I want to carry out a structural change test on exchange rate data. I have a zoo series named m with several exchange rates. I first create a window m1.
m1 <- window(m, start = as.Date("1998-12-31"), end = as.Date("2010-12-31"))
The first column of m1 looks like the following.
head(m1$mbel)
1998-12-31 1999-01-31 1999-02-28 1999-03-31 1999-04-30 1999-05-31
1.2346 1.2278 1.2269 1.2259 1.2328 1.2357
mbel is the variable of interest currently. For mbel, I want to test parameter stability for a simple linear model, mbel ~ mbel(lag1).
I first combine the level and first lags of logs of mbel data.
cb <- cbind(log(m1$mbel),lag(log(m1$mbel),k = -1))
colnames(cb) <- c("rate","ratelag1")
cb <- window(cb, start = as.Date("1999-01-31"), end = as.Date("2010-12-31"))
head(cb)
rate ratelag1
1999-01-31 0.2052240 0.2107470
1999-02-28 0.2044907 0.2052240
1999-03-31 0.2036753 0.2044907
1999-04-30 0.2092880 0.2036753
1999-05-31 0.2116376 0.2092880
1999-06-30 0.2190552 0.2116376
Now with library strucchange, I use the following tests.
r <- Fstats(cb$rate~cb$ratelag1, data = cb )
re <- efp(cb$rate~cb$ratelag1,data = cb,type="RE")
plot(re)
sctest(re)
The following error is showing up:
Error in eval(attr(terms(formula), "variables")[[2]], data, env) :
numeric 'envir' arg not of length one
What am I missing here? Please help.
