How to interpret GARCH output?

[ayubrabia]

How do u interpret this result??Plz let me know asap!

Dependent Variable: STOCKRET
Method: ML - ARCH (Marquardt) - Normal distribution
Date: 03/16/12 Time: 14:21
Sample: 1 263
Included observations: 263
Failure to improve Likelihood after 54 iterations
Presample variance: backcast (parameter = 0.7)
GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1)


Coefficient Std. Error z-Statistic Prob.


FX_ab -11615.87 677.9465 -17.13391 0.0000
DR_cd -346.2585 12.52274 -27.65037 0.0000
C 13116.92 456.9621 28.70462 0.0000


Variance Equation


C 208883.7 65232.49 3.202142 0.0014
RESID(-1)^2 1.248536 0.324457 3.848078 0.0001
GARCH(-1) -0.423042 0.114009 -3.710589 0.0002


R-squared 0.178363 Mean dependent var 4582.827
Adjusted R-squared 0.162378 S.D. dependent var 1341.754
S.E. of regression 1227.996 Akaike info criterion 16.01994
Sum squared resid 3.88E+08 Schwarz criterion 16.10143
Log likelihood -2100.622 Hannan-Quinn criter. 16.05269
F-statistic 11.15803 Durbin-Watson stat 0.042817
Prob(F-statistic) 0.000000

[Nau2306]

Hello

Even I am getting the same message: failure to improve likelihood after ... iterations. Why is it so?

[EViews Glenn]

At what value is your convergence tolerance set? What do the gradients of the equation look like at this value. There is a discussion of these issues in the manual.