Models as Approximations - A Conspiracy of Random Regressors and Model Deviations Against Classical Inference in Regression

Loading...
Thumbnail Image
Penn collection
Statistics Papers
Degree type
Discipline
Subject
Ancillarity of regressors
Misspecification
Econometrics
Sandwich estimator
Bootstrap
Physical Sciences and Mathematics
Funder
Grant number
License
Copyright date
Distributor
Related resources
Author
Buja, Andreas
Berk, Richard A
Brown, Lawrence D
George, Edward I
Pitkin, Emil
Traskin, Mikhail
Zhao, Linda
Zhang, Kai
Contributor
Abstract

More than thirty years ago Halbert White inaugurated a “modelrobust” form of statistical inference based on the “sandwich estimator” of standard error. It is asymptotically correct even under “model misspecification,” that is, when models are approximations rather than generative truths. It is well-known to be “heteroskedasticity-consistent”, but it is less well-known to be “nonlinearity-consistent” as well. Nonlinearity, however, raises fundamental issues: When fitted models are approximations, conditioning on the regressor is no longer permitted because the ancillarity argument that justifies it breaks down. Two effects occur: (1) parameters become dependent on the regressor distribution; (2) the sampling variability of parameter estimates no longer derives from the conditional distribution of the response alone. Additional sampling variability arises when the nonlinearity conspires with the randomness of the regressors to generate a 1/ √ N contribution to standard errors. Asymptotically, standard errors from “model-trusting” fixedregressor theories can deviate from those of “model-robust” randomregressor theories by arbitrary magnitudes. In the case of linear models, a test will be proposed for comparing the two types of standard errors

Advisor
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
2015-07-01
Journal title
Statistical Science
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation
Collection