I’m getting explosive IRFs for a second-order approximation to a model in dynare++, which is chracteristic of 3rd and 4th order terms getting introduced from multiplication of 1st and 2nd order ones. Of course, the problem could be with my model somewhere, but it’d be nice to rule out the possibility of higher order terms creeping in. So does anyone know if dynare++ already has a pruning algorithm in it, or if there’s any other diagnostic I could use to see if this is the source of the problem?
And in advance of the inevitable spec-related questions, I’m using version 1.3.5 with the windows binaries available on the website.
Thanks,
Philip