Home Writing Projects Setup Protocol CV

MacKenzie Scott's Yield Giving network has made over $26 billion in 2,700+ gifts since 2019. I model that as $26.3 billion and wanted one number for the health impact: quality-adjusted life-years, the unit health economists use to compare a death averted against years lived in better health.

The dollar total sets the scale linearly. The hard part is estimating health per dollar: the cost-per-QALY of each cause and how credibly the evidence behind it identifies cause and effect. So this is a GiveWell-style model: 13 intervention archetypes, each cost-per-QALY drawn where possible from a published causal estimate (Medicaid mortality, community health centers, supportive housing, collaborative-care depression), each effect then shrunk toward zero in proportion to how well its study identifies causation, and the whole thing run live through thousands of Monte Carlo draws each time you move a slider.

I built the model with Claude; every estimate below is a model output, not a measured fact. The Python package, tests, and sources are on GitHub; this page runs a checked TypeScript implementation in the browser, reading the exported parameter file.

The single most important control is evidence stance. Slide it from skeptical to credulous and the central estimate moves from ~98,000 QALYs (weight each effect by study quality) to ~238,000 (trust every cited effect at face value). That gap — not the dollar figure's last decimal — is the real uncertainty.

Assumptions

Drag to see the estimate move. The model reruns in your browser.

skeptical

Skeptical weights each effect by how well its study identifies causation. Credulous trusts every cited effect at face value.

0.80

Fraction of the studied effect a marginal unrestricted grant actually delivers.

$26.3B

Lifetime total. Default $26.3B; Yield Giving says over $26B in 2,700+ gifts.

3.0%

Annual discount on future life-years.

Allocation across causes
6%
5%
4%
8%
7%
8%
8%
18%
22%
4%
3%
4%
3%
Running Monte Carlo…

How it works

Each Monte Carlo draw allocates the giving across the archetypes (a Dirichlet around a best-guess split, since Scott doesn't publish dollars-by-cause), assigns each a cost-per-QALY, and multiplies by two independent discounts:

The "vs. global frontier" figure compares Scott's blended cost-per-QALY to the best global-health buys, handicapping the frontier with the same realization and credibility so it's like-for-like. GiveWell's current impact estimates put malaria nets around $5,500 per life saved; an older AMF/GiveWell summary reports about $3,340 per life saved and about $78-$100 per DALY averted. I model that as one $50-$150/QALY-equivalent benchmark, treating one DALY averted as approximately one QALY gained under the one-year-of-full-health convention.

What this doesn't capture

A QALY is a health metric. Most of Scott's giving targets economic mobility, education, and equity, whose value is largely non-health — income, opportunity, rights, wellbeing. The model therefore understates her total social impact; it answers one specific question. The largest dollar buckets (equity & justice at ~22%, education at ~18%) contribute little health precisely because no credible study ties those grants to QALYs, not because the giving lacks value.

Key sources

Full annotated bibliography, parameter file, and the 34-test Python suite: github.com/MaxGhenis/mackenzie-scott-qaly.