Can Dynare Handle a Model with Sub-Periods?


I have been working on a model that has two sub-periods within one period. My question is whether or not Dynare (and DSGE models in general) can theoretically handle this sort of set up if I correctly write up the timing conventions? My initial thoughts are that I do not see why they couldn’t, but I would like some confirmation/guidance.



Dear Ben,

I am not sure to understand your problem… Are you referring to logical or historical time? In the first case there is obviously no problem. A lot of things can happen in a period, but only the periods matter for the dynamic (a lot of models have this kind of arguments, an extreme example is the Walras auctioneer fiction in a general equilibrium model). If you are referring to historical time (i.e. the sub periods matters in the the definition of the dynamic) then you just have to write the model in the highest frequency and add equations to define the variables at lower frequencies.


To be clear, I mean that there are two markets that operate sequentially within a period. When one market is operational, it takes the current state of the other market as exogenous. For concreteness, say there is a fruit market and a clothes market, but the fruit market is only open in the morning and closes by noon while the clothes market does not open until the evening . In one day (a period):

  1. The fruit market opens, agents take the current state of the clothes market as given (so they are using the aggregates from the day before) and any expectations that involve clothes market variables are realized later that same day when the clothes market opens.

  2. The fruit market closes.

  3. The clothes market opens. Agents takes the current state of the fruit market aggregates as given, and all expectations about fruit market variables are realized the following day (the following period)

  4. The clothes market closes. The day ends (the period ends).

I think this is what you mean by historical time and this is what I am interested in. From your answer, it seems like this is possible so long as I write the variables with the correct timing conventions?


I am not sure that I would call this historical time, i.e. I don’t think you really need to write explicitly the sub periods in the mod file. It’s more a matter of how you write the programs of the agents taking decisions in these markets.


Okay I think I understand, thanks!