The Mackenzie Method: How to Waste Billions and Influence Nobody

Philanthropy can be good. Although free market capitalism has lifted billions out of poverty and massively enhanced human welfare, it is not optimized to save lives or enhance well-being. Unpriced externalities, both positive and negative, present opportunities for philanthropists to step in and do precisely that. But back chaining from desirable outcomes is much harder, as proponents of central planning have realized time and again.

This is because capitalism only needs one thing to work - rational self interest. If each actor can be presumed to want to be better off, and the path to making oneself better off necessarily flows through making others better off through consensual transactions, then “betterment” is sort of embedded in its natural logic. Each actor need only worry about what makes them individually better off.

On the contrary, effective philanthropy necessitates not just identifying market failures but also estimating how different solutions will affect aggregate social welfare. Companies and startups receive real time feedback whether they like it or not. People either buy your product or they don’t. Nonprofits have to make sure they are measuring the right thing and measure it, even when there’s no “profit” incentive to do so for anyone involved. 

When we enter the domain of important but relatively hard-to-measure problems (eg. improving governance or institutions), one has to rely on a priori reasoning and guard carefully against cognitive biases. But it can still be worth trying, if the risk of unintended consequences is low enough. In other words, philanthropy can be great, if you try really hard to follow reason and evidence, and set up structures, incentives and cultures that make it easy for others to do so as well.

After spending some time looking through how Mackenzie Bezos chose to give away $17B, I’m convinced that she very likely did not do this. My claim is that Mackenzie Scott clearly spent her money sub-optimally for any conceivable set of consistent values and goals. I believe it was a mistake to give at the speed she gave, because it’s virtually impossible to find enough cost effective opportunities that can absorb $17B over the course of two years. I also think there’s a reasonable chance that it would have been better for her to do nothing.

However, I want to start by giving Mackenzie some credit. I think it’s generally good to be public about your philanthropy, since it likely encourages others to give more. Humility might be a selfish, overrated impulse in this domain. I’m also glad she made her giving public, which I believe is a good precedent to set. But it’s not enough that we encourage a culture of giving, since there’s nothing inherently good about giving. Giving is good if your giving ends up helping people more than if you hadn’t given. And that’s the only type of giving worth encouraging.

How do we know MacKenzie is granting on “vibes”?

This twitter thread by Alex does a good job of making the prima facie case. Mackenzie’s process involves “quiet research” and open calls. If you’re confused about the former, it means doing desktop research and coming to a grant decision before establishing contact with the organization, in order “to avoid diverting them from their work”. In other words, no detailed due diligence. So what does this desktop research entail:

Evaluate organizations through careful analysis of criteria specific to their size, geography, and mission for indicators of high potential for sustained positive impact, including stable finances, multi-year track records, measurement and evidence of outcomes, and experienced leadership representative of the community served 

Mackenzie’s evaluation criteria

They evaluate organizations. What do they evaluate them on? On important criteria. Important for what? Important for sustained positive impact. Ok, I’m being unfair. They do mention four indicators, but three of these have serious problems or inconsistencies:

1. Stable finances

The meaning of "stable" in this context is unclear. If it refers to being "well-funded" or having fewer funding shortfalls, it's questionable why this criterion is at the top of her list. In theory, philanthropists should seek out opportunities with high counterfactual impact—projects that wouldn't have been adequately funded otherwise. However, the presence of other funders can also serve as a positive signal, indicating that the organization has undergone due diligence and gained the trust of other donors.

This is similar to how venture capitalists often co-invest alongside other reputable funds, relying on the collective wisdom of the market. However, there's a crucial difference between the incentives of philanthropists and investors. While investors are primarily motivated by financial returns, philanthropists may have various reasons for giving, such as personal fulfillment, social status, or genuine altruism. As a result, following other philanthropists' lead may not always align with maximizing impact for beneficiaries.

Moreover, as one of the world's wealthiest individuals, MacKenzie Scott has the unique opportunity to identify and support promising but underfunded organizations. By conducting thorough research and due diligence, she could have uncovered hidden gems and catalyzed their growth, encouraging other donors to follow her lead.

2. Measurement and evidence of outcomes

Great, Mackenzie seems to want to fund organizations that diligently measure and track impact. However, it’s very unlikely that they screened organizations rigorously on this metric. For this to be possible, these monitoring and evaluation practices would have had to be public or easily available in some other way.

