Page 1 of 1

system of equations

Posted: Sun Jan 10, 2010 9:24 am
by Dini33
Hi! I have such system of equations

X(t+1)=a1*X(t) / (1+b1*X(t)+c1*Y(t))
Y(t+1)=a2*Y(t) / (1+b2*Y(t)+c2*X(t))

I have estimated parameters a1,b1,c1,a2,b2,c2 using FIML (is it right?)
And now I need to forcast future behavior. I created the model object and added my system. In example from User s Guide (Chapter 36,II,EViews6.0) similar system is solved solving each equation separately. I don't think that this approach is good in my case. But I followed further instructions from Guide. Consequently my dynamic forecast was a smooth curve without any fluctuation. What do I wrong? And what variables are endogenous, but what are exogenous in my case?