ALEXANDER DANIEL ALVAREZ BERARDI
14/6/2019
library(readr)
library(dplyr)
library(stargazer)
library(lmtest)
library(car)
library(tidyr)
library(kableExtra)
ejemplo_regresion<-read.csv("C:\\Users\\AD_be\\Desktop\\Econometria\\ejemplo_regresion.csv")
modelo_lineal<-lm(Y~ï..X1+X2,data = ejemplo_regresion)
stargazer(modelo_lineal,title="modelo estimado",type="text")##
## modelo estimado
## ===============================================
## Dependent variable:
## ---------------------------
## Y
## -----------------------------------------------
## ï..X1 0.237***
## (0.056)
##
## X2 -0.0002***
## (0.00003)
##
## Constant 1.564***
## (0.079)
##
## -----------------------------------------------
## Observations 25
## R2 0.865
## Adjusted R2 0.853
## Residual Std. Error 0.053 (df = 22)
## F Statistic 70.661*** (df = 2; 22)
## ===============================================
## Note: *p<0.1; **p<0.05; ***p<0.01
u_i<-modelo_lineal$residuals
data_prueba_white<-as.data.frame(cbind(u_i,ejemplo_regresion))
regresion_aux<-lm(I(u_i^2)~ï..X1+X2+I(ï..X1^2)+I(X2^2)+ï..X1*X2,data=data_prueba_white)
sumario<-summary(regresion_aux)
n<-nrow(data_prueba_white)
R_2<-sumario$r.squared
LM_W<-n*R_2
gl=2+2+1
p_value<-1-pchisq(q=LM_W,df=gl)
VC<-qchisq(p=0.95, df=gl)
salida_white<-c(LM_W,VC,p_value)
names(salida_white)<-c("LMw","valor critico","p value")
stargazer(salida_white,title = "Resultados de la prueba de White",type = "text",digits = 6)##
## Resultados de la prueba de White
## ===============================
## LMw valor critico p value
## -------------------------------
## 3.690182 11.070500 0.594826
## -------------------------------
prueba_white<-bptest(modelo_lineal,~I(ï..X1^2)+I(X2^2)+ï..X1*X2,data=ejemplo_regresion)
print(prueba_white)##
## studentized Breusch-Pagan test
##
## data: modelo_lineal
## BP = 3.6902, df = 5, p-value = 0.5948
dwtest(modelo_lineal,alternative = "two.sided",iterations = 1000)##
## Durbin-Watson test
##
## data: modelo_lineal
## DW = 1.9483, p-value = 0.5649
## alternative hypothesis: true autocorrelation is not 0
durbinWatsonTest(modelo_lineal,simulate=TRUE,reps=1000)## lag Autocorrelation D-W Statistic p-value
## 1 -0.04366918 1.948305 0.524
## Alternative hypothesis: rho != 0
data_prueba_BG<-cbind(u_i,ejemplo_regresion)%>%as.data.frame()%>%mutate(Lag_1=dplyr::lag(u_i,1),lag_2=dplyr::lag(u_i,2))%>% replace_na(list(Lag_1=0,Lag_2=0))
kable(head(data_prueba_BG,6)) | u_i | ï..X1 | X2 | Y | Lag_1 | lag_2 |
|---|---|---|---|---|---|
| 0.0734697 | 3.92 | 7298 | 0.75 | 0.0000000 | NA |
| -0.0033412 | 3.61 | 6855 | 0.71 | 0.0734697 | NA |
| -0.0391023 | 3.32 | 6636 | 0.66 | -0.0033412 | 0.0734697 |
| -0.0621832 | 3.07 | 6506 | 0.61 | -0.0391023 | -0.0033412 |
| 0.0162403 | 3.06 | 6450 | 0.70 | -0.0621832 | -0.0391023 |
| 0.0124247 | 3.11 | 6402 | 0.72 | 0.0162403 | -0.0621832 |
Calculando la regresión auxiliar y el estadístico LMBP
regresion_aux_BG<-lm(u_i~ï..X1+X2+Lag_1+lag_2,data=data_prueba_BG)
sumario_BG<-summary(regresion_aux_BG)
R_2_BG<-sumario_BG$r.squared
n1<-nrow(data_prueba_BG)
LM_BG<-n1*R_2_BG
gl1=2
p_value1<-1-pchisq(q = LM_BG,df = gl1)
VC1<-qchisq(p = 0.95,df = gl1)
salida_bg<-c(LM_BG,VC1,p_value1)
names(salida_bg)<-c("LMbg","Valor Crítico","p value")
stargazer(salida_bg,title = "Resultados de la prueba de Breusch Godfrey",type = "text",digits = 6)##
## Resultados de la prueba de Breusch Godfrey
## ===============================
## LMbg Valor Crítico p value
## -------------------------------
## 3.667591 5.991465 0.159806
## -------------------------------
bgtest(modelo_lineal,order = 2)##
## Breusch-Godfrey test for serial correlation of order up to 2
##
## data: modelo_lineal
## LM test = 3.3052, df = 2, p-value = 0.1916
El test BG puede usarse también para verificar la autocorrelación de 1° orden
library(lmtest)
bgtest(modelo_lineal,order = 1)##
## Breusch-Godfrey test for serial correlation of order up to 1
##
## data: modelo_lineal
## LM test = 0.051063, df = 1, p-value = 0.8212