Shocks generated do not match underlying formula

Stochastic simulations
1

I am trying to generate a process where oil price (F) reverts to a time variant mean (eta) where eta itself is a moving average process. I am using the following lines of commands. Surprisingly, the series that Dynare generates do not match the underlying formula for F.

var F eta;
varexo sh;
parameters ka Fbar;
ka=0.1;
Fbar=0.841261;

model;
log(F)=ka*eta(-1)+(1-ka)*log(F(-1))+sh;
eta=eta(-1)+1.5*(log(F(-1))-eta(-1));
end;

initval;
F=Fbar;
eta=log(Fbar);
sh=0;
end;

shocks;
var sh=0.021;
end;

stoch_simul(periods=50000, nograph);
2

What do you mean with

?

3

Dynare generated series for F do not match the formula : log(F)=ka*eta(-1)+(1-ka)*log(F(-1))+sh.

When I plug in the obtained numbers in this formula, the equation does not hold.

4

The log you are using is a nonlinear function. Dynare will have to approximate it.

5

I see. Thank you for the clarification.