Imagine, just for a moment, that you happened upon a magic lamp and rubbed it. As the stories go, a genie popped out and offered you three wishes-
-warning you sternly that no shenanigans will enable you to gain infinite wishes-
and asks you for your first wish.
Being a virtuous sort, you wish for homelessness to be solved (with all the required caveats so the solution goes well for everyone).
The genie waves their arms around and—poof! - homelessness is solved. Nobody is homeless anymore.
(Maybe new houses magically appeared, in the names of homeless people, and they were all transported to their new houses, and all the issues that rendered them homeless to begin with were magically fixed. It doesn’t really matter what the solution is, only that you observe that homelessness in society is durably done with.)
The question then is:
What happens to all the charities and nonprofits and companies that have fixing homelessness as their aim?
Institutional Commensalism
In biology, commensalism is defined as:
an association between two organisms in which one benefits and the other derives neither benefit nor harm.
If we think about an institution (here defined as “an organization created to solve a problem or address an issue”) as one organism, and the problem or issue that institution was created to solve as another organism, I claim that the relationship between the two is not adversarial, but commensal.
Why Not Adversarial?
In the above hypothetical—a genie magically solves an issue or problem—what happens to the institutions that purport to address that problem?
Do they stick around? Are they still funded?
Who is going to donate money to a homeless shelter, if literally no one was homeless anymore? What government bureaucrat or elected leader will spend time or political capital on a solved problem?
If a problem gets solved, the institutions addressing that problem must either pivot to a different problem, or die out.
Institutions need their problems to justify their existence, and they need a justification to survive. They can’t actually be opposed to the problems they profess to solve, or they would be for their own destruction.
If you doubt this, simply ask yourself: when was the last time you saw a first-world commercial, pledge drive, donation box, for-profit venture, charity, nonprofit, or politician that addressed rickets, scurvy, or smallpox?
Why Commensal?
The Institution Benefits
An institution benefits from its problem.
For one, it literally only exists because of that problem, and if that problem goes away, the institution is in peril.
For two, an institution gets more resources the bigger its problem is perceived as. Homelessness gets a certain amount—in the US, a recent budget proposal earmarked $3.6 billion of the federal budget to this problem. Climate change, in the poorly named Inflation Reduction Act, got $783 billion.
If the climate was magically fixed tomorrow—perhaps by the same genie from above—how much money would governments spend on it the next day? What about private donors?
So institutions benefit from their problems, but that’s only half of commensalism. The other half is that the problems derive neither benefit nor harm from the institutions created to solve them.
The Problem Is Neither Benefited nor Harmed
There are more than 11,000 homeless shelters and other associated housing centers in America.
Homelessness is still a problem.
Environmental groups have been around at least since the 1960s, and the environment is still in peril (at least according to said groups).
The War on Drugs has raged for decades, and drugs have not gone away.
All the charities and nonprofits and government agencies and companies targeting these problems have not vanquished them.
The Caveat: Some problems have been solved—smallpox, for instance—but whenever a problem is solved, it tends to be reframed (if it wasn’t framed that way to begin with) as a mere battle in the larger war. Smallpox was eradicated, but there are many other infectious diseases. Acid rain was stopped, but pollution or deforestation or climate change endures. This can be seen as an example of institutions pivoting to new problems if their old one goes away.
The Implications
So what are the implications, for those of us who want effective institutions to genuinely solve problems?
We can see that institutions thrive the bigger and more salient the problems they’re supposed to solve are. This means the institution is incentivized to never actually solve its problem.
Thus if we want genuine solutions we have to create institutions that are incentivized to solve their problems.
This is a difficult task, but not an impossible one.
A first step might be to create institutions with a built-in expiration date—a time some years in the future at which the institution is automatically disbanded. That way the institution will be less concerned with ensuring its own survival, and might actually focus on permanent solutions to the problems it was created to address.
The Caveat: expiration dates could be gamed by people just setting up a new institution with a different name the next day. Periods when institutions expire could also be tumultuous, with negative consequences for people depending on them.
Another possible avenue could be to reward the institution, not for how big or important its problem is, but for how solved its problem is. Imagine, for instance, if a charity for the homeless was funded based on how many people in its locality weren’t homeless.
A formula could be devised measuring the progress of the institution towards its problem being solved, and it gets paid for that progress.
