I disagree that EA only looks for ways to do good only if you have money. I think it also gives you ways to do good if you have time too. If you don't have any of the two I think your mission/ prioritization should be trying to move from that situation and earn time (or money) somehow.
When it comes to the context of your cause, I don't think I have valuable information to give you, because it's a complex systemic issue that many countries/ areas face, maybe you can research about success cases.
Ah, good point, thank you! Something I wrote might give the impression that it's only for wealthy people; I'm sorry, that wasn't my intention. I meant that, in general, it's aimed at people with financial resources.
And yes, I've been looking for success stories, asking around in groups, and I haven't met anyone from a poor country (and who is a poor person) who has managed to progress and help more effectively over time. If you know someone who has and would like to quote me, that would be a great help!
1
Introduction
We work with vulnerable youth in a community association in Brazil. We want these young people to become allies in improving the neighborhood: reducing violence and helping others.
However, we've noticed a problem. Many children participate when they're young, but during adolescence they stop coming and start getting involved in higher-risk environments, like drug use or crime. We see this in those who have already approached us; we don't know what happens to those who never came, so it may be a real pattern or just what we're able to observe.
We don't know exactly why this happens. Perhaps they're looking for quick rewards. Perhaps living with so much uncertainty makes it very difficult to think long-term, and organized crime can offer physical protection, territorial status, immediate respect, and a tribal identity. We don't observe that crime forces young people into it with violence; they seem to accept the risks and rewards.
We still don't know which of these explanations is more accurate. It's possible that several are true at the same time.
Whatever the explanation, the challenge remains the same: how do we make coming back the following week more appealing than abandoning the project?
Nor do we want to answer that question with intuition. We want to build a method that allows us to compare different alternatives and discard those that don't work.
If we discover an inexpensive way to get a young person to stay in the program and build a different future, the benefit to him, his family, and the neighborhood can be far greater than the cost of trying.
That's why this work doesn't begin by asking what the best intervention is. It begins by asking something more basic:Is it worth looking for?
We don't know the best way to cooperate. That's why, instead of starting by advocating for a solution, we'll build a decision tree.
Each step will answer a simple question. Some ideas will be discarded because they are too expensive, too risky, or because the evidence doesn't support them. Others will be strengthened when they survive those tests.
Imagine a tree where, at each fork, we must answer a question. Each answer eliminates some paths and brings us closer to a more promising intervention. We're not looking for the perfect intervention. We're looking for the one that best withstands the questions we ask it.
The goal isn't to prove we're right from the start. It's to make as few mistakes as possible before investing time and resources.
At the end of this process, we hope to retain only the alternatives that offer the best balance between cost, risk, and benefit. In a sense, we want to create a practical guide so that someone with limited resources can decide when to help, whom to help, and how to do so with the greatest possible impact, at least in our neighborhood; whether it works elsewhere is something we will determine with data.
Some questions will be philosophical ("is it worth helping?"), others biological ("what strategies favor survival?"), others mathematical ("when does the investment cease to be profitable?") and others practical ("which intervention is easiest to implement?"). They all form part of the same problem seen from different perspectives.
So, how do we decide what to do?
Branches of the decision tree
Passos, being poor:
Steps: Each step will be explained by:
0. Helping while being poor
a. Why? Is it intelligent for a poor person to help someone worse off?
The problem
You're walking along a shallow lake. In the middle, a child is drowning. Saving him will cost you a pair of wet pants and two minutes of your day. Will you help him?
In 1972, Peter Singer [1] published an argument that disturbed us because it was difficult to refute. He proposed a simple scene: you walk past a small child drowning in a shallow puddle. Saving him requires you to get wet and be late for work; it seems easy.
However, life isn't always that simple. Can we generate such a large impact at such a low cost? At the price of a mere expense? Intuition tells us we should help. But we still don't know if that help generates the greatest possible benefit. To answer that, we need to compare scenarios.
For example: In emergency rescue situations, if I get involved without equipment or resources, I could hinder those who are prepared and put myself at risk. Therefore, the idea that "it's always good to help" doesn't apply; we would have to carefully assess the scenario and the potential consequences. We might be competing with schools; it could be interesting to know if they continue studying without failing.
And I also have to consider whether saving him will harm me too much, especially since I've already been late three times this month, and I find it hard to believe the fourth time was because I was saving a child. It's really difficult to know if it's worth taking the risk without calculating the scenarios, even though I want to avoid doing the calculations.
