Eliciting and Utilizing Willingness to Pay: Evidence from Field Trials in Northern Ghana

Abstract

We utilize the Becker-DeGroot-Marschak (1964) mechanism to estimate the willingness to pay for clean drinking water technology in northern Ghana. The BDM mechanism has attractive properties for empirical research, allowing us to directly estimate demand, compute heterogeneous treatment effects, and study the screening and causal effects of prices with minor modifications to a standard field experiment setting. We demonstrate the implementation of BDM along these three dimensions, compare it to the standard take-it-or-leave it method for eliciting willingness to pay, and discuss practical issues for implementing the mechanism in true field settings.

One‐Off Subsidies and Long‐Run Adoption—Experimental Evidence on Improved Cooking Stoves in Senegal

“…The reasons likely relate to very high transaction costs on rural markets that have to be added to the in‐town market price. Likewise, while generally the BDM method is found to be a very accurate reflection of the true WTP (see Berry, Fischer, and Guiteras 2019), one might speculate that the WTP observed in our set‐up is higher than what people reveal on regular markets. More specifically, due to the door‐to‐door marketing feature that is implicit to the BDM method, especially male household members (oftentimes the financial decision makers) probably dedicated more attention to the offer than they usually do to regular market offers.…”

“…There is thus no indication that the BDM elicitation approach imposed unrealistic cognitive demands, a common problem with stated WTP approaches for environmental non‐market products (Gregory, Lichtenstein, and Slovic 1993). The WTP elicited by BDM is widely seen as a precise approximation of a real‐life WTP because of its incentive‐compatible features (see Berry, Fischer, and Guiteras 2019 for a discussion of the BDM method)…”

Demand for self‐tests: Evidence from a Becker–DeGroot–Marschak mechanism field experiment

“…Existing BDM studies that report the distribution on which they made the random draw typically used a range of prices for the random draw that is broadly similar to our range. For example, Berry et al (2015) drew prices from a discrete uniform distribution between US$0 and approximately US$8.60. In a seemingly poorer study setting than ours, Adams et al (2016) where WTP i is the WTP by individual i, X i is the correlate of interest (e.g., primary school completion), Z 0 i is a vector of controls (e.g., in all specifications, we include indicator variables for the three demographic groups on which we stratified our sampling and for the facility at which the respondent was recruited), and ε i is an idiosyncratic error term.…”

“…Over 83% of individuals we approached to participate completed the BDM auction, suggesting our sample may be at least somewhat representative of facility attendees. 2 As in previous BDM field experiments (e.g., Adams, Lybbert, Vosti, & Ayifah, 2016;Berry, Fischer, & Guiteras, 2015;Hoffmann, Barrett, & Just, 2009), participants completed a practice BDM auction for a small token item using funds provided by the surveyor. 3,4 Participants then bid on HIV self-tests in the main BDM auction.…”

“…Fourth, we appear to provide the first evidence on demand for self-tests (or testing of any kind) that simultaneously is incentive compatible and exact, bolstering the reliability of our findings and offering insights on demand without restricting the range over which demand is estimated. 12,13 In doing so, we expand the small body of BDM field experiments measuring demand in poor countries, including studies of bed nets (Hoffmann et al, 2009), nutritional inputs (Adams et al, 2016;Alphonce & Alfnes, 2017;Meenakshi et al, 2012), and water filters (Berry et al, 2015).…”