BIROn - Birkbeck Institutional Research Online

    Posterior means and precisions of the coefficients in linear models with Highly Collinear Regressors

    Hashem Pesaran, M. and Smith, Ron P. (2017) Posterior means and precisions of the coefficients in linear models with Highly Collinear Regressors. Working Paper. Birkbeck, University of London, London, UK.

    [img]
    Preview
    Text
    26712.pdf - Draft Version

    Download (303kB) | Preview

    Abstract

    When there is exact collinearity between regressors, their individual coe¢ cients are not identified, but given an informative prior their Bayesian posterior means are well defined. The case of high but not exact collinearity is more complicated but similar results follow. Just as exact collinearity causes non-identification of the parameters, high collinearity can be viewed as weak identification of the parameters, which we represent, in line with the weak instrument literature, by the correlation matrix being of full rank for a finite sample size T, but converging to a rank deficient matrix as T goes to infinity. This paper examines the asymptotic behaviour of the posterior mean and precision of the parameters of a linear regression model for both the cases of exactly and highly collinear regressors. We show that in both cases the posterior mean remains sensitive to the choice of prior means even if the sample size is su¢ ciently large, and that the precision rises at a slower rate than the sample size. In the highly collinear case, the posterior means converge to normally distributed random variables whose mean and variance depend on the priors for coe¢ cients and precision. The distribution degenerates to fixed points for either exact collinearity or strong identification. The analysis also suggests a diagnostic statistic for the highly collinear case, which is illustrated with an empirical example.

    Metadata

    Item Type: Monograph (Working Paper)
    Additional Information: BCAM 1707; ISSN 1745-8587
    Keyword(s) / Subject(s): Bayesian identification, multicollinear regressions, weakly identified regression coefficients, highly collinear regressors
    School: School of Business, Economics & Informatics > Economics, Mathematics and Statistics
    Research Centres and Institutes: Applied Macroeconomics, Birkbeck Centre for, Innovation Management Research, Birkbeck Centre for
    Depositing User: Administrator
    Date Deposited: 21 Mar 2019 16:20
    Last Modified: 21 Jul 2020 20:04
    URI: https://eprints.bbk.ac.uk/id/eprint/26712

    Statistics

    Downloads
    Activity Overview
    139Downloads
    268Hits

    Additional statistics are available via IRStats2.

    Archive Staff Only (login required)

    Edit/View Item Edit/View Item