1
$\begingroup$

I am trying to run a Johansen-Procedure in a set of macroeconomic variables (GDP, credit outstanding and industrial production). I am working with them in level.

How should I interpret the following result? Which "r" should I use next to run the model?

###################### 
# Johansen-Procedure # 
###################### 

Test type: maximal eigenvalue statistic (lambda max) , without linear trend and constant in cointegration 

Eigenvalues (lambda):
[1]  6.999273e-01  3.188663e-01  1.303763e-01  5.099467e-02 -6.157574e-17

Values of teststatistic and critical values of test:

          test 10pct  5pct  1pct
r <= 3 |  3.04  7.52  9.24 12.97
r <= 2 |  8.10 13.75 15.67 20.20
r <= 1 | 22.27 19.77 22.00 26.81
r = 0  | 69.82 25.56 28.14 33.24

[...]

My goal is to estimate the best VEC model so I can forecast those variables. I choose the number of lags from the VARselect function (saved as def), AIC(n) criteria, and then run a vec model.

jo.eigen <- ca.jo(training, type='eigen', K=def$selection[1], 
                ecdet='const',
                spec='transitory',
                season=4)

vec <- cajorls(jo.eigen, r= ???)
vec.level <- vec2var(jo.eigen, r= ???)

Thank you very much!

$\endgroup$

1 Answer 1

2
$\begingroup$

The maximum eigenvalue test is used to determine the rank of the coefficient matrix associated with the cointegration term. It is a sequential test starting from the hypothesis $$H_0: r=0 \\ H_1: r=1.$$ The test statistic at r=0 is 69.82 and the 1% critical value is 33.24, meaning we reject $H_0$. Now we test $$H_0: r=1 \\ H_1: r=2.$$ Now the test statistic is rejected at the 5% level, but not at the 1% level. If you find this threshold 'acceptable' you move on to $$H_0: r=2 \\ H_1: r=3.$$ Here the test breaks down, we cannot reject $H_0$ at any of the typically used significant levels (1%, 5%, or 10%), and accept r=2.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.