Está en la página 1de 16

d.

Interprete los resultados del modelo VEC (corrección del


equilibrio) determinando cual es la relación de equilibrio o de
largo plazo, el coeficiente de ajuste de cada variable, etc.
Determine si alguna de las dos variables es débilmente exógena
en la ecuación de la otra.
Estimación con MCE
Vector Error Correction Estimates
Date: 01/19/15 Time: 07:53
Sample (adjusted): 1990Q4 2004Q2
Included observations: 55 after adjustments
Standard errors in ( ) & t-statistics in [ ]

Cointegrating Eq: CointEq1

CONSUMO(-1) 1.000000

PIB(-1) -0.684675
(0.03682)
[-18.5927]

C -313501.0

Error Correction: D(CONSUMO) D(PIB)

CointEq1 -0.408693 -0.284186


(0.19096) (0.27104)
[-2.14018] [-1.04850]

D(CONSUMO(-1)) -0.940522 -1.179921


(0.12608) (0.17896)
[-7.45942] [-6.59332]

D(CONSUMO(-2)) -0.548644 -0.712941


(0.12963) (0.18399)
[-4.23233] [-3.87486]

D(PIB(-1)) 0.165252 -0.540455


(0.14697) (0.20860)
[ 1.12439] [-2.59085]

D(PIB(-2)) 0.604274 0.084048


(0.07936) (0.11264)
[ 7.61419] [ 0.74616]

C 35061.58 112132.8
(12375.8) (17565.4)
[ 2.83308] [ 6.38374]

R-squared 0.944929 0.938086


Adj. R-squared 0.939309 0.931768
Sum sq. resids 2.73E+11 5.49E+11
S.E. equation 74595.60 105876.3
F-statistic 168.1510 148.4831
Log likelihood -691.9560 -711.2165
Akaike AIC 25.38022 26.08060
Schwarz SC 25.59920 26.29958
Mean dependent 28186.91 47436.02
S.D. dependent 302797.0 405325.6

Determinant resid covariance (dof adj.) 4.64E+19


Determinant resid covariance 3.68E+19
Log likelihood -1395.012
Akaike information criterion 51.23679
Schwarz criterion 51.74774

,la estimación del vector de cointegración. En este caso hay un único vector de
cointegración igual a CONSUMO = 313501.0+ 0.684675*pib(-1) (0.03682) [-
18.5927] .Donde el valor entre paréntesis es el error estándar estimado y el
valor entre corchetes el estadístico t.
El modelo de corrección del error (MCE). Para la ecuación que relaciona el
consumo con el pib, el parámetro que indica la velocidad de corrección es -
0.408693, con un error estándar de (0.19096)
y estadístico t de [-2.14018] . El resto de los parámetros indican el
comportamiento a corto plazo de las variables. En este caso el único parámetro
significativo (0.07936) es el del incremento de la D(PIB(-2)) .

9.
10.

[20:24, 23/1/2018] Econometris 2: El 9 y el 10

Calcular la F CON LAS SITGUIENES FORMULAS


La prueba F conjunta es igual a (0.2153-0.1397)/k-1/(1-0.2153)/n-k-1

0.0756/0.00121=62,48

Prueba F del arch (3) f=0.10756/0.00121=88.9669

F de tablas = 3,00475

Fcalculado=88.9669>f de tablas=3.00475

Con la prueba de hipotesis.


Ho: B2=0,b3=0,b4=0

H1: al menos alguno de los B es distinto de cero.

Por lo tanto se rechaza la hipotesis nula por lo cual se elige el modelo arch (3 ).

[20:24, 23/1/2018] Econometris 2: El 12 y el 13 no entiendo cuales son los incissos

11.

Null Hypothesis: IBM has a unit root


Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=26)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 0.682290 0.9918


Test critical values: 1% level -3.432743
5% level -2.862483
10% level -2.567317

*MacKinnon (1996) one-sided p-values.


