How to interpret 2 cointergration in VECM Model

[andrewwang]

Hi, I'm now doing thesis for my degree, My topic is easy which is "The determinance of Government Debt in High debt country - Case of Japan".
I pick out Government debt as dependent variable, with independent variable of inflation, real interest rate, government expense and tax revenue.
Observation is 30 obs from 1980-2009, best with data availability.

I did detrend as suggest by professor as during 2008 and 2009 the debt is change aggressively.
I form y=c+t for all the variables, extract out the residuals and form them 5 in a detrend model.
Then I run unit root test of ADF and PP on all, all is stationary in I(1), where residuals are stationary in I (0).
After that, I go to JJ test and get result of, 2 cointegration at 2 lags.

My problem is that, how do I interpret them in VECM models? As in VECM comes out with 2 equations with 2 coinTeq.
Base on results of VECM with 2 cointergration, I formed 2 equations as follow.
Equation (4): DEBT=7.42E+12INF-9.23E+12RINT-2.5245TAXREV+6.23E+12
Equation (5): GOVEXP=3.72E+12INF-6.53E+13RINT+3.2031TAXREV-6.89E+12

My questions is how to interpret as they are 2 VECM results, and my Government expenditure (GOVEXP) not in my Equation 4 and became another dependent variables of another equation.

And I also did run Granger causality test base on VECM, the result quite consistent with theory and equation 4, but also wonder the ECT should use coinTeq(1) or coinTeq(2)?

Any helps will be appreciate, my thesis deadline is coming to end. Btw, my university didn't offer Time Series Analysis class for me during my thesis is taken, so self study and hopefully some people would give some guilds, Thank you very much, God Bless.