Benefits transfer of a third kind: an examination of structural benefits transfer: George Van Houtven, Subhrendu K. Pattanayak, Sumeet Patil, and Brooks Depro
Introduction The policy community frequently uses benefits transfer methods because they offer a practical and low cost way to provide benefit estimates for benefit-cost analyses, natural resource damage assessments, and other natural resource policy and management analyses. These methods take and adapt results from existing primary valuation studies and apply them to assess the benefits of selected policy changes. For the most part, benefits transfer approaches fall into two categories – “unit value” transfers or “value function” transfers – where the key distinction between the two approaches is the degree to which differences between the study and policy contexts are formally accounted for in the transfer. In unit value transfers, a single value or range of values, such as the value per recreation day or per unit change in water quality, is usually transferred with little or no adjustment for differences between the two settings. With benefit, or value, function transfers, information from existing studies is used to identify a functional relationship between the value of interest and the factors that may influence the magnitude of the value (e.g. using meta-regression analysis). This functional relationship allows the analyst to account for differences between the two settings and adapt the transfer estimates accordingly. One of the main limitations of these traditional approaches is that they do not always make full and best use of existing evidence from stated preference and revealed preference studies. To the extent that they do use evidence from both revealed preference and stated preference studies, they typically combine them in an ad hoc manner. That is, they do not explicitly impose consistency with the economic theory that is assumed to underlie the value estimates drawn from different nonmarket valuation techniques. To address these limitations, a third kind of benefits transfer – “structural benefits transfer” (or “preference calibration”) – has been proposed in which the transfer methodology is directly tied to utility theory via the preference structure (Smith et al. 2002; Bergstrom and Taylor 2006). Structural benefits transfer is in essence a form of benefit function transfer; where the functional form is specifically derived from an assumed utility function. Although this third approach has
the potential to improve and strengthen benefits transfers, it has thus far only been applied and evaluated in a limited number of examples. The purpose of this paper is to further examine and evaluate structural benefits transfer as an alternative transfer method, by extending existing applications in two main directions. First, using a single set of benefit estimates from the nonmarket valuation literature, we apply the preference calibration approach using several different utility function specifications, and we compare their implications for predicting benefits. Through these applications we show that the calibration approach can be generalized, in that a distinct set of preference parameters can be calibrated for each alternative specification. We also show that the calibration framework provides a number of opportunities for comparing the internal consistency and plausibility of the different assumed preference structures. For a given set of benefit estimates reported in the literature, these checks allow us to evaluate which specifications are most suitable for representing the conditions that produced the estimates. Second, whereas existing applications have focused on use-related values for environmental improvements, we explicitly include nonuse values in the preference specifications. This addition is particularly important for combining evidence from revealed and stated preference methods. Although value estimates from the two methods may be derived with the same underlying preference structure, only stated preference estimates – in this case, value estimated drawn from a contingent valuation study – will shed light on the nonuse component of this structure. This chapter begins in the next section by providing a background discussion of the structural benefits transfer approach. The third section then introduces and describes the preference specifications that will be applied, and subsequent section discusses how estimates from different nonmarket valuation methods can be directly linked to these preference specifications. Following that, we present a case study application focusing on water quality changes using the five preference specifications. The results and implications of these applications are then discussed in the final section along with suggested directions for future research.