The Specification of Price in Studies of Consumer Demand under Block Price Scheduling: Comment
评论Polzin（1984）关于阶梯定价下价格变量设定的实证研究，指出其使用加总数据的问题，主张用个体数据比较不同价格设定的效果。
The problems of price variable specification in demand models of commodities priced by block schedules are well known and in 1975 Taylor suggested a simple, but yet substantially correct procedure . This procedure, when modified by Nordin (1976), required both marginal price and an expenditure difference variable to be included in demand models. Unfortunately, it is often difficult to obtain data for these variables and much debate, but little evidence, about the empirical consequences of misspecification has resulted. In this comment we focus on Polzin's (1984) recent attempt to expand the empirical evidence. We find several problems with the evidence presented by Polzin and as a result we are unsatisfied with his conclusion that, despite its theoretical and conceptual elegance, the Taylor/Nordin (price) specification may not be the best description of consumer behavior when faced with a block-rate price schedule (1984, 309). The major problem with Polzin's approach stems from his use of aggregate data. We argue that comparisons of alternative price specifications must be based on disaggregate data for actual as opposed to representative or households. Some of the problems with data for typical households in the case of block pricing can be demonstrated by assuming two consumers, A and B, reacting to the hypothetical natural gas rate structure below: