| bh6lrtest {urca} | R Documentation |
This function estimates a restricted VAR, where some restrictions are placed on r1 cointegrating relations which are chosen in the space of the matrix H. The test statistic is distributed as χ^2 with (p-s-r2)r1 degrees of freedom, with s equal to the number of columns of \bold{H}, r1 the number of cointegrating relations in the first partition and r2 the number of cointegrating relations in the second partition which will be estimated without any restrictions.
bh6lrtest(z, H, r, r1, conv.val = 0.0001, max.iter = 50)
z |
An object of class |
H |
The (p \times s) matrix containing the known cointegration relations. |
r |
The count of cointegrating relationships; |
r1 |
The count of cointegrating relationships in the first
partition of the cointegration space; |
conv.val |
The convergence value of the algorithm. (see details); |
max.iter |
The maximal number of iterations. |
Please note, that the following ordering of the dimensions should be
obeyed: r1 ≤q s ≤q p - r2. A two-step algorithm is used to
determine the eigen values of the restricted model. Convergence is
achieved if the quadratic norm of the eigen values is smaller than
conv.val.
An object of class cajo.test.
Bernhard Pfaff
Johansen, S. (1995), Likelihood-Based Inference in Cointegrated Vector Autoregressive Models, Oxford University Press, Oxford.
Johansen, S. and Juselius, K. (1992), Testing structural hypotheses in a multivariate cointegration analysis of the PPP and the UIP for UK, Journal of Econometrics, 53, 211–244.
ca.jo, alrtest, ablrtest,
blrtest, bh5lrtest, cajo.test-class,
ca.jo-class and urca-class.
data(UKpppuip) attach(UKpppuip) dat1 <- cbind(p1, p2, e12, i1, i2) dat2 <- cbind(doilp0, doilp1) H1 <- ca.jo(dat1, type='trace', K=2, season=4, dumvar=dat2) H6 <- matrix(c(1,0,0,0,0, 0,1,0,0,0, 0,0,1,0,0), c(5,3)) bh6lrtest(z=H1, H=H6, r=2, r1=1, conv.val=0.0001, max.iter=50)