A Model
As effective philanthropists, we are concerned with where to deploy our scarce resources to do the most good. We are also faced with a question of when to deploy these resources. We must decide between donating now and investing to donate later.
Below, I present a simple toy model of global health and development giving which suggests patient philanthropy - investing to donate later - is more effective at improving lives given current conditions.
Let's imagine the target population for our giving has an annual consumption growing at an annual rate of . Their consumption over time is then
If we assume logarithmic utility with respect to consumption, the utility conferred to this population by a donation of size at time is
So for a constant donation size, the utility provided to the target population by the donation decreases over time - just as we would expect for a population that is becoming richer. However, if the size of the potential donation increases over time with investment return , the utility provided by the donation increases over time as long as .
This suggests that as long as investment returns outstrip the rate of income growth among the global poor, saving to donate later will lead to greater benefit than donating today. Over the last 35 years the compound annual growth rate of the real threshold income for the poorest decile of the world population was 2.62% per year,[1] while real returns on global equities have been approximately 5.2% annual over the last 125 years[2]. As long as these conditions persist, keeping your charitable funds invested in stocks makes them increasingly valuable for future giving to the global poor compared with donating them today.
Questions
- Should I take this model and its conclusions at all seriously? Obviously, there are a lot of idealized assumptions here - which ones are most unrealistic?
- If an individual donor wanted to take a patient philanthropy strategy, what is the best way for them do that? One could simply save money themselves and disperse it upon death, but if they wanted even longer compounding horizons they could consider donating to a patient fund such as the Founders Pledge Patient Philanthropy Fund or even setting up their own trust or foundation - but those options come with their own risks.
- Does this argument apply to existential risk reduction giving? Obviously we have no simple model or empirical finding for how 'cost per risk amount reduced' evolves over time. Some of these questions are explored in Phil Trammell's 'Patient Philanthropy in an Impatient World' and its discussion of how the 'hingyness' of the present moment may evolve over time.
- ^
World Bank. Poverty and Inequality Platform. 2025.
- ^
Dimson, E., Marsh, P., & Staunton, M. (2025). Report: Stocks have far outperformed over the past 125 years. Cambridge Judge Business School. https://www.jbs.cam.ac.uk/2025/report-stocks-have-far-outperformed-over-the-past-125-years/
Good post!
One possible complication I didn’t see addressed is the role of cause saturation.
Suppose that the most effective global health charities (or interventions more generally) are likely to become saturated over time as EA grows, more money flows in, and people “get their act together.” If that saturation happens on, say, a 5-year timeline, then delaying donations means missing the chance to fund the top-marginal opportunities now. Even if you plan to donate to the “next best” charity later, there’s a real cost: the best opportunities would have had 5 extra years of impact that are now lost.
In other words, for waiting to be better, the investment return advantage must outweigh not just the growth in baseline incomes, but also the lost value of funding the most cost-effective opportunities before they close.
Im not sure how likely this sorta thing is in practice, but i thought it was worth a note.
In theory, this seems important and worth considering. Another effect that might pull in the opposite direction:
As we learn more about effective causes we are able to identify more effective solutions/issue areas.
It's not obvious which effect (or something else) will dominate. One way we might be able to acertain the answer to this is to look at the effectiveness of Givewell's top charities across time. My understanding is this hasn't moved much, but also that their definitions of "life saved" has changed across time. Unsure which direction that might affect things.
Yep, agreed. Good point
yes, I definitely think this is a complication here. The toy model in this post assumes the only cause is something like direct cash transfers. I think this makes sense as a baseline (for the same reason GiveWell uses cash transfers as a baseline) but of course we can and do find global health interventions more promising than cash transfers and it is possible the effectiveness of these interventions diminishes over time faster than investment returns. However, I do not think this is what we have seen so for in practice. In 2015, GiveWell had 3 non-cash charities they estimated to be 5-10x more effective than cash transfers, but by 2018 they had 7 which they estimated to be 5-15x more effective than cash[1].
https://www.openphilanthropy.org/research/givewells-top-charities-are-increasingly-hard-to-beat/
I don't think I have a good objection here.
1) You could make an objection about value drift and this should influence you to donate now, but I don't think this gets to the heart of the issue.
2) If now is the "hinge of history", maybe it is a uniquely good time to do longtermist philanthropy.
However, if we believe neartermist work is pressing enough to justify funding as well, it seems like patient philanthropy is pretty much a pareto improvemnt over normal neartermist philanthropy.
Would any justification for neartermist philanthropy change this?
I think value drift (along with risk of losing control of your assets due to any number of scenarios) can just be modeled by discounting the expected growth rate for risk of loss. If you think you have a 1% chance of losing your money each year, you might treat your investment growth rate as 4.2% instead of 5.2%, for example.
This is an interesting analysis!
The two parts that seem most unrealistic to me:
good points!
with respect to the utility function I originally chose log because it's what Open Philanthropy uses. However I now see that GiveWell sometimes uses an isoelastic utility function with η=1.59, which is faster diminishing returns than log-utility (η=1).
Redoing the math with a general isoelastic utility function, for η>1 you get an optimal point at which to donate, with that point depending on the parameters of the model (income and investment growth rates, the rate of diminishing returns, etc.). This optimal donation time is also dependent on the size of the population you can donate too, so to get a more accurate model you would need to incorporate that (as well as a bunch of currently-unaccounted-for factors like the changing effectiveness of non-cash charities).