JPM Morgan Chase
Comportamiento del precio de la cción de JP MORGAN
Figura 1. Precio de Cierre de JPM: enero 2015 - marzo 2020
Fuente: elaboración propia con salida de R.Fuente: elaboración propia con salida de R.
Gráfica de la serie en rendimientos
Figura 2. Rendimientos en logaritmos de JPM: enero de 2015 a marzo 2020
Fuente: elaboración propia con salida de R.
Fuente: elaboración propia con salida de R.
Tabla 1. Prueba de efectos ARCH
| Prueba | Valor p | Ho | Resultado |
|---|---|---|---|
| ARCH test | 2.20E-16 | La serie No tiene efectos ARCH | RECHAZO H0 |
Modelos ARCH
El primer modelo a implementar es un ARCH 1
La modelación de la varianza con el ARCH(1) se presenta en la figura 4.
Figura 4. ARCH(1) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(1,0)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.00013 0.000007 19.6184 0
alpha1 0.48825 0.051236 9.5295 0
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.00013 0.000011 11.4236 0.0e+00
alpha1 0.48825 0.122354 3.9905 6.6e-05
LogLikelihood : 3812.676
Information Criteria
------------------------------------
Akaike -5.6918
Bayes -5.6841
Shibata -5.6918
Hannan-Quinn -5.6889
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.3456 0.5566
Lag[2*(p+q)+(p+q)-1][2] 0.8921 0.5346
Lag[4*(p+q)+(p+q)-1][5] 2.6037 0.4842
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 1.398 2.370e-01
Lag[2*(p+q)+(p+q)-1][2] 4.912 4.294e-02
Lag[4*(p+q)+(p+q)-1][5] 21.789 6.227e-06
d.o.f=1
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[2] 7.006 0.500 2.000 8.124e-03
ARCH Lag[4] 24.091 1.397 1.611 1.083e-06
ARCH Lag[6] 33.041 2.222 1.500 1.125e-08
Nyblom stability test
------------------------------------
Joint Statistic: 2.3735
Individual Statistics:
omega 0.2517
alpha1 2.0239
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 0.61 0.749 1.07
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.7378 0.46079
Negative Sign Bias 1.9836 0.04751 **
Positive Sign Bias 0.9546 0.33997
Joint Effect 5.3147 0.15015
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 115.9 6.554e-16
2 30 130.8 6.898e-15
3 40 148.9 9.490e-15
4 50 158.1 2.124e-13
Elapsed time : 0.2336578
Fuente: elaboración propia con salida de R.
El siguiente modelo a implementar es un ARCH 2
La caracterización de la varianza con el ARCH(2) se presenta en la figura 5.
Figura 5. ARCH(2) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(2,0)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000111 0.000007 17.0252 0e+00
alpha1 0.393574 0.050832 7.7426 0e+00
alpha2 0.147801 0.033450 4.4185 1e-05
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000111 0.000011 10.5212 0.000000
alpha1 0.393574 0.099650 3.9496 0.000078
alpha2 0.147801 0.038840 3.8053 0.000142
LogLikelihood : 3837.756
Information Criteria
------------------------------------
Akaike -5.7278
Bayes -5.7161
Shibata -5.7278
Hannan-Quinn -5.7234
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.7025 0.4020
Lag[2*(p+q)+(p+q)-1][2] 1.3783 0.3903
Lag[4*(p+q)+(p+q)-1][5] 2.5054 0.5047
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.8359 0.36056
Lag[2*(p+q)+(p+q)-1][5] 4.7784 0.17170
Lag[4*(p+q)+(p+q)-1][9] 12.9480 0.01098
d.o.f=2
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[3] 1.917 0.500 2.000 0.166145
ARCH Lag[5] 5.280 1.440 1.667 0.088524
ARCH Lag[7] 12.657 2.315 1.543 0.004314
Nyblom stability test
------------------------------------
Joint Statistic: 1.7094
Individual Statistics:
omega 0.2027
alpha1 0.9643
alpha2 0.6719
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 0.846 1.01 1.35
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.4947 0.62091
Negative Sign Bias 2.1710 0.03011 **
Positive Sign Bias 1.0825 0.27925
Joint Effect 7.2536 0.06424 *
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 94.59 5.043e-12
2 30 112.40 9.252e-12
3 40 118.89 5.230e-10
4 50 125.49 1.233e-08
Elapsed time : 0.192493
Fuente: elaboración propia con salida de R.
