.. Estimation of Regression Models with Autocorrelated Errors Suppose that we want to estimate the coefficients of the regression model y: = Bo + B1X11 + B2X2 + + BKXK+ + 8 when the error term er is autocorrelated. This can be accomplished in two stages, as follows: x7. Estimate the model by least squares, obtaining the Durbin-Watson d st tistic and hence the estimate d r = 1 2 of the autocorrelation parameter. 2. Estimate by least squares a second regression in which the dependent variable is (y: - ryt-1) and the independent variables are (X11 1X1,4-1 (x2+ - rx2,-1),..., (XK+ rxk,t-1). The parameters B1, B2, ..., Bk are estimated regression coefficients from th second model. An estimate of Bo is obtained by dividing the estimated inter cept for the second model by (1 r). Hypothesis tests and confidence inte vals for the regression coefficients can be carried out using the output from the second regression. Example 13.8 Time-Series Regression Model (Regression Analysis with Correlated Errors) In this extended example we demonstrate how to carry out a regression analysi ing Minitab, when the errors are autocorrelated. In this example we wish to deve model that predicts the aggregate consumption of durable goods as a function o posable income and the federal funds interest rate