Hi EA Community, I’d like to introduce my high-impact startup and spark a conversation. I’m also seeking potential co-founders or funders to help get our idea off the ground. I introduced myself last week.
- Summary - The Logical Foundation is a 501(c)(3) nonprofit startup founded in 2022. Its mission is to ensure a prosperous future for humanity by unleashing the massive anti-poverty potential of guaranteed income. This post uses an ITN (importance, tractability, neglectedness) framework to analyze our intervention, strategy, and cause area.
- This post gives my understanding of Guaranteed Income’s potential to cost-effectively reduce poverty and homelessness at scale in developed nations, and why our Maximum Impact Pilot focuses on homelessness in Arizona. It’s based on my research on guaranteed income, homelessness, technological disruption, and some ideas from Doing Good Better.
I will discuss
- The massive anti-poverty potential of guaranteed income in developed nations, and how we’re stuck in Pilotland.
- Our Big Idea: Building a platform to unleash the massive anti-poverty potential of guaranteed income.
- Our Theory of Change. How can we scale to the levels needed to meaningfully reduce poverty and end homelessness?
- ITN Framework: Our grand plan to escape Pilotland and reduce poverty with guaranteed income.
- The implications of guaranteed income for humanity and the Effective Altruism Movement.
- Longtermist implications of guaranteed income.
Our Startup Plan: Maximum Impact in Developed Nations.
- Experimental results support the cost-effectiveness of guaranteed income against homelessness.
- Arizona is by far the deadliest place in the U.S. for homeless people.
- QALYs (Quality-Adjusted Life Years) expected from Foundation's Maximum Impact Pilot
My recommendation? Effective Altruists should strongly consider helping us grow.
- What help do we need to (hopefully) make it happen, and how EA people and organizations can increase our chance of success?
The massive anti-poverty potential of guaranteed income in developed nations, and how we’re stuck in Pilotland.
Guaranteed income is a type of cash transfer program that provides continuous unconditional cash transfers to individuals or households.
Cash transfers are the most widely researched intervention in the world, and also the most consistently positive intervention. While we know that niche global health interventions can outperform cash transfers at small (<$1B/Year) scales, direct cash is the only intervention that can effectively meet the $100+ Billion a year scale of global poverty.
Aid Effectiveness = Value Transfer Efficiency
Aid programs are designed (in theory) to bring as much value as possible to people in need. Conventional aid programs help beneficiaries by providing them with goods and/or services.
Unfortunately, it’s expensive to provide goods and services. Paperwork is usually an undignified hassle, and beneficiaries rarely get exactly what they need most in developed nations. Oh, and many interventions are ineffective or actively harmful.
In the developing world, there are a few niche opportunities to distribute basic, cheap, neglected health interventions that can save a lot of lives, more than cash would, at small scales. That said cash assistance is the only non-niche intervention cost-effective enough to end global poverty at scale. Our focus is exclusively on developed nations, and how aid works there.
Guaranteed income is architecturally superior to other forms of aid.
Fundamentally, poor people are way better than nonprofit ‘experts’ at knowing what they need and getting it. For example, if someone needs a car loan and the aid program provides food stamps, they have to launder their food stamps to get the cash for their car loan. Guaranteed income empowers poor people with liquid aid.
The image above shows why guaranteed income is intrinsically more efficient than other forms of aid.
While some still argue that direct cash transfers could harm recipients, by, for example, causing them to gamble or use drugs, there is extremely strong evidence to the contrary from every form of cash transfer experiment. Most guaranteed income proposals in developed nations top out at $1,000 per month, enough to get basic needs, reduce stress levels, and break down harmful stereotypes about people living in poverty, but not enough to live comfortably.
Essentially, funding guaranteed income is the way to spend money to help people in developed developed countries. Here's the situation (Focusing on the U.S. first). There are two important categories of guaranteed income: policy, and the nonprofit industry.
Guaranteed Income (Universal Basic Income) Policy
50% of Americans don’t have $400 in cash on hand to deal with an emergency. That’s a crazy but true statistic, so a national guaranteed income policy would do a massive amount of good. This was proven when the expanded child tax credit (a temporary UBI for families with kids) briefly drove child poverty down by over 40%.
A great article by the Washington Post laid out what’s going on from a policy (government funding) approach in the guaranteed income movement.
“If empirical evidence ruled the world, guaranteed income would be available to every poor person in America, and many would no longer be poor. But empirical evidence does not rule the world.”
