Predicting the performance of conservation tenders when information on bidders's costs is limited

Buying environmental services from private landholders using tendering mechanisms are usuallysubject to a budget constraint. Auction theory has mostly focused on target-constrained auctionsand is not well developed for this type of auction. This paper examines the predictive capacity of asimple model developed for budget-constrained tenders, already used to design new conservationprograms, by submitting it to controlled lab experiments. We study the capacity of the model topredict both experimental bids and the performance of the auction institution, based on the kind oflimited information typically available to a conservation agency. We conclude there exists anoptimal level of information on bidders’ costs, neither too large nor to small, making the tenderworth considering as a policy option as well as allowing an ex-ante assessment of its economicperformance.