Sunday, December 30, 2012

Being a Utilitarian, Part 2: Conventional Charities

This is the second post in a series on actually being a utilitarian in the world; for the first post, look here.  Also, for a more theoretical series on utilitarianism, look here.


So, say that you're a utilitarian, and you're wondering what to do with your life.  (Even if you're not a utilitarian but are wondering what to do with your life, most of this will apply.)  What should you do?  What, in the current society, can an individual do to make the world a better place?  And what causes should you care about?

Is there anything you can do with your life to make the world a better place?

There are some complicated and hard to evaluate things you could do, like decide to become a politician and influence congress, or try to become famous and use your fame to convince people to do something meaningful with their lives.

But there are also more conventional things you can do to try to make the world a better place, which are much easier to evaluate: you could go work for a charity, or you could donate money to one.  Which of these two things is more efficient largely depends on how much money you could donate; but, for instance, if you think that if you were to work for a charity your work would be worth $50,000 per year for them even if you were to work for free (an estimate of what it would cost them to hire someone to do your work), then donating $50,000 a year to that charity would do as much good as working for the charity.  Note that working directly for them is even worse, compared to donating, when you take into account the fact that, even if they hire you, they'll have to pay you--you're only actually adding value to the extent that you're more productive than the marginal hire for them.

Also, note that many people decide to donate a little bit of money to a bunch of charities.  This, to put it bluntly, is stupid, unless your donating an amount of money comparable to the entire operating budget of the charities.  Unless you expect significantly decreasing marginal value of donations to an charity--something that should only happen if you're donating something on the same order of magnitude of the entire charity's income--whatever charity you think it best for the first $100 that you donate should also be the charity which you think is best for the next $100 that you donate.  (The same logic applies to time you spend working or volunteering for charities.)

Note, by the way, that if you think you could make significantly more than $50,000 per year--and donate more than that--it can be more efficient to spend your life working in a high-paying profession and donating a lot of your money than working directly for charities, a strategy often called earning to give.

But whether you're donating to a charity, working for a charity, or something else, you can evaluate--at least roughly--how much value you're providing them, in units of dollars.

Which then brings us to the question: well, what causes are important?  Or, put another way, what charities have the highest impact per dollar donated?


I should probably preface this post by saying that I don't think any cause I mention in this article represents the most pressing concern for this world or the most efficient use of time or money--I'm instead going to talk about conventional charities, the kinds of charities no one would ever look askance at you for supporting.  But when evaluating how important a cause is it's important to have some sort of baseline to compare it to: if a friend came up to you and told you he was starting a charity that would help one drug addict to stay clean for the rest of their lives for every $10,000 donated, what would you think of that charity?  Is it efficient?  Is it a total waste of money?  Is it an order of magnitude better than any other charity?

Or, to take a real example, is it a good idea to donate money to Komen in order to try to fight breast cancer?  How about donating to the Berkeley Repertory Theater to support a local theater company?

Just how much impact can a dollar have?

Well, if you want a one sentence answer, here it is: one two thousandth of a life.

That's the answer that GiveWell has reached when evaluating the Against Malaria Foundation, one of their three top rated charities.

GiveWell is an organization that attempts to determine exactly how much impact charities per dollar donated, and they've estimated that the Against Malaria Foundation--a group dedicated to providing anti-mosquito bed nets to developing countries--saves a life for every roughly $2,300 donated.  It's of course an imperfect process but I suspect that this number is at least right to an order of magnitude.

How does this compare to the drug addiction charity, or Komen, or the Berkeley Repertory Theater?

Well, let's start with that drug addiction charity.  They claim that every $10,000 donated keeps a drug addict clean for the rest of their life.  That's about four times the cost that it takes to save a life by donating to the Against Malaria Foundation, and it doesn't really save someone's life--just make it better--so it's unlikely to be as efficient a way to donate your money as anti-mosquito bed nets are, but it's in the same ballpark.

