The Cochrane-Orcutt procedure:

**One note**: In step 5 when it says to use the estimated betas obtained in step 4 in equation (9.5), this means to go back to the origanal equation (9.5) and find u_{t} = Y_{t} - b_{1} - b_{2}X_{2t} - ... - b_{k}X_{kt} where the b's are the estimated betas using the transformed ("starred") variables in step 4.

How do you get the coefficients and intercept for the new relation in terms of the old variables?

Let me explain. Suppose you have a simple linear regression between Y and X:

Y = b0 + b1 X

You then calculate rho^ for the residuals, transform the variables, regress, find rho^ again, etc., until you have:

Y* = b*0 + b1* X*

How do you find b0 and b1 valuables so that the equation can be expressed in terms of the original Y and X?

Posted by: Barton Paul Levenson | Wednesday, January 28, 2009 at 07:00 AM

"Values," not "valuables." It's early. I'm not really awake yet.

Posted by: Barton Paul Levenson | Wednesday, January 28, 2009 at 07:01 AM