Michel,
trend coefficients are not working for me. There are two issues:
1. The auxiliary parameters defined in pound (#) lines are not properly parsed and are omitted from expressions in trend_coeffs, I mean omitted in expressions when options_.trend_coeff{tmp1} is set in model_file.m.
2. Some observable variables have no trends (trend equal to zero). I am used to omit them. However, the code at dynare_estimation.m:121 pads with zeros only trailing part of the trend_coeffs cell array. This means that trend_coeffs for observable variables without (zero) trend which are before some variable with a trend (non-zero) get uninitialized. This causes a matlab error.
There is an obvious workaround, so I can live with it. However, it would be very nice if you can fix these problems.
This report corresponds to revision 876.
many thanks,
Ondra K.