Cluster-Robust Bootstrap Inference in Quantile Regression Models

Andreas Hagemann of the University of Michigan developed a wild bootstrap procedure for cluster-robust inference in linear quantile regression models. The method is designed for settings with a large number of small, heterogeneous clusters and is shown to perform well even with few clusters relative to sample size. An application to Project STAR data is provided.

Use Cases

• Testing hypotheses on the entire quantile regression process based on the described bootstrap method.
• Estimating asymptotic covariance functions for quantile regression processes with clustered data.
• Implementing cluster-robust inference in econometric applications where the number of clusters is small relative to sample size.
• Applying the proposed method to educational intervention data like Project STAR.

Strengths

• Method is described as easy to implement.
• Procedure is shown to perform well even when the number of clusters is much smaller than the sample size.
• Supplementary materials for the article are available online.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

University of Michigan

Collection Method

Methodological development with an application to Project STAR data.