Consider Girls First Fund, a grantee that received $15M from Mackenzie (which is on the larger side of the check size spectrum). Girls First Fund, in turn, writes checks to smaller grassroots organizations, with the aim of reducing child marriages, primarily in South Asia and Africa. Look through their annual report, and you’ll find no evidence of their impact on the ultimate beneficiaries - how many women they reached, how these women were helped, and how many child marriages were averted in expectation.

The “evidence” is instead centered around how much money they managed to spend and how many “women-led” organizations they funded. That gets us no closer to understanding how many child marriages they are likely to avert in the future, which is what we should care about, correct? Even if the outcomes are difficult to measure, the annual report could have but did not contain other qualitative signals that they take impact seriously - perhaps how they vet the grassroots organizations or a fleshed out theory of change section. So either the quiet research wasn’t very quiet or it wasn’t particularly rigorous.

3. Experienced leadership representative of the community served

Visit Mackenzie’s website and you’ll see this point retierated in several places. Below is an excerpt from her essay referring to a previous, smaller round of giving:

On this list, 91% of the racial equity organizations are run by leaders of color, 100% of the LGBTQ+ equity organizations are run by LGBTQ+ leaders, and 83% of the gender equity organizations are run by women, bringing lived experience to solutions for imbalanced social systems.

It's hard not to notice that focusing on leadership representation is a great tool for signaling one's commitment to diversity and inclusion but what is its likely impact on beneficiaries? Even if you concede that in some theoretical sense, organizations serving “underrepresented” communities are best served by organizations led by someone from precisely that demographic, making this an important scoring criterion seems like a bad idea.

This should follow fairly straightforwardly from the fact that you think these communities are “underrepresented”. Presumably, this has implications for the preponderance of human capital, especially senior management talent. It’s preposterous to claim that systemic racism keeps people from skills acquisition and being represented in the upper echelons of a variety of professions, while insisting that these communities must produce the leaders with the requisite skills, experience, and networks to run all the organizations that meaningfully impact these communities.

Such insistence necessarily implies trade-offs that lower the expected impact of your philanthropic portfolio. If you’re hiring someone to run your business and you only hire from, or give extra points to, applicants from your family, community or even country, it’s less likely you’ll find the best person for the job. Insofar as one believes CEOs and senior management have significant impact on the success of a nonprofit, imposing additional scoring criteria that don’t track probability of success will detract from that goal substantially. This is especially true if the community in question is underrepresented, because there are fewer such people, per capita.

If anything, a preference for a significant proportion of employees or field staff being drawn from the beneficiaries’ community is more reasonable. One could argue that this has important implications for building trust within the community, which can impact the quality of feedback and the ability to deliver services that the beneficiaries are actually happy with. Even if you believe leaders from other ethnic backgrounds couldn’t possibly care as much, the obsession with representation in leadership makes no sense in a world in which people can be paid and incentivized to act as fiduciaries.

Mackenzie’s focus areas

As Alex points out, one of the bigger clues for “vibes” based granting comes from looking at the top focus areas (by $ amounts). It looks an awful lot like what someone would fund if their only source of information about the world’s most important problems came from corporate social responsibility (CSR) brochures. At the level of identifying and allocating to different causes, this particular allocation seems inconsistent with any coherent set of values.

If Mackenzie’s goal was alleviating suffering or mitigating the worst injustices of the birth lottery, a lion’s share of the funds should have gone to beneficiaries in the third world. Out of Mackenzie’s disclosed grants of $12B, $10.5bn (~87%) went to charities with US only operations. To put things in context, consider this - India was recently in the news for eliminating extreme poverty. Only 2% of Indians - 28M people - now live on less than 2$ a day or $700 a year (purchasing power adjusted). In contrast, in America, only about 200,000-300,000 fall below that threshold (some claim even that figure is an overestimate). More than 700M people live below that threshold globally, most of them concentrated in Sub-Saharan Africa and South Asia. If you want to help the have-nots, I would think that allocating roughly 90% of your funds towards helping Americans would be a mistake, even if you do have some special obligation to help Americans more.