Guaranteed Income has a long and storied history in the U.S., which I will quickly skim over as it is not what this post is about. Regarded by historians as the Father of the American Revolution, Thomas Paine, advocated unsuccessfully for a form of universal cash payment in Agrarian Justice. Dr. Martin Luther King Jr. advocated it as the only obvious solution to truly ending poverty right up until his assassination. President Nixon nearly passed a Universal Basic Income but was thwarted ironically by a democratic senate that wanted larger payments and a 200-year-old fraudulently administered study. Guaranteed Income then lay mostly dormant until presidential candidate Andrew Yang decided to make a Universal Basic Income his #1 policy priority.
Since then, a ton of publicly and privately funded guaranteed income pilots have proven just how powerful and effective guaranteed income is at helping people. Unfortunately, it seems like getting a guaranteed income passed federally or in any state is highly unlikely for now.
Several organizations exist to push guaranteed income policy, such as the Economic Security Project, and Mayors for Guaranteed Income, so we’ve decided to pursue a different neglected route to proliferating guaranteed income.
Our work won’t get roadblocked by politics and presents an opportunity with a very high chance of success. As we scale, our work will accelerate the progress of all the existing organizations working towards a guaranteed income policy.
Our Focus: The Nonprofit Industry
The Guaranteed Income Movement is living in Pilotland. Over 100 guaranteed income pilots have recently been announced or completed in cities across the U.S. There’s a problem though. As new pilots get announced, old pilots end and fizzle out. We keep seeing the same cycle across the country:
- A new pilot is announced
- The pilot reports amazing results with a huge impact
- The pilot ends, often without renewal or expansion
There are few if any cases of pilots leading to the establishment of long-term guaranteed income programs. This is a shame because guaranteed income has proven much more consistently impactful than all other anti-poverty interventions.
Americans gave $484.85 billion in 2021. The nonprofit industry brings in hundreds of billions of dollars each year. Not governments, just private foundations, corporations, and philanthropists.
Why aren’t they spending their money on guaranteed income?
The clearest answer is that they do when given the chance. But that only happens when there is a guaranteed income pilot in their area. Most nonprofits and funders in general have geographic limits on where they spend their money. Many also have limits on the sorts of people they are allowed to help (youth vs LGBTQ vs African American vs veterans, etc...). If there isn’t a guaranteed income pilot in a specific town, none of the funders that live there will be spending any of their money on guaranteed income.
They are not free to spend their money on whatever thing, anywhere on earth, has the largest impact potential. They can only spend money within their limits, and guaranteed income will almost always be the most impactful thing they could spend it on because almost all of them are limited to developed nations.
Imagine you’re a funder living in Pilotland, where tons of varying experiments are going on all over the place. If you’re looking to make an impact, the odds are strong that there are not any guaranteed income pilots within your geographic limits helping the people you exist to help. You can either create a program yourself or deal with the less effective options on your plate.
Our Big Idea: Building a platform to unleash the massive anti-poverty potential of guaranteed income.
Although there are logistical complications, cash is easier to administer than any other intervention, enabling us to create a platform for fundraising, and disbursing guaranteed income within all geographic areas and for many kinds of populations in need. The idea is two-fold.
- Empower as many funders to fund guaranteed income as possible. Funders often have geographic limits and/or population limits, so we will build a donation platform that empowers them to spend their funds on guaranteed income within their limits. Doing so maximizes their potential for impact, and also enables us to construct an incredibly powerful fundraising machine. Since guaranteed income is proven to be the most effective (not to mention dignified and empowering) anti-poverty intervention in developed nations, our fundraisers will
- Know that the money they are asking for could, literally, not be spent better. This should be extremely motivating and fulfilling (from personal experience, it is).
- Have the strongest data-backed solicitations in the fundraising industry. Combining powerful evidence with highly motivated fundraisers (that are treated well and paid fairly) should result in a world-class fundraising machine.
- Disburse guaranteed income as effectively as possible. Luckily, organizations such as AidKit have already built most of the infrastructure needed to disburse cash at scale. This should be the easy part. As funding rolls in from various places and with various limitations, we will open highly automated online applications for guaranteed income. Guaranteed income will become available (with limits automatically applied) as soon as enough funding rolls in to provide guaranteed income in each location. We will ensure that follow-ups, audits, and rigorous Randomized Controlled Trials (RTCs) are baked into the platform. This creates a win-win-win for:
- Funders that want to maximize their impact against poverty and get detailed impact reports they can trust.
- Fundraisers that want to maximize their impact
- People in need. They will be empowered to fulfill their basic needs, regain dignity, and pursue fulfillment.