In other words, it seems they could only know by analyzing the figures and the scenario before trying to calculate the aid and the cost-benefit. I'll try to see what survival strategies exist and see if I can escape the math. Perhaps with a good survival strategy or a simulation that demonstrates that helping is worthwhile, I can know if it's worth helping even if I only have two hours.
I want to see, first and foremost, if helping a homeless person can be a good strategy for a poor person. If it isn't something entirely impossible: to help and be good, and increase the satisfaction and decrease the suffering of someone poorer than myself.
Survival
Is it always wrong to help someone vulnerable or in a worse situation, or is it possible that helping can be a good strategy if we consider extreme cases, such as survival in a tribe?
There's a common idea that nature selects for the best skills and behaviors, but what about the difference between a poor person and a destitute person? For a time, thinking about behaviors selected for survival was very appealing in terms of group selection. It's as if a poor group had a greater chance of survival if it helped other, even poorer groups.
What if modifying the environment leads to more risks?
Animals don't just adapt to their environment; they modify it to facilitate their survival. A beaver doesn't build a dam out of altruism toward other beavers, but because swimming in deep water is the only way to escape wolves. By building the dam, it tames the environment. If we invest in these at-risk youth with limited resources, we can't expect them to return the favor directly. You are building the dam. You are reducing the toxicity, violence, and unpredictability of the Samambaia neighborhood. The reward is living in an environment that allows you to sleep peacefully.
And there's another problem: in nature, the "wolves" hunt elsewhere, but organized crime isn't always so passive. If they perceive that the organization is taking away their recruits, they might actively sabotage the project. We haven't seen any signs of that yet, but inevitably, we would have to cancel the project and find a safer location.
But how do we calculate if we can actually build this dam?
What if sharing has different value for each person?
In biology, altruism makes more mathematical sense when resources have different values for the giver and the receiver. Vampire bats that manage to feed one night regurgitate blood for those that didn't. For the donor, this represents perhaps 5% of their energy (a tolerable safety margin). For the receiver, it means 100% of their survival that night. In our case, the free time of some is very valuable to young people because they possess what they themselves lack:stability and long-term vision.The exchange isn't "I'll give you 10 reais, you give me back 10 reais tomorrow." It's more like "I'll give you stability, and tomorrow you won't become a predator in my neighborhood." But how do you know how long the effect actually lasts, or if that 5% is just wasted?
Let's consider two types of groups throughout human history. The first is concerned only with the poor, ignoring the poorest and investing resources only in those who are already less poor. The second also seeks to rescue its most needy members. Which group is more likely to survive?
In evolutionary economics, there are simulations that demonstrate precisely that.
Imagine 100 individuals.
Each person receives food at random.
Without sharing:
Some die from bad luck.
Exchange:
Average food consumption remains virtually unchanged.
But the death rate is plummeting.
The variance decreases.
This result appears repeatedly.
This mechanism is practically an insurance policy.
Apparently, the group that helps is doing so, but it's difficult to know for sure, because some who don't help could also survive within the group if they disguise their selfishness well.
This individual who "doesn't help" benefits from the group without paying anything in return. In this case, the group would be supporting a parasite; all those who help the needy are supporting someone who doesn't help.
And, on top of everything else, we would be supporting a parasite that spends less and earns more. And this parasite, with more resources than us, would have a greater chance of surviving and perpetuating itself. This suggests a major problem: if anyone can benefit from cooperation without contributing, how can cooperation be sustained over time?
Let's consider two people from the same community throughout history. One only cooperates with those who are already strong enough to reciprocate immediately. The other also chooses to help those going through difficulties, even without expecting anything in return. In a population where the same people meet repeatedly (not in a single encounter, but in a neighborhood), which of the two strategies becomes more prevalent over time?
In other words, on an individual level, why should a specific person (me, you, a volunteer, a donor) invest time and money in a young person in the group if the return isn't guaranteed for us and may never directly benefit the person who helped? I see three ways to justify one poor person helping another:
So far we've learned that helping isn't always a good idea, but it's not always a bad one either. It all depends on the context and the opportunity cost. The next question is no longer whether we should help, but when helping stops being a good investment.
b. How? Evaluate using method or intuition?
