Dear all


I use gretl 1.7.4 to estimate a GARCH model by using default menu function (i.e., \Model\Time series\GARCH ) and by the command in console. But they generate different results. The details of iterations are as follows. Any idea?

==== results by menu=====

Automatic initialization of parameters

 Regression coefficients:
  theta[0] = 0.0511825
  theta[1] = 1.08508

 Variance parameters:
  alpha[0] = 0.1
   beta[0] = 0.9

Iteration 1: Log-likelihood = -168.929078042
Parameters:      0.11038      1.0851     0.10000     0.90000
Gradients:       -26.738      17.074      481.64      14.243

Iteration 2: Log-likelihood = -139.605272823 (steplength = 0.0016)
Parameters:     0.067597      1.1124     0.87063     0.92279
Gradients:       -4.4061     0.82275     -15.175     -8.1703

Iteration 3: Log-likelihood = -138.361721087 (steplength = 0.04)
Parameters:     -0.13607      1.1652     0.87232     0.62335
Gradients:        14.823    -0.51238     -14.088     -6.9177

Iteration 4: Log-likelihood = -137.371894269 (steplength = 0.008)
Parameters:     -0.12485      1.1848     0.87847     0.47302
Gradients:        13.892     -1.0676     -13.148     -5.2299

Iteration 5: Log-likelihood = -137.147777193 (steplength = 0.04)
Parameters:     -0.12168      1.1207     0.88177     0.43192
Gradients:        13.651     0.40209     -12.854     -4.5731

Iteration 6: Log-likelihood = -136.065550003 (steplength = 0.2)
Parameters:    -0.055939      1.1709     0.88165     0.29403
Gradients:        7.6020     -1.0743     -11.185    -0.64410

Iteration 7: Log-likelihood = -135.589722350 (steplength = 1)
Parameters:     0.054295      1.1305     0.86108     0.32088
Gradients:       -3.2400    -0.48930     -11.296     -1.6096

Iteration 8: Log-likelihood = -135.332923963 (steplength = 1)
Parameters:     0.044387      1.1121     0.84171     0.26964
Gradients:       -2.1320   -0.080313     -9.7947     0.99314

Iteration 9: Log-likelihood = -134.612248244 (steplength = 1)
Parameters:     0.014804      1.0933     0.76024     0.27015
Gradients:       0.83895     0.60996     -7.3230      3.3510

Iteration 10: Log-likelihood = -133.876720666 (steplength = 1)
Parameters:    -0.034438      1.0938     0.46226     0.40605
Gradients:        6.1180     0.99752      10.528      6.9157

Iteration 11: Log-likelihood = -133.215756120 (steplength = 0.008)
Parameters:     0.014505      1.1018     0.54648     0.46137
Gradients:      -0.55270     0.72059     -1.2327     0.43556

Iteration 12: Log-likelihood = -133.185323523 (steplength = 0.04)
Parameters:     0.010238      1.1316     0.52856     0.49641
Gradients:      -0.25051    -0.14061    -0.71733    -0.28480

Iteration 13: Log-likelihood = -133.185060907 (steplength = 0.008)
Parameters:     0.011028      1.1308     0.52769     0.49709
Gradients:      -0.36195    -0.11592    -0.65345    -0.28715

Iteration 14: Log-likelihood = -133.185045046 (steplength = 0.04)
Parameters:    0.0087095      1.1249     0.52597     0.50187
Gradients:     -0.053499    0.085396    -0.65223    -0.41383

Iteration 15: Log-likelihood = -133.181789118 (steplength = 1)
Parameters:    0.0081963      1.1273     0.52065     0.49465
Gradients:     -0.016645  -0.0072318    0.043542 -0.00045134

Iteration 16: Log-likelihood = -133.181774062 (steplength = 1)
Parameters:    0.0081232      1.1271     0.52128     0.49410
Gradients:   -0.00074809 -0.00035241  -0.0012373  0.00028761

Iteration 17: Log-likelihood = -133.181774041 (steplength = 1)
Parameters:    0.0081164      1.1271     0.52126     0.49413
Gradients:   6.7268e-006 4.3038e-006-1.6833e-005-1.9424e-005

