Here is a comparison you have seen: proposed taxes on the wealthy set against American charitable giving, $617 billion in 2025, a record, the first year over $600 billion. Against that denominator, $60–150 billion in new tax revenue looks modest. A fraction of American generosity.
Why disrupt a system that gives so much?
The comparison is wrong, and it is wrong in a specific, checkable way: the $617 billion belongs to millions of households, foundations, corporations, and bequests. It is not the billionaires’ number.
The billionaires’ number is $20–40 billion a year. The tax counterfactual is $60–150 billion. Compared like-for-like, the comparison flips… the taxes raise two to five times what billionaires give.
The billionaire figure is smaller than most people assume because most people are anchored on the total. Megagifts, individual gifts of $600 million or more, came to $19.2 billion in 2025, and a single donor, MacKenzie Scott, accounted for roughly a third of that1. The annual total swings wildly because a handful of very large gifts dominate it.
Philanthropy at this scale is not a system. It is weather.
Against it, two tax changes with mainstream revenue estimates: restoring a 1974-style 70% top marginal rate ($50–100B/year, depending on design and behavioral response) and a 25% tax on equity-backed borrowing, the buy-borrow-die channel, at roughly $10–50B/year. Every input is a contested range, which is why I built the comparison as a model instead of a table.
The page shows each estimate as an uncertainty band, not a point. Three sliders let you set your own assumptions, pick the most conservative tax estimates and the most generous philanthropy estimate, and see whether the ratio still holds. (It does. That is the point of publishing the sliders.)
The deeper argument on the page is about substitution, because that is where the defense of philanthropy actually lives.
The defense has two parts:
- private allocation is more efficient than public
- voluntary giving substitutes for what the forgone revenue would have funded.
The second claim is load-bearing; without it, “let them give voluntarily” just means “let these things go unfunded.”
The substitution claim is currently being tested in public. As funding for education, infrastructure, and health retreats, philanthropy is not filling the gap and structurally cannot.
The scale is wrong: $20–40 billion against drawdowns running to hundreds of billions. The direction is wrong: donor money flows to named buildings, endowed institutions, and legible prestige projects, not water systems, Title I schools, or reimbursement rates. Giving follows donor preference; the gaps follow political neglect. The two maps do not overlap.
Note what this argument does not need. It does not need to show that government allocates better than donors. A defender can retreat to the efficiency claim, but efficiency without substitution is not a defense of philanthropy as a system; it is a general complaint about government that would hold whether billionaires gave or not. What remains is revenue against revenue.
And the ratio answers that.
Ultimately, if we accept that donor preference is not a substitute for democratic allocation, we are left with a simple question of revenue.
The Generosity Ratio provides the answer.
Sources: Giving USA 2026: The Annual Report on Philanthropy for the Year 2025 — $617.2B total giving, $19.2B in megagifts. Tax revenue ranges are estimates under mainstream assumptions; they are inputs to be contested, which is why the model exposes them as sliders rather than burying them in a methodology note.