I don't think you read the summary. The mechanism is pretty clear:
1. Compared with a flat tax, a progressive tax regime expands the economy because poor people spend more of their money than rich people.
2. This expansionary effect gets multiplied because this spending recirculates around and around (the "standard fiscal multiplier" in the summary).
3. Rich people accrue most of the benefits of the expansionary effect.
4. These benefits are actually larger than the initial increase in tax burden in step 1), so while everyone ends up better off, rich people end up _more_ better off than poor people. Therefore the progressive tax scheme has the net effect of boosting income inequality.
It's not a perpetual motion machine, the gain starts because rich and poor people having different consumption rates. You could only do this until you run out of tax dollars being paid by poor people.
Really, though, what the paper is proposing is saying is that rich people benefit quite a lot from expansionary policies, so much so that they can even afford to pay to get them started. I don't think this is very controversial or even counterintuitive. The thing to realize is that the paper starts with the claim that _progressive taxation makes the economy bigger_. If you accept that, it makes sense that the rich would benefit the most from it.
I've read the summary and it does not change anything.
> These benefits are actually larger than the initial increase in tax burden in step 1
This isn't clear at all. I suspect this is only true in a specific set of circumstances, not as a universal rule.
> You could only do this until you run out of tax dollars being paid by poor people
The article is talking about the 50th and 99th percentiles. If what you're saying is true, the government should reduce taxes on the 50th percentile household all the way to 0%, and fund this tax cut entirely by increasing taxes on those who are more wealthy. And if what you're saying is true, such a change in tax policy would increase disposable income for the 99th percentile.
There isn't a single serious economist who believes the above. Even people like Sanders who would advocate for such a tax policy, are arguing for it from a fairness perspective. Not a "this will make the rich even richer" argument, which is what you're implying.
I think the trick of this trickle up is that giving money from the 90th percentile (read: affords saving for retirement) to the <50th percentile (read: doesn't afford saving for retirement) effectively gives money to the >99th percentile (read: affords saving into the blackhole of offshore accounts)
1. Compared with a flat tax, a progressive tax regime expands the economy because poor people spend more of their money than rich people.
2. This expansionary effect gets multiplied because this spending recirculates around and around (the "standard fiscal multiplier" in the summary).
3. Rich people accrue most of the benefits of the expansionary effect.
4. These benefits are actually larger than the initial increase in tax burden in step 1), so while everyone ends up better off, rich people end up _more_ better off than poor people. Therefore the progressive tax scheme has the net effect of boosting income inequality.
It's not a perpetual motion machine, the gain starts because rich and poor people having different consumption rates. You could only do this until you run out of tax dollars being paid by poor people.
Really, though, what the paper is proposing is saying is that rich people benefit quite a lot from expansionary policies, so much so that they can even afford to pay to get them started. I don't think this is very controversial or even counterintuitive. The thing to realize is that the paper starts with the claim that _progressive taxation makes the economy bigger_. If you accept that, it makes sense that the rich would benefit the most from it.