Enhancing the effectiveness of the policy design of Thailand's national parks using random utility model: case study of Khao Yai National Park

Abstract

This dissertation consists of three self-contained essays dealing with three separate issues, based on the policy objectives and the methodologies used to approach them. The ultimate goal is to provide policy makers the necessary information for better informed decisions when designing environmental policies in the context of national parks (where Khao Yai National Park is used as a representative).

The first essay applied RUM model to analyze Thai recreationists‟ site choices for ten study waterfall sites in and around Khao Yai National Park. The panel trip data on previous waterfall recreation trips collected on intercept respondents at six intercept sites was used in analysis. In addition to the trip costs, three broad classes of site characteristics: physical characteristics; quality/quantity levels of facility provision; and recreational activities, were included in the model as they were hypothesized to have impacts on recreationists‟ site choices. The analysis focused on day-trip recreationists who employed personal car in travelling to the waterfall sites. An estimator for “intercept and follow” sample proposed by McFadden (1996) was applied to correct the sampling bias arising from choice-based sampling. Two major conclusions were drawn from empirical results. First, the site characteristics suggested by Thai literatures were all significant to recreationist‟s site choice but not all of them are preferred by recreationists. Second, the choice-based sampling data introduced substantial bias onto the estimates of the coefficients and, consequently, the welfare effects. The results of sensitivity analysis indicated the complexity in anticipating the effects of applying wrong population proportions

The second essay aimed to design a desirable Payments for Environmental Services (PES) program to be implemented in Bu Pram sub-district in Prachinburi Province in eastern Thailand. Landowners' preferences on program factors were both identified and quantified using a choice experiment. The results showed that all hypothetical programs attributes were statistically significant to landowners‟ participation decisions, except free-to-choose in-kind benefits. On average, landowners preferred the land use option that allowed them to cultivate and manage the enrolled land under the program. The length of contract adversely affected landowners‟ participation in the hypothetical PES program. The other three program attributes of unrestricted amount of minimum land to enroll into the program and two in-kind benefits (advisory services and ecotourism-related job training) had positive impacts on landowners‟ participation decisions. Willingness to accept (WTA) calculations suggested that higher monetary incentives could help induce landowners to participate in PES programs that would generate desirable environmental benefits. Moreover, the WTA estimates showed that in-kind benefits could enhance the attractiveness of a PES program and significantly reduce the need for monetary incentives.

The third essay aimed to validate the use of stated preference (SP) data collected from self-administered internet survey in modeling recreational demand. Variety of tests and measures were conducted to test whether the internet SP data would yield consistent information with in-person interview SP data. The probabilistic conditional logit model was used to analyze waterfall site choices of daytrip recreationists. For preference homogeneity test, the underlying preference structure of the internet SP data was not statistically different from that observed from the RP data, whereas the underlying preference structure observed from the SP data – which was a part of the RP survey – was not always the same as that observed from the RP data. For predictive ability test, variety of tests and measures indicated that the in-person interview SP data was not superior to the internet SP data. In contrary to what is often assumed, the findings of this study suggested that the data quality of SP study administered in-person was not always superior. Although the findings might have been confounded by different sample frames, the study indicates a great potential of internet survey as an alternative survey mode for the hypothetical study of recreational demand.