关键词: 分位数回归, 检查函数, 填补技术, 随机缺失

Abstract: There exist methods to treat the missing data of mean-regression models, but they cannot be applied to quantile regression with missing data because of inconsistency. The pairwise deletion is commonly adopted in quantile regression. However, it incurs inefficiency due to the loss of available information of deleted observations. The proposed check function-based imputation can capture the information, therefore achieves efficiency improvement, and its estimation results may find a more useful explanation of economic phenomena.

Key words: Quantile Regression, Check Function, Imputation, Missing at Random.