A PARTIR DE LA INFORMACIÓN SUMINISTRADA EN LAS SIGUIENTES HOJAS, CONTRASTAR EL CUMPLIMIENTO DE LAS HIPÓTESIS DEL MBRL Y JUSTIFICAR LAS SUCESIVAS ELECCIONES DE LAS REGRESIONES 1ª, 2ª, 3ª Y DEFINITIVA. DETERMINAR LA VALIDEZ FINALMENTE SELECCIONADO DEL MODELO Econometría I. 3º LADE Prof. Rafael de Arce Enero 2003 [email protected] Regresión inicial: Dependent Variable: @PCH(FBCK) Method: Least Squares Sample(adjusted): 1981:1 2002:1 Included observations: 85 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(-1)) @PCH(EBEF) MIBOR3M(-4) @PCH(FBCK(-1)) -0.000849 -0.082837 0.003012 0.000145 0.966078 0.003030 0.222276 0.106613 0.000159 0.073065 -0.280249 -0.372675 0.028247 0.912442 13.22216 0.7800 0.7104 0.9775 0.3643 0.0000 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.872550 0.866177 0.005968 0.002849 317.2840 0.602030 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.010078 0.016314 -7.347858 -7.204172 136.9238 0.000000 2ª Regresión propuesta Dependent Variable: @PCH(FBCK) Method: Least Squares Sample(adjusted): 1981:1 2002:1 Included observations: 85 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(-1)+FBCK(-1)) @PCH(EBEF) MIBOR3M(-4) -0.012515 1.866697 0.385611 6.88E-05 0.003183 0.112127 0.114533 0.000201 -3.932263 16.64812 3.366813 0.341568 0.0002 0.0000 0.0012 0.7336 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.793133 0.785472 0.007556 0.004625 296.6992 0.466342 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.010078 0.016314 -6.887040 -6.772092 103.5189 0.000000 3ª Regresión propuesta Dependent Variable: @PCH(FBCK) Method: Least Squares Sample(adjusted): 1981:1 2002:2 Included observations: 86 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(-1)+FBCK(-1)) @PCH(EBEF) MIBOR3M(-4) FIC923 -0.040247 2.276896 0.556332 0.001323 0.014296 0.007098 0.139777 0.110873 0.000346 0.003306 -5.670539 16.28946 5.017732 3.824841 4.324401 0.0000 0.0000 0.0000 0.0003 0.0000 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.830869 0.822517 0.006833 0.003781 309.3484 0.726528 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.010059 0.016218 -7.077869 -6.935174 99.47968 0.000000 INFORMACIÓN DE APOYO PARA LA TOMA DE DECISIONES A. Correlación entre las variables explicativas: @PCH(FBCK) @PCH(FBCK) @PCH(GTOHOGK(-1)) @PCH(EBEF) MIBOR3M(-4) @PCH(FBCK(-1)) @PCH(GTOHOG K(-1)) 0.709218 1.000000 -0.359500 -0.402549 0.778941 1.000000 0.709218 0.020301 -0.250831 0.931744 @PCH(EBEF) MIBOR3M(-4) @PCH(FBCK(-1)) 0.020301 -0.359500 1.000000 0.428392 -0.006646 -0.250831 -0.402549 0.428392 1.000000 -0.307170 0.931744 0.778941 -0.006646 -0.307170 1.000000 B. Gráficos de ajuste y de residuos: 1ª Regresión propuesta 2ª Regresión propuesta 0.04 0.04 0.02 0.02 0.00 0.00 -0.02 0.02 -0.02 0.02 0.01 -0.04 -0.04 0.01 0.00 0.00 -0.01 -0.01 -0.02 -0.02 -0.03 82 84 86 88 90 92 Residual 94 Actual 96 98 00 02 82 84 Fitted 86 88 90 Residual 92 94 Actual 96 98 00 02 Fitted C. Regresión parcial 1 Dependent Variable: @PCH(GTOHOGK(-1)) Method: Least Squares Sample(adjusted): 1981:1 2002:2 Included observations: 86 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(EBEF) MIBOR3M(-4) @PCH(FBCK(-1)) 0.009105 -0.246827 -1.96E-05 0.266731 0.001109 0.045424 7.80E-05 0.020973 8.209129 -5.433812 -0.251240 12.71762 0.0000 0.0000 0.8023 