As the scale of guaranteed income programs increases, each incremental staff member can manage more and more participants. A $100M program, for example, funding 4,166 participants at $1K per month for two years -- with automation -- should require less than 1 staff hour spent per incremental participant. That equates to less than 1 full-time worker over 2 years. The vast majority of that time is required right at the beginning, however, so onboarding participants must be effectively spread out over time. As such, a reasonably sized team with staff specializing in specific roles could manage well over 1,000 additional participants per incremental staff member.
One lean team of ~20 could, we believe, manage to disburse $100M+ a year. At that scale, the bank interest from funds pending disbursement could theoretically exceed the amount needed to operate the platform.
Many details are still in development, and we will publish “Foundation’s Basic Income Network Whitepaper” when it is ready. We have had this concept reviewed by over a dozen nonprofit and financial professionals, so we are confident that it is likely to work. If we can get it off the ground (<$5M we estimate), it should be able to unleash the massive anti-poverty potential of guaranteed income.
Why should we be the ones to build this?
We started a new nonprofit to exclusively focus on this for a few reasons. First, I realized that while GiveDirectly could build the platform, it might go against their #1 value of "putting the poor in control of how aid money is spent". If we want to maximize funders, we have to work within their bounds, and I think GiveDirectly might not be ok with sending money to people that are not desperately poor.
And that could be a very good thing! GiveDirectly probably should focus on international cash transfers, because building this platform would quickly make their extremely important global poverty and crisis response work a very tiny fraction of their overall funding. Such a situation could lead to GiveDirectly (unintentionally, of course) losing focus on eliminating global desperate poverty and crisis response. They are doing fantastic work developing crisis response, and our platform won’t be designed for that sort of need.
I think it would be better for us to focus entirely on developed countries ourselves (with Aidkit or some other infrastructure contractor), sending all funds that we get from funders without geographic limits to GiveDirectly (after fully funding GiveWell's top charity fund). Doing so would enable us to focus on the platform, and help accelerate GiveDirectly’s work. That said, we would love to collaborate with them on fundraising and improve our system with their experience.
Our Theory of Change. How can we scale to the levels needed to meaningfully reduce poverty and end homelessness?
I posted a blog last week introducing myself and where my expertise lies. I've put in ~10,000 hours analyzing & predicting disruptive systems (eg, understanding emerging technologies and shaping them to make the future better).
I hypothesize that guaranteed income is such a system, although it is not a hardware-based technology like solar or the internet. I’ll first explain how I see us scaling our impact, and then I’ll try to explain some larger principles with the help of the Seba Technology disruption framework.
Our theory of change
Every successful company, technology, and societal change has come about through the creation of powerful virtuous cycles. We believe that our system can leverage guaranteed income to create increasingly large direct impacts while establishing cash as the baseline for nonprofit interventions in developed nations.
- Bring in funding from donors looking to maximize their impact against poverty
- Ensure at least 80% of funding goes directly to people in need.
- Greatly help participants in our programs
- Start first with our maximum impact pilot because it is an exceptional opportunity for a startup nonprofit to prove extremely impactful with a very small budget.
- Conduct rigorous RTC studies and, hopefully, prove consistently powerful effects.
- Spend all extra money on expanding our fundraising capacity and educating people about why guaranteed income is so powerful.
- Repeat 1-5 while expanding our scope to every geographic region, every population in need, and therefore funder.
- Direct funding that does not have a geographic or population-specific area to GiveWell's top charities & GiveDirectly.
- Establish guaranteed income as the baseline default intervention for helping people.
- (respectfully) encourage all other nonprofits to prove through rigorous RTC studies that their interventions are more impactful than guaranteed income.
- (respectfully) point out to funders that many (perhaps all) of the other nonprofit interventions they are funding (at least those in the “helping people” department) are either unproven or being proven less impactful than guaranteed income by rigorous studies.
Our Theory of Change + The Seba Technology Disruption Framework
Potential Weak Points in our Theory of Change
The biggest weak point in our plan could be our ability to fundraise for guaranteed income at scale cost-effectively. We'll have to convince lots of donors and funders of all kinds of the accurate but (potentially) counter-intuitive notion that directly giving cash to people in need is an extremely effective way to help them. Although we have overwhelming evidence, age-old misconceptions about the lazy or undeserving poor may make our expansion plans underperform. Our ability to make a truly transformational impact is reliant upon the capability of our fundraising machine. Having experienced fundraisers and narrative-change professionals on our team would greatly mitigate this weak point.