Iteration 17: Log-likelihood = -133.181774041 (steplength = 1)
Parameters:    0.0081164      1.1271     0.52126     0.49413
Gradients:   6.7268e-006 4.3038e-006-1.6833e-005-1.9424e-005


--- FINAL VALUES:
Log-likelihood = -133.181774041 (steplength = 6.4e-005)
Parameters:    0.0081164      1.1271     0.52126     0.49413
Gradients:   6.7268e-006 4.3038e-006-1.6833e-005-1.9424e-005


theta[0]:     0.00381941 (0.0406031)
theta[1]:       0.530368 (0.0834569)
theta[2]:       0.115430 (0.0264921)
theta[3]:       0.494127 (0.192625)

Function evaluations: 47
Evaluations of gradient: 17

Model 11: GARCH estimates using the 99 observations 1980:02-1988:04
Dependent variable: Y
Standard errors based on Hessian

      VARIABLE       COEFFICIENT        STDERROR      T STAT   P-VALUE

  const                 0.00381941       0.0406031     0.094   0.92506
  Y_1                   0.530368         0.0834569     6.355  <0.00001 ***

  alpha(0)              0.115430         0.0264921     4.357   0.00001 ***
  alpha(1)              0.494127         0.192625      2.565   0.01031 **

  Mean of dependent variable = 0.0519415
  Standard deviation of dep. var. = 0.544934
  Unconditional error variance = 0.22818
  Log-likelihood = -58.5562
  Akaike information criterion (AIC) = 127.112
  Schwarz Bayesian criterion (BIC) = 140.088
  Hannan-Quinn criterion (HQC) = 132.362


==== results by the command in console=====
? garch 0 1; Y Y(-1) --verbose

Automatic initialization of parameters

 Regression coefficients:
  theta[0] = 0.0511825
  theta[1] = 1.08508

 Variance parameters:
  alpha[0] = 0.1
   beta[0] = 0.9

Iteration 1: Log-likelihood = -270.911199903
Parameters:      0.11038      1.0851     0.10000     0.90000
Gradients:       -125.16     -125.16      1010.7      30.388

Iteration 2: Log-likelihood = -170.537739660 (steplength = 0.0016)
Parameters:    -0.089875     0.88483      1.7171     0.94862
Gradients:       -25.749     -25.749     -8.7587     -8.4495

Iteration 3: Log-likelihood = -161.761522666 (steplength = 0.008)
Parameters:     -0.29926     0.67545      1.6788     0.88214
Gradients:       -11.183     -11.183     -10.840     -8.8715

Iteration 4: Log-likelihood = -158.391619816 (steplength = 0.04)
Parameters:     -0.31167     0.66303      1.6896     0.51071
Gradients:       -10.407     -10.407     -10.879     -8.1069

Iteration 5: Log-likelihood = -157.154680098 (steplength = 0.04)
Parameters:     -0.31888     0.65583      1.6928     0.36062
Gradients:       -10.000     -10.000     -10.621     -6.6638

Iteration 6: Log-likelihood = -155.742679661 (steplength = 0.2)
Parameters:     -0.35479     0.61992      1.6858     0.22820
Gradients:       -7.1755     -7.1755     -10.191     -3.8497

Iteration 7: Log-likelihood = -154.844791246 (steplength = 1)
Parameters:     -0.48452     0.49018      1.6370     0.10641
Gradients:        5.3640      5.3640     -8.7104      3.1207

Iteration 8: Log-likelihood = -154.573226547 (steplength = 1)
Parameters:     -0.45621     0.51850      1.6312     0.19907
Gradients:        2.1978      2.1978     -10.062     -2.3733

Iteration 9: Log-likelihood = -154.393792855 (steplength = 1)
Parameters:     -0.41196     0.56275      1.6201     0.14415
Gradients:       -2.0434     -2.0434     -9.3522     0.38397

Iteration 10: Log-likelihood = -153.968693027 (steplength = 1)
Parameters:     -0.42290     0.55181      1.5786     0.15683
Gradients:      -0.97944    -0.97944     -9.3680     0.21026

Iteration 11: Log-likelihood = -150.992406742 (steplength = 0.04)
Parameters:     -0.46207     0.51263      1.2039     0.16524
Gradients:        3.4088      3.4088     -6.1548      7.6249

