Estimating Demand with Consumer Heterogeneity: an Application to Wholesale Price Regulation in Retail Gasoline Markets

这个研究项目的目的是双重的。第一个目的是为当前关于消费者需求离散选择模型中消费者偏好规范的文献做出贡献。该研究项目开发了零售汽油需求的随机系数离散选择模型，该模型允许具有不同产品特征的相关偏好的异质消费者类型。这种灵活性将允许在需求估计中使用诸如品牌忠诚度之类的偏好模式——这种偏好模式与独立和相同分布的特殊消费者偏好的假设不相容。然后，本研究计划将使用南加州零售站品牌标识的离散变化来测试哪种建模假设能够更准确地预测由外生品牌标识变化引起的均衡价格的实际变化。本研究的第二个目的是为分析批发价格管制和汽油行业竞争的潜在影响提供一个框架。州和联邦层面的政策制定者已经提议对批发价格进行监管，作为恢复汽油零售市场“竞争性”零售价格模式的一种手段。公平批发价格立法（FWP）将防止批发汽油供应商在零售加油站之间实行价格歧视。炼油商将被要求对所有零售商收取相同的价格，而不是像目前的做法那样，对不同的加油站收取不同的批发价格。本研究项目将使用上述离散选择需求估计的结果来模拟FWP立法的影响。这项模拟活动将用于回答以下与政策相关的紧迫问题：平均零售价格会因FWP而上涨还是下降？哪些市场因素会影响预期效应的方向和幅度？FWP是一项倒退还是进步的政策——它会导致处于不利社会经济地位的社区的价格上涨还是下降？是否有其他政策可以增加零售汽油市场的竞争，从而降低价格和减少价格差异？这项对FWP立法影响的详细而严格的实证估计将指导州和联邦一级的政策制定者，因为他们考虑立法旨在增加零售汽油市场的竞争。

The purpose of this research program is two-fold. The first purpose is to contribute to the current literature on the specification of consumer preferences in discrete choice models of consumer demand. The research program develops a random coefficients discrete choice model of demand for retail gasoline that allows for heterogeneous consumer types with correlated preferences across product characteristics. This flexibility will allow for preference patterns such as brand-loyalty in the demand estimation - a preference pattern that is incompatible with an assumption of independently and identically distributed idiosyncratic consumer preferences. This research program will then use discrete changes in the brand identities of retail stations in Southern California to test which modeling assumptions are able to more accurately predict actual changes in equilibrium prices resulting from exogenous changes to brand identities. The second purpose of this study is to provide a framework for analyzing the potential impacts of wholesale price regulation and competition in the gasoline industry. Policy makers at the state and federal levels have proposed wholesale price regulation as a means to restore 'competitive' retail price patterns in retail gasoline markets. Fair Wholesale Price legislation (FWP) would prevent wholesale gasoline suppliers from price discriminating between retail stations. Instead of charging different wholesale prices to different stations, as is currently the practice, refiners would be required to charge the same price to all retailers. This research project will use the results from the discrete choice demand estimation developed above to simulate the impact of FWP legislation. This simulation exercise will be used to answer the following pressing policy-relevant questions: Would average retail prices increase or decrease as a result of FWP? What market factors influence the direction and magnitude of the estimated effect? Is FWP a regressive or progressive policy - will it lead to higher or lower prices in disadvantaged socioeconomic communities? Is there an alternative policy might increase competition in retail gasoline markets, leading to lower prices and decreased price dispersion? This detailed and rigorous empirical estimation of the effects of FWP legislation will guide policy makers at the state and federal level as they consider legislation aimed at increasing competition in retail gasoline markets.