El siguiente modelo es un ARCH 3
Figura 6. ARCH(3) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(3,0)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000101 0.000007 14.8682 0.000000
alpha1 0.324079 0.048113 6.7358 0.000000
alpha2 0.108861 0.032812 3.3178 0.000907
alpha3 0.132955 0.037866 3.5112 0.000446
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000101 0.000010 9.7624 0.000000
alpha1 0.324079 0.084679 3.8272 0.000130
alpha2 0.108861 0.039962 2.7241 0.006447
alpha3 0.132955 0.044586 2.9820 0.002864
LogLikelihood : 3850.734
Information Criteria
------------------------------------
Akaike -5.7457
Bayes -5.7301
Shibata -5.7457
Hannan-Quinn -5.7399
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.5675 0.4513
Lag[2*(p+q)+(p+q)-1][2] 1.4381 0.3756
Lag[4*(p+q)+(p+q)-1][5] 2.8593 0.4334
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.9395 0.3324
Lag[2*(p+q)+(p+q)-1][8] 5.3405 0.3086
Lag[4*(p+q)+(p+q)-1][14] 11.0601 0.1342
d.o.f=3
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[4] 0.5476 0.500 2.000 0.4593
ARCH Lag[6] 3.4024 1.461 1.711 0.2532
ARCH Lag[8] 5.5575 2.368 1.583 0.1952
Nyblom stability test
------------------------------------
Joint Statistic: 1.0558
Individual Statistics:
omega 0.1105
alpha1 0.4989
alpha2 0.3983
alpha3 0.3114
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 1.07 1.24 1.6
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.3115 0.75550
Negative Sign Bias 2.2108 0.02722 **
Positive Sign Bias 0.9959 0.31949
Joint Effect 7.7707 0.05100 *
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 86.38 1.445e-10
2 30 93.99 8.755e-09
3 40 119.85 3.748e-10
4 50 127.21 7.135e-09
Elapsed time : 0.2862518
Fuente: elaboración propia con salida de R.
Hasta el momento, los modelos ARCH(1)y ARCH(2) y ARCH(3) cumplen con todas las condiciones; la sumatoria de los parametros es menro a 1, son significativos y no son negativos.
El siguiente modelo es un ARCH 4
Figura 7. ARCH(4) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(4,0)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000096 0.000007 14.4623 0.000000
alpha1 0.284531 0.048930 5.8151 0.000000
alpha2 0.106493 0.032363 3.2905 0.001000
alpha3 0.076284 0.034896 2.1860 0.028815
alpha4 0.099062 0.032821 3.0182 0.002543
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000096 0.000010 9.4685 0.000000
alpha1 0.284531 0.081528 3.4900 0.000483
alpha2 0.106493 0.039803 2.6755 0.007462
alpha3 0.076284 0.038312 1.9911 0.046471
alpha4 0.099062 0.040060 2.4728 0.013405
LogLikelihood : 3858.277
Information Criteria
------------------------------------
Akaike -5.7555
Bayes -5.7360
Shibata -5.7555
Hannan-Quinn -5.7482
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.3846 0.5351
Lag[2*(p+q)+(p+q)-1][2] 1.2797 0.4158
Lag[4*(p+q)+(p+q)-1][5] 2.5724 0.4907
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.2761 0.5993
Lag[2*(p+q)+(p+q)-1][11] 4.5918 0.6430
Lag[4*(p+q)+(p+q)-1][19] 7.8446 0.6931
d.o.f=4
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[5] 0.4591 0.500 2.000 0.4981
ARCH Lag[7] 2.9612 1.473 1.746 0.3264
ARCH Lag[9] 5.3293 2.402 1.619 0.2338
Nyblom stability test
------------------------------------
Joint Statistic: 1.2491
Individual Statistics:
omega 0.07541
alpha1 0.23239
alpha2 0.44287
alpha3 0.33667
alpha4 0.61847
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 1.28 1.47 1.88
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.1755 0.86072
Negative Sign Bias 1.9717 0.04885 **
Positive Sign Bias 0.7206 0.47127
Joint Effect 5.9959 0.11181
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 78.49 3.380e-09
2 30 91.07 2.485e-08
3 40 113.64 3.202e-09
4 50 115.56 2.675e-07
Elapsed time : 0.3626602
Fuente: elaboración propia con salida de R.
