# ================================================================
# ANÁLISIS DE DATOS DE PANEL: EFECTOS FIJOS Y ALEATORIOS
# ================================================================
# 1. Configuración inicial
knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = FALSE)
# Paquetes necesarios
library(readxl) # Lectura de Excel
## Warning: package 'readxl' was built under R version 4.4.3
library(dplyr) # Manipulación de datos
## Warning: package 'dplyr' was built under R version 4.4.3
##
## Adjuntando el paquete: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
library(plm) # Modelos de datos de panel
## Warning: package 'plm' was built under R version 4.4.2
##
## Adjuntando el paquete: 'plm'
## The following objects are masked from 'package:dplyr':
##
## between, lag, lead
library(lmtest) # Pruebas estadísticas
## Warning: package 'lmtest' was built under R version 4.4.2
## Cargando paquete requerido: zoo
## Warning: package 'zoo' was built under R version 4.4.2
##
## Adjuntando el paquete: 'zoo'
## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numeric
library(sandwich) # Errores robustos
## Warning: package 'sandwich' was built under R version 4.4.2
library(car) # Multicolinealidad
## Warning: package 'car' was built under R version 4.4.2
## Cargando paquete requerido: carData
## Warning: package 'carData' was built under R version 4.4.2
##
## Adjuntando el paquete: 'car'
## The following object is masked from 'package:dplyr':
##
## recode
library(tseries) # Normalidad de residuos
## Warning: package 'tseries' was built under R version 4.4.3
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
library(ggplot2) # Visualización
## Warning: package 'ggplot2' was built under R version 4.4.3
# ------------------------------------------------
# 2. Lectura y preparación de datos
# ------------------------------------------------
datos <- read_excel("Panel_datos.xlsx") %>%
mutate(
Año = as.numeric(Año),
Pais = as.factor(Pais),
ln_exports = log(Exportaciones),
ln_tc_real = log(tc_real),
ln_credito = log(credito),
ln_pibpc = log(pibpc)
)
# Declarar como datos de panel
panel_data <- pdata.frame(datos, index = c("Pais", "Año"))
# ------------------------------------------------
# 3. Estimación de modelos
# ------------------------------------------------
modelo_fe <- plm(
ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc,
data = panel_data, model = "within"
)
modelo_re <- plm(
ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc,
data = panel_data, model = "random"
)
cat("\n=== RESULTADOS EFECTOS FIJOS ===\n")
##
## === RESULTADOS EFECTOS FIJOS ===
print(summary(modelo_fe))
## Oneway (individual) effect Within Model
##
## Call:
## plm(formula = ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc,
## data = panel_data, model = "within")
##
## Balanced Panel: n = 6, T = 20, N = 120
##
## Residuals:
## Min. 1st Qu. Median 3rd Qu. Max.
## -0.3513868 -0.1088991 -0.0076645 0.0885470 0.3527776
##
## Coefficients:
## Estimate Std. Error t-value Pr(>|t|)
## ln_tc_real 0.3579758 0.1317055 2.7180 0.007632 **
## ln_credito 0.5229497 0.1128161 4.6354 9.854e-06 ***
## ied 0.0117813 0.0067906 1.7349 0.085553 .
## ln_pibpc 1.9510590 0.1830115 10.6609 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Total Sum of Squares: 13.202
## Residual Sum of Squares: 2.8403
## R-Squared: 0.78486
## Adj. R-Squared: 0.76725
## F-statistic: 100.321 on 4 and 110 DF, p-value: < 2.22e-16
cat("\n=== RESULTADOS EFECTOS ALEATORIOS ===\n")
##
## === RESULTADOS EFECTOS ALEATORIOS ===
print(summary(modelo_re))