0.0000 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.732506 0.722720 0.002967 0.000722 380.5656 0.454711 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.006011 0.005634 -8.757339 -8.643184 74.84976 0.000000 D. Regresión parcial 2 Dependent Variable: @PCH(FBCK(-1)) Method: Least Squares Sample(adjusted): 1981:1 2002:2 Included observations: 86 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(-1)) @PCH(EBEF) MIBOR3M(-4) -0.016824 2.487796 0.719925 -0.000373 0.004177 0.195618 0.140904 0.000235 -4.027306 12.71762 5.109325 -1.587571 0.0001 0.0000 0.0000 0.1162 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.701748 0.690837 0.009060 0.006731 284.5504 0.435285 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.009906 0.016295 -6.524429 -6.410273 64.31190 0.000000 E. Regresión parcial 3 Dependent Variable: @PCH(FBCK) Method: Least Squares Sample(adjusted): 1981:1 1992:2 Included observations: 46 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(1)+FBCK(-1)) @PCH(EBEF) MIBOR3M(-4) -0.053147 2.622747 0.010536 0.199072 -5.044215 13.17484 0.0000 0.0000 0.646715 0.001849 0.133171 0.000519 4.856264 3.563080 0.0000 0.0009 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.862086 0.852235 0.006348 0.001693 169.5614 0.619415 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.011589 0.016515 -7.198321 -7.039309 87.51244 0.000000 F. Regresión parcial 4 Dependent Variable: @PCH(FBCK) Method: Least Squares Sample: 1992:3 2002:2 Included observations: 40 Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(1)+FBCK(-1)) @PCH(EBEF) MIBOR3M(-4) -0.025198 1.765159 0.005754 0.223043 -4.379188 7.913980 0.0001 0.0000 0.985561 0.000598 0.268364 0.000473 3.672480 1.265389 0.0008 0.2139 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.829528 0.815322 0.006831 0.001680 144.8036 0.919089 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.008300 0.015895 -7.040179 -6.871291 58.39261 0.000000 G. Tabla F-Snedecor (para el 95% de confianza) Grados libertad 4,77 Valor de tablas 13,65 H. Descripción de las variables @PCH(FBCK) @PCH(GTOHOGK(-1)) @PCH(EBEF) MIBOR3M(-4) @PCH(FBCK(-1)) FIC923 FICJUN92 Cto. Intertrimestral Form Bta. Capital Fijo Cto. Intertrimestral Consumo privado período precedente Cto. Intertrimestral Excedente Bruto de Explotación Tipo de Interés MIBOR 3 meses hace cuatro trimestres Endógena desplazada un período Ficticia de escalón: unos desde 1992:3 y ceros resto Ficticia de impacto: todo ceros y un uno en 1992:2: EXPO Y BARCELONA REGRESIÓN SELECCIONADA DEFINITIVAMENTE Dependent Variable: @PCH(FBCK) Method: Least Squares Sample(adjusted): 1981:1 2002:2 Included observations: 86 after adjusting endpoints Variable Coefficient Std. Error t-Statistic Prob. C @PCH(GTOHOGK(-1)+FBCK(-1)) @PCH(EBEF) MIBOR3M(-4) FIC923 FICJUN92 -0.026858 2.054360 0.213218 0.001191 0.010534 -0.019206 0.006780 0.130507 0.119023 0.000305 0.002998 0.003834 -3.961215 15.74133 1.791407 3.906844 3.513836 -5.009263 0.0002 0.0000 0.0770 0.0002 0.0007 0.0000 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat 0.871252 0.863205 0.005998 0.002878 321.0795 0.716334 Mean dependent var S.D. dependent var Akaike info criterion Schwarz criterion F-statistic Prob(F-statistic) 0.010059 0.016218 -7.327429 -7.156196 108.2737 0.000000