Skip to content

Frage uber package strucchange

2 messages · Matthieu Stigler, Achim Zeileis

Original
Matthieu Stigler
# 
Hello

I'm using the nice package strucchange to see whether there is a 
structural change in a regression, and especially, if the change is in 
the slope or intercept coefficient.

The "Vignette" with the package (see fig 4 page 8) gives an example to 
obtain a plot of the empirical process for each coefficient. However, I 
don't obtain it, even with the code given in example...

Do you have also only one plot? Do you see what I do wrong, and how I 
could obtain the same plot as in example?

Thanks  a lot!!

The code in example is:

library("strucchange")
data("USIncExp")
USIncExp2 <- window(USIncExp, start = c(1985, 12))

coint.res <- residuals(lm(expenditure ~ income, data = USIncExp2))
coint.res <- lag(ts(coint.res, start = c(1985, 12), freq = 12),
    k = -1)
USIncExp2 <- cbind(USIncExp2, diff(USIncExp2), coint.res)
USIncExp2 <- window(USIncExp2, start = c(1986, 1), end = c(2001,
    2))
colnames(USIncExp2) <- c("income", "expenditure", "diff.income",
    "diff.expenditure", "coint.res")
ecm.model <- diff.expenditure ~ coint.res + diff.income


me <- efp(ecm.model, type = "ME", data = USIncExp2, h = 0.2)

plot(me, functional = NULL)
1 day later
Achim Zeileis
#
On Sun, 2 Mar 2008, Matthieu Stigler wrote:

            

      
      
      
    

        
Hmmm, I don't know what could go wrong for you. For me, the code you
attached below produces a plot with three panels (one for each parameter).
Unsurprisingly, this looks like the vignette because it was generated with
the same code.

I've got no idea what might go wrong on your machine...

One additional remark: The function gefp() (instead of efp) has some more
flexible plotting functionality. You can look at

  Achim Zeileis (2006). Implementing a Class of Structural Change
  Tests: An Econometric Computing Approach. Computational Statistics
  and Data Analysis, 50, 2987-3008.

for details. A preprint version of that paper is available from my Web
page.

Best,
Z