Augmented Dickey-Fuller Test Equation
Dependent Variable: D(IBM)
Method: Least Squares
Date: 01/20/18 Time: 17:26
Sample: 1/02/1990 12/31/1999
Included observations: 2528

Variable Coefficient Std. Error t-Statistic Prob.

IBM(-1) 0.000509 0.000746 0.682290 0.4951


C 0.014462 0.034357 0.420953 0.6738

R-squared 0.000184 Mean dependent var 0.033105


Adjusted R-squared -0.000212 S.D. dependent var 1.047080
S.E. of regression 1.047191 Akaike info criterion 2.930891
Sum squared resid 2770.036 Schwarz criterion 2.935507
Log likelihood -3702.646 Hannan-Quinn criter. 2.932566
F-statistic 0.465519 Durbin-Watson stat 1.990778
Prob(F-statistic) 0.495118

Null Hypothesis: D(IBM) has a unit root


Exogenous: Constant
Lag Length: 4 (Automatic - based on SIC, maxlag=26)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -25.60344 0.0000


Test critical values: 1% level -3.432743
5% level -2.862483
10% level -2.567317

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation


Dependent Variable: D(IBM,2)
Method: Least Squares
Date: 01/21/18 Time: 07:02
Sample: 1/02/1990 12/31/1999
Included observations: 2528

Variable Coefficient Std. Error t-Statistic Prob.

D(IBM(-1)) -1.131415 0.044190 -25.60344 0.0000


D(IBM(-1),2) 0.133498 0.039255 3.400819 0.0007
D(IBM(-2),2) 0.155068 0.033910 4.572929 0.0000
D(IBM(-3),2) 0.129911 0.027949 4.648184 0.0000
D(IBM(-4),2) 0.104001 0.019814 5.248870 0.0000
C 0.037501 0.020765 1.806002 0.0710

R-squared 0.503872 Mean dependent var -0.000344


Adjusted R-squared 0.502889 S.D. dependent var 1.476989
S.E. of regression 1.041368 Akaike info criterion 2.921317
Sum squared resid 2734.974 Schwarz criterion 2.935167
Log likelihood -3686.545 Hannan-Quinn criter. 2.926343
F-statistic 512.2736 Durbin-Watson stat 1.997048
Prob(F-statistic) 0.000000
IBM
600

500

400
Frequency

300

200

100

0
10 20 30 40 50 60 70 80 90 100 110 120 130 140

Log Differenced IBM


700

600

500
Frequency

400

300

200

100

0
-.20 -.16 -.12 -.08 -.04 .00 .04 .08 .12 .16
12.

Dependent Variable: DLIBM


Method: ML - ARCH (Marquardt) - Normal distribution
Date: 11/21/15 Time: 07:43
Sample: 1/02/1990 12/31/1999
Included observations: 2528
Convergence achieved after 17 iterations
Presample variance: backcast (parameter = 0.7)
GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1)

Variable Coefficient Std. Error z-Statistic Prob.

C 0.000877 0.000349 2.512691 0.0120


AR(1) 0.020549 0.020572 0.998857 0.3179

Variance Equation

C 1.03E-05 1.81E-06 5.659100 0.0000


RESID(-1)^2 0.061218 0.006648 9.208711 0.0000
GARCH(-1) 0.914263 0.009556 95.67694 0.0000

R-squared -0.000082 Mean dependent var 0.000602


Adjusted R-squared -0.000478 S.D. dependent var 0.019072
S.E. of regression 0.019077 Akaike info criterion -5.138853
Sum squared resid 0.919274 Schwarz criterion -5.127311
Log likelihood 6500.510 Hannan-Quinn criter. -5.134665
Durbin-Watson stat 2.011595

Inverted AR Roots .02

13.

Null Hypothesis: SPX has a unit root


Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=26)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 1.751784 0.9997


Test critical values: 1% level -3.432743
5% level -2.862483
10% level -2.567317

*MacKinnon (1996) one-sided p-values.