The Caveat: Goodhart’s demon could run rampant. Any metric can be gamed.
The Conclusion
If I had access to such a genie and used it to solve homelessness, I would expect all of the institutions that address homelessness to die out.
Institutions need their problems the way that superheroes need their supervillains.
Who would Batman be without the Joker?
Superman without Lex Luthor?
And so on.
If we want the problems and challenges facing our societies and our world solved, we need institutions to solve them—but we fail to design our institutions in such a way that they’re actually incentivized to solve the problems they’re created to solve.
Fixing this—coming up with a better way to design institutions—could make humanity far more effective at addressing the challenges in front of us today.
I don’t think any association committed to solving homelessness ever gets to a point where they go “hmm, yes, we could take some action that solves homelessness for good but that would leave us all unemployed, therefore we shan’t do that”. I get your point but I just don’t believe any such organization operates even close to a level of effectiveness where that sort of thing would be even an option, and thus, the perverse incentives would begin mattering. I’m not sure what “solving homelessness” even looks like (even assuming you give houses to every homeless person now, how do you guarantee things stay that way?), but whatever it is, it’s probably some massive country-wide operation that no single small time charity could ever hope to organize. So basically, for most of these problems, I don’t think the issue arises. If it did, I don’t think the people involved would have such a hard time pivoting to different causes; in most cases there is very little going on that makes them specialized in solving THAT problem rather than having general organizational skills that can easily be applied to a different one. So honestly this feels like a theoretical worry that isn’t a real problem in practice.
While I haven’t looked at the data lately, there are a lot of institutions in the US, as I use the term. Surely of the many social ills they address there are some that solvable/solved?
While I used ending homelessness as an example, the salience of an issue matters too. Climate change organizations receive lots of funding because their cause is seen as an important priority. If that changes, their funding dries up. So they have an incentive on the margin to overemphasize the importance of their associated problem—they benefit from the problem, while generally not solving it. Hence, commensalism.
There being a lot means resources are split and coordination may be harder, it doesn’t make this easier. “Solving climate change” for example is nothing short of a monumental task that requires a complete overhaul of our current industrial system. No tiny group of activists will have the power to do that, they don’t need to keep the problem around to stay relevant. And again, I think this widely misreads the people part of these things: these are people who mostly do want to solve the problems. They may not be optimal at it for various reasons and I get how incentives can create pressures that feed this sort of commensalism even without it being an intentional strategy, but it’s absolutely not obvious to me how the process of “could solve the problem but doesn’t” would manifest here. Perhaps you could argue it for some causes like e.g. animal rescues: lots of people have refuges for stray cats/dogs but their general motivation is probably just that they enjoy having the little fuzzballs around and helping them, so they’re not really tackling the problem of stray animals head on with systematic spaying and neutering etc (though that is absolutely also a thing that they do). Also, I’d make a distinction between straight up commensalism and simple “are not cold utilitarian minimizers of suffering”, because lots of people aren’t the latter because they have other values they think are too important to trade off, so they don’t just go at the problem as hard as they could, but that’s not the same as keeping the problem intentionally around to justify their own work.
I don’t see this as a conscious choice people make to not solve the problems the institution they’re a part of is supposed to address. I agree that many of the individuals within the institution are working in good faith and genuinely care.
The issue is that the incentives of the people are not the same as the incentives of the institution itself, which are to grow and attract more status and money, which happens when the problem is seen as harder and more important.
Yes, Climate Change is obviously not solvable by a few activists, but there’s a finite amount of time/energy/money in the world, and it’s not clear to me at all that it’s optimally distributed between cause areas. More time/energy/money going into solving climate change means less going elsewhere.
If those resources were freed up, couldn’t a lot of homelessness charities pivot to some other problem? You’re not wishing their funders away.
Some things would be different, but a lot of the skills are the same (research, running organizations, running political campaigns, marketing, etc.).
They would need another problem to pivot to.
Also, I suspect that such a pivot on an institutional scale is difficult to pull off. People often prioritize altruistic work because they’re passionate about a specific cause—maybe they were homeless in the past, or they were a cancer survivor, etc. That wouldn’t necessarily translate.
It has been done. March of Dimes was originally an anti-polio organization; after the polio vaccines basically solved their problem, they generalized...
Thanks! I’ve been pointed to them by others as well; it’s a good example of an institution surviving the death of their problem.
