------------------------------------------------------------------------------------------------------------- . * Osa 1. DVECH mudel . mgarch dvech (rt rn =), arch(1/1) garch(1/1) distribution(t) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = -9698.5285 Iteration 1: log likelihood = -9698.2518 (not concave) Iteration 2: log likelihood = -9556.7809 Iteration 3: log likelihood = -9547.6989 Iteration 4: log likelihood = -9538.4814 Iteration 5: log likelihood = -9520.7678 Iteration 6: log likelihood = -9504.3261 Iteration 7: log likelihood = -9431.2693 (switching technique to nr) Iteration 8: log likelihood = -9350.6089 (not concave) Iteration 9: log likelihood = -9321.6573 (not concave) Iteration 10: log likelihood = -9274.6379 (not concave) Iteration 11: log likelihood = -9255.3739 (not concave) Iteration 12: log likelihood = -9243.7809 (not concave) Iteration 13: log likelihood = -9234.3453 (not concave) Iteration 14: log likelihood = -9229.1268 (not concave) Iteration 15: log likelihood = -9223.7799 (not concave) Iteration 16: log likelihood = -9212.061 (not concave) Iteration 17: log likelihood = -9208.8092 (not concave) Iteration 18: log likelihood = -9207.0542 (not concave) Iteration 19: log likelihood = -9205.5353 (not concave) Iteration 20: log likelihood = -9204.5258 (not concave) Iteration 21: log likelihood = -9203.9181 (not concave) Iteration 22: log likelihood = -9203.3621 (not concave) Iteration 23: log likelihood = -9202.7779 (not concave) Iteration 24: log likelihood = -9202.3016 (not concave) Iteration 25: log likelihood = -9201.8639 (not concave) convergence not achieved Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = -9201.8639 Iteration 1: log likelihood = -9182.4338 Iteration 2: log likelihood = -9167.3624 Iteration 3: log likelihood = -9162.4645 Iteration 4: log likelihood = -9157.4441 (switching technique to nr) Iteration 5: log likelihood = -9155.1731 Iteration 6: log likelihood = -9153.7973 Iteration 7: log likelihood = -9153.5173 Iteration 8: log likelihood = -9153.5158 Iteration 9: log likelihood = -9153.5158 Diagonal vech MGARCH model Sample: 04jan2012 thru 15mar2023 Number of obs = 2,817 Distribution: t Wald chi2(.) = . Log likelihood = -9153.516 Prob > chi2 = . ------------------------------------------------------------------------------ | Coefficient Std. err. z P>|z| [95% conf. interval] -------------+---------------------------------------------------------------- rt | _cons | .0296031 .0199701 1.48 0.138 -.0095375 .0687438 -------------+---------------------------------------------------------------- rn | _cons | .0051163 .0233806 0.22 0.827 -.0407087 .0509414 -------------+---------------------------------------------------------------- /Sigma0 | 1_1 | .0256486 .0089615 2.86 0.004 .0080843 .0432129 2_1 | .0122987 .0049306 2.49 0.013 .0026349 .0219625 2_2 | .0148148 .0066724 2.22 0.026 .001737 .0278925 -------------+---------------------------------------------------------------- L.ARCH | 1_1 | .0388459 .0075047 5.18 0.000 .0241369 .0535549 2_1 | .0335271 .0065298 5.13 0.000 .0207289 .0463252 2_2 | .0342681 .0068158 5.03 0.000 .0209093 .0476268 -------------+---------------------------------------------------------------- L.GARCH | 1_1 | .9490213 .0106504 89.11 0.000 .928147 .9698957 2_1 | .9573288 .009346 102.43 0.000 .939011 .9756467 2_2 | .9609887 .0082335 116.72 0.000 .9448514 .977126 -------------+---------------------------------------------------------------- /df | 4.905408 .3170689 