0

How can you run a Breusch-Pagan test in R for SUR models? I am using systemfit to estimate the model but don't know how to test the degree of correlation (homoskedasticity) that exists between the endogenous variables. In Stata, you can do it by adding “corr” to the end of the command line.

hsb2 <- 
read.dta("https://stats.idre.ucla.edu/stat/stata/notes/hsb2.dta")
r1 <- read ~ female + as.numeric(ses) + socst
r2 <- math ~ female + as.numeric(ses) + science
fitsur <- systemfit(list(readreg = r1, mathreg = r2), data = hsb2)
summary(fitsur)

The test's null hypothesis is that there is no homoskedasticity so a significant finding of the test indicates that the amount of correlation present is significantly not zero; thus, we should use a SUR model rather than OLS.

Sally Sharif
  • 21
  • 3
  • Can you make a [reproducible example](https://stackoverflow.com/q/5963269/5325862) of what you're trying to do and with what data? – camille Jul 29 '21 at 19:37
  • Yes, I embedded it in the text. The test shows whether OLS is efficient or if SUR should be used. Thank you! – Sally Sharif Jul 29 '21 at 23:08

0 Answers0