Design and Analysis of Contingent Valuation Surveys Using the Nested Tobit Model

Abstract

Contingent valuation surveys have become an important tool in placing monetary values on non-market goods and amenities. Many policy issues involve evaluation of several alternatives such as different environmental quality levels, different levels of risk, etc. Contingent valuation then involves asking the respondent a sequence of nested questions. Asking and analyzing a nested sequence of questions is an efficient approach to data gathering and preference revelation. The resultant sequentially censored data set cannot be efficiently analyzed with the standard regression models like the Tobit or nested logit models. The nested Tobit model is proposed as an efficient and consistent method of estimating regressions using sequentially censored data. An empirical application suggests greater efficiency in comparison to the Heckman two-stage procedure. Copyright 1994 by MIT Press.