Null Hypothesis: D(SPX) has a unit root
Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=26)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -50.41026 0.0001


Test critical values: 1% level -3.432743
5% level -2.862483
10% level -2.567317

*MacKinnon (1996) one-sided p-values.

Estimation Command:
=========================
ARCH(1,0,BACKCAST=0.7,DERIV=AA) DLSPX C AR(1) AR(2)

Estimation Equation:
=========================
DLSPX = C(1) + [AR(1)=C(2),AR(2)=C(3)]

GARCH = C(4) + C(5)*RESID(-1)^2

Substituted Coefficients:
=========================
DLSPX = 0.000597163104963 + [AR(1)=0.0392133724296,AR(2)=0.00903188438966]

GARCH = 6.42454683855e-05 + 0.193560120286*RESID(-1)^2

Dependent Variable: DLSPX


Method: ML - ARCH (Marquardt) - Normal distribution
Sample (adjusted): 1/04/1990 12/31/1999
Included observations: 2526 after adjustments
Convergence achieved after 20 iterations
Presample variance: backcast (parameter = 0.7)
GARCH = C(4) + C(5)*RESID(-1)^2

Variable Coefficient Std. Error z-Statistic Prob.

C 0.000597 0.000188 3.179740 0.0015


AR(1) 0.039213 0.018806 2.085199 0.0371
AR(2) 0.009032 0.016243 0.556051 0.5782

Variance Equation

C 6.42E-05 1.76E-06 36.54671 0.0000


RESID(-1)^2 0.193560 0.021994 8.800418 0.0000

R-squared -0.000343 Mean dependent var 0.000558


Adjusted R-squared -0.001136 S.D. dependent var 0.008885
S.E. of regression 0.008890 Akaike info criterion -6.645274
Sum squared resid 0.199413 Schwarz criterion -6.633725
Log likelihood 8397.980 Hannan-Quinn criter. -6.641083
Durbin-Watson stat 2.041518

Inverted AR Roots .12 -.08

Estimation Command:
=========================
ARCH(BACKCAST=0.7,DERIV=AA) DLSPX C AR(1) AR(2)

Estimation Equation:
=========================
DLSPX = C(1) + [AR(1)=C(2),AR(2)=C(3)]

GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1)

Substituted Coefficients:
=========================
DLSPX = 0.00059879827307 + [AR(1)=0.0471076309555,AR(2)=0.0157562905396]

GARCH = 5.81568633892e-07 + 0.0539564201378*RESID(-1)^2 + 0.939389114987*GARCH(-1)

Dependent Variable: DLSPX


Method: ML - ARCH (Marquardt) - Normal distribution
Sample (adjusted): 1/04/1990 12/31/1999
Included observations: 2526 after adjustments
Convergence achieved after 19 iterations
Presample variance: backcast (parameter = 0.7)
GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1)

Variable Coefficient Std. Error z-Statistic Prob.

C 0.000599 0.000160 3.735535 0.0002


AR(1) 0.047108 0.021665 2.174391 0.0297
AR(2) 0.015756 0.019710 0.799411 0.4241

Variance Equation

C 5.82E-07 1.39E-07 4.183000 0.0000


RESID(-1)^2 0.053956 0.005467 9.869837 0.0000
GARCH(-1) 0.939389 0.006233 150.7115 0.0000

R-squared -0.000980 Mean dependent var 0.000558


Adjusted R-squared -0.001773 S.D. dependent var 0.008885
S.E. of regression 0.008893 Akaike info criterion -6.808809
Sum squared resid 0.199540 Schwarz criterion -6.794950
Log likelihood 8605.526 Hannan-Quinn criter. -6.803780
Durbin-Watson stat 2.055921

Inverted AR Roots .15 -.10

[20:24, 23/1/2018] Econometris 2: Y el 14 tambken


R_FTSE
350

300

250
Frequency

200

150

100

50

0
-.05 -.04 -.03 -.02 -.01 .00 .01 .02 .03 .04 .05 .06

Null Hypothesis: R_FTSE has a unit root


Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=27)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -47.56851 0.0001


Test critical values: 1% level -3.432663
5% level -2.862447
10% level -2.567298

*MacKinnon (1996) one-sided p-values.

