Econometria

DEFINICION DE LOS MODELOS DE REGRESION
1. MODELO 1
LOGDOWt=β1+β2LOGIPIt+β3LOGICONFt+β4DESt+β5INTREF+εt

Dependent Variable: LOGDOW | | |
Method: Least Squares | | |
Date: 04/30/10 Time: 18:24 | | |
Sample: 2000M04 2010M01 | | |
Included observations: 118 | | |
| | | | |
| | | | |
Variable | Coefficient | Std. Error | t-Statistic | Prob.   |
| | || |
| | | | |
C | -3.421151 | 1.017740 | -3.361518 | 0.0011 |
LOGIPI | 2.526653 | 0.202108 | 12.50147 | 0.0000 |
LOGICONF | 0.154248 | 0.033616 | 4.588570 | 0.0000 |
DES | 0.035337 | 0.008961 | 3.943463 | 0.0001 |
INTREF | 0.013373 | 0.006173 | 2.166279 | 0.0324 |
| | | | |
| | | | |
R-squared | 0.747128 |     Mean dependent var | 9.245351 |
Adjusted R-squared| 0.738176 |     S.D. dependent var | 0.140865 |
S.E. of regression | 0.072079 |     Akaike info criterion | -2.380660 |
Sum squared resid | 0.587077 |     Schwarz criterion | -2.263258 |
Log likelihood | 145.4590 |     F-statistic | 83.46639 |
Durbin-Watson stat | 0.389544 |     Prob(F-statistic) | 0.000000 |

INTERPRETACION
SIGNIFICANCIA INDIVIDUAL
Con un nivel de significanciadel 5%, se rechazan todas las hipótesis nulas, las cuales establecen que los coeficientes son estadísticamente iguales a cero. Dado esto las cuatro variables y la constante son significativas ya que su probabilidad siempre es menor al 5 %

SIGNIFICANCIA CONJUNTA
INTERPRETACION DE LOS COEFICIENTES
ANALISIS DE VARIANZA

GRAFICAS QUE EXPLICAN EL MODELO.

MATRIZ DE CORRELACIONES ENTRE LASVARIABLES INDEPENDIENTES

| LOGIPI | LOGICONF | DES | INTREF |
LOGIPI |  1.000000 |  0.293861 | -0.556363 |  0.576334 |
LOGICONF |  0.293861 |  1.000000 | -0.766373 |  0.719480 |
DES | -0.556363 | -0.766373 |  1.000000 | -0.784787 |
INTREF |  0.576334 |  0.719480 | -0.784787 |  1.000000 |

MATRIZ DE COVARIANZAS ENTRE LAS VARIABLES INDEPENDIENTES

| LOGIPI | LOGICONF | DES |INTREF |
LOGIPI |  0.001932 |  0.004450 | -0.035361 |  0.049579 |
LOGICONF |  0.004450 |  0.118690 | -0.381764 |  0.485099 |
DES | -0.035361 | -0.381764 |  2.090715 | -2.220770 |
INTREF |  0.049579 |  0.485099 | -2.220770 |  3.830082 |

2. MODELO 2

LOGDOWt=β1+β2LOGIPIt+β3LOGICONFt+β4DESt+β5INTREF+β6εt-1+εt
Este modelo que presentamos agrega una nueva variable que es la mediamóvil, que quiere decir que los hechos inesperados o aleatorios, afectan el comportamiento del dow jones, porque hay noticias relevantes y situaciones significativas del ayer que afectan actualmente el comportamiento del Dow Jones.

Dependent Variable: LOGDOW | | |
Method: Least Squares | | |
Date: 04/30/10 Time: 19:25 | | |
Sample: 2000M04 2010M01 | | |
Included observations:118 | | |
Convergence achieved after 20 iterations | |
Backcast: 2000M03 | | |
| | | | |
| | | | |
Variable | Coefficient | Std. Error | t-Statistic | Prob.   |
| | | | |
| | | | |
C | -2.374845 | 1.192204 | -1.991979 | 0.0488 |
LOGIPI | 2.325968 | 0.239396 | 9.715979 | 0.0000 |
LOGICONF | 0.131217 | 0.036145 | 3.630288 | 0.0004 |
DES | 0.031464 |0.009900 | 3.178273 | 0.0019 |
INTREF | 0.017112 | 0.007299 | 2.344561 | 0.0208 |
MA(1) | 0.672341 | 0.070193 | 9.578490 | 0.0000 |
| | | | |
| | | | |
R-squared | 0.867085 |     Mean dependent var | 9.245351 |
Adjusted R-squared | 0.861151 |     S.D. dependent var | 0.140865 |
S.E. of regression | 0.052490 |     Akaike info criterion | -3.006887 |
Sum squared resid |0.308580 |     Schwarz criterion | -2.866004 |
Log likelihood | 183.4063 |     F-statistic | 146.1287 |
Durbin-Watson stat | 1.296321 |     Prob(F-statistic) | 0.000000 |
| | | | |
| | | | |
Inverted MA Roots |      -.67 | | |
| | | | |
| | | | |

INTERPRETACION
SIGNIFICANCIA INDIVIDUAL
Con un nivel de significancia del 5%, se rechazan todas las hipótesis...