Estimating dynamic logit models with unobserved individual heterogeneity and with application in household brand choices

被引：0
|
作者：
Liu, Changbiao [1 ]
机构：
[1] China-ASEAN School of Statistics, Guangxi University of Finance and Economics, Nanning, China
来源：
Computational and Mathematical Organization Theory | 2024年 / 30卷 / 04期
关键词：
In this paper; we propose a new method for dynamic discrete choice logit models with panel data; which capture both unobserved individual heterogeneities and the state dependence on purchase behaviors. The consistency and asymptotic normality of this estimators are studied in detail. Comparing with the estimators developed by Honore and Kyriazidou (Econometrica 68:839–874; 2000); the pseudo conditional likelihood estimators proposed by Bartolucci and Nigro (J Econom 170:102–116; 2012) and the modified profile likelihood estimators given by Bartolucci et al. (Economet Rev 35:1271–1289; 2016); simulations show the proposed estimators have some advantages on the mean bias and root mean squared error. As a byproduct; another estimator for static logit models is given and comparable with that developed by Chamberlain (in: Griliches Z; Intrilligator MD (eds); Handbook of econometrics; vol; 2; North-Holland; Amsterdam; 1984); Last; the proposed approach is applied to the panel data on household detergent purchases and concludes that there exists significant dynamic relationship on household detergent purchases. © The Author(s); under exclusive licence to Springer Science+Business Media; LLC; part of Springer Nature 2024;
D O I：
10.1007/s10588-024-09391-0
中图分类号：
学科分类号：
摘要：
引用
收藏
页码：321 / 349
相关论文
共 50 条
12345