Answer to Question #232640 in Economics for KAT

Question

Explain Cochrane-Orcut transformation so that no autocorrelation is present?

Answer 1

Due to the fact that the presence in the regression model of autocorrelation between the residuals of the model can lead to negative results of the entire process of estimating the unknown coefficients of the model, namely:

a) an increase in the variances of the estimates of the model parameters;

b) the bias of the estimates obtained by the OLS;

c) reducing the significance of parameter estimates,

autocorrelation of residuals should be eliminated.

Cochrane-Orcutt method. includes the following steps:

1. Applying OLS to the original regression equation, initial estimates of the parameters a0 and a1 are obtained; calculate the residuals

2. As an estimate of the parameter ρ, its least-squares estimate in regression is used.

3. Applying the Generalized LSM to the transformed equation, new estimates of the parameters a0 and a1 are obtained.

4. A new residual vector is constructed and the process returns to step 2.

The stages are alternated until the required accuracy is achieved, i.e., until the difference between the previous and subsequent estimates ρ becomes less than any predetermined number.