NESTED LOGIT OR RANDOM COEFFICIENTS LOGIT? A COMPARISON OF ALTERNATIVE DISCRETE CHOICE MODELS OF PRODUCT DIFFERENTIATION

• 成果类型: Article
• 署名作者: Grigolon, Laura; Verboven, Frank
• 署名单位: McMaster University; KU Leuven; Centre for Economic Policy Research - UK
• 刊物名称: REVIEW OF ECONOMICS AND STATISTICS
• ISSN/ISSBN: 0034-6535
• DOI: 10.1162/REST_a_00420
• 发表日期: 2014-12
• 页码: 916-935
• 关键词: european car market; eat cereal industry; demand; automobiles; performance; mergers; POLICY

摘要

We propose a random coefficients nested logit (RCNL) model to compare the tractable nested logit (NL) model with the more complex random coefficients logit (RC) model. After a simulation study, we use data on the European automobile market. Both the NL and RC models are rejected against the RCNL model. The RC model results in different substitution patterns and a wider market definition than the NL and RCNL models. Nevertheless, the predicted price effects from mergers are robust across models. Our findings stress the importance of accounting for discrete sources of market segmentation not captured by continuous product characteristics.