Dependent Variable: R_FTSE


Method: ML - ARCH (Marquardt) - Normal distribution
Date: 11/21/15 Time: 08:41
Sample: 1/01/1990 12/31/1999
Included observations: 2610
Convergence achieved after 8 iterations
Presample variance: backcast (parameter = 0.7)
GARCH = C(3) + C(4)*RESID(-1)^2

Variable Coefficient Std. Error z-Statistic Prob.

C 0.000434 0.000192 2.253740 0.0242


AR(1) 0.075194 0.019209 3.914483 0.0001

Variance Equation

C 7.39E-05 2.11E-06 35.07479 0.0000


RESID(-1)^2 0.161291 0.020231 7.972485 0.0000

R-squared 0.004944 Mean dependent var 0.000391


Adjusted R-squared 0.004563 S.D. dependent var 0.009398
S.E. of regression 0.009377 Akaike info criterion -6.524781
Sum squared resid 0.229296 Schwarz criterion -6.515789
Log likelihood 8518.839 Hannan-Quinn criter. -6.521523
Durbin-Watson stat 2.001995

Inverted AR Roots .08

Estimation Command:
=========================
ARCH(BACKCAST=0.7,DERIV=AA) R_FTSE C AR(1)

Estimation Equation:
=========================
R_FTSE = C(1) + [AR(1)=C(2)]

GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1)

Substituted Coefficients:
=========================
R_FTSE = 0.000461729814631 + [AR(1)=0.0625343790523]

GARCH = 8.22083214083e-07 + 0.0508775126148*RESID(-1)^2 + 0.940245848504*GARCH(-1)

Dependent Variable: R_FTSE


Method: ML - ARCH (Marquardt) - Normal distribution
Date: 11/21/15 Time: 08:43
Sample: 1/01/1990 12/31/1999
Included observations: 2610
Convergence achieved after 13 iterations
Presample variance: backcast (parameter = 0.7)
GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1)

Variable Coefficient Std. Error z-Statistic Prob.

C 0.000462 0.000169 2.731890 0.0063


AR(1) 0.062534 0.020697 3.021458 0.0025

Variance Equation

C 8.22E-07 2.42E-07 3.392187 0.0007


RESID(-1)^2 0.050878 0.006660 7.638899 0.0000
GARCH(-1) 0.940246 0.007975 117.8980 0.0000

R-squared 0.004868 Mean dependent var 0.000391


Adjusted R-squared 0.004486 S.D. dependent var 0.009398
S.E. of regression 0.009377 Akaike info criterion -6.648590
Sum squared resid 0.229314 Schwarz criterion -6.637351
Log likelihood 8681.411 Hannan-Quinn criter. -6.644519
Durbin-Watson stat 1.977721

Inverted AR Roots .06


14.

15.

16.

17.

La variable UNEM no es significativa para ningún año y si lo fuera seria al 30 %

Además al aumentar variables esto no cambia, serie ideal ver el análisis con un modelo de datos
de panel con efectos fijos.

Una posibilidad es explicar los datos con el modelo de efectos fijos considera que existe un
término constante diferente para cada individuo, y supone que los efectos individuales son
independientes entre sí.

Con este modelo se considera que las variables explicativas afectan por igual a las unidades de
corte transversal y que éstas se diferencian por características propias de cada una de ellas,
medidas por medio del intercepto. Es por ello que los N interceptos se asocian con variables
dummy con coeficientes específicos para cada unidad, los cuales se deben estimar.

También podría gustarte