Willingness to pay (WTP) for payments for environmental services (PES) can be temporarily reliable if contingent valuation (CV) studies are embedded in an accurate survey population, yield low measurement errors, and are based on a correct assumption of no change in sociodemographic factors affecting buyer preferences. These pre-conditions are assumed in PES schemes applying temporal reliability of WTP. This study tests these conditions in CV-PES studies from 2001, 2003, and 2011 in Lombok, Indonesia, by comparing them with a new CV-PES study in 2015. Our results show that the CV-PES studies would not meet the pre-conditions due to inclusion of non-PES buyers, potential measurement errors implied by a lack of validating information and high WTP estimate, and/or failure to test the condition of no change of socio-economic factors. Results contribute to identifying pragmatic challenges and lessons for applying temporal reliability of WTP to PES implementation.