The problem
You're out at sea, at night, with no stars or coastal lights. You have no compass, only what your body feels: the wind on your cheek, the rocking of the hull, the certainty that land is "out there." For centuries, this is how people navigated, and for centuries, ships were lost precisely because they relied on this method. The body that feels the wind was trained by generations of people who also relied solely on the wind; it wasn't adapted to a world with as many variables as today's. But someone observed this and demonstrated the power of doing the opposite.
Once upon a time, there was a man convinced he had a Great Idea. In fact, the more he pondered it, the more he realized it wasn't just a great idea, but the most wonderful idea of all time. The Great Idea would unlock the mysteries of the universe, supplant the authority of the corrupt and flawed system, grant almost magical powers to whoever adopted it, feed the hungry, heal the sick, make the world a better place, and so on and so forth.
The man was Francis Bacon, the Great Idea was the scientific method, and he was the only eccentric in all of history who claimed such a level of benefit to humanity and proved to be absolutely right. (Scene inspired by Eliezer Yudkowsky, "Resist the Spiral of Happy Death", 2007. Yudkowsky himself acknowledges, in a footnote, that Bacon did not invent science alone; perhaps he was only the first to communicate its power).
In the 17th century, Francis Bacon warned that a mind left to its own devices follows an erratic course. Four centuries later, and by a different path, the community that inherited this same distrust arrived at a modern version of the warning: we trust too much in what we see with our own eyes and too little in what others have tested and replicated. The eye, too, can deceive us. The instrument, calibrated by those who have erred before us, is less deceptive.
The body, instinct, good intentions: they can be right, they're not always wrong, but they could complement each other. As if, on their own, they didn't know when they were wrong. That's what our protocol is for: not to replace the guide, but to offer them something beyond their own body that they can verify.
It's important to clarify that the captain at sea doesn't construct the compass during a storm. The cartographers (the project designers) have already done the complex calculations on land. The volunteer is only given the compass and a simple ruler. It tends to be very helpful to explicitly state:"The complex method is for those who design the intervention. For the volunteer in the field, the method must be so simple that applying it consumes less than 5% of their energy (the bat's blood), transforming theory into quick golden rules."
Good intentions without a plan come at a price. It would be like caulking a ship's hull with paper: water will leak in anyway, and you still paid for the caulking. Systematic reviews show that social-emotional learning programs without sufficient evidence not only waste resources but also produce a slightly negative effect, worse than nothing. The volunteer who improvises doesn't achieve the same results as the one who follows a plan. And the plan is just as important as being present.
Survival
Imagine two tribes. In the first, the healer knows how to heal because "he learned it from his grandfather." When the grandfather dies, the knowledge is lost. In the second, the healer left the recipe written down: quantity, sequence, contraindications. When he dies, the next person reads it and continues the tradition.
Clarify the nature of that recipe. The healer didn't leave a treatise on molecular biology; he left direct instructions:"Boil these two leaves for 10 minutes"The method transfers knowledge, but obscures the complexity. This justifies why young people are taught step by step, giving them small, tangible victories (status, cost-cutting aid) that act as the short-term dopamine hit they previously sought in the criminal environment.
Humanity took a qualitative leap by abandoning individual trial-and-error learning and beginning to transmit codified knowledge. This method guarantees the solution's longevity beyond the individual. The idea behind the protocol is to move beyond a charismatic volunteer and complement it with rational methods.
Facts in favor of having a method:
Cognitive psychology refers to this phenomenon as ecological rationality (Gigerenzer & Todd, 1999; Todd & Gigerenzer, 2012). Heuristics, mental shortcuts, intuitions, work well when the environment in which they are calibrated remains stable and fail when it changes. A volunteer's intuition may be deficient or have been formed in a specific set of situations and simply not encompass the full range of situations they will encounter in practice.
The field of effective altruism arose precisely from this observation: good intentions without a method are wasted resources with a low probability of return. Therefore, those dedicated to this field focus on finding more effective ways to generate a greater positive impact on the world.
More specifically, Heller et al. (2017) estimate a cost-benefit ratio of between 5 and 30 times for the type of intervention we analyze at the end of this document. This value corresponds to the multiplier of the difference between helping with a method and helping without it.
Facts against having a method:
Methods have a design cost. They require thinking before acting, testing before scaling, and correcting when data contradicts intuition. They are slower at first. Without good intuition about when to act and when to program, we could get stuck in an endless cycle of thinking.
