Very cool, it seems the researchers took into account all sorts of things like state sales tax, state income tax, corporate tax, welfare / handout type programs, and spending + savings habits.
From what I can tell the key inflection points in the graph come from four issues.
1.) How much can one contribute to retirement accounts to avoid paying income tax that year?
2.) Where are the key discontinuous cutoff points in the tax system? (social security maxes out, certain payments to the needy disappear, etc)
3.) What is the typical marginal propensity to consume?
4.) How much wealth has one accumulated by a certain age in life, which can be invested, which is only taxed at lower capital gains / dividend rates?
My thoughts on the first are mixed. Yes, we don't pay taxes this year on money we stash into a non-roth 401k, but I hope they aren't doing the bad accounting of treating that as 0% tax... we will eventually pay SOME tax when we withdraw on it, just not this year.
The second point is very valid... there is a magic range around $110-$160K income right now where your marginal tax rate from paying social security is almost 12% lower than the guys making $80K-$110K. In the US, you and your employer each pay 6.2% of your earned income to social security, but it caps at $110K (the rationale being you won't get credit for more than that at retirement, so they won't take more).
The final points are a bit bogus in terms of providing us guidance, but they are more of the author's observations from the data they studied. Poor people consume more as a percentage of their income and hence pay more sales tax each year. Wealthier people buy more, but less as a percentage of their income, so they pay less sales tax, and save more over time which can be turned into investment income.
http://people.bu.edu/kotlikoff/Does%20It%20Pay%20to%20Work%2...