I am trying to simulate a simple RBC model in which the standard deviation of the innovations to TFP follows an AR(1) process (i.e. time-varying volatility) but I am running into a problem. I want to perform a third-order approximation so that the policy function estimates take into account time-varying volatility but Dynare does not seem to calculate policy functions. It appears to skip that step and go right to the simulation (in which all series are zero everywhere). However, it does not report any errors. I am using the “use_dll” option in the model specification. When I run a second-order approximation of the model (with or without the “use_dll” option), Dynare appears to do everything correctly. I have tried performing a third-order approximation with the “use_dll” option on a number of example mod files (without time-varying risk) and it always works fine.
I am using Dynare 4.1.2 with 64-bit Matlab 2009b on Ubuntu 10.04LTS (Lucid Lynx). I have attached the relevant mod file and the results Dynare generates for second and third-order approximations. Any help would be appreciated.