Forest related ecosystem services provide benefits at multiple geographic scales (i.e. watersheds, habitat, and carbon sequestration), motivating the development of payment for ecosystem services (PES) programs, which provide monetary incentives for forest conservation. In this study, we apply a dichotomous choice, contingent valuation study to investigate stated willingness to accept (WTA) for access to forest resources amongst residents of Kam Cha i, Thailand. We consider the endogeneity of perceived consequentiality by estimating a joint bivariate probit model of consequentiality and WTA. Compared to a subset of consequential respondents, we find upward bias in mean WTA for decreased access to forest resources. In the standard probability model of WTA, the inclusion of a consequentiality dummy variable magnifies this bias. This indicates that failure to account for the endogeneity when conditioning using the consequentiality dummy variable may in fact exacerbate bias in some cases. Our joint bivariate probit model of consequentiality and WTA reduces mean WTA, results which are in line with the subset of consequential respondents. Overall, our evidence supports the hypothesis of positive hypothetical bias in studies of WTA for reductions in access to public goods.