OK, how about Komen?  Some legitimate concerns exist about Komen, but let's ignore those for a second.  Let's also postulate--much more favorably than could possibly be true--that if Komen's budget were permanently multiplied by ten, they could permanently cure breast cancer but that breast cancer would otherwise go uncured.  This is about as favorable as one could possibly be to Komen.  How would it stack up as a charity, then?  Well, it's budget is around $400 million per year, so ten times that is $4 billion per year.  Worldwide about 450,000 people die of breast cancer each year, meaning that it cost about $10,000 per life saved.  So even under absurdly optimistic assumptions, it seems unlikely that this is the most efficient place to donate your money.

And how about your local theater company?  Should you donate to them?  Well, in 2009 (the latest year for which the theater put an annual report online) the Berkeley Repertory Theater sat approximately 160,000 audience members total, while spending about $10,000,000 putting on their plays--a total cost of about $63 per audience member.  So, for the cost of saving a life in a developing country, you could subsidize about 30 tickets to a play.

These calculations, of course, aren't perfect, but they provide a reasonable baseline for how to evaluate how much good you can do.

In future posts I'll talk about the other things a utilitarian could do with their life, and why I'm not planning on donating any money to the Against Malaria Foundation.  But I think it's useful to have a relatively easily quantifiable baseline to compare things to, and a pretty good one is: $2,300 to safe a life by donating money to the Against Malaria Foundation.


  1. There is a contradiction in your logic that I can't get around. If we value each life equally (say $2000 to save a life, as you state in your baseline), the fact remains that different people earn different amounts of money despite each having only one life. A person who earns $20 million over a lifetime could (potentially) save ten times as many lives as one who earns $2 million over a lifetime, if all the earnings went to charity. Should we not then value the lives of higher-earning people more? But then this violates the starting assumption.

    If you believe that lifetime earnings are purely a function of luck, then there is no problem, but I think this belief is untenable.

    1. " A person who earns $20 million over a lifetime could (potentially) save ten times as many lives as one who earns $2 million over a lifetime, if all the earnings went to charity. Should we not then value the lives of higher-earning people more? But then this violates the starting assumption."

      I mean, if we thought that the people earning $20 million were going to donate their money to a good cause then you could consider saving their life an investment in the amount of good they'd do, but the problem is that most people don't donate all of their money to the best cause--most donate almost none of it and that which they do donate is going to be, mostly, to basically worthless causes, and so what you get from saving their life is less the money they will then donate than the amount of good from having them live. (Which, honestly, is only dubiously positive to first order right now for most people because most people inadvertently cause a lot of harm e.g. by eating meat, but that's a different story.)

      There's a slightly different way to interpret your point, though, which is not that rich people will donate more money to charity but that rich people, generally, are people who have provided more value to society. That certainly is going to have issues in that there is a non-equivalence of the dollar between, for instance, here and Africa and so if someone produces twenty loaves of bred per day they will have a lot less money in Africa than they would here, but you could argue, for instance, that saving Steve Jobs should have been a pretty big priority because he will produce lots of good products for the world. Which, I think, is possibly over-interpreting the extent to which material goods are what make Americans' lives better, but is still a valid point.

  2. If you had a sane tax code, and a government that did some good, it wouldn't be very hard to extract value from rich people. Certainly it's easier than relying on voluntary donations. Regardless, I'm not convinced that counting individual lives is the correct utility function. How about something broader like maximizing the expected survival time of the human species?

    1. So, first, I agree that a sane tax code could do a lot to help this--particularly taxing rich people to fund the kinds of things that save peoples' lives.

