The Cochrane-Orcutt procedure is a two-step estimation technique designed to handle first-order serial correlation in the errors of a linear regression model. This method uses the ordinary least squares residuals to estimate the first-order autocorrelation coefficient and then rescale the variables to eliminate serial correlation in the errors.
