Johansen cointegration and granger causality test

[Chuvaekek]

Hi!
Good day.

I'm from the Philippines. Im currently doing my thesis. my study is all about the impact of macroeconomic factors to stock market. here are my variables:

Independent:

interest rate, inflation rate, forex and m2

dependent:

financials index, industrial index and property index.

this is what i did:

first: i check for unit roots and it turned out that may variable are nonstationary at level and stationary in first difference. then I use the stationary data to estimate VAR for lag then i check for autocorrelation.then i run cointegration test then granger.

my questions is:

1. did i do the right step in testing johansen and granger tests?

2. i run the data in group (e.g 1 dependent and 3 independent). but when i run it into pairs (e.g interest rate and financials index) they dont have cointegrating relation but granger test results says that interest rate grander cause financials index though they dont have cointegration. what is the right thing to do?.



Thank you!
God bless!

[Chuvaekek]

pls someone reply