15.47 0.000 4.283964 5.526851 ------------------------------------------------------------------------------ . * Keskväärtuse mudelites konstant puudub . mgarch dvech (rt rn =, noconstant), arch(1/1) garch(1/1) distribution(t) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = -9700.4326 Iteration 1: log likelihood = -9700.1626 (not concave) Iteration 2: log likelihood = -9557.8039 Iteration 3: log likelihood = -9548.6568 Iteration 4: log likelihood = -9528.7289 Iteration 5: log likelihood = -9527.7221 (backed up) Iteration 6: log likelihood = -9527.2198 (backed up) Iteration 7: log likelihood = -9526.969 (backed up) (switching technique to nr) Iteration 8: log likelihood = -9526.9611 (backed up) Hessian is not negative semidefinite fatal numerical error encountered when searching for starting values try constraining some parameters and using alternative starting values r(498); . * t-jaotuse vabadusastmete arv ette antud . mgarch dvech (rt rn =, noconstant), arch(1/1) garch(1/1) distribution(t 4.9) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = -9436.1129 Iteration 1: log likelihood = -9436.0106 (not concave) Iteration 2: log likelihood = -9413.2862 Iteration 3: log likelihood = -9323.6757 Iteration 4: log likelihood = -9264.7131 Iteration 5: log likelihood = -9185.1281 Iteration 6: log likelihood = -9169.6443 Iteration 7: log likelihood = -9161.7099 (switching technique to nr) Iteration 8: log likelihood = -9158.6549 Iteration 9: log likelihood = -9155.628 Iteration 10: log likelihood = -9154.958 Iteration 11: log likelihood = -9154.8695 Iteration 12: log likelihood = -9154.8694 Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = -9154.8694 Iteration 1: log likelihood = -9154.8694 (backed up) Diagonal vech MGARCH model Sample: 04jan2012 thru 15mar2023 Number of obs = 2,817 Distribution: t(4.9) Wald chi2(.) = . Log likelihood = -9154.869 Prob > chi2 = . ------------------------------------------------------------------------------ | Coefficient Std. err. z P>|z| [95% conf. interval] -------------+---------------------------------------------------------------- /Sigma0 | 1_1 | .0261947 .0091028 2.88 0.004 .0083535 .0440358 2_1 | .0125284 .0049909 2.51 0.012 .0027464 .0223105 2_2 | .0149445 .0066939 2.23 0.026 .0018247 .0280642 -------------+---------------------------------------------------------------- L.ARCH | 1_1 | .039094 .0075109 5.20 0.000 .0243729 .0538151 2_1 | .0336755 .0065287 5.16 0.000 .0208795 .0464716 2_2 | .0343346 .0067875 5.06 0.000 .0210312 .0476379 -------------+---------------------------------------------------------------- L.GARCH | 1_1 | .9485068 .0107782 88.00 0.000 .927382 .9696317 2_1 | .9570335 .0094177 101.62 0.000 .9385752 .9754918 2_2 | .9608922 .008238 116.64 0.000 .944746 .9770384 ------------------------------------------------------------------------------ . * Osa 3. vaatluste lisamine ja prognooside leidmine . tsappend, add(100) . predict var*, variance dynamic(2819) . * Tingimustetea dispersioonide ja kovariatsiooni arvutus . display [/Sigma0]1_1/(1-[L.ARCH]1_1 - [L.GARCH]1_1) 2.1126145 . display [/Sigma0]2_2/(1-[L.ARCH]2_2 - [L.GARCH]2_2) 3.1308778 . display [/Sigma0]2_1/(1-[L.ARCH]2_1 - [L.GARCH]2_1) 1.3484591 . twoway (tsline var_rt_rt) (tsline var_rn_rt) (tsline var_rn_rn) , legend(position(6)) . gen corr = var_rn_rt/(sqrt( var_rt_rt* var_rn_rn)) . twoway (tsline corr) --------------------------------------------------------------------------------------------------------------------