I’m not claiming that alleviating poverty in the developing world is the only noble philanthropic goal. Since innovation is concentrated in the United States and a handful of other countries, and innovation has large positive externalities/spillover effects, you could argue that it’s in humanity’s long term interest to be funding R&D efforts in the first world, especially those that the market wouldn’t otherwise fund. That would be a strategy that’s much less quantifiable relative to increasing vaccination rates in Africa but not clearly inferior (conditional on the details of execution and which technologies get funded). However, Mackenzie doesn’t seem to have done much of this. The keyword ‘innovation’ finds 35 matches in Mackenzie’s giving database, and all of those pertain to “innovations in equity” or funding “innovation centers” for underrepresented communities, not funding projects at the frontiers of science and technology.

Perhaps Mackenzie focussed on the US because she believes social and political stability in the US is the important cause, lest the whole word be thrown into disarray; and that social justice and racial equity are pre-requisites for a stable soceity as she alludes to here:

Like many, I watched the first half of 2020 with a mixture of heartbreak and horror. Life will never stop finding fresh ways to expose inequities in our systems; or waking us up to the fact that a civilization this imbalanced is not only unjust, but also unstable.

When it comes to pursuing systemic goals like “improving social mobility” or “strengthening democracy”, people argue that our inability to measure it accurately should not be a reason to not pursue them. I mostly agree. But the problem is that we don’t really know how to bring about systemic changes. This is almost by definition since “systemic” necessarily implies a web of complex and interdependent processes and relationships. Our democratic and bureaucratic institutions are large scale coordination problems (or prisoner’s dilemmas if you will) among thousands of stakeholders with different interests and preferences. Systemic change involves being able to move everyone towards a new equilibrium.

The answer to this need not be apathy. It’s ambitious but reasonable to say “I have $17B and I’ll pay people to figure this out”. However, intent is not enough. One has to be able to decompose the problem to make it tractable. Perhaps this means identifying the most tractable levers for change; or honing in on policies and laws that cause the most harm to the communities you want to help. If anything, this undertaking requires more focus on prioritization and a willingness to research, experiment, monitor and iterate.

This post would’ve been a lot shorter if all I disagreed with was Mackenzie’s priorities. But even if all that Mackenzie cared about was making African Americans better off, I don’t see strong evidence of a resolve to bring that future to fruition. Mackenzie sure has written checks to a bunch of black colleges and nonprofits focused on issues that affect the black community. But she doesn’t seem to have a broader theory of change or strategy that underpins her donations. In all her essays and across the content on her website, I didn’t encounter any material on prioritizing between issues that affect these underrepresented communities or discussion of the relative tractability of different problems and methods.

I’m not convinced that Mackenzie would have a good answer to a question like “Why did you donate $50M to housing reform and $100M to education? And within housing, why did you prioritize an org that works on X policy over Y policy?” My guess is her answer would take us back to her evaluation criteria - of track record, representative leadership etc. But anyone who’s committed to doing good should be able to draw a straight line between the world they want to see and how their donations will bring that to fruition. That seems blatantly absent.

Mackenzie’s insistence on giving fast

The question of philanthropic timelines, at least in its abstract form, is mostly a macroeconomic one. On the one hand, deferring philanthropic grants means they could be larger grants in the future, if your money grows with the market. On the other hand, philanthropic opportunities to do good will (hopefully) be scarcer and more expensive in the future, as diseases are cured, wealth is generated and the secrets of well-being are unveiled. But this argument is more about whether one should try to give away money in this generation or save it for humanity’s future problems. (I’d recommend Phil Trammel’s work on this). But even if you do decide to give away your fortune within your lifetime, it’s very unlikely you can give away over $6B per year cost-effectively for two or three years, as Mackenzie has.

In 2022, GiveWell, the organization that’s widely known as the gold standard for identifying and directing funding towards cost-effective charities in the global health and well-being space, set out to raise $600M but identified $900M of highly cost-effective opportunities. The remainder is what GiveWell calls its funding gap - or the additional amount it can move cost-effectively but couldn’t due to lack of funds. GiveWell is dedicated to finding new cost-effective opportunities every year, but it seems unlikely that they would have been able to move even a fraction of Mackenzie’s 17B over three years.

This is because virtually no high impact organization can absorb arbitrarily large amounts of cash and maintain their cost effectiveness. (much like successful hedge funds that cap the size of their funds). This is mostly a function of diminishing returns. If an organization has been implementing a cost effective SMS campaign to raise awareness about the benefits of vaccination, perhaps the first tranche of additional funding would help them expand and cover the entire country they’re operating in. Once they’ve sent text messages to every family in the country, they need to figure out a different way to reach people who weren’t already persuaded, which is likely something higher touch and more expensive (and less cost effective). At some point, they may expand to a different country and seek to improve vaccination rates already at a much higher base and so on. Moreover, as charities get larger and expand, the problems of coordination and bureaucracy chip away at cost effectiveness.

