Hello,
I'm using Stata 14 and monthly time-series data for January 2000 to December 2015. My thesis is economics-related.
I am trying to establish the long-run and short-run relationship between various retail rates (mthtd, dddr, savr, alvr, etc) and monetary policy rate (mpr).
where,
mthth - 3-month deposit rate
dddr - demand deposit rate
savr - savings rate
avlr - average lending rate
Each of these variables is used as a dependent variable in a series of regressions with the independent variables being:
npl - non-performing loans
crr - reserve requirement coefficient
xcrate - exchange rate
vix - VIX volatility index
embi - EMBI spread
a dummy for inflation targeting period is included in all regresions.
The result after testing for unit root and co-integration led me to run ECM estimation, an example is shown below:
In the long run the results showed an incomplete pass-through effect (1.0 is considered a perfect pass-through).
I intend to test for asymmetric co-integration. That is, if the monetary policy rate is increased, the other retail rates also increase, and decrease when the monetary policy rate is decreased. There exists is symmetric co-integration if they respond in similar timeframe, and asymmetric if the change in one direction is faster than in the opposite direction.
The literature mentions the use of TAR and M -TAR models but try as i may, i haven't found a way of doing this kind of estimation.
Can anyone please help me with how to run a TAR or M-TAR specification or its equivalent in Stata?
Thank you.
I'm using Stata 14 and monthly time-series data for January 2000 to December 2015. My thesis is economics-related.
I am trying to establish the long-run and short-run relationship between various retail rates (mthtd, dddr, savr, alvr, etc) and monetary policy rate (mpr).
where,
mthth - 3-month deposit rate
dddr - demand deposit rate
savr - savings rate
avlr - average lending rate
Each of these variables is used as a dependent variable in a series of regressions with the independent variables being:
npl - non-performing loans
crr - reserve requirement coefficient
xcrate - exchange rate
vix - VIX volatility index
embi - EMBI spread
a dummy for inflation targeting period is included in all regresions.
The result after testing for unit root and co-integration led me to run ECM estimation, an example is shown below:
Code:
. vec avlr mpr crr vix xcrate dummyit
Code:
Vector error-correction model
Sample: 1960m4 - 1976m4 Number of obs = 193
AIC = 10.43915
Log likelihood = -954.378 HQIC = 10.80199
Det(Sigma_ml) = .0007949 SBIC = 11.33512
Equation Parms RMSE R-sq chi2 P>chi2
D_avlr 8 2.10038 0.2955 77.60149 0.0000
D_mpr 8 2.86624 0.2867 74.34033 0.0000
D_crr 8 .713627 0.2752 70.25058 0.0000
D_vix 8 3.00484 0.1464 31.73876 0.0001
D_xcrate 8 .078599 0.1338 28.5696 0.0004
D_dummyit 8 .073178 0.0093 1.738625 0.9880
Coef. Std. Err. z P>z [95% Conf. Interval]
D_avlr
_ce1
L1. .0269756 .0209137 1.29 0.197 -.0140145 .0679657
avlr
LD. -.3544857 .1617355 -2.19 0.028 -.6714815 -.03749
mpr
LD. -.047585 .1181624 -0.40 0.687 -.279179 .1840089
crr
LD. .6698698 .1950973 3.43 0.001 .2874861 1.052254
vix
LD. -.0210711 .0502179 -0.42 0.675 -.1194965 .0773543
xcrate
LD. .082976 1.891667 0.04 0.965 -3.624624 3.790576
dummyit
LD. .0003657 2.121369 0.00 1.000 -4.157442 4.158174
_cons -.0276036 .1556347 -0.18 0.859 -.332642 .2774347
D_mpr
_ce1
L1. .0605919 .0285395 2.12 0.034 .0046554 .1165283