ITN Framework: Poverty+Guaranteed Income
Importance (Scale of the Problem in America alone)
Poverty is the UN’s #1 SDG for a reason. Arguably, poverty is the biggest problem in the world and the root of almost every other (non-existential risk) problem humanity faces.
Child poverty alone cost $1.03 Trillion in 2015, a massive sum that is equivalent to approximately 1/3 of the entire federal budget. During the coronavirus, the expanded child tax credit (a small UBI for families with kids) drove child poverty down by over 40%, and then it rebounded when the program ended.
Other forms of poverty across various groups also have negative economic impacts. And above the arbitrary federal poverty line, there are far more negative opportunity cost impacts that dwarf the 1 Trillion per year figure. While the official poverty rate is 11.6%, 7 in 10 Americans struggle with at least one aspect of financial stability, and 40% are struggling to pay for basic needs.
Additionally, a surprisingly large amount of Americans have very unpredictable changes in fortune year after year. More than half of the population experiences a median swing of 7.5 points (out of a 100-point score of financial health). Rob Levy, vice president of research and measurement at the Financial Health Network said, "We didn't think people's lives would change that much," but it turns out that, “most Americans' financial prospects are "like a game of 'Chutes and Ladders,' which has implications for stress.”
The large-scale interventions we’re currently using (the welfare system and a fractured assortment of low-impact nonprofits) are not effectively reducing poverty. We need to leverage the most effective intervention against poverty that is also highly impactful at any scale.
Essentially, guaranteed income can solve poverty. Research by Karl Widerquist of Georgetown University shows that it would cost only $539 Billion, less than 3 percent of the U.S. GDP, to permanently end poverty with Universal Basic Income. Widerquist says the $539 billion per year is 2.95 percent of America’s GDP is about one-sixth of the cost of commonly circulated estimates, and that this amount is less than 25 percent of current entitlement programs.
Widerquist’s research used U.S. Census Bureau data for 2015 to examine an estimated poverty-level UBI of $12,000 per adult and $6,000 per child. It also found that some 43.1 million people (including 14.5 million children) would benefit from this increased income, reducing the poverty rate from 13.5 percent of the population to zero.
That said, driving policy change is difficult, so guaranteed income likely won’t happen until the political will exists among special interests and major political donors. Sadly, the preferences of the average American appear to have only a minuscule, near-zero, statistically non-significant impact upon public policy.
Without relying on finicky government policy to proliferate guaranteed income, we believe that our platform can empower the nonprofit industry to fund guaranteed income at scale, ending the homelessness crisis and taking a decent % chunk out of poverty. Over time, should the government fail to pass guaranteed income policy, our platform could be able to reach the total scale of poverty in the U.S.
Should the government succeed, we'd still be able to do a great deal of good by making the nonprofit industry spend its aid money on (extra) guaranteed income.
There are a lot of organizations and people working to reduce poverty, and trillions are spent each year fighting it. Of course, our focus is on the neglected intervention with the potential to solve it.
There are a few guaranteed income organizations already dedicated to creating the policy change that could establish a Universal Basic Income. It seems that almost all organizations in the guaranteed income movement are focused exclusively on policy change, one niche community, or one niche population in need. That’s not to say policy change funding isn’t neglected, just that there are already many organizations and people working on that.
There are no other organizations, however, with the vision to build what we want to build. All nonprofit guaranteed income providers in the developed world seem to be limited by geographic region, the population they are focused on helping, or some combination of the two.
Unleashing the full existing funding potential for guaranteed income is a completely neglected yet important part of the solution to poverty.
The implications of guaranteed income for humanity and the Effective Altruism Movement.
We believe the growth of our guaranteed income platform in the medium term could have very large positive impacts against poverty and for the EA movement such as
- Over time, we will establish guaranteed income (direct cash) as the baseline for all interventions looking to help people. Essentially it will force all charities and fundraisers to prove that their interventions are more cost-effective than cash, doing so would cause a culture shift in the nonprofit industry at large toward more impact-focused analysis of interventions and funding based on rigorous studies.
- Reducing poverty and income inequality at scale would mitigate the mental health crisis and the housing crisis, reduce crime, and improve education outcomes for future generations. It would change power dynamics, enabling victims to escape abusive relationships and workers to escape exploitative jobs.
- Reducing poverty at a large scale would free millions of poor people to do things other than work to survive day-to-day (such as seek out impactful careers). At the same time, our plan, growth, and hopeful success should be extremely newsworthy. Educating our participants, donors, and the general public while we scale about effective altruism could lead to significant movement expansion.
