Valuation on the Frontier: Calibrating Actual and Hypothetical Statements of Value

Abstract

The lack of robust evidence showing that hypothetical behavior directly maps into real actions remains a major concern for proponents of stated preference nonmarket valuation techniques. This article explores a new statistical approach to link actual and hypothetical statements. Using willingness-topay field data on individual bids from sealed-bid auctions for a $350 baseball card, our results are quite promising. Estimating a stochastic frontier regression model that makes use of data that any contingent valuation survey would obtain, we derive a bid function that is not statistically different from the bid function obtained from subjects in an actual auction. If other data can be calibrated similarly, this method holds significant promise since an appropriate calibration scheme, ex ante or ex post, can be invaluable to the policy maker that desires more accurate estimates of use and nonuse values for nonmarket goods and services. Benefit-cost analysis remains the dominant paradigm used throughout the public sector. Yet, a recurring issue in properly estimating the total benefits of nonmarket goods and services is whether hypothetical statements map into real behavior. Some recently published studies provide evidence that suggests important differences exist between responses from real and hypothetical valuation questions. 1 This observation triggered a search for an ex ante and ex post procedure to correct the systematic bias between intentions and actions in valuation exercises. Recent technology using ex ante procedures has produced some strong evidence that hypothetical bias can be overcome by using a “cheap talk” scheme (see, e.g., Cummings and Taylor; List). Ex post calibration has also shown some signs of promise, as work due to Blackburn, Harrison,