Science1 publisher2 min readPublished
A Lancet model scores ten proposed taxes against the mortality cost of aid cuts
The ISGlobal-led study puts avertable deaths at 6.6 to 29.5 million by 2030, and that four-and-a-half-fold spread turns on which tax is levied and at what rate rather than on any statistical bound.
The Scientist · Science desk
What happened
- The Lancet study, led by ISGlobal with economic and health research partners, fitted the aid-mortality relationship across 59 low-income and lower-middle-income countries from 2002 to 2021 and projected it to 2030.
- Its baseline scenario extends current cuts in official development assistance to 2030 and projects about 7.6 million additional deaths, of which 1.4 million would be children under five.
- Ten revenue measures already in international discussion were assessed, including levies on billionaires, multinationals, financial transactions, carbon emissions, cryptocurrencies and sovereign debt interest.
- Every modelled tax scenario raised enough to offset part of the projected harm, and in most cases enough to offset all of it.
- The authors state that their purpose was to estimate the health impact of these measures rather than to endorse any particular one.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- decision A donor ministry trimming assistance is now arguing against a published mortality counterfactual rather than a budget line, which raises the price of making the cut quietly.
- constraint None of the ten mechanisms is collecting at the modelled rates, so the range describes revenue that would have to be legislated before any fund could draw on it.
- precedent Health-framed revenue arguments now have a citable death count, which makes mortality modelling a standard exhibit in tax negotiations whether or not the underlying elasticity is causal.
The whole ledger rests on one fitted relationship. Where development assistance was higher, overall mortality was 24% lower and mortality among children under five was 33% lower [3]. That is observed variation across countries and years, not an experiment: aid flows toward places that also differ in conflict, income and administrative capacity, and a single aid coefficient absorbs whatever those differences did. The projections multiply that coefficient forward to 2030, so every averted-death figure inherits it. Halve the causal share of the association and you halve the lives.
That makes the internal comparisons more informative than the top-line number. A 3% tax on billionaire wealth carries the largest estimate, up to 29.5 million deaths averted by 2030 [8], roughly four times the deaths attributed to continued aid reductions [1]. So the top of the range prices an aid budget substantially larger than the one being withdrawn, extrapolated along the same curve. The bottom of the range, 6.6 million, is about 87% of the projected additional deaths [2], meaning the weakest scenario modelled still leaves a residue behind.
Sufficiency does not pick out billionaires. Financial transaction taxes were associated with around 24 million deaths averted, and a global minimum tax on large multinationals with about 20 million [9][10], which is 3.2 and 2.6 times the projected shortfall respectively [3]. The case for large-fortune taxation put by Lucio Exposito, senior economist on the study, is a political one: with donor governments more indebted and reallocating toward military spending, he calls funding humanitarian assistance through taxes on large fortunes "one of the most viable strategies" [11]. Viability and magnitude are separate tests, and these numbers only settle the second.
What the published account withholds matters for reading the range: no confidence intervals on the mortality estimates, and no revenue total for any of the ten mechanisms [15]. Without the revenue, the step where dollars become deaths cannot be checked, and neither can the question of how much of a 3% levy's notional yield survives avoidance. That is the joint where estimates of this kind usually fail, and it sits outside the model, as does the assumption that new money buys the same programmes at the same effectiveness as the assistance it replaces. The transferable part of the work, for the G20, OECD and Seville financing discussions where these measures are live [14], is the conversion rate from money to mortality rather than the figure attached to any one tax.
What to watch
- Independent re-analysis of the aid-mortality relationship using post-2021 data from the current round of donor cuts.
- Whether the OECD/G20 Inclusive Framework turns discussion of minimum and wealth taxation into an agreed rate.
- Any treasury publishing net revenue projections, after avoidance, for a billionaire wealth levy at the modelled 3%.