Absorbency is a concern - perhaps even more so - if your philanthropic interests are in areas that are far less measurable or quantifiable. If you’re funding research into some new promising area, you can only spend funds effectively if you’re sensitive to feedback loops that tell you whether certain research directions are promising. This probably means a phased deployment and waiting for results of Phase 1 before you move on to Phase 2. Hiring 10,000 researchers right away to work on a new promising area is likely to be mostly wasteful if all it took to uncover bottlenecks were 5 researchers and the incremental funds were best spent on engineering teams to work on those bottlenecks.

So what could be a justification for ignoring these concerns and giving tens of billions within a couple of years? One possibility is the belief that the scale of human suffering currently is so unconscionable that we can’t afford to think about the future or anything else. As I pointed out earlier, if that were true, there is no good reason to prioritize America over the rest of the world. You should be almost exclusively focused on the developing world. Another possibility could be a focus on catastrophic risks (eg . risks from pandemics or artificial intelligence). If one is convinced these are significant and imminent, it could make sense to pour all your funds into it. But you still have to find opportunities to effectively turn money into risk mitigation. 

Given her allocation however, it appears that Mackenzie believes the suffering of black, hispanic and female Americans is so intense and urgent that it overrides every other problem in the world. Even if you accept that premise, you have to ask what the best ways of helping these groups are. Unless you're willing to just wire "deserving" Americans their share of your fortune and call it a day (which might’ve been better than what Mackenzie did), you can’t keep this pace unless you prioritize giving fast rather than giving effectively.

Would it be better if Mackenzie did nothing?

You can’t stuff $17B under your mattress. So doing nothing here entails investing in a diversified portfolio - primarily bonds and stocks, but also likely including smaller allocations to real estate, hedge funds and private equity.

It’s not exactly fashionable to tell billionaires they should hoard their money. But absent great philanthropic opportunities, they probably should, at least for now. This is true even though the additional dollar has virtually no impact on Mackenzie or any billionaire’s well-being, because capital is the fuel that the process of capitalism relies on to keep producing welfare as its side-effect. All else equal, when Mackenzie’s advisor buys Company A’s bonds, it helps company A borrow at a slightly lower rate that helps them take slightly higher risks and produce slightly cheaper products for consumers. More so because capital is a renewable form of fuel. Insofar as Mackenzie’s competent advisors make her a solid risk-adjusted return, her wealth will keep growing, and can be reinvested ad-infinitum, contributing to the capitalistic engine. To top it off, if amazing philanthropic opportunities come by, she can liquidate some of these investments to write even more generous checks.

I can feel the retort coming. Well, when Mackenzie writes checks to these non-profits, even if the non-profits don’t end up helping the beneficiaries in some tangible way, isn’t she helping pay employees and transferring wealth from herself to people who need it more than she does?

Yes, she is. But virtually anything she does with this money is likely to do that. Assuming the wealth generated will at some point be consumed, this consumption will go towards employee’s salaries and vendor’s business accounts before it continues to circulate further in the economy. Imagine Mackenzie gave her son some of her money and he decided to buy a Lamborghini with it. Her son thinks it’ll make him happy but it doesn’t. But the money still flows through the company’s income statement and pays for employees and vendors before the surpluses end up ultimately being distributed to the shareholders.In fact, there’s a case to be made that buying products from companies with profitable business models is better than funding non profits that have little impact but that’s for another day. 

I can’t prove to you that Mackenzie would have been better off buying Gucci bags and Lamborghinis. I’m also not claiming that Mackenzie is worse than the median philanthropist. I’m glad that she donated $50M to GiveDirectly for example, one of the few organizations that can absorb large amounts of capital without diminishing returns. But these decisions seem like a natural consequence of spreading her money across thousands of nonprofits, some of whom will no doubt have a positive impact on the world. When you have $17B at your disposal, you have the ability to envision a new world, use your resources to answer questions that have never been answered, and pay the right people to turn those answers into better lives. Unfortunately, giving fast and giving publicly took precedence over that admittedly more difficult yet impactful path.