Willingness-to-pay estimation using generalized maximum-entropy: A case study

• To analyze the willingness to pay, latent variable models are often used.
• In many cases, it is difficult to specify the error distribution of these models.
• The generalized maximum entropy (GME) method was used as a solution.
• The GME method proved robust to misspecification of the error distribution.

Estimation of potential customers' willingness-to-pay provides essential information for setting the price of new products. When no market data are available, one usually has to resort to customer surveys. To avoid biases encountered when directly asking respondent how much they would be willing to pay for some products, a useful strategy is to propose some tentative prices and ask the customers whether they would agree to buy the product at those prices. The resulting data can then be analyzed using latent variable models. However, it is often very difficult to specify the error distribution for such models. In this paper, we investigate the use of generalized maximum-entropy (GME) approach as a solution to this problem. Using simulations, this method is shown to be robust to misspecification of the error distribution. As an illustration, the approach is then applied to the determination of the entrance fee to the Royal Park Rajapruek in Chiang Mai, Thailand.