss specific aspects of recreational activities that are of interest to the residents. For example, the specific location of the observation tower, date and hours of guided tours or additional features for the camping facilities can be discussed in public meetings. Collaborative participation can enhance forest sustainable management plans.

The methods used in this paper are generalizable to forests with high ecological value in other locations (local or international). For example, the WTP and cost–benefit methods can be used to analyze the economic viability of developing other recreational opportunities in the region, such as zip-lining and mountain biking, which were identified as relevant during the focus group sessions. Significant alterations to the RHCF are prohibited by an agreement with the US Forest Service. However, recreational opportunities with notable alterations to land use could be developed in other forests of the region. To develop and implement sustainable management plans, environmental, social and economic objectives must be met; therefore, assessing economic viability is important and necessary.

This study was conducted with 157 respondents. Although choice experiment studies have been previously conducted with fewer observations [38,51], the sample size may be considered relatively low for choice experiments. During the data collection phase, we became aware of recent crime activities in areas near the RHCF. These illegal activities combined with the COVID-19 pandemic dissuaded the residents from participating in this study.