Indeterminacy in a Model of Deep Habits


I have put together a model of Deep Habits with capital which largely follows Zubairy (2010). However, it also incorporates a durable goods sector. See the attached code.

The error message I get is ‘Blanchard Hahn Conditions are not satisfied: Indeterminacy due to Rank Failure’. The command check tells me I have ‘9 eigenvalues larger than 1 in modulus for 9 forward looking variables. The rank condition isn’t verified’.

When I run model_diagnostics, I get no output (I am running Dynare 4.4.2).

I gather the problem could be timing, parameterisation or collinearity (although I don’t get a warning). I have investigated each of these.

The pre-determined variables (durable sector capital, non-durable sector capita, the stock of durables, the shadow value of durables and the stock of habits) all have, to my eyes, the correct Dynare timing. I have played with the parameter which traditionally results in indeterminacy which is the parameter governing the accumulation of habits (bc and bd), and they don’t seem to affect the results. Plus I cannot identify a co-linear relationship between equations (also model_diagnostics doesn’t give me a warning).

Have I made a mistake or is there another source of the error message I am unaware of.

I recognise that this is a common problem, and I have searched the form for answers but I am just trying to get a little clarity.

Any help would be greatly appreciated.

Deep_Habits_Capital.mod (2.89 KB)

Try to simplify your model and work from there. In particular, does your model work with only one sector? Potential sources of the problem could be the use of the strange auxiliary variable

and the potentially non-unique split between capital in the two sectors. A more common way would be to have one unique predetermined capital stock that is split up frictionlessly in every period rather than having two independent predetermined capital stocks whose sum enters the production function.