Dependent Variable: S   
Method: ML - ARCH (Marquardt) - Normal distribution   
Date: 08/02/09   Time: 21:49   
Sample (adjusted): 1954 1998   
Included observations: 45 after adjustments   
Convergence achieved after 36 iterations   
Variance backcast: ON   
GARCH = C(6) + C(7)*RESID(-1)^2   
   
Coefficient Std. Error z-Statistic Prob.  
   
X 0.712936 0.035440 20.11674 0.0000
X(-1) -0.492487 0.072880 -6.757513 0.0000
X(-2) -0.278308 0.048102 -5.785762 0.0000
S(-1) 0.899183 0.075283 11.94401 0.0000
S(-3) 0.158552 0.044658 3.550379 0.0004
   
Variance Equation   
   
C 0.000114 5.35E-05 2.133031 0.0329
RESID(-1)^2 1.648891 0.661189 2.493828 0.0126
   
R-squared 0.998755     Mean dependent var  1.983578
Adjusted R-squared 0.998558     S.D. dependent var  1.178477
S.E. of regression 0.044745     Akaike info criterion  -4.388976
Sum squared resid 0.076080     Schwarz criterion  -4.107940
Log likelihood 105.7520     Durbin-Watson stat  2.370300
请问为什么在做ARCH的时候，结果始终无法出现ARCH？全是RESID(-1)^2