- Greatly reducing problems in first-world countries would enable far more funding to be allocated to fighting poverty overseas. Leveraging funders we find that don’t have geographic limits, we access enough money to fully fund all Givewell’s interventions, and enable GiveDirectly to end desperate poverty globally.
- Expanding guaranteed income globally could materially improve international relations and thus reduce some existential risks.
- If guaranteed income was spread out internationally, it could reduce tensions, and mitigate the risk of global nuclear annihilation.
- The same could be true for reducing global catastrophic biological risks. Not only could reducing international tensions reduce the incentives for governments to develop bio-weapons, but guaranteed income would also greatly increase the resilience of society to future pandemics. Guaranteed income as a safety net would be extremely resilient against pandemics, so it could be very important for enabling us to make it through a worse future pandemic.
- As guaranteed income generally improves quality of life, we could have a lot more time, energy, and capability to focus on reducing all of the existential risks we are aware of. This is a likely result of significant EA Movement expansion.
The likelihood of this? Given the research, I've outlined, and the input that I've received from a bunch of unbiased professionals in the nonprofit industry, I think the likelihood is closer to 40% if we do not get any support from EA, and over 70% if we do get tons of support.
Longtermist implications of guaranteed income
This is the most speculative part of the post, so it may be best to just skip it.
Guaranteed income at scale, universal basic income (UBI), could be the logical foundation (see what we did there), for human civilization. If money remains a thing and capitalism continues to look like the most efficient system for getting people the things they want, then guaranteed income should be one of the fundamental positives in the long-term future of humanity. Think of all the impacts outlined in the above sections, but repeated for tens, or thousands of generations to come. Global (or civilizational, should we escape Earth) guaranteed income could be an important part of permanently ending resource-based conflict and enabling humans to become more aligned.
In the same vein, establishing UBI as a human right could make it much harder for a (non-AI) actor to enslave humanity forever, reducing several S-risks. Given the dispersed power that UBI distributed evenly across society, people would be much more prepared for and highly resilient to malicious actors looking to entrench power.
If the disruption enabled by guaranteed income is inevitable. Our impact could therefore be accelerating the disruption while ensuring that guaranteed income isn’t used for negative purposes. Dictatorial governments could, for example, tie social credit scores to the amount of a basic income. In the wrong hands, guaranteed income could be an extremely powerful tool to entrench social stratification. We can help to ensure that the growth of basic income is safe, effective, and dignified.
We also have an idea for a Basic Income Endowment, which would enable funders to generate guaranteed income in perpetuity. It would be the most impactful way of establishing a meaningful permanent legacy for any wealthy person. Over time, it could become very large, perhaps large enough to provide guaranteed income to all of society.
This potential future excites me, and I’d like to see what other longtermist effective altruists think about it.
Our Startup Plan: Maximum Impact in Developed Nations.
We've found a niche opportunity to build the most cost-effective anti-poverty program in America (an ~$50M opportunity before cost-effectiveness starts to decrease). We believe that funding capacity-building for our Maximum Impact Pilot could outperform even GiveWell's top charities in cost-effectiveness.
Experimental results support the cost-effectiveness of guaranteed income against homelessness.
Importantly, the fact that we’ve identified this problem in the U.S. means that EA doesn't need to fund it (just help us get established). There's tons of local funding available which we intend to pursue, so EA capacity-building funding would result in 10X the impact of spending on the intervention. Furthermore, all of the local funders are not willing or able to spend outside of their geographic limits so the funding we receive for this intervention wouldn't reduce funding for global EA interventions.
A small but growing collection of experiments have been showing that guaranteed income, giving large structured sums of unconditional cash to people suffering from homelessness, can help most unhoused people (chronic sufferers and recently unhoused people) escape the streets and regain economic security. Compared to the other effective intervention against homelessness, housing first (which is theoretically 100% effective), guaranteed income experiments have demonstrated >66% housing success rates at a tiny fraction of the cost per beneficiary (see California’s $694M program that costs $50,000 to $277,000 per person). We believe a large-scale well-structured guaranteed income program could make homelessness solvable.
Experiment results and implications
Based on the results of these experiments, we’ve concluded the following. There is a:
- ~50% likelihood that guaranteed income can help >66% of homeless people regain stability. The biggest study so far had a 75% success rate, but the total size of that study wasn’t massive so we’re going with a conservative estimate.
- >90% likelihood that it can help >50% of homeless people regain stability. The least successful experiment to date had a 66% success rate.
- >99% likelihood that a broad guaranteed income would prevent people from falling into homelessness in the first place, ending homelessness permanently over time.