Modelos GARCH
ahora se hace el ajuste con un GARCH (1,1):
Figura 8. GARCH(1,1) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(1,1)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000026 0.000006 4.4567 8e-06
alpha1 0.188195 0.033774 5.5721 0e+00
beta1 0.689322 0.052813 13.0521 0e+00
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000026 0.000010 2.5775 0.009953
alpha1 0.188195 0.070477 2.6703 0.007578
beta1 0.689322 0.098338 7.0097 0.000000
LogLikelihood : 3866.252
Information Criteria
------------------------------------
Akaike -5.7704
Bayes -5.7587
Shibata -5.7704
Hannan-Quinn -5.7660
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.2801 0.5966
Lag[2*(p+q)+(p+q)-1][2] 1.1682 0.4468
Lag[4*(p+q)+(p+q)-1][5] 2.5839 0.4883
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.5384 0.4631
Lag[2*(p+q)+(p+q)-1][5] 1.0764 0.8421
Lag[4*(p+q)+(p+q)-1][9] 1.7800 0.9296
d.o.f=2
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[3] 0.2327 0.500 2.000 0.6296
ARCH Lag[5] 0.6630 1.440 1.667 0.8347
ARCH Lag[7] 0.7267 2.315 1.543 0.9535
Nyblom stability test
------------------------------------
Joint Statistic: 0.6693
Individual Statistics:
omega 0.04325
alpha1 0.31492
beta1 0.14052
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 0.846 1.01 1.35
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.08139 0.93514
Negative Sign Bias 2.55215 0.01082 **
Positive Sign Bias 0.39786 0.69080
Joint Effect 8.95182 0.02994 **
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 71.59 4.995e-08
2 30 74.50 7.182e-06
3 40 94.88 1.477e-06
4 50 98.98 3.123e-05
Elapsed time : 0.2206969
Fuente: elaboración propia salida de R
Ahora, se hace el ajuste con un GARCH (1,2)
Figura 9. GARCH(1,2) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(1,2)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000029 0.000007 4.4768 0.000008
alpha1 0.218100 0.038720 5.6327 0.000000
beta1 0.410163 0.137022 2.9934 0.002759
beta2 0.232942 0.115062 2.0245 0.042919
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000029 0.000010 2.9426 0.003255
alpha1 0.218100 0.073558 2.9650 0.003027
beta1 0.410163 0.153942 2.6644 0.007712
beta2 0.232942 0.143354 1.6249 0.104175
LogLikelihood : 3867.661
Information Criteria
------------------------------------
Akaike -5.7710
Bayes -5.7554
Shibata -5.7710
Hannan-Quinn -5.7651
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.2605 0.6098
Lag[2*(p+q)+(p+q)-1][2] 1.1463 0.4531
Lag[4*(p+q)+(p+q)-1][5] 2.5641 0.4924
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.0692 0.7925
Lag[2*(p+q)+(p+q)-1][8] 0.6035 0.9936
Lag[4*(p+q)+(p+q)-1][14] 2.1165 0.9888
d.o.f=3
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[4] 0.1691 0.500 2.000 0.6809
ARCH Lag[6] 0.4102 1.461 1.711 0.9166
ARCH Lag[8] 1.0083 2.368 1.583 0.9231
Nyblom stability test
------------------------------------
Joint Statistic: 0.7541
Individual Statistics:
omega 0.04212
alpha1 0.31142
beta1 0.13532
beta2 0.12666
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 1.07 1.24 1.6
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.1302 0.89641
Negative Sign Bias 2.3115 0.02096 **
Positive Sign Bias 0.6606 0.50900
Joint Effect 7.9348 0.04738 **
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 70.87 6.581e-08
2 30 81.66 6.593e-07
3 40 95.35 1.273e-06
4 50 102.41 1.217e-05
Elapsed time : 0.631649
Fuente: elaboración propia salida de R
Ahora se el ajuste con un GARCH(2,1)
Figura 10. GARCH(2,1) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(2,1)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000026 0.000009 2.974601 0.002934
alpha1 0.188201 0.036540 5.150517 0.000000
alpha2 0.000000 0.056880 0.000001 0.999999
beta1 0.689148 0.087514 7.874731 0.000000
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000026 0.000017 1.494756 0.134978
alpha1 0.188201 0.065719 2.863720 0.004187
alpha2 0.000000 0.096592 0.000001 0.999999
beta1 0.689148 0.182404 3.778138 0.000158
LogLikelihood : 3866.051
Information Criteria
------------------------------------
Akaike -5.7686
Bayes -5.7530
Shibata -5.7686
Hannan-Quinn -5.7627
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.2684 0.6044
Lag[2*(p+q)+(p+q)-1][2] 1.1362 0.4561