Iteration 12: Log-likelihood = -148.359103161 (steplength = 0.008)
Parameters:     -0.43062     0.54409     0.75562     0.32000
Gradients:       -2.3822     -2.3822      1.9263      6.8635

Iteration 13: Log-likelihood = -148.097032678 (steplength = 0.008)
Parameters:     -0.45911     0.51560     0.71943     0.36973
Gradients:       0.75946     0.75946      1.8664      4.7097

Iteration 14: Log-likelihood = -147.949161098 (steplength = 1)
Parameters:     -0.45122     0.52348     0.75142     0.44913
Gradients:      -0.66911    -0.66911     -1.6394    -0.55287

Iteration 15: Log-likelihood = -147.921370789 (steplength = 1)
Parameters:     -0.45617     0.51854     0.69072     0.47194
Gradients:      -0.69320    -0.69320     0.78234    -0.15366

Iteration 16: Log-likelihood = -147.911096499 (steplength = 1)
Parameters:     -0.46083     0.51388     0.71098     0.45901
Gradients:       0.15242     0.15242   -0.086460    0.020146

Iteration 17: Log-likelihood = -147.910849840 (steplength = 1)
Parameters:     -0.45975     0.51495     0.70935     0.46012
Gradients:    -0.0033672  -0.0033672   -0.010763  -0.0083438

Iteration 18: Log-likelihood = -147.910848410 (steplength = 1)
Parameters:     -0.45979     0.51491     0.70914     0.46007
Gradients:    0.00068169  0.00068169  0.00046165 -0.00074704

Iteration 19: Log-likelihood = -147.910848385 (steplength = 1)
Parameters:     -0.45979     0.51492     0.70917     0.46004
Gradients:   8.9567e-005 8.9567e-005 8.2642e-006 -0.00017120

Iteration 20: Log-likelihood = -147.910848385 (steplength = 0.008)
Parameters:     -0.45979     0.51492     0.70917     0.46004
Gradients:   1.1075e-005 1.1075e-005 4.6432e-005 -0.00012421

Iteration 21: Log-likelihood = -147.910848385 (steplength = 0.008)
Parameters:     -0.45979     0.51492     0.70917     0.46004
Gradients:   5.2761e-005 5.2761e-005 3.4521e-005-8.0059e-005

Iteration 22: Log-likelihood = -147.910848385 (steplength = 0.008)
Parameters:     -0.45979     0.51492     0.70917     0.46004
Gradients:   4.1306e-005 4.1306e-005 5.3314e-006-9.2102e-005

Iteration 23: Log-likelihood = -147.910848385 (steplength = 1)
Parameters:     -0.45978     0.51492     0.70917     0.46003
Gradients:  -2.4463e-006-2.4463e-006 2.1778e-006 7.0214e-006

Iteration 23: Log-likelihood = -147.910848385 (steplength = 1)
Parameters:     -0.45978     0.51492     0.70917     0.46003
Gradients:  -2.4463e-006-2.4463e-006 2.1778e-006 7.0214e-006


--- FINAL VALUES:
Log-likelihood = -147.910848385 (steplength = 0.0016)
Parameters:     -0.45978     0.51492     0.70917     0.46003
Gradients:  -2.4463e-006-2.4463e-006 2.1778e-006 7.0214e-006


theta[0]:      -0.216364 (0.171258)
theta[1]:       0.242311 (0.0635179)
theta[2]:       0.157041 (0.0421745)
theta[3]:       0.460033 (0.177473)

Function evaluations: 58
Evaluations of gradient: 23

Model 10: GARCH estimates using the 99 observations 1980:02-1988:04
Dependent variable: Y
Standard errors based on Hessian

      VARIABLE       COEFFICIENT        STDERROR      T STAT   P-VALUE

  const                -0.216364         0.171258     -1.263   0.20645
  Y_1                   0.242311         0.0635179     3.815   0.00014 ***

  alpha(0)              0.157041         0.0421745     3.724   0.00020 ***
  alpha(1)              0.460033         0.177473      2.592   0.00954 ***

  Mean of dependent variable = 0.0519415
  Standard deviation of dep. var. = 0.544934
  Unconditional error variance = 0.290835
  Log-likelihood = -73.2853
  Akaike information criterion (AIC) = 156.571
  Schwarz Bayesian criterion (BIC) = 169.546
  Hannan-Quinn criterion (HQC) = 161.82