Expected Consumer's Surplus and the Welfare Effects of Price Stabilization
证明预期补偿变差不能有效衡量价格稳定政策的福利效果，因为它无法反映消费者风险态度，仅在严格偏好条件下才适用。
Suppose a consumer faces a price which is uncertain ex ante. A question which often arises is whether the consumer would benefit from a stabilization policy under which the price would then be fixed and known with certainty. Many authors have addressed this question by calculating the expected compensating variation of the price change; others have used expected Marshallian consumer's surplus, often with the explanation that it is a good approximation to expected compensating variation.2 An important unresolved issue, however, is whether expected compensating variation does in fact provide a valid ranking of stochastic prices against stabilized prices, i.e., a ranking in agreement with that given by expected utility. In Section 2, we demonstrate that expected compensating variation is not in general a valid utility indicator. In essence, while utility comparisons under uncertainty require the use of the cardinal properties of (von NeumannMorgenstern) utility functions, the expected surplus measures depend only on ordinal preference rankings. Expected compensating variation, then, is completely insensitive to, and cannot reflect, the precise factor of interest in suchstudies: the consumer's attitudes toward risk. We then derive a set of restrictions on the utility function which are necessary in order for expected compensating variation to be a valid measure. We see that price stabilization policies can only be evaluated with compensating variation if consumer preferences are assumed to satisfy quite stringent requirements. This is in constrast to the familiar certainty case in which compensating variation always provides correct rankings.