> gdp<-read.csv(file.choose(), sep = ",", header = TRUE)
> attach(gdp)
> View(gdp)
> gdp$t<-seq(1:35)
> gdp$lGDP<-log(GDP)
> attach(gdp)
# Scatter Plot
> library(ggplot2)
> ggplot(gdp, aes(YEAR,GDP))+geom_line(col="red")+labs(title="GDP Growth of Ethiopia")
# Whole Period Growth Rate
> model0<-lm(lGDP~t, data = gdp)
> summary(model0)
# Chow Test
>library(strucchange)
>sctest(lGDP~t, type = "Chow", point = 12)
# II. Parallel Regression where only the intercepts in the two regressions are different but the slopes are the same
# Creating time dummy
> gdp$D<-ifelse(YEAR>=1992,1,0) 
> attach(gdp)
> gdp$D<-factor(D, levels = c(0,1), labels = c("pre","post"))
> attach(gdp)
> model1<-lm(lGDP~t+D, data = gdp)
> summary(model1)
> gdp<-cbind(gdp, pred=predict(model1))
> library(ggplot2)
> ggplot(data=gdp, mapping=aes(x=t, y=lGDP, color=D)) + geom_point() + geom_line(mapping=aes(y=pred))

# Concurrent Regression where the intercepts in the two regressions are the same, but the slopes are different
> attach(gdp)
> gdp$D1<-ifelse(YEAR>=1992,1,0)
> attach(gdp)
> gdp$tD<-t*D1
> attach(gdp)
> model2<-lm(lGDP~t+tD,data = gdp)
> summary(model2)
> curve((22.404+0.054*x), xlab = "t", ylab = "", xlim = c(1,35), col="red")
> curve(22.404+0.071*x, xlab = "t", ylab = "", xlim = c(1,35), col="blue", add = TRUE)
> legend( "bottomright", legend = c("post", "pre"), fill = c("red", "blue") )
# Dis-similar Regression (both the intercepts and slopes are different)
> model3<-lm(lGDP~t+D+t*D, data = gdp)
> summary(model3)
> curve(21.9+0.0732*x, xlab = "t", ylab = "", xlim = c(1,35), col="red")
> curve(22.808+0.018*x, xlab = "t", ylab = "", xlim = c(1,35), col="blue", add = TRUE)
> legend( "topleft", legend = c("post", "pre"), fill = c("red","blue") )