I thought this would be a one time error, however i've tested it on 3 different time series data bases, and the conclusions is that after a VECM with trend, the commands of vecmarl and vecnorm, jbera cannot estimate the results of the test.
The error is general for what i've been testing, other VECM estimations with (restricted constant, no constant, drift) have no troubles.
I'll leave an example of what the error looks like.
Code:
vec log_PIB_r log_Military_S_r, lag(3) trend(trend) Vector error-correction model Sample: 1953 - 2014 Number of obs = 62 AIC = -5.013312 Log likelihood = 170.4127 HQIC = -4.811255 Det(Sigma_ml) = .000014 SBIC = -4.498682 Equation Parms RMSE R-sq chi2 P>chi2 ---------------------------------------------------------------- D_log_PIB_r 7 .029485 0.7528 167.4535 0.0000 D_log_Military~r 7 .145817 0.3055 24.19635 0.0011 ---------------------------------------------------------------- ------------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------------+---------------------------------------------------------------- D_log_PIB_r | _ce1 | L1. | -.0369507 .0182902 -2.02 0.043 -.0727988 -.0011026 | log_PIB_r | LD. | .4199505 .1275807 3.29 0.001 .1698969 .6700041 L2D. | .2141882 .1306261 1.64 0.101 -.0418343 .4702107 | log_Military_S_r | LD. | .0265396 .0272462 0.97 0.330 -.026862 .0799411 L2D. | .0391178 .0261992 1.49 0.135 -.0122317 .0904674 | _trend | -.0000329 .0002105 -0.16 0.876 -.0004455 .0003797 _cons | .0150881 .0099484 1.52 0.129 -.0044104 .0345866 -------------------+---------------------------------------------------------------- D_log_Military_S_r | _ce1 | L1. | -.2698643 .0904541 -2.98 0.003 -.4471511 -.0925775 | log_PIB_r | LD. | .3414482 .63095 0.54 0.588 -.8951912 1.578088 L2D. | .6698477 .6460111 1.04 0.300 -.5963108 1.936006 | log_Military_S_r | LD. | -.0068011 .1347459 -0.05 0.960 -.2708983 .257296 L2D. | -.1033091 .1295682 -0.80 0.425 -.3572581 .1506399 | _trend | 4.51e-06 .0010411 0.00 0.997 -.002036 .002045 _cons | .0265753 .0491999 0.54 0.589 -.0698546 .1230053 ------------------------------------------------------------------------------------ Cointegrating equations Equation Parms chi2 P>chi2 ------------------------------------------- _ce1 1 13.45587 0.0002 ------------------------------------------- Identification: beta is exactly identified Johansen normalization restriction imposed ---------------------------------------------------------------------------------- beta | Coef. Std. Err. z P>|z| [95% Conf. Interval] -----------------+---------------------------------------------------------------- _ce1 | log_PIB_r | 1 . . . . . log_Military_S_r | 1.284193 .3500859 3.67 0.000 .5980377 1.970349 _trend | -.1041789 . . . . . _cons | -19.4083 . . . . . ---------------------------------------------------------------------------------- . veclmar, mlag(2) error computing temporary var estimates r(111); . vecnorm, jbera error computing temporary var estimates r(111);
0 Response to Reporting an error with assumptions revision after VECM trend estimation.
Post a Comment