Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

PRACTITIONERS’ CORNER: Computing Robust Standard Errors for Within‐groups Estimators

PRACTITIONERS’ CORNER: Computing Robust Standard Errors for Within‐groups Estimators M. Arellano I. THE MODEL AND THE ESTIMATOR The purpose of this note is to explain how to use standard packages to calculate heteroskedasticity and serial correlation consistent standard errors for within-groups estimators of a linear regression model from panel data. We are concerned to discuss the model y,=x;ß+e1+ui E(u, Ix11, (t=l,...,T;i=l,...,N) . . (1) where x1j is a k X 1 vector of exogenous variables such that . , Xc', e.) = O and e1 is an unobservable permanent effect potentially correlated with x1. The u11 are assumed to be independently distributed across individuals but no restrictions are placed on the form of the autocovariances for a given individual: u, ¡x11 ,...,X17.', e) = wsj thus allowing for heteroskedasticity and serial correlation of arbitrary form. Alternatively model (1) can be written as (2) y1=X113+e11+u1 (i=l,...,N) TX1 TXkICX1 TX1 u1 ' i.d. (O, £2) where 1 is a T X 1 vector of ones. We assume that T is small and N is large, as is often the case with household or company data, thus considering asymptotic results as N - oo for fixed T. Transforming the variables in (2) into deviations from time means eliminates the ej's. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Oxford Bulletin of Economics & Statistics Wiley

PRACTITIONERS’ CORNER: Computing Robust Standard Errors for Within‐groups Estimators

Loading next page...
 
/lp/wiley/practitioners-corner-computing-robust-standard-errors-for-within-nlXzBCSnnE

References (7)

Publisher
Wiley
Copyright
© 1987 Blackwell Publishing Ltd
ISSN
0305-9049
eISSN
1468-0084
DOI
10.1111/j.1468-0084.1987.mp49004006.x
Publisher site
See Article on Publisher Site

Abstract

M. Arellano I. THE MODEL AND THE ESTIMATOR The purpose of this note is to explain how to use standard packages to calculate heteroskedasticity and serial correlation consistent standard errors for within-groups estimators of a linear regression model from panel data. We are concerned to discuss the model y,=x;ß+e1+ui E(u, Ix11, (t=l,...,T;i=l,...,N) . . (1) where x1j is a k X 1 vector of exogenous variables such that . , Xc', e.) = O and e1 is an unobservable permanent effect potentially correlated with x1. The u11 are assumed to be independently distributed across individuals but no restrictions are placed on the form of the autocovariances for a given individual: u, ¡x11 ,...,X17.', e) = wsj thus allowing for heteroskedasticity and serial correlation of arbitrary form. Alternatively model (1) can be written as (2) y1=X113+e11+u1 (i=l,...,N) TX1 TXkICX1 TX1 u1 ' i.d. (O, £2) where 1 is a T X 1 vector of ones. We assume that T is small and N is large, as is often the case with household or company data, thus considering asymptotic results as N - oo for fixed T. Transforming the variables in (2) into deviations from time means eliminates the ej's.

Journal

Oxford Bulletin of Economics & StatisticsWiley

Published: Nov 1, 1987

There are no references for this article.