Furthermore, seeking a structure can generate a false sense of security: a poorly followed protocol, when rigorously applied, can be worse than a bad protocol because it consumes more resources and more time.
Furthermore, looking for a structure can generate a false sense of security: a poorly followed protocol, when applied rigorously, can be worse than a bad protocol because it consumes more resources and more time.
Choosing intervention based on our good will is dangerous. Well-intentioned interventions can generate iatrogenesis (undamage greater than the absence of intervention). For example, exhaustive reviews of Scared Straight-type deterrence programs (Petrosino et al., 2013) consistently demonstrated that young people enrolled in these programs had higher rates of criminal recidivism than those from the control group who did not receive any attention. Choosing the wrong route on the ground wastes the boat's gold, which accelerates its destruction. But it can be the most effective when we are in practice.But it can be most effective in practice.
Mental image:
Two volunteers, same panel, same amount of time. In the first scenario, the volunteer arrives tired from work and has to decide, using their intuition at that moment, what's best for the group of young people. In the second, the volunteer arrives with a single mission pre-approved by the organization: "Today's goal is for you to assemble this circuit in 30 minutes and take the result with you." In the first case, the volunteer spends 80% of their energy deciding what to do. In the second, they spend only 5% on the method and 95% on genuinely connecting with the young person.
To help external readers (or future designers within the association) understand this asymmetry, it might be useful to include a summary table in that same section outlining the responsibilities:
Actor
Decision environment
Tool used
Expected cognitive cost
Association (Designers)
Cold, away from the field.
The complete compendium (Philosophy, Mathematics, Decision Tree).
Stop. The strain is assumed to protect the volunteer.
Volunteers and Young Monitors
In the heat of the moment, direct interaction.
Field protocol (Micro-steps, immediate rewards, status).
Low (≤ 5% of your daily energy).
What remains open to interpretation?
I want to know if the compass is pointing north; can I use the sun as a reference? In the case of effective altruism, what is our sun? How do we measure if the compass is working correctly?That's what we'll see next.
c. How much? What is the limit?
Here we must answer two different questions to avoid confusing the objective. The first is macro: What is the minimum probability of success that makes this project financially profitable for society? The second is micro: What is the maximum cost that a poor volunteer can afford to invest without ruining their own life?
I. Society: The expected value of the project (Savings for society)"
To calculate whether the project is a financially prudent investment, we don't need to "save" every young person permanently; we only need to prevent a fraction of the time that, statistically, they would spend inside the penal system or in criminal activity.
If the operating cost of the project for 10 young people isR$ 21,600 annuallyThe calculation for the association to be self-financing is measured intime deprivation of liberty or criminal activity avoided:
This means it is not necessary to radically transform all 10 young people simultaneously. If the project manages to prevent a cumulative total of9.2 months (0.77 years) of criminal activity or prison timeAmong the 10 participants over the course of a year, community investment has already generated a neutral financial return for the state. Any result exceeding this threshold represents net savings and a pure social benefit.
2. Consideration of real costs (Externalities)
The cost of a young person resorting to crime exceeds the administrative cost of a jail cell. If we conservatively add the opportunity cost (lost GDP), the expenses of the judicial system, and the damages caused by victimization, the figure increases significantly:
Recalculating the break-even point with this comprehensive figure makes the goal even more achievable:
Under this more realistic approach, the project is self-financing if it manages to avoid just a few minor inconveniences.4 months (0.34 years) of accumulated criminal activity among the entire group of participants over a year. This makes the intervention an asymmetrical bet in our favor: even if the initial effectiveness is modest, preventing the equivalent of a quarter of criminal activity for all the young people together recoups the total cost of the investment.
We're looking for the expected cost-benefit argument under uncertainty: given that we don't know exactly what will work, what is the cost of trying something reasonable compared to the cost of not trying anything? And considering this, we can think in terms of graphs.
"The iceberg reveals what is hidden. The graphic below shows when weight sinks the ship."
Savings from other interventions:
In contrast, Heller et al. [3] studied the BAM (Becoming A Man) program in Chicago and estimated a social return of $5 to $30 for every dollar invested in structured support for young people. The asymmetry between the cost of intervention and the cost of future harm is so great that, even under high uncertainty about the effectiveness of our intervention, the expected value of attempting something structured far outweighs that of doing nothing.
