我做了一个MV-BEKK模型，三元的，但是用BHHH估计方法和BFGS估计方法估计出来的参数结果有很大出入，BFGS估计出来的参数T检验量显著的，在BHHH下却不显著了。。 以下是程序，不对之处还望指正！
calender(sevenday)2010 6 18
allocate 2012:12:20
open data "D:\process\WinRATS Pro 8.012121\数据 \ ersz.xlsx "
data(format=xlsx,org=columns) / er szsb
set er = log(er)
set sz = log(sz)
set sb = log(sb)
set err = er - er(1)
set szr = sz - sz(1)
set sbr = sb - sb(1)

system(mode1=var1)
variables err szr sbr
lags 1
det constant
end(system)
garch(p=1,q=1,mode1=var1,mv=bekk,pmethod=bhhh,piters=10)

MV-GARCH, BEKK - Estimation by BHHH
NO CONVERGENCE IN 16 ITERATIONS
LAST CRITERION WAS0.0000000
SUBITERATIONS LIMIT EXCEEDED.
ESTIMATION POSSIBLY HAS STALLED OR MACHINE ROUNDOFF IS MAKING FURTHER PROGRESS DIFFICULT
TRY HIGHER SUBITERATIONS LIMIT, TIGHTER CVCRIT, DIFFERENT SETTING FOR EXACTLINE OR ALPHA ON NLPAR
RESTARTING ESTIMATION FROM LAST ESTIMATES OR DIFFERENT INITIAL GUESSES MIGHT ALSO WORK
Daily(7) Data From 2010:06:19 To 2012:02:15
Usable Observations 607
Log Likelihood 2373.4645
Variable Coeff Std Error T-Stat Signif
*************************************************************************************
1.ERR{1} 0.421464 1.703904 0.247350.80463567
2.SZR{1} 0.023056 0.229144 0.100620.91985452
3.SBR{1} 0.018195 0.212509 0.085620.93176754
4.Constant -0.004035 0.061473 -0.065640.94766183
5.ERR{1} -0.222139 22.521542 -0.009860.99213029
6.SZR{1} 0.514110 2.657763 0.193440.84661664
7.SBR{1} 0.295234 2.576193 0.114600.90876147
8.Constant -0.021454 0.311075 -0.068970.94501673
9.ERR{1} -0.847583 21.812389 -0.038860.96900370
10. SZR{1} 0.386388 2.816724 0.137180.89089137
11. SBR{1} 0.472043 2.673667 0.176550.85985986
12. Constant 0.005877 0.362987 0.016190.98708205
13. C(1,1) 0.008213 0.051687 0.158890.87375551
14. C(2,1) 0.021845 0.464668 0.047010.96250298
15. C(2,2) 0.024733 0.208680 0.118520.90565393
16. C(3,1) 0.006185 0.512205 0.012070.99036614
17. C(3,2) 0.028309 0.263522 0.107420.91445231
18. C(3,3) -0.0000111570.053739 -7.18296e-0090.99999999
19. A(1,1) 0.127543 1.302694 0.097910.92200631
20. A(1,2) 0.167066 7.014196 0.023820.98099759
21. A(1,3) -0.055264 7.283187 -0.007590.99394583
22. A(2,1) 0.008711 0.196294 0.044380.96460359
23. A(2,2) 0.132445 1.177238 0.112500.91042311
24. A(2,3) -0.007510 1.068526 -0.007030.99439209
25. A(3,1) -0.028882 0.248033 -0.116450.90729955
26. A(3,2) -0.093375 1.771254 -0.052720.95795744
27. A(3,3) 0.065484 1.807029 0.036240.97109224
28. B(1,1) 0.919466 0.723912 1.270140.20403637
29. B(1,2) -0.343400 4.556516 -0.075360.93992463
30. B(1,3) -0.175175 5.238847 -0.033440.97332552
31. B(2,1) 0.002106 0.076471 0.027540.97803062
32. B(2,2) 1.012776 0.439478 2.304500.02119482
33. B(2,3) 0.037890 0.376051 0.100760.91974287
34. B(3,1) 0.006374 0.064562 0.098730.92135432
35. B(3,2) -0.014998 0.378949 -0.03958 0.96842904
36. B(3,3) 0.974313 0.484321 2.011710.04425049

接下来是BFGS方法下的估计结果：
system(mode1=var1)
variables err szr sbr
lags 1
det constant
end(system)
garch(p=1,q=1,mode1=var1,mv=bekk,pmethod=bfgs,piters=10)

MV-GARCH, BEKK - Estimation by BFGS
NO CONVERGENCE IN 200 ITERATIONS
LAST CRITERION WAS0.0027069
Daily(7) Data From 2010:06:19 To 2012:02:15
Usable Observations 607
Log Likelihood 7039.8533
Variable Coeff Std Error T-Stat Signif
************************************************************************************
1.ERR{1} 0.99719910 0.00237387 420.073450.00000000
2.SZR{1} 0.00073980 0.00105177 0.703380.48181834
3.SBR{1} -0.00085640 0.00081491 -1.050920.29329727
4.Constant -0.00017141 0.00014307 -1.198110.23087303
5.ERR{1} 0.04690045 0.02826964 1.659040.09710782
6.SZR{1} 0.98720530 0.01102580 89.535920.00000000
7.SBR{1} 0.00611300 0.00876629 0.697330.48559602
8.Constant 0.00165852 0.00175239 0.946430.34392766
9.ERR{1} -0.01309313 0.02830064 -0.462640.64361941
10. SZR{1} 0.05574097 0.01234377 4.515720.00000631
11. SBR{1} 0.95120835 0.00960919 98.989430.00000000
12. Constant 0.00830550 0.00171896 4.831700.00000135
13. C(1,1) -0.00020299 0.00017421 -1.165190.24394142
14. C(2,1) 0.01004779 0.00119406 8.414820.00000000
15. C(2,2) 0.00513893 0.00210055 2.446470.01442622
16. C(3,1) -0.00059902 0.00061592 -0.972570.33076787
17. C(3,2) -0.00038448 0.00080127 -0.479830.63134614
18. C(3,3) -0.00000622 0.00102912 -0.006050.99517512
19. A(1,1) -0.07963097 0.05321638 -1.496360.13455932
20. A(1,2) -1.68833904 0.58046311 -2.908610.00363043
21. A(1,3) 1.24440565 0.61317632 2.029440.04241331
22. A(2,1) -0.06318163 0.00490437 -12.882730.00000000
23. A(2,2) 0.28050534 0.09136633 3.070120.00213975
24. A(2,3) 0.01202312 0.07593032 0.158340.87418569
25. A(3,1) 0.00659181 0.00349265 1.887340.05911463
26. A(3,2) 0.08012758 0.04132739 1.938850.05251969
27. A(3,3) -0.11419219 0.04628369 -2.467220.01361655
28. B(1,1) -0.34952232 0.10604955 -3.295840.00098128
29. B(1,2) -1.45993399 0.86873152 -1.680540.09285321
30. B(1,3) -10.55373366 0.75674619 -13.946200.00000000
31. B(2,1) -0.05568243 0.00562023 -9.907500.00000000
32. B(2,2) -0.12636183 0.16666580 -0.758170.44834629
33. B(2,3) 0.23189906 0.11179770 2.074270.03805387
34. B(3,1) 0.02493584 0.00715660 3.484320.00049340
35. B(3,2) 0.05414030 0.06964698 0.777350.43695043
36. B(3,3) 0.08461779 0.08101339 1.044490.29625812

两种算法下的估计结果明显是不一样的，这是怎么回事呢？ 我的编程没有错吧！求大神解惑，不胜感激！！！