Re: [Gretl-users] pre-release testing for gretl 1.8.0

Hello, I tried Engle-Granger Cointegration Test GUI on cvs build Jan 23 in my Windows XP computer. If I choose test-down from maximum lag order, the ADF test result is not shown. I use australia data and lag order 4. Is it only in my computer? I've also tried on 1.7.9, but it runs OK. Thanks. Yuniarto Hadiwibowo The result I get: Step 1: testing for a unit root in lpus Step 2: testing for a unit root in lpau Step 3: cointegrating regression Cointegrating regression - OLS estimates using the 77 observations 1972:1-1991:1 Dependent variable: lpus coefficient std. error t-ratio p-value --------------------------------------------------------- lpau 0.935025 0.00299947 311.7 7.78e-120 *** Mean dependent var 4.324348 S.D. dependent var 0.364635 Sum squared resid 1.133150 S.E. of regression 0.122106 R-squared 0.999219 Adjusted R-squared 0.999219 Log-likelihood 53.16570 Akaike criterion -104.3314 Schwarz criterion -101.9876 Hannan-Quinn -103.3939 rho 0.988215 Durbin-Watson 0.011277 Log-likelihood for PUS = -279.809 Step 4: Dickey-Fuller test on residuals There is evidence for a cointegrating relationship if: (a) The unit-root hypothesis is not rejected for the individual variables. (b) The unit-root hypothesis is rejected for the residuals (uhat) from the cointegrating regression.