Should we diversify our charitable giving?

Citing Steve Landsburg, Tim Harford argues:

Someone with $100 to give away and a world full of worthy causes should
choose the worthiest and write the check.  We don’t.  Instead, we give $5
for a LiveStrong bracelet, pledge $25 to Save the Children, another $25
to AIDS research, and so on.  But $25 is not going to find a cure for
AIDS.  Either it’s the best cause and deserves the entire $100, or it’s
not and some other cause does.  The scattershot approach simply proves
that we’re more interested in feeling good than doing good.

Many people are unconvinced by this argument–which I owe to Steven Landsburg–because
they are used to diversifying their financial investments (a bit of
Google stock and a bit of Exxon, too) and varying their choices
(vanilla ice cream AND bananas). But those instincts are selfish: They
are not intended to benefit both Google and Exxon, nor both the
ice-cream company and the banana growers. With charity, the logic is
different, and a truly selfless donor would bite the bullet and put his
entire donation behind one cause. That we find that so hard to imagine
is just one more indication of how hard it is for us to think ourselves
into a truly selfless view of the world.

We can think of charitable projects, at least in ex ante terms, as aligned along a continuum of expected returns.  The highest-return project is just a wee bit better than the runner-up candidate.  In that setting, it is hard, as always, to evaluate the efficiency consequences of differing distributions of wealth.  But in a Rawlsian sense — what would a poor person want if he did not know which group he would end up assigned to? — the poor would prefer that any particular gift is diversified.  Even if the dollar rate of return falls by a small amount, the insurance value of that giving rises. 

Keep in mind that a single donation is itself supporting a bundle of projects, not a single giving opportunity.  (What would a truly specialized donation look like?)

I agree with Harford’s point in a different regard.  The fixed costs of processing a donation are relatively high, if only because the charity will send further letters asking for more money.  For that reason it may be better to focus our giving on a single charity.


A rational investor with $100 to invest in a world full of worthy investments
diversifies rather than betting it all on one investment. Why shouldn't the
same thing be true with giving to charity? How would a donor know which charity
is the worthiest? Should the Gates Foundation invest all its money in one

The Gates Foundation is in a TOTALLY different category from someone with $100 because the marginal benefit of the hundred millionth dollar for the recipients of a given charity is usually substantially lower than that of the ten millionth dollar. The marginal benefit of the hundredth dollar is essentially never lower than that of the tenth.

The Rawlesian argument only holds that we should only invest in those charities that help the poorest poor (with due attention to the macro effects). That and hedonics research make me think that Operation Smile or Smile Train are good choices, as, of course, is microcredit. The Rawlesian argument does NOT hold that we should diversify our charitable giving because the charities that we give to will give constant sized chunks of aid in any event, they will simply give more units of this aid with more money, increasing the likelihood that a given member of the poorest poor will recieve aid.

For a rational agent, there is NO purely altruistic argument for diversification of small funds. However, there may be such an argument for a human. Much of the actual reason for giving to charity is that the person is purchasing an identity. Purchasing diverse charitable goods may, for some people, lead them to acquire more of the identity product than purchasing more of fewer goods, and a more giving identity may enable the person to give more total than they would otherwise actually give.

Michael Vasser completely misses the point I was making, which is that
both "rational" investors and "rational" donors diversify. The sums
are irrelevant (although someone with only $100 to invest would likely
put it into a money market fund, which is a diversified mutual fund invested
in fixed-income assets).

The Rawlesian argument is irrelevant and anytime I see his name invoked I
know I'm in for a hornswoggling. Why are the poor worthier than the
non-poor? Rawles gave no good reason.

There is also no such thing as altruism, pace Rand and biologists, because
all action is self-interested, including giving to charity.

The best argument for multiple donations, as I see it, is that the political and moral strength of some charitable organizations depends on how many members they can say they have, so you might want to let multiple organizations be able to claim you as a member.

Following up on David Balan's point about discontinuous returns, there are a wide variety of 'challenge grants' floating around, some of them with caps. Picking up 'free money' in this fashion, or by playing with employer donation-matching programs may justify diversified giving.

Transaction costs, as Tyler alludes to, are also important for small donations. It ultimately my cost a charity $1-$10 to process a donation, whether it is a $100 or $1000 donation. If you split your donation among many charities, you are already increasing the total transaction costs and reducing the efficiency of your donation.

To put it another way, a ration investor with $100 will *not* diversify if it costs $19.99 to get into each asset class.

Conpetitive Altruism

One of my colleagues - Gilbert Roberts - may have suggested part of the reason for scattershot donations: competitive altruism -

The idea is that altruism, including charity giving, can function as a signal of fitness ('I have enough and to spare') which is sexually attractive.


If people give in order to feel good about themselves, then presumably awareness of the argument against diversification should change the ability of different types of donation to make them feel good. I would feel like an idiot giving $100 to a charity that provided stuffed animals for Katrina victims, or supporting live ballet or opera performances for the rich, because I know that same money could buy deworming medicines or mosquito nets to save the lives of multiple African children. (Incidentally, I follow the 100% concentration model in my charitable contributions.) There is such an enormous variation in the utilitarian impact of different charities (orders of magnitude), that reallocation of contributions from ineffective to comparatively effective charities is likely to outweigh any reduction in total giving.