## Oneway (individual) effect Random Effect Model
## (Swamy-Arora's transformation)
##
## Call:
## plm(formula = ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc,
## data = panel_data, model = "random")
##
## Balanced Panel: n = 6, T = 20, N = 120
##
## Effects:
## var std.dev share
## idiosyncratic 0.02582 0.16069 0.046
## individual 0.53777 0.73333 0.954
## theta: 0.9511
##
## Residuals:
## Min. 1st Qu. Median 3rd Qu. Max.
## -0.3306956 -0.1338540 -0.0049222 0.1136873 0.4276227
##
## Coefficients:
## Estimate Std. Error z-value Pr(>|z|)
## (Intercept) 2.3334213 1.5968571 1.4613 0.14394
## ln_tc_real 0.3457925 0.1413622 2.4461 0.01444 *
## ln_credito 0.5273136 0.1196138 4.4085 1.041e-05 ***
## ied 0.0105340 0.0073063 1.4418 0.14937
## ln_pibpc 1.9145143 0.1950666 9.8147 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Total Sum of Squares: 13.651
## Residual Sum of Squares: 3.4453
## R-Squared: 0.74762
## Adj. R-Squared: 0.73885
## Chisq: 340.67 on 4 DF, p-value: < 2.22e-16
# ------------------------------------------------
# 4. Pruebas de especificación
# ------------------------------------------------
hausman_test <- phtest(modelo_fe, modelo_re)
bp_test <- bptest(modelo_re) # Heterocedasticidad
bg_test <- pbgtest(modelo_re) # Autocorrelación
vif_test <- vif(lm(ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc, data = datos))
jb_test <- jarque.bera.test(residuals(modelo_re)) # Normalidad de residuos
reset_test <- resettest(lm(ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc, data = datos))
plm_bp_test <- plmtest(modelo_re, type = "bp") # Efectos significativos
cat("\n=== PRUEBA DE HAUSMAN ===\n")
##
## === PRUEBA DE HAUSMAN ===
print(hausman_test)
##
## Hausman Test
##
## data: ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc
## chisq = 0.64189, df = 4, p-value = 0.9583
## alternative hypothesis: one model is inconsistent
cat("\n=== PRUEBAS ADICIONALES ===\n")
##
## === PRUEBAS ADICIONALES ===
list(
BP_Heterocedasticidad = bp_test,
BG_Autocorrelacion = bg_test,
VIF = vif_test,
Jarque_Bera = jb_test,
RESET = reset_test,
Breusch_Pagan_PL = plm_bp_test
)
## $BP_Heterocedasticidad
##
## studentized Breusch-Pagan test
##
## data: modelo_re
## BP = 36.48, df = 4, p-value = 2.305e-07
##
##
## $BG_Autocorrelacion
##
## Breusch-Godfrey/Wooldridge test for serial correlation in panel models
##
## data: ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc
## chisq = 80.45, df = 20, p-value = 3.292e-09
## alternative hypothesis: serial correlation in idiosyncratic errors
##
##
## $VIF
## ln_tc_real ln_credito ied ln_pibpc
## 1.536036 1.306944 1.085246 1.491416
##
## $Jarque_Bera
##
## Jarque Bera Test
##
## data: residuals(modelo_re)
## X-squared = 4.2218, df = 2, p-value = 0.1211
##
##
## $RESET
##
## RESET test
##
## data: lm(ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc, data = datos)
## RESET = 2.5288, df1 = 2, df2 = 113, p-value = 0.08426
##
##
## $Breusch_Pagan_PL
##
## Lagrange Multiplier Test - (Breusch-Pagan)
##
## data: ln_exports ~ ln_tc_real + ln_credito + ied + ln_pibpc
## chisq = 611.59, df = 1, p-value < 2.2e-16
## alternative hypothesis: significant effects
# ------------------------------------------------
# 5. Errores robustos (Arellano, agrupados por país)
# ------------------------------------------------
robust_vcov <- vcovHC(modelo_re, method = "arellano", type = "HC1", cluster = "group")
robust_results <- coeftest(modelo_re, vcov = robust_vcov)
cat("\n=== COEFICIENTES CON ERRORES ROBUSTOS ===\n")
##
## === COEFICIENTES CON ERRORES ROBUSTOS ===
print(robust_results)
##
## t test of coefficients:
##
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.3334213 2.7958623 0.8346 0.4056744
## ln_tc_real 0.3457925 0.3175605 1.0889 0.2784734
## ln_credito 0.5273136 0.2585487 2.0395 0.0436906 *
## ied 0.0105340 0.0030586 3.4440 0.0008004 ***
## ln_pibpc 1.9145143 0.3327688 5.7533 7.364e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# ------------------------------------------------
# 7. Visualizaciones
# ------------------------------------------------
# Evolución temporal de exportaciones
ggplot(datos, aes(x = Año, y = ln_exports, color = Pais)) +
geom_line(linewidth = 1) +
geom_point() +
labs(title = "Evolución logarítmica de Exportaciones por país",
x = "Año", y = "ln(Exportaciones)") +
theme_minimal()
# Distribución del crédito
ggplot(datos, aes(x = Pais, y = ln_credito, fill = Pais)) +
geom_boxplot(alpha = 0.7) +
labs(title = "Distribución del Crédito Doméstico (% PIB) por país",
y = "ln(Crédito)", x = "País") +
theme_minimal()
# Relación Crédito vs Exportaciones
ggplot(datos, aes(x = ln_credito, y = ln_exports, color = Pais)) +
geom_point() +
geom_smooth(method = "lm", se = FALSE) +
labs(title = "Relación entre Crédito Doméstico y Exportaciones",
x = "ln(Crédito)", y = "ln(Exportaciones)") +
theme_minimal()
# Distribución de PIB per cápita
ggplot(datos, aes(x = ln_pibpc, fill = Pais)) +
geom_density(alpha = 0.5) +
labs(title = "Distribución de PIB per cápita (log) por país") +
theme_minimal()
