[R-SIG-Finance] Granger's causality test

[Matthieu Stigler]

Rafael

I don't know about the function found on the test, but you can rely on 
the function grangertest!

Note also you could use causality() from vars pkg, which gives exactly 
the same output:
library(vars)
v<-VAR(cbind(tacc, value))

causality(v, cause="value")$Granger
grangertest(tacc ~ value, order = 1)

Note that, although the F stat is the same, the df and p-values are not 
exactly the same: grangertest() is considered as a univariate 
regression, while causality() considers you have been estimating a full 
VAR. See Lutkephol (2006) p. 103 for a discussion on the choice of df. 
The advantage of the causality(9 function is that it allows to use HC 
covariance estimators for the test.

Hope this helps

Matthieu


Le 17/11/2011 02:57, rafleon at highstat.cl a écrit :
> Hi all,
> I've got several time series as these below
>
> tacc=c(106, 117, 125, 140, 150, 150, 135, 135, 102, 102, 112, 124, 128,
> 128)
> value=c(1273, 1464, 1799, 1440,  963,  817,  794,  791,  638,  954, 1301,
> 2007, 1998, 2000)
>
> #######
> Additional information
> tacc= Total Allowable Commercial Catch, In this case rock lobster in New
> South Wales, Australia.
> Value= The tacc is split into Individual Transferable Quotas (ITQs), which
> can be leased or sold. In this case Value is the sale price]
> The tacc is set every year and every stakeholder (commercial, recreational
> fisher association, conservationists, aborigines, fisheries manager) can
> suggest a figure for the tacc and the final decision is made by the
> Minister.
> I'm expecting that lower tacc cause higher value. Why? because lower tacc
> produce stock rebuilding, increasing catch rates, which in turn reduces
> fishing costs and improve future catches expectation and finally this
> increase the quota prices.
> ######
>
> I'm traying to find 'causality' between these variables. Well, I used the
> function grangertest
>
> library(lmtest)
> grangertest(tacc ~ value, order = 1)
> grangertest(value ~ tacc , order = 1)
>
> Additionally, I found the below fucntion on internet that return different
> results.
> Please, could someone of you tell me which function is more reliable? And,
> I don't understand the k parameter.
> I would appreciate deeply any help.
> Cheers,
>
> Rafael
>
>
>
> ts=cbind(value,tacc)
> #Note that the results are different if ts=cbind(tacc,value)
>
> "granger"<-function(d, L, k=1 )
> {
> #d is a bivariate time-series:  regress d[,k] on L lags of d[,1] and
> d[,2].
> #This is a modified version for R, in which the command ts.matrix was
> substituted by ts.intersect.
> names.d<- dimnames(d)[[2]]
> D<- d
> for(i in 1:L)
> {
> D<-ts.intersect(D, lag(d,  - i))
> #don't you find -1 bizarre here and it annoying to need the loop
> }
> dimnames(D)[[2]]<- paste(rep(names.d, L + 1), "_", rep(0:L, times =
> rep(2, L + 1)), sep = "")
> y<- D[, k]
> n<- length(y)
> x1<- D[,  - (1:2)]
> x0<- x1[, ((1:L) * 2) - (k %% 2)]
> z1<- lm(y ~ x1)
> z0<- lm(y ~ x0)
> S1<- sum(z1$resid^2)
> S0<- sum(z0$resid^2)
> ftest<- ((S0 - S1)/L)/(S1/(n - 2 * L - 1))
> list(ftest = ftest, p.val = 1 - pf(ftest, L, n - 2 * L - 1), R2 =
> summary(z1)$r.squared)
> }
>
> granger(ts,L=1, k=2)
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.