The results from the R strucchange package do not match the SAS proc autoreg results when I evaluate structural breaks using BIC and Perron (1998, 2003). The number of gaps found is not at all the same. My data has about 20,000 observations, so it's hard for me to recreate the problem here. But can anyone explain why there is a difference in the number of breaks found? The following is an example of R and SAS code that I use to evaluate gaps.
R code:
library(lmtest)
library(sandwich)
library(strucchange)
data.bp <- breakpoints(formula, data)
coeftest(data.bp)
coeftest(data.bp, vcov = NeweyWest)
SAS Code:
proc autoreg data=data;
model formula / BP=(EPS=0.15, PRINTEST=BIC);
source
share