Thread by @wwwojtekk

I'll post a longer thread with substantial comments later, but for now I'll just describe where the bodies are buried in the comparison of the refereed QJE DINA paper and the book/NYT animation

https://twitter.com/wwwojtekk/status/1182769865683279872

Tl;dr - here is my animation
/thread

I'll show everything for top 0.01%. Because the QJE stops in 2014, I'll stop there too and I'll show it starting in 1962 when micro files start and QJE has details.

Note the milder pattern in the blue line and a strong trend in the red one - that's where we are heading.

There are two important things that changed. One is assumptions about incidence. Most importantly, corporate incidence is assumed on shareholders only, rather than all non-housing capital (nothing is on labor in either case)

This mechanically concentrates the burden at the top and makes it sensitive to changes in corp revenue. Here is the impact of removing the changes in corporate incidence

There are other incidence assumptions changes that have changed, the two biggest ones are for sales tax and income but their impact is about constant. If we just remove mean changes in incidence of sales and income tax over the whole period, we are almost there

But there is still a gap in some years. What to do about that? The gap is there in years with high capital gains realizations. Why don't we then adjust in an ad hoc way when capital gains are large?

Here is where it hits ,starting from the original graph:

And let's show it altogether again. The really big one is the corporate tax incidence assumption. The rest is mostly a level shift and some volatility adjustment (capital gains).

I'll comment later tonight with thoughts about what I think about all of it /end

And here is my take on how to think about all of that

https://twitter.com/wwwojtekk/status/1183953723825020928?s=19