Conditional Quantile Functions for Zero-Inflated Longitudinal Count Data

• Published in: Econometrics and statistics Vol. 31; pp. 49 - 65
• Main Authors: Lamarche, Carlos, Shi, Xuan, Young, Derek S.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2024
• Subjects: Generalized linear mixed models; Quantile models; Subject heterogeneity; Zero-inflated count data; Subject heterogeneity; Generalized linear mixed models; Quantile models; Zero-inflated count data
• ISSN: 2452-3062; 2452-3062
• DOI: 10.1016/j.ecosta.2021.09.003

The identification and estimation of conditional quantile functions for count responses using longitudinal data are considered. The approach is based on a continuous approximation to distribution functions for count responses within a class of parametric models that are commonly employed. It is first shown that conditional quantile functions for count responses are identified in zero-inflated models with subject heterogeneity. Then, a simple three-step approach is developed to estimate the effects of covariates on the quantiles of the response variable. A simulation study is presented to show the small sample performance of the estimator. Finally, the advantages of the proposed estimator in relation to some existing methods is illustrated by estimating a model of annual visits to physicians using data from a health insurance experiment.