I do think that the case underlines how important problems are for institutions, in a sort-of “exception that proves the rule” kind of way.
This is tricky, because it seems like the things that will be useful if the problem remains for a long time, are exactly the things that will prevent the institution from disbanding when the problem is solved.
Like, imagine that an oracle told you that homelessness is a serious problem that will require at least 200 years to be solved. (Imagine a world without a looming AI apocalypse.) In that case, if you are serious about solving the problem, you would want to have: institutions that specialize on homelessness research, maybe to teach it as a subject at universities, people whose full time job it is to solve homelessness, secure funding for such people, etc.
But those are exactly the things that would cause trouble when homelessness is magically gone. People who spent their careers studying something that is now useless. People losing their jobs. Universities that need to update their lessons. -- It would be too tempting to redefine the word “homeless” to refer to something that still exists; and thus people can keep their jobs and funding. Like, maybe “many 18 years olds cannot afford to buy a new house” will become the new operational definition of homelessness. Or it will be considered horrible that the former homeless cannot live in the center of the most expensive city. Or that their new houses do not have gardens.
From certain perspective this is not necessarily a bad thing. The fact that we improved society doesn’t mean we cannot improve it further. Maybe everyone should have a cheap huge house with a garden at exactly the place they want to be, if we could somehow afford it. I mean, our standards of living are already amazing from the perspective of people who lived centuries ago, and we still keep thinking about improving them.
The actual problem is more about priorities. It is okay to expand the definition of homelessness (or anything else) after the original problem was solved. But the new problem is less urgent than the original one; and therefore deserves less resources, etc. But if the existing organization and people want to keep their funding and jobs, they need to insist that the new problem deserves the same priority as the old one. Which means they have an incentive to exaggerate it. Lies lead to misallocation of resources.
A quick idea: You need to make a measurable definition of “homelessness” and publish a graph of homelessness visibly (on your organization’s homepage, in every publication you print). You must keep the original definition; or if you decide to change it, your must keep drawing the line that corresponds to the old definition, and add a new line using different color for the new definition.
An obvious problem: Graph alone does not distinguish between situations “homelessness was reduced by 50% because of ongoing efforts of our organization, therefore it is critical to keep funding it” and “homelessness was reduced by 50% for reasons completely unrelated to our organization, therefore the situation is less urgent”.
This was well thought-out, thank you.
You’re right about redefining the word/problem. I’ve been referring to this as “The Pivot” in my head.
It would still be better if we found a way to form institutions such that, once they had solved a problem, their resources were efficiently allocated to the solution of the next-most-pressing problem.
I think the claim here could be summarized as: charities may have a vested interest in the problem they’re trying to solve (conflict on interest). However, it’s good to observe this isn’t always the case.
For example, I know volunteers that help the homeless. Everyone in the org is a volunteer, except the cook. If homelessness disappeared tomorrow, they could just take a rest day or go to the park etc.. This is a first prevention mechanism (minimizing vested interests).
Sometimes though you may need employees so that your organization is effective. In this case, they kind of need to pay a cook. It’s hard work, specialized work, workers are usually not high income and need the money. Maybe it would be ideal to find a volunteer cook, alas. In that case, there is still ethics. If the organization and people are effectively ethical, then they should not respond to the incentive by increasing homelessness (in any case… I think it’s fairly difficult to increase homelessness on purpose, and even more difficult to do such a way as to make a personal difference). This is a second mechanism (ethical reflection).
But conflicts of interest are extremely important to keep an eye on. On everyday discussions, political and social cases. (see: scout mindset)
As for why the problems haven’t been solved, I think it could be that it’s just not that simple. It’s like asking a farmer “If fertilizers worked well, why do you need to keep fertilizing the soil after all those years?”. Some problems may demand constant, permanent attention. Don’t volunteer trying to solve homelessness, instead volunteer trying to make the life of homeless people better. Hopefully that one day lowers or eliminates homelessness as well, but we shouldn’t condition help on that.
I use homelessness as an example, but I believe the logic generalizes. You’re right that in many cases, the incentives facing an institution aren’t powerful enough to matter, or the people involved could/would just go do other things.
But there are also a lot of cases (see: almost all nonprofits) where people’s jobs depend on the existence and salience of the problem, in which case I think the incentives do start to matter.