II. Participant - The survival rule or Stop-Loss
The volunteer's loss limit (The 5% rule)
The state saving R$76,500 does not cover the volunteer's rent. Therefore, whoever operates the base needs a non-negotiable stop-loss. The assistance cannot cost more than 5% of the bat's blood (its free energy), and that metric changes depending on who the volunteer is.
If the intervention requires the operator, whether adult or young, to sacrifice their basic survival, altruism fails because the helper becomes a victim.
Heller et al. estimate a return on investment of between 5 and 30 times the investment for structured follow-up programs with vulnerable youth: for every dollar invested, between 5 and 30 dollars of social costs are avoided. The asymmetry between the cost of an effective intervention and the cost of not intervening is large enough that, even under conditions of uncertainty, the expected value of attempting something reasonable exceeds the expected value of doing nothing.
Facts against:
Helping comes at a real cost: time, attention, money… and resources may not be used where they should be, resulting in wasted resources. The field of social intervention includes well-intentioned programs with negligible or slightly negative effects. A 2024 systematic review examined 17 social-emotional development programs for adolescents and found none with sufficient evidence to recommend them with confidence, and some produced effects only slightly worse than no intervention. Good intentions do not guarantee good results. Sometimes, they hinder progress. It seems we need to find ways to evaluate and document very well.
Mental image:
Two merchant captains must cross a strait infested with privateers. The first captain invests his gold in reinforcing the ship's hull and buying more cannons. The second captain invests the same amount of gold in hiring the privateer's leader as an expert navigator of the route, offering him a share of the cargo if they reach port safely. Which ship crosses faster and with less wear and tear? If the privateer accepts the responsibility, the second captain has transformed a deadly threat into an allied escort at the same cost as the cannons. If the privateer refuses and decides to attack, the first captain was right. The dilemma of this step is precisely this: to ascertain, at the lowest possible cost, whether the young man accepts the offer of "expert navigator" (status + responsibility) before he decides to scuttle the ship.
So far we know how much it can cost us to be wrong and how much we gain if we're right. But we still don't know how to choose between different interventions to start the crossover. That will be the next question.
So far we know how costly it can be to make a mistake. But we still don't know how to choose between different interventions. That will be the next question.
Bibliographic references
Singer, P. (1972). "Hunger, Plenty and Morality".Philosophy and Public Affairs, 1(3), 229-243.
https://www.lesswrong.com/posts/QsMJQSFj7WfoTMNgW/the-tragedy-of-group-selectionism, "The tragedy of group selectionism", Eliezer Yudkowsky, November 7, 2007.
https://heritage.umich.edu/stories/the-prisoners-dilemma/
Aktipis, C.A. (2004). "Knowing when to withdraw: contingent movement and the evolution of cooperation."Journal of Theoretical Biology, 231(2), 249-260.
Costs of Crime in Brazil: IPEA (Institute for Applied Economic Research). (2018). "Economic Costs of Crime in Brazil." This report demonstrates that Brazil spends approximately 4.3% of its GDP on crime, detailing expenses related to public and private security, judicial costs, and the loss of productive capacity (the minimum wage that is no longer generated).
Costs of incarceration:
SENAPPEN (National Secretariat of Criminal Policies). (2024). "Penitentiary Information Report". Official data that establish the national average cost per inmate between R$ 2,500 and R$ 4,000 per month, varying according to state infrastructure.
Effective socio-emotional interventions and cognitive-behavioral therapy:
Heller, S. B., Shah, A. K., Guryan, J., Ludwig, J., Mullainathan, S., & Pollack, H. A. (2017). "Thinking, Fast and Slow? Some Field Experiments to Reduce Crime and School Dropout in Chicago." Quarterly Journal of Economics, 132(1), 1-54. A seminal study on the Becoming a Man (BAM) program, which uses cognitive-behavioral therapy for vulnerable youth, demonstrating reductions of up to 50% in arrests for violent crimes and social returns of up to 30:1 per dollar invested.
The risk of poorly designed interventions (negative effect):
Petrosino, A., Turpin-Petrosino, C., Hollis-Peel, M. E., & Lavenberg, J. G. (2013). "Well-intentioned programs can have detrimental effects: lessons from the 'Scared Straight' experiments and other youth awareness programs." Crime & Delinquency. A Cochrane Collaboration systematic review showing that "Scared Straight" type programs (which expose young people to the harshness of prison) have failed and, in fact, increased recidivism compared to the control group.