- Logarithmic rate of success cost curve, which depends on the amount and duration of the guaranteed income. Not enough data is available yet to determine the optimal amount or duration. ($1,000/month is likely more effective than $500/month over the same duration). More work is needed and ongoing to find the optimal guaranteed income schedule.
- Known fact: America currently spends significantly more than $110,341 per homeless person in America each year.
Contrary to some popularly held assumptions, our research points strongly towards almost all homeless people spending guaranteed income transfers on meeting their basic needs and quickly working to regain economic security when given the opportunity, (USC Literature Review and Our Research).
Furthermore, there is extremely low downside potential. There are almost no recorded cases where guaranteed income has caused harm (to homeless people or in general), despite there being a very large amount of data available for cash transfers in general. Even in cases where one could anticipate that guaranteed income would enable participants to do bad behavior, they have not done so.
Almost all spend resources on basic needs, followed by rent, housing supplies, and other things that help them get ahead. It may even be an incredibly counter-intuitive yet cost-effective way to reduce drug usage. Of course, a lot more research needs to be done before rolling out guaranteed income as a drug abuse reduction intervention.
It's important to note that the studies are based on surveys given to the participants. Depending on how the surveys are presented, and the biases of the participants, they could feel obliged to not tell researchers about their drug usage. This is a commonly known effect in psychology and sociology.
That said, it's important to remember that the general public is also full of drug users (have you used alcohol or cannabis in the last year?). We judge ourselves and our friends at a lower standard than we do people forced to live on the streets. Furthermore, there's a big difference between drug use and drug addiction, and drug use, in moderation, is not inherently negative. Finally, it's nearly impossible to accurately determine who does how many drugs without extremely invasive, undignified, & expensive oversight that would result in few participants wanting to engage in the program. The best way to measure whether participants are spending their money 'right', is by recording their quality of life before, during, and after guaranteed income.
We’ve spoken with local advocates from groups such as FeedPhoenix and AZ Hugs for the Houseless, representatives from groups studying guaranteed income, a wide range of unbiased nonprofit professionals (Stuart Turgel, Carol Farabee, Rodney Houston, etc...), and organizations doing research into this intervention. During our conversations, people with experience fighting homelessness have agreed that guaranteed income at scale could help around 2/3 of homeless people regain stability, diminishing the crisis to manageable levels.
Ending homelessness at a fraction of the cost
We believe that if a small fraction of the current amount spent on homelessness is allocated to a well-run guaranteed income program, it would likely help the majority of homeless people get off the streets. Guaranteed income makes homelessness solvable and could turn the homelessness industrial complex from an infinite dehumanizing money sink to an opportunity to invest in people while they re-enter society.
Keeping people on the streets costs between 45-180K (accounting for inflation) because they, “randomly ricochet through very expensive services” such as hospitals, addiction treatment services, police arrests, jail time, and court time. The federal government is spending $51 Billion a year and California alone is spending over $10 Billion. $61B / 552,830 homeless people = $110,341.
We’re spending significantly more than $110,341 per homeless person in America per year. We could afford to provide a guaranteed income of $1,000 a month to all of them for around 10% of our current spending. More accurately, we can’t afford not to.
Guaranteed income, the intervention with the potential to solve the problem, is neglected. Not because people know about it and have decided not to fund it, but because they have not heard about it at all. Currently, a few organizations are experimenting with guaranteed income for fighting homelessness, but only one (startup) nonprofit is focusing on Arizona or using a platform strategy to reach the full scale of the problem. That’s us.
We believe that even a very small investment to help establish our nonprofit (a few million dollars), would be sufficient to build a team that can fund a program that meets the scale of the homelessness crisis at scale within a few years.
Arizona is by far the deadliest place in the U.S. for homeless people.
Homeless people in Arizona are among the most vulnerable people in the world. Primarily because of Arizona’s summer heat, the preventable death rate among Arizona’s homeless population is 7% annually. I couldn’t find any other population with a higher preventable death rate outside of war or genocide after searching through the Human Mortality Database. I’m sure the population of stage 4 cancer patients, for example, also has a high death rate, but those are not easily preventable deaths.
Arizona’s (Maricopa County’s at least) Homeless Death Rate from 2020:
The overall homeless mortality rate in Arizona is ~8%. After subtracting the natural 1% U.S. mortality rate, we find the mortality rate increase due to homelessness in AZ is 7%. This is, of course, an educated estimation based on the best numbers available.