Lag[4*(p+q)+(p+q)-1][5] 2.5504 0.4953
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.5474 0.4594
Lag[2*(p+q)+(p+q)-1][8] 1.4180 0.9365
Lag[4*(p+q)+(p+q)-1][14] 3.0367 0.9535
d.o.f=3
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[4] 0.2860 0.500 2.000 0.5928
ARCH Lag[6] 0.5599 1.461 1.711 0.8751
ARCH Lag[8] 1.0547 2.368 1.583 0.9161
Nyblom stability test
------------------------------------
Joint Statistic: 2.4494
Individual Statistics:
omega 0.04305
alpha1 0.31578
alpha2 0.81065
beta1 0.14070
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 1.07 1.24 1.6
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.09385 0.92524
Negative Sign Bias 2.57255 0.01020 **
Positive Sign Bias 0.40405 0.68624
Joint Effect 9.05758 0.02854 **
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 71.38 5.414e-08
2 30 73.20 1.094e-05
3 40 94.28 1.778e-06
4 50 97.26 4.955e-05
Elapsed time : 0.203861
Fuente: elaboración propia salida de R
Finalmente, se hace el ajuste con un GARCH(2,2)
Figura 11. GARCH(2,2) vs rendimientos
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(2,2)
Mean Model : ARFIMA(0,0,0)
Distribution : norm
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
omega 0.000029 0.000028 1.033695 0.301279
alpha1 0.218105 0.044913 4.856165 0.000001
alpha2 0.000000 0.219360 0.000002 0.999998
beta1 0.410125 0.862834 0.475323 0.634557
beta2 0.232988 0.540734 0.430873 0.666561
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
omega 0.000029 0.000063 0.460572 0.64511
alpha1 0.218105 0.083156 2.622834 0.00872
alpha2 0.000000 0.487437 0.000001 1.00000
beta1 0.410125 1.803834 0.227363 0.82014
beta2 0.232988 1.069582 0.217831 0.82756
LogLikelihood : 3867.661
Information Criteria
------------------------------------
Akaike -5.7695
Bayes -5.7501
Shibata -5.7695
Hannan-Quinn -5.7622
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 0.2605 0.6098
Lag[2*(p+q)+(p+q)-1][2] 1.1463 0.4531
Lag[4*(p+q)+(p+q)-1][5] 2.5641 0.4924
d.o.f=0
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 0.06919 0.7925
Lag[2*(p+q)+(p+q)-1][11] 1.42195 0.9874
Lag[4*(p+q)+(p+q)-1][19] 3.08935 0.9942
d.o.f=4
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[5] 0.2876 0.500 2.000 0.5918
ARCH Lag[7] 0.3203 1.473 1.746 0.9434
ARCH Lag[9] 2.0213 2.402 1.619 0.7544
Nyblom stability test
------------------------------------
Joint Statistic: 1.6161
Individual Statistics:
omega 0.04215
alpha1 0.31165
alpha2 0.45997
beta1 0.13546
beta2 0.12679
Asymptotic Critical Values (10% 5% 1%)
Joint Statistic: 1.28 1.47 1.88
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
t-value prob sig
Sign Bias 0.1302 0.89641
Negative Sign Bias 2.3114 0.02096 **
Positive Sign Bias 0.6606 0.50897
Joint Effect 7.9346 0.04738 **
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 70.87 6.581e-08
2 30 81.66 6.593e-07
3 40 95.35 1.273e-06
4 50 102.41 1.217e-05
Elapsed time : 0.2999051
Fuente: elaboración propia salida de R
Selección de modelo y simulación de los rendimientos
| MODELO | \(\omega\) | \(\alpha_{1}\) | \(\alpha_{2}\) | \(\alpha_{3}\) | \(\alpha_{4}\) | \(\beta_{1}\) | \(\beta_{2}\) | AKAIKE | BAYES |
|---|---|---|---|---|---|---|---|---|---|
| ARCH(1) | 0.00013 | 0.48825 | -5.6918 | -5.6841 | |||||
| ARCH(2) | 0.000111 | 0.393574 | 0.147801 | -5.7278 | -5.7161 | ||||
| ARCH(3) | 0.000101 | 0.324079 | 0.108861 | 0.132955 | -5.7457 | -5.7301 | |||
| ARCH(4) | 0.000096 | 0.284531 | 0.106493 | 0.076284 | 0.099062 | -5.7555 | -5.736 | ||
| GARCH(1,1) | 0.000026 | 0.188195 | 0.689322 | -5.7704 | -5.7587 | ||||
| GARCH(1,2) | 0.000029 | 0.218100 | 0.410163 | 0.232942 | -5.7710 | -5.7554 | |||
| GARCH(2,1) | 0.000026 | 0.188201 | 0.000000 | 0.689148 | -5.7686 | -5.753 | |||
| GARCH(2,2) | 0.000029 | 0.218105 | 0.000000 | 0.410125 | 0.232988 | -5.7695 | -5.7501 |
La figura 12 muestra los resultados de la simulación
Figura 12. Simulación del ARCH(4) y GARCH(1,1) vs rendimientos
Fuente: elaboración propia de R
Conclusión final
Referencias
J.P.Morgan. (s.f.). J.P.Morgan en México . Obtenido de https://www.jpmorgan.com.mx/es/about-us JP Morgan Chase. (s.f.). Obtenido de https://es.wikipedia.org/wiki/JP_Morgan_Chase