Exact welfare measurement: theory and application to recreation economics

Graduation date: 1986Although the limitations of consumer surplus have become widely known, there exists a lack of studies which\ud present in a coherent framework recent developments in estimating Hicksian welfare measures. This study attempts to fill partially this gap by analyzing some theoretical and empirical aspects in the estimation of exact welfare measures using market data. Analytical expressions of the "exact" welfare measures, compensating variation (CV) and equivalent variation (EV), for some frequently used single demand equations as well as flexible functional form demand systems are presented. The different magnitudes of benefit estimates generated for the same valuation problem have been a source of important criticism of market related valuation studies. This study compares different functional forms consistent with the restrictions required for the aggregation of consumers and commodities. The results for steelhead rivers in Oregon indicate that the definitional differences in welfare measure (ordinary versus compensated measures) are quite small, which is consistent with the small budget shares that these activities account for a representative household. The results also suggest that the welfare measures may be very sensitive to the formulation of the model. However, the almost ideal demand system (AIDS) and translog indirect utility function demand system (TLOG) produce very close welfare estimates when catch rate, trips and a composite\ud good are considered the commodities yielding utility. The treatment of endogenous quality variables has been\ud a widely discussed issue in the recreation economics and labor economics literatures. In this study, unlike that of Bockstael and McConnell (1981), the demand system is specified as a function of exogenous parameters rather than endogenous implicit prices, avoiding the simultaneity of the demand system and the identification problem. Our empirical results support the comparative static predictions on the price and income elasticities for catch rate per trip and number of trips. A limitation of this study is the restrictive form assumed for the production technology. However, following Strong (1984), our model can be extended to take into account substitution among inputs, non constant returns to scale, and exogenous quality variables. Future research should be directed to study the properties of "exact welfare measures associated with different flexible functional forms. For large economic welfare changes, a global approximation to the underlying utility function as provided by Fourier series might be preferable. The comparisons among different model specifications should be done using rigorous statistical methods including the construction of confidence intervals for the welfare estimates