Importance, Neglectedness, Tractability (Effective Altruism Definitions Sequence)

One of the most distinctive features of effective altruism is the use of the importance, tractability, and neglectedness framework for evaluating charities. The ITN framework brings together everything we’ve talked about in the series so far.

Although it might look like a list of criteria, the ITN framework is in fact an equation. Here’s 80,000 Hours:

Although they say “scale” and “solvability,” it’s the same framework.

Here we see the typical effective altruist quantitative mindset. You (usually) can’t put exact numbers on tractability or neglectedness; very few people actually try to solve the equation. But effective altruists still think about doing good as a math problem. In principle, if you were a nigh-omniscient god of social science, you could plug in numbers and solve for morality.

Importance is, well, how important a problem is. In general, for effective altruists, importance has two components:

• How many beings are affected.
• How severely these beings are affected.

Here we see the maximizing, welfarist consequentialism. We define “important” actions as those that create a good state of the world, not as those that (say) involve a particular set of rules. We want the world to be as good as possible; we don’t want to settle for a ‘good enough’ action. If something is good, it is good for someone; we don’t value intact ecosystems or the Mona Lisa all on their own.

“How many beings are affected” is perhaps the simplest term in this whole equation. However, effective altruists do typically care about many more beings than most people do, including people in foreign countries, animals, and digital minds. Therefore, effective altruists are more likely to believe that (say) eating meat is bad because chickens suffer, not just because of your health or the effects on the environment.

In principle, effective altruists care about all the ways it’s possible for something to be good or bad for someone. But many of the ways something can go well or poorly are difficult to measure. Typically, effective altruists pick some easy-to-measure subset of the outcomes we care about. For example, the charity evaluator GiveWell typically looks at disability-adjusted life years (a measure of how much its charities improve people’s health and longevity) and consumption (the amount of goods and services a person has access to).^[1] Its charities have other benefits: for example, an insecticide-treated bednet keeps you from being bitten by mosquitoes at night. But GiveWell believes that the vast majority of the benefits of its charities come in the form of either health or wealth, so that’s what it measures.

This is not saying that the only way to improve people’s lives is by making them richer or healthier! You can also, say, improve someone’s life by giving them the right to vote. But the Against Malaria Foundation has a negligible effect on voting rights, so that doesn’t appear in GiveWell’s calculations.

Health and wealth are good in themselves, but they’re also good as proxies for other things we care about. All things equal—especially in poor societies—people who have more money are more educated, have more fulfilling work, are more able to choose who they marry, have more free time to pursue their interests, and have more of everything else one might care about. Similarly, it’s much harder to get an education, fall in love, pursue your dream career, and raise children if you’re sick with chronic malaria. We care about more than health and wealth, but health and wealth are still good ways to measure the broad set of things we care about.

I’m discussing GiveWell specifically as an example, but similar considerations apply to all other effective altruist causes: we have to figure out some way of measuring how much of a positive effect our actions have, even if it’s missing some things or is just a rough proxy for a wider array of values.

Tractability is how much it is possible to make progress on a problem. It would be nice to create a perpetual motion machine, but unfortunately you can’t defeat the laws of physics. Perpetual motion machine design is important but not at all tractable.

In my experience, tractability is one of the most controversial items on this list. You can look up a few numbers and roughly calculate how important or neglected a cause area is. But tractability depends tremendously on your understanding of how the world works. Here are some common disputes about tractability:

• Is it possible for people in the developed world to take actions that make the developing world rich, or should we just provide medicine and hope they figure it out for themselves?
• If you keep children in poor countries from dying of malaria, will this increase the population such that poor countries wind up even poorer and more miserable than they started out?
• How hard is it to convince people to share your moral beliefs?
• How much can you influence governments to get them to implement better policies? What strategies work best to achieve this goal?
• How can we make science work better?
• How hard is it to teach AI systems ethical reasoning?
• Is it possible to get any significant portion of the population to be vegan?
• Do short-term welfare improvements for farmed animals help farmed animals in the long run, or do they make farmed animals worse off (e.g. by making people think it’s okay to eat meat)?
• Will we develop cultivated meat before unaligned superintelligences release a genetically engineered superbug and kill all humans, thus solving factory farming permanently?

These questions are hard. Answering them involves carefully studying complex data and making difficult judgment calls about ambiguous situations. In my experience, importance tends to govern effective altruists’ broad cause prioritization (global health, factory farming, AI safety, etc.). But disagreements on strategy and tactics within a cause tend to depend on tractability, and these disagreements can get vicious.

Neglectedness is a subtle concept. People often think of “neglectedness” as “no one else is working on it,” because… that’s what the word “neglected” means…^[2] In reality, in effective altruist thought, “neglectedness” means how much good would be done by an additional person working on the problem.^[3]

I’m going to talk about donations, because money is fungible, which makes the arguments simpler. The same arguments apply for career choice, but in a more complicated way, because you have to account for what kinds of jobs you’re good at.

Many people think of their donations as going to the average thing a nonprofit does: “I donated to the Foundation to Cure Rare Diseases In Cute Puppies, which is doing all of this medical research about photogenically tragic diseases!” In reality, your donations go to what the nonprofit spent money on because you donated, that it wouldn’t have been able to spend money on otherwise. (Your donation is called the “marginal donation.”) The Foundation to Cure Rare Diseases In Cute Puppies, it turns out, is flush with cash. It would be able to fund medical research with or without your donation. Your donation went to hire an additional professional puppy cuddler, to combat the serious issue of touch-starved canines. Still good, but not the photogenically tragic diseases you thought you were helping with.

