OLS estimation with the QR decomposition and, for some options, computation of variance- covariance matrices
ols(y, x, untransformed.residuals=NULL, tol=1e-07, LAPACK=FALSE, method=3, ...)numeric vector, the regressand
numeric matrix, the regressors
NULL (default) or, when ols is used with method=6, a numeric vector containing the untransformed residuals
numeric value. The tolerance for detecting linear dependencies in the columns of the regressors, see the qr function. Only used if LAPACK is FALSE
logical, TRUE or FALSE (default). If true use LAPACK otherwise use LINPACK, see the qr function
an integer, 1 to 6, that determines the estimation method
further arguments (currently ignored)
A list with items depending on method
method = 1 or 2 only returns the OLS coefficient estimates together with the QR-information. method = 1 is slightly faster than method=2. method=3 returns, in addition, the ordinary variance-covariance matrix of the OLS estimator. method=4 returns the White (1980) heteroscedasticity robust variance-covariance matrix in addition to the information returned by method=3, whereas method=5 does the same except that the variance-covariance matrix now is that of Newey and West (1987). method=6 undertakes OLS estimation of a log-variance model, see Pretis, Reade and Sucarrat (2018, Section 4).
W. Newey and K. West (1987): 'A Simple Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix', Econometrica 55, pp. 703-708.
F. Pretis, J. Reade and G. Sucarrat (2018): 'Automated General-to-Specific (GETS) Regression Modeling and Indicator Saturation for Outliers and Structural Breaks', Journal of Statistical Software 86, Issue 3, pp. 1-44, DOI: https://doi.org/10.18637/jss.v086.i03
H. White (1980): 'A Heteroskedasticity-Consistent Covariance Matrix and a Direct Test for Heteroskedasticity', Econometrica 48, pp. 817-838.