      But I also agree that while this is a reasonable metric, it's not ultimately the best one. Obviously in the end the metric is maximizing utility, but for intermediate metrics I'm not sure that expected survival time of the human species is a very good one for the following reason: that's a good metric in the limit where the percentage deviations in possible utilities of possible future worlds is small compared to the deviations in lengths of the worlds, but I don't think that applies here. In particular, I think that the current utility of the world isn't even positive (number of animals is much greater than number of humans and they have shitty lives), so trying to maximize length without regard to value of utility at each point in time might lead to a really really bad solution. Also, the possibility of things in the future that could lead to tons and tons and tons of happiness (genetic modifications, etc.), as well as things that could lead to lots and lots of pain, mean--I think--that in fact we're probably living in the other extreme, where deviations in the length of survival time of humans--while large--may in fact be really small compared to deviations of the impact that we could have on the universe.

  3. I have two problems with your math.
    First, curing breast cancer applies across many years. So over 5 years it costs $2000 per life saved and over 50 it costs $200 per life saved. Ad infinitum. Those numbers are spread out, but spreading out $4 billion with the Anti Malaria foundation to save 400,000 per year, you spend 800 million each year, saving 3.2 billion. Investing that 3.2 billion gets you .06*3.2 billion = 192 million, which doesn't cover the 800 million you spend. So investing in Komen over 10 years gets you more return for your investment than the Anti Malaria Foundation.

    Second, you're talking about saving lives, but as a utilitarian you should really care about the pain that their deaths cause to their loved ones. So for the Berkeley Rep charity, you are comparing like things: the pain of deaths on loved ones versus the joy from a Berkeley Rep performance. Now I still think maybe 4 grieving loved ones (for one death) is worse than 30 happy theatergoers, but that's the calculation that matters.

    1. Sam's scenario had it costing 4 billion per year *permanently* for Komen to cure breast cancer, not just in the first year. Even if they discovered a cure in the first year, the cure would cost money to administer. Cancer is complicated, and any cure would probably be more expensive than simpler matters like not getting bitten by mosquitoes or staying hydrated when you have a digestive infection. So I expect developing-world health to be the cheapest way to save lives (until health standards come way up there).

    2. Yup, that's the assumption I was making about the cost to cure breast cancer. It's obviously arbitrary but I'm not convinced it's the wrong one. (Also, all of the other assumptions are really favorable to Komen.)

      On your second point, Gabe: when I say that for $2,300 you can save a life, what I mean is that for that money you get all of the results associated with someone who would otherwise die living. That means (a) they live out the rest of their life, (b) their loved ones don't grieve for their untimely death, (c) they contribute to society, and, perhaps most importantly,(d)--the society that they live in becomes one were fewer people die young, which has all sorts of implications for its shape: people plan for the future more, live their lives less in fear, people have more time to establish their lives and livelihoods which makes society much more productive, etc. It's very hard to end the cycles of political corruption, violence, and poverty in those regions if large sections of the population are dying too young to do anything about it. And in addition to all of that, you have to consider the implications of saving a life on things like meat consumption and the environment, both of which are going to make a life in a developed country more negative than one in a developing one.

      All of which is to say that it's actually nontrivial to calculate the impact of saving a life, which is why I just left things in those units.

    3. On Komen:
      I don't think it's trivial to do the new math, which finds that the yearly costs of breast cancer treatment would have to be less than the difference between the $800 million spent on equivalent malaria prevention and the $192 million interest on the original $4 billion investment. So if the administrative costs are below $608 million, then breast cancer treatment is still a better investment. I'm not saying that they would be - polio treatments cost $1.5 billion worldwide. But now we're on a similar order of magnitude, so it's probably worth getting more accurate numbers instead of discounting Komen from the first-order approximation.

      On pain/death:
      I concede that point. My argument was admittedly weak from the start.

  4. Small point:
    "Also, note that many people decide to donate a little bit of money to a bunch of charities. This, to put it bluntly, is stupid, unless your donating an amount of money comparable to the entire operating budget of the charities. "

    I disagree with the exact wording, just because there are many other reasons to donate than for the sole purpose of the money itself helping those charities. For example, one could donate to feel less pressured, or to provide emotional support, or to get into cool events, or to signal.