Overwhelming demographic data as well as medical analysis make it evident that living on the streets directly accounts for most, though not all, of the massive mortality rate increase. There is a causal relationship between living on the streets and high death rates, especially in Arizona due to the high summer heat. California, which has a much more temperate climate, has ~5,000 homeless deaths within a population of 173,800, a 2.8% mortality rate.
What does this mean for 2022?
2022 Point-In-Time count of homeless individuals data:
There are at least 13,553 homeless people in Arizona.
2022 Estimated AZ Homeless Deaths = 13,553 * 8% = 1,084
2022 Estimated Preventable Homeless Deaths in AZ = 13,553 * 7% = 948.5
The true outcome could differ by a few hundred up or down, but this number will also increase because the homelessness crisis is getting consistently worse year over year.
QALYs (Quality-Adjusted Life Years) expected from Foundation's Maximum Impact Pilot
We’ve designed our pilot to become the most impactful anti-poverty program in America. To do so, we’ll be helping homeless individuals in Arizona with guaranteed income.
We will be giving ~24 homeless participants in our pilot $6,000 over 6 months ($1,000 per month) while helping each of them connect to the local services they need to regain stability. We chose 24 because it resulted in the smallest budget where we could still send 80% of the budget directly to people in desperate poverty. This assistance will greatly help nearly every participant over the first year, and our goal is to help at least 50% of the participants achieve permanent housing stability. We consider this to be a conservative goal because every homelessness cash-assistance pilot to date has helped at least 66% of participants achieve stability.
living in homelessness reduces one's QALY (quality-adjusted life year) from 1.0 to 0.434, so helping someone stay out of homelessness would save 1 - 0.434 = 0.566 QALYs per year they would otherwise be homeless. (Academic Source: https://pubmed.ncbi.nlm.nih.gov/34629422/)
According to William MacAskill in Doing Good Better, Health economists estimate that the benefit of “preventing a preventable death” is equivalent to saving 36.5 QALYs.
With approximately 8% of the homeless population dying every year (7% higher than the normal 1% rate), each Participant we help will have their death likelihood reduced to the mean (by 7%) during the pilot year. Then there will be lasting reductions in mortality rates for participants that permanently exit homelessness. (7% per year housed vs. the counterfactual)
So, conservatively, how many QALYs should we save versus the counterfactual?
1. Life quality impacts from helping 24 people have a far better life (improving from 0.434 to 0.9) during the Pilot year: 24 * (0.9-0.434) = 11.184
2. Life quality impacts from helping 50% of participants permanently escape homelessness (counting only three years ahead): 24 * 50% * 0.566 * 3 = 20.376
3. Prevention of death impacts by greatly reducing the likelihood of death for 24 people during the pilot year: 24 * 7% * 36.5 = 61.32
4. Prevention of death impacts from helping 50% of participants avoid Arizona’s high likelihood of death by escaping homelessness (counting only three years ahead): 24 * 50% * 7% * 36.5 * 3 = 91.98
Conservative Estimate: Total QALYs saved by Foundation's Maximum Impact Pilot: 11.184 + 20.376 + 61.32 + 91.98 = 184.86 QALYs
Pilot Budget: $182,000
$182,000 / 184.86 = $984.5 per QALY
GiveWell estimates its top charities can save a life for every ~$5,000 donated. Converted into estimated QALYs, $5,000 divided by 36.5 is $136.9/QALY. $984.5 divided by $136.9 means our pilot should be about 7X less impactful per dollar. Therefore directly funding this intervention would not be a good use of EA funds.
However, capacity-building funding for us to build a fundraising team and get self-sufficient could empower us to raise $10 to $20 with every incremental dollar spent by an EA Funder. Given that 10-20X multiplier, helping us build fundraising capacity could legitimately outperform GiveWell’s top charities.
Our Conservative Assumptions:
- We have factored zero QALY improvements for friends, families, or anyone else that knows participants and would personally benefit from them not being homeless.
- Unquestionably, the quality of life for the general public goes up as homelessness goes down. We have factored zero QALYs for this.
- Quality of life while homeless in Arizona is substantially worse than in any other place in America (if not the world) because Arizona summers are hellishly hot. It’s easier to warm up when cold than cool down in high heat. A more accurate QALY for homeless people in Arizona could be <0.35. We went with the conservatively high QALY for homeless people on average of .434 (aka saving .566 QALYs vs. the counterfactual) because it’s an official number from an academic study.
- In every experiment to date, guaranteed income has helped at least 66% of participants regain economic stability. We are using an unreasonably low estimate of 50% for two reasons.
- There have only been three experiments completed as of November 2022; it’s not the biggest data set.