avlr
LD. -.2348009 .2207096 -1.06 0.287 -.6673837 .197782
mpr
LD. -.2578047 .1612483 -1.60 0.110 -.5738455 .0582361
crr
LD. .6963436 .2662363 2.62 0.009 .1745301 1.218157
vix
LD. -.043976 .0685291 -0.64 0.521 -.1782905 .0903385
xcrate
LD. -.8315905 2.581432 -0.32 0.747 -5.891104 4.227923
dummyit
LD. .3711811 2.894891 0.13 0.898 -5.302701 6.045063
_cons -.0323948 .2123842 -0.15 0.879 -.4486602 .3838706
D_crr
_ce1
L1. -.008501 .0071057 -1.20 0.232 -.0224279 .0054259
avlr
LD. .0720169 .0549515 1.31 0.190 -.035686 .1797199
mpr
LD. -.0124776 .040147 -0.31 0.756 -.0911643 .0662092
crr
LD. -.438733 .0662866 -6.62 0.000 -.5686523 -.3088137
vix
LD. .0192707 .0170621 1.13 0.259 -.0141705 .0527119
xcrate
LD. .4796758 .6427159 0.75 0.455 -.7800242 1.739376
dummyit
LD. .2226528 .7207598 0.31 0.757 -1.190011 1.635316
_cons -.019047 .0528787 -0.36 0.719 -.1226873 .0845932
D_vix
_ce1
L1. -.1396939 .0299195 -4.67 0.000 -.1983351 -.0810527
avlr
LD. .4147958 .2313819 1.79 0.073 -.0387044 .868296
mpr
LD. -.3731921 .1690453 -2.21 0.027 -.7045149 -.0418693
crr
LD. -.595865 .27911 -2.13 0.033 -1.14291 -.0488195
vix
LD. .0705053 .0718427 0.98 0.326 -.0703039 .2113145
xcrate
LD. -2.792197 2.706256 -1.03 0.302 -8.09636 2.511967
dummyit
LD. -.0592391 3.034872 -0.02 0.984 -6.007478 5.889
_cons -.0182203 .2226539 -0.08 0.935 -.4546139 .4181733
D_xcrate
_ce1
L1. -.0001194 .0007826 -0.15 0.879 -.0016533 .0014145
avlr
LD. .0001613 .0060524 0.03 0.979 -.0117011 .0120238
mpr
LD. .0007353 .0044218 0.17 0.868 -.0079313 .0094019
crr
LD. .001499 .0073008 0.21 0.837 -.0128104 .0158084
vix
LD. .0009432 .0018792 0.50 0.616 -.00274 .0046264
xcrate
LD. -.2977891 .0707891 -4.21 0.000 -.4365331 -.1590451
dummyit
LD. -.0264471 .0793849 -0.33 0.739 -.1820386 .1291443
_cons .0236027 .0058241 4.05 0.000 .0121877 .0350177
D_dummyit
_ce1
L1. .0005587 .0007286 0.77 0.443 -.0008695 .0019868
avlr
LD. -.000468 .005635 -0.08 0.934 -.0115123 .0105763
mpr
LD. .0004337 .0041168 0.11 0.916 -.0076352 .0085026
crr
LD. -.0019113 .0067973 -0.28 0.779 -.0152338 .0114112
vix
LD. -.0004187 .0017496 -0.24 0.811 -.0038479 .0030105
xcrate
LD. -.0158339 .0659068 -0.24 0.810 -.1450089 .1133411
dummyit
LD. .0017346 .0739098 0.02 0.981 -.1431259 .1465951
_cons .0055801 .0054224 1.03 0.303 -.0050477 .0162078
Cointegrating equations
Equation Parms chi2 P>chi2
_ce1 5 47.47435 0.0000
Identification: beta is exactly identified
Johansen normalization restriction imposed
beta Coef. Std. Err. z P>z [95% Conf. Interval]
_ce1
avlr 1 . . . . .
mpr -1.555906 .2998338 -5.19 0.000 -2.143569 -.9682423
crr -.6836976 1.468203 -0.47 0.641 -3.561322 2.193927
vix 1.511256 .2750964 5.49 0.000 .9720773 2.050435
xcrate 5.275661 2.164469 2.44 0.015 1.03338 9.517943
dummyit -17.77735 4.573427 -3.89 0.000 -26.7411 -8.813596
_cons 30.2893 . . . . .
I intend to test for asymmetric co-integration. That is, if the monetary policy rate is increased, the other retail rates also increase, and decrease when the monetary policy rate is decreased. There exists is symmetric co-integration if they respond in similar timeframe, and asymmetric if the change in one direction is faster than in the opposite direction.
The literature mentions the use of TAR and M -TAR models but try as i may, i haven't found a way of doing this kind of estimation.
Can anyone please help me with how to run a TAR or M-TAR specification or its equivalent in Stata?
Thank you.
