matdehaven
commented
3 hours ago First draft of the slides.
General structure: introduce VFCI, introduce VFCI in the VAR, show that we can match the BCA shock.
Some difficulty getting the figures / story to work when working with the VAR in levels. I found it works much better when estimated in differences (this used to not be the case, but I have corrected the VFCI estimation since I last tried differences).
Presentation Draft in Differences (You have to download the html file, then open it. Won't display within Github.) Presentation Draft in Levels
VAR Charts in Differences Our standard pdf booklet of VAR charts VAR Charts in Levels
In particular I like that when estimating the model in differences we get a convincing figure both in the IRFs and in the historical shocks contribution to unemployment. And, looking at the PDF of VAR Charts, we can explain up to 57.7% of the variation in unemployment (compared to 88.7% from the BCA shock) and we have a correlation up to 0.8 with the BCA shock.
A downside right now is that the IRFs are not as tidy looking when the VAR is estimated in differences (the lines visually jump around more). One solution to this would be to show the IRFs as cumulative responses, at least for specific variables (output, consumption, etc.)
@fernando-duarte, please let me know your thoughts on the slides so far. This is just a first pass at getting...
- The right model specifications
- The right figures to show
Do you have a preference between the differences vs. levels versions?
Notes and updates on preparing slides for Brown Macro Breakfast Presentation on November 20, 2024.