- Because of the logarithmic rate of the success cost curve, we could have miscalculated the optimal duration and/or the monthly amount needed. If we set the bar too high for ourselves and miss slightly, our pilot could look like a failure even if it’s extremely impactful. If only a third or quarter of participants regained housing stability, our pilot would still be an order of magnitude more cost-effective compared to other homelessness interventions
Final Notes on QALYs
This pilot project is not just a small but powerful intervention, it is an essential prerequisite for us to bootstrap organization growth and test our platform at scale. Our platform's beta test will take this same intervention (with a few tweaks), and scale it up to the size of homelessness in Arizona ($50-100M). There were 13,553 homeless people in Arizona as of Jan 24, 2022. It’s higher now and growing. Conservatively, here’s how many QALYs are on the table for $984.5 per QALY (if we only succeed in helping solve 50% of homelessness in AZ, and don’t account for the efficiency benefits from scaling).
184.6/24= 7.7025 QALYs per person
As mentioned above, there are 13,553 people (and growing) homeless in Arizona as of 2022.
13,553 * 7.7025= 104,391.98 Quality-Adjusted Life Years
Since the crisis is growing quickly, and our estimates are conservative, the counterfactual damage from homelessness in AZ could be much worse. Our pilot will be crucial for enabling our expansion, so some fraction of the QALYs from enabling future impacts have to be factored in... right?
Even a 5% allocation would make our pilot potentially more impactful per dollar than any GiveWell top charity (before considering the 10-20X fundraising multiplier).
$182,000 / (104,391.98 * 5%) = $34.87/QALY
I could easily go a lot further by considering expansion beyond Arizona and just homeless people. We could greatly reduce poverty and homelessness overall in the U.S. and among all developed nations by realizing our vision.
Realizing our vision could meaningfully improve the prosperity of human civilization, generating nearly incalculable impacts for early growth accelerators.
My recommendation? Effective Altruists should strongly consider helping us grow.
How can the EA community increase our chances of success and help us make this massive impact happen?
I do not think that large quantities of general EA funding should be diverted toward guaranteed income in developed countries. Money sent to developing nations simply has a greater impact at scale.
We only need help from EA groups to get off the ground because we plan to bring in funding from sources that would otherwise generate much less impact than through guaranteed income. Almost all of our funding sources long term will not be funding sources that could otherwise spend money on Effective Altruism cause areas, and therefore we will not be diverting funds that could be spent better on global poverty or other EA cause areas.
Here are some potentially powerful actions that I believe EA organizations and effective altruists should strongly consider.
- Charity Entrepreneurship (the EA nonprofit) should strongly consider us for their incubation program to help us get off the ground. I applied to be a part of their 2023 pool, and I think it could be incredibly helpful to help us further improve our strategy, establish a strong team, and help fund our pilot.
- I am going to apply for the EAGX conference in Berkley this December and the EAGlobal conference early next year. Helping me go to those conferences and make connections could be highly impactful.
- Fundraising professionals that want to have a large impact should consider reaching out to me. We anticipate raising money for guaranteed income could likely be an extremely impactful career path. Our work in Arizona should save QALYs for under $1k each, and professional fundraisers can each bring in over $500K per year.
- EA Funders such as EA Grantmakers, and Effective Altruists that have decided to earn to give should strongly consider helping us get off the ground by providing up a few million dollars one time. I do not think that millions of dollars of EA funds are necessary (though millions would certainly help accelerate our progress), as we could become baseline self-sufficient with $300K-$500K to build a strong fundraising team and get to work. At this early stage, helping us grow could be more impactful than any other intervention period. As we grow, we will divert non-limited funds to GiveDirectly or other globally impactful nonprofits.
- Funders that are limited to spending in the U.S. should be very excited. If there are EA-oriented funders (philanthropists or organizations) dedicated to maximizing impact specifically in the U.S., our work in Arizona is likely the best opportunity for impact per dollar in the U.S.
I’m looking for funders and passionate co-founders with expertise in fundraising. If you are or could help with any of those things, please reach out!
Oh, and I’m also looking for strong arguments against everything I’ve laid out in this long article. This post relies somewhat on imperfect data, so it's intended to be in the correct ballpark and directionally accurate. I've researched disruptive systems, guaranteed income, homelessness interventions, and talked to over two dozen unrelated experts, none of whom have strongly argued that our pilot or grand plan would not work. I’m happy to discuss all points of contention in the comments and over Zoom or on Twitter spaces should there be enough interest. The goal, as always, is to make a large impact.