Cluster-robust inference: A guide to empirical practice

MacKinnon, James G.; Nielsen, Morten Ørregaard; Webb, Matthew D.

doi:10.1016/j.jeconom.2022.04.001

Cited by 127 publications

(63 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The standard errors of our variables of interest increase slightly but the statistical significance of the estimates remains unaffected. We also show the robustness of our event study results to alternative levels of clustering, including two-way clustering, clustering at the district level, and wild bootstrapped clustering as suggested by MacKinnon et al (2022) in Appendix Table B.2.…”

Section: Robustness Of the Resultsmentioning

confidence: 71%

No Surprises, Please: Voting Costs and Electoral Turnout

Alipour

Lindlacher

2022

SSRN Journal

View full text Add to dashboard Cite

Section: Robustness Of the Resultsmentioning

confidence: 71%

No Surprises, Please: Voting Costs and Electoral Turnout

Alipour

Lindlacher

2022

SSRN Journal

View full text Add to dashboard Cite

“…The asymptotic theory for clusterrobust inference has been analyzed recently by Djogbenou et al (2019) and Hansen and Lee (2019) under the assumption that G → ∞. In general, the quality of the asymptotic approximation is determined by the number of clusters G and the extent of heterogeneity of the score vectors (MacKinnon et al, 2022a). The more the distributions of the scores vary across clusters, the worse the asymptotic approximation will likely be.…”

Section: The Linear Regression Model With Clusteringmentioning

confidence: 99%

“…When any of the N g is large, this can take a very long time, or even fail because Stata runs out of memory. For example, it failed with the samples used in MacKinnon et al (2022a), where the largest values of N g were either 35, 995 or 144, 914. To circumvent this problem, one could use a random sample of the data for each cluster to compute an approximation to G * j (ρ), but with (38) that is no longer necessary.…”

Section: What Should Be Reportedmentioning

confidence: 99%

Leverage, Influence, and the Jackknife in Clustered Regression Models: Reliable Inference Using summclust

MacKinnon¹,

Nielsen²,

Webb³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Cluster-robust inference is widely used in modern empirical work in economics and many other disciplines. When data are clustered, the key unit of observation is the cluster. We propose measures of "high-leverage" clusters and "influential" clusters for linear regression models. The measures of leverage and partial leverage, and functions of them, can be used as diagnostic tools to identify datasets and regression designs in which cluster-robust inference is likely to be challenging. The measures of influence can provide valuable information about how the results depend on the data in the various clusters. We also show how to calculate two jackknife variance matrix estimators, CV 3 and CV 3J , as a byproduct of our other computations. All these quantities, including the jackknife variance estimators, are computed in a new Stata package called summclust that summarizes the cluster structure of a dataset.

show abstract

“…Additionally, we allow for a second level of clustering at the year level (across firms). Finally, we compare our estimates against a randomized assignment within the matched firm pairs with a randomization inference estimator (Barrios et al, 2012;MacKinnon et al, 2021). Across these methods, the standard errors vary and two-way clustering and randomization inference lead to higher p-values.…”

Section: Inferencementioning

confidence: 99%