Of course, it’s nearly impossible to know what your specific donation funds. But, in general, your marginal donation goes to something more important if a nonprofit has less money. The people who run nonprofits tend to pay for important things like medical research first, and only then pay for puppy cuddlers.

It should be obvious how the “give to nonprofits without much money” heuristic fails. Some nonprofits spend money in dumb ways and cut the important programs first. Bigger nonprofits can do large, ambitious programs that smaller nonprofits can’t afford—your marginal dollar goes farther with the Against Malaria Foundation than with Jane’s Artisanal Bednets (she buys nets at a sporting goods store, flies to Nigeria, and hands them out on a street corner).^[4] And if you pick a cause that absolutely no one donates to—maybe you’re very concerned about schist welfare—then there won’t be any nonprofits for you to donate to in the first place.

Nevertheless, in general, if less money is being spent on a cause, then the money is going to more important things. The annual budget of the Biological Weapons Commission—the international organization that enforces the ban on the development of biological weapons—is $1.8 million, smaller than that of the average McDonalds restaurant. It’s safe to assume that, if a country gave more to the Biological Weapons Commission, the money would go to vitally important monitoring and diplomacy. On the other hand, the annual budget of GLAAD, which lobbies for better representation of LGBT+ people in the media, is about $24.8 million. Marginal donations to GLAAD probably go to blog posts whining about iconic lesbian video game Life is Strange—if not putting a chandelier in the CEO’s home office.

Compared to how much lobbying for LGBT+ representation costs, a lot of money is spent on it—not to mention the numerous people who will whine about iconic lesbian video games for free. So the low-hanging fruit has already been picked. LGBT+ representation is important, but additional money probably won’t help very much. On the other hand, control of biological weapons is underfunded—a marginal dollar there goes to very important programs.

People commonly make some mistakes about neglectedness analyses. One is to choose the wrong level of analysis. Climate change isn’t neglected: there was about $600 billion in climate-related spending in 2020. But specific programs can still be neglected, such as supporting right-wing anti-climate-change advocacy, which most dedicated climate change activists find morally repugnant. Similarly, although hundreds of billions of dollars are spent each year to help the global poor, much of that money is misspent, so you can still help a lot of people by donating to malaria prevention or Vitamin A supplementation.

Another common mistake is not to think about what kind of resources a cause needs. For example, like I said above, very neglected causes just don’t have any nonprofits you can donate to. It’s common for young causes—such as wild animal welfare a few years ago—to primarily need competent people who can start new organizations, and to benefit surprisingly little from more money. Similarly, effective altruist organizations suffer from a perennial shortage of managers and mid-career professionals, and a perennial oversupply of eager twenty-two-year-olds who are still deciding between generalist research and community-building. This dynamic leads to bitterness when young effective altruists struggle to get hired and yet see constant complaints about talent constraints in effective altruism.

I think neglectedness is actually the core disagreement between effective altruists and many leftists. Socialist Nathan J. Robinson writes:

Effective Altruism’s focus has long been on philanthropy, and its leading intellectuals don’t seem to understand or think much about building mass participation movements. The Most Good You Can Do and Doing Good Better, the two leading manifestos of the movement, focus heavily on how highly-educated Westerners with decent amounts of cash to spare might decide on particular career paths and allocate their charitable donations. Organizing efforts like Fight For 15 and Justice For Janitors do not get mentioned…

I have to say, my own instinct is that all of this sounds pretty damned in-effective in terms of how much it is likely to solve large-scale social problems, and both MacAskill and Singer strike me as being at best incredibly naive about politics and social action, and at worst utterly unwilling to entertain possible solutions that would require radical changes to the economic and political status quo.

As Robinson says, leftists emphasize organizing and mass participation. From a leftist perspective, all things equal, the fact that a movement is big is a point in favor of joining it. Leftists believe that nothing you do is going to do much good unless it’s part of a broad, coordinated effort to permanently shift the balance of power. Anything else will be co-opted by the powerful to maintain their power. But to fundamentally alter power relations in a society, you need a lot of people. A large movement is a potential locus of working-class power which could be used to overthrow the status quo and create a just society for everyone.

From an effective altruist perspective, all things equal, the fact that a movement is big is a point against joining it: if lots of people are working on janitorial justice, probably the problem is already well-handled. Effective altruists imagine creating a good society, not as a revolution, but as a long series of small discrete tasks. If someone else has scratched fighting for 15 off the to-do list, the effective altruist smiles and sets about making the right-wing case for environmentalism herself.

Failure to understand this difference in assumptions leads to a lot of miscommunication. I hope my explication of “neglectedness” will progress the dispute to a more sophisticated level.

1. ^{^}

   Technically, it’s the logarithm of consumption, which for non-math-people basically means that giving $10 to someone who makes $10 a year is much better than giving $10 to someone who makes $100,000 a year.

2. ^{^}

   Why did 80,000 Hours rename “importance” and “tractability”, but not “neglectedness”, which is by far the most confusing one?

3. ^{^}

   Either directly or indirectly (through donations).

^{^}

Another way of putting this, for the economics-brained, is that doing good has fixed costs.

Show all footnotes