Dear Dynare forum users,
I’m a phD student and (at the moment) a rookie using Dynare.
Currently I’m doing a research that seeks to assess the economic impact of a shock over the GDP.
I already achieved to build an IRF with a shock taking place at time t=1. What we want now is, instead of evaluate a single shock taking place at time t, see impact of an event that increases over time.
For example, climate change. The effect of climate change is not ocasional. It is permanent and cumulative. How can I try to simulate the effect of this kind of events?
Sorry because I know that the post is not as descriptive as the forum seeks but if someone could give me some guidelines I would appreciate it.
Thank you very much in advance.
Do you mean something like a conditional forecast? Keeping the shock at a particular level through a period of time, and seeing what happens to the rest of the economy?
Reuben
Hi @punnoosejacob thanks for your reply!
I’m not sure if is exactly a conditional forecasting what we are looking for… If I understand properly, conditional forecasting is useful to assess the implementation of scenarios.
What we have is the next. It’s demonstrated that an increase in climate change will increase the frequency and intensity of (for example) fires. We already achieved to see (at t=1) the implications of climate change, and, subsequently, the increase of fire events over forest sector GDP (figure below).
We are seeking to accumulate effects. Fires are not only at time t=1. First simulated fire and its consequences over GDP will be key during a second fire event happening at t=2 and so on.
I hope that is clear now, or at least, better than I did before…
Thanks!
The big questions are whether the later shocks are a surprise or anticipated (i.e. stochastic vs. perfect foresight model) and whether the model is supposed to be nonlinear.
For stochastic models, the simult_-function allows you to feed in any sequence of shocks. See https://github.com/JohannesPfeifer/DSGE_mod/blob/master/RBC_news_shock_model/RBC_news_shock_model.mod