Expectations operator problem - Lucas Fuerst model

[timuy]

Hi,

The mod file attached is a very simple implementation of a Lucas-Fuerst CIA model. The problem basically lies in the statement

bet*expectation(-1)(i/exp(mu(+1))*p(+1)*c(+1)) = (1/expectation(-1)(p*c));

which is a first order condition that states that the portfolio variable n must be chosen before current period uncertainty is resolved, i.e. given time t-1 info set.

If I change it to

bet*i/expectation(-1)(exp(mu(+1))*p(+1)*c(+1)) = (1/expectation(-1)(p*c));

the code runs fine or when I remove the expectations operator. Of course the results change dramatically and it really is the original statement above with everything inside the expectations operator that is correct. I am using Dynare 4.2.0.

Thanks in advance and look forward to your prompt response.