Thirty Years of Heteroskedasticity-Robust Inference

By James G. MacKinnon. 2013

132 citations recorded by the source corpus.

White (Econometrica, 48:817–838, 1980) marked the beginning of a new era for inference in econometrics. It introduced the revolutionary idea of inference that is robust to heteroskedasticity of unknown form, an idea that was very soon extended to other forms of robust inference and also led to many new estimation methods. This paper discusses the development of heteroskedasticity-robust inference since 1980. There have been two principal lines of investigation. One approach has been to modify White’s original estimator to improve its finite-sample properties, and the other has been to use bootstrap methods. The relation between these two approaches, and some ways in which they may be combined, are discussed. Finally, a simulation experiment compares various methods and shows how far heteroskedasticity-robust inference has come in just over 30 years.

Subfields: Economics and Econometrics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/1d875a928aa1cd859b43e122c0f04b41bad384f0

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Statistics and Probability

Citations in the corpus

Inference in Linear Regression Models with Many Covariates and Heteroscedasticity (2015, methodology): “, MacKinnon (2012). Last but not least, ` =HCK uses Σ̂n,` = Σ̂ n , our proposed new inference method that is valid under conditional heteroskedasticity and many covariates.”
Asymptotic theory for clustered samples (2019, unknown): “Surveys can be found in Wooldridge (2003), Cameron and Miller (2011, 2015), MacKinnon (2012, 2016), and textbook treatments in Angrist and Pischke (2009) and Wooldridge (2010).”
Rdrobust: Software for Regression-discontinuity Designs (2017, methodology): “This method applies ideas from leastsquares estimation and inference; see MacKinnon (2012), and Cameron and Miller”
Gravity models, PPML estimation and the bias of the robust standard errors (2019, methodology): “Approaches similar to those available for linear models and surveyed by MacKinnon (2013) could possibly bring improve-”
On the Effect of Bias Estimation on Coverage Accuracy in Nonparametric Inference (2015, background): “Next, motivated by the fact that least-squares residuals are on average too small, we appeal to the HCk class of estimators (see MacKinnon (2013) for a review), which are defined as follows.”
Optimal bandwidth choice for robust bias-corrected inference in regression discontinuity designs (2018, background): “First, motivated by the fact that the least-squares residuals are on average too small, we propose HCk heteroskedasticity consistent estimators; see MacKinnon (2013) for details and a recent review. Calonico, Cattaneo, Farrell, and Titiunik (2019) discuss how they can be applied in the context of lo”
Agnostic notes on regression adjustments to experimental data: Reexamining Freedman's critique (2012, background): “MacKinnon (2013) discusses bias-corrected sandwich estimators and improved confidence intervals based on the wild bootstrap.”
OLS and IV estimation of regression models including endogenous interaction terms (2019, methodology): “the jackknife estimator in MacKinnon and White (1985).7 The robust variance estimator in (2.”
Small-Sample Methods for Cluster-Robust Variance Estimation and Hypothesis Testing in Fixed Effects Models (2016, methodology): “ downward and associated hypothesis tests can have Type-I error rates considerably in excess of nominal levels when based on samples with only a small or moderate number of clusters (e.g., Webb & MacKinnon 2013). Cameron & Miller (2015) provide a thorough review of this literature, including”
Resurrecting weighted least squares (2017, unknown): “…finite-sample inference is HC3; for example, see MacKinnon and White (1985), Long and Ervin (2000), Angrist and Pischke (2009, Section 8.1), and MacKinnon (2012).9 It is also the default option in several statistical software packages to carry out HC estimation, such as in the R function…”
Small-Sample Adjustments for Tests of Moderators and Model Fit Using Robust Variance Estimation in Meta-Regression (2015, methodology): “Thus, we have not considered iterative procedures such as bootstrapping, although some recent work in econometrics has proposed promising techniques for bootstrapping in settings with cluster-dependent observations (Webb & MacKinnon, 2013). Neither have we considered corrections such as saddlepoint ”
Margins of labor market adjustment to trade (2019, background): “Column (3) uses heteroskedasticity weights and presents heteroskedasticity-robust standard errors, which are likely understated in this small sample (MacKinnon, 2011).”