Tyler asked what a truly specialized or specific donation would look like. The answer can be seen on . We have created a system for people to give directly (with the least possible intermediation) to specific projects, mostly in the developing world. We charge a flat 10% for processing costs, and even when transfer fees are included, we can usually get at least 87% of the money directly to the project. The site has information on each project, including the people running it and the expected outcomes. Project leaders are given incentives to post progress reports because their projects appear more prominently on the page if they do. We just introduced a feedback mechanism that allows users to provide feedback on the progress reports. This is only the first step in a more robust feedback system that will distribute the monitoring and evaluation function, and enable projects to get new ideas from anyone in the world. The goal is to enable grassroots groups to build a track record (brand, reputation....) and help the cream of the projects to rise to the top.


The Mercatus Center is actually a pretty good charity: their operating costs are small, but stopping an onerous regulation can produce societal benefits in the hundreds of millions or billions of dollars. On this point, I agree with Matt Yglesias, who thinks that people interested in a cause will get more bang-for-the-buck by trying to influence other actors (such as governments, corporations, and wealthy individuals) through idea production than by acting directly. I target my own giving towards the Singularity Institute for Artificial Intelligence on this theory (and because the area in which it operates is tremendously underfunded relative to potential gains and losses that dwarf even those associated with issues like HIV or malaria).

The piece below notes another practical reason not to diversify, namely that charities are reluctant to sell the contact information of large donors (who they wish to hoard), but will sell the info of small donors, whose donations are largely consumed by transaction costs.

There are numerous smart comments here, but none of you have solved for the right answer. The general problem is which risks should count in which contexts. Kenneth Arrow, in a different context, got the problem wrong, so it is tough.

Doesn't this just show that charitable giving, like tipping, primarily is for the benefit of the giver (or tipper). We are advertising to others and to ourselves that we are good, generous people. The returns are measured in our increased sense of self-worth and in the increase in esteem we are held in by our peer group.

This is just normal human behavior, we are constantly jostling for increased social stature, and there are many thinly veiled reasons we have for justifying our behavior. Fashion, car purchases, travel, hobbies, charity, voting, etc, etc.

I'm exaggerating with the use of the term "primarily", but a scrambling for social status is a large part of almost everything people do. It's not good or bad, it's just human nature.

Further to Carl Shulman's post about the roulette wheel ... should I give $10 to fight malaria, or should I give $10 to an organization -- say, the Cato Institute -- that promotes liberty and free markets? The first way, I am likely to prevent one case of malaria. The second way, perhaps I have one chance in a million that my donation will help tens of millions of people get rich enough to eliminate malaria forever.

What is the marginal benefit of my $10 to Cato's cause? Is it more or less than the marginal benefit of the $10 sent to fight malaria now? And if it's more, is it still a good idea?

You seem to be concerned with minimizing the chances that you give nothing to the optimal charity, but why is it so bad to take a risk that provides the best expected benefit to humanity and get snake eyes?

There is a systemic problem not being addressed. In today's current environment, it may make sense to give 100% to a charity.

However, if all donors were like this, there would also be a corresponding rise in charities that are fraudulent or inefficient, but succeed by "big game hunting." A fraudulent or inefficient charity could go pretty far by gladhanding just a few donors, who may or may not have time to put in an investment of time commensurate with their investment of money.

In a diversified system, though, the fact that charities have to attract a greater number of small donations creates greater overall accountability. The greater the number of donors, the greater the scrutiny a charity will be placed under.

This is, of course, just a hunch, but the fact that the prizes get bigger may make for a greater incentive to try and "cheat."

There is also no such thing as altruism, pace Rand and biologists, because
all action is self-interested, including giving to charity.

Can this proposition be expressed in any falsifiable way? If not, I am inclined to classify it as a statement of faith, and thus one that tells more about its adherents than the world.

Do people here REALLY believe that there's a one-dimensional measure of worthiness? Isn't it more likely that each person has multiple objectives (e.g., we really have a multiobjective optimization problem)? Spreading contributions could then be entirely logical - as long as the selection is "noninferior".

What the hell?!?

Nobody has gotten this answer correct! You should donate $X to a charit[y|ies] exactly the same way and for exactly the same reasons as you would invest an equivalent sum. For $100, the cheapest trade you'll find is $7 at Scottrade. 7% of your money is going down the drain, but that's a heck of a lot better than buying 10 stocks each for $10 each, leaving only $3 to actually buy the stock. For $10,000, you'll have the problem of evaluation. You could reasonably invest $1,000 in ten stocks, but which ten? How do you know they're the best ones to invest in? You might do a good job, but for that amount of money it still makes sense to invest in a mutual fund or an index fund such as an ETF. For $100,000, it starts to make sense to evaluate twenty stocks, and pick ten.

The problem is, as always, that the transaction costs kill you.

"The Economic Case Against Charity"

There is a man in one town who is dying of a horrible disease. Some scientists are working on a cure but this is an ongoing project and they can always use more money.

There is another town somewhere without many doctors. Children there need more medical care. But the town is poor and charitable donations are necessary to fund doctors working there.

A fool would think, "Gee, my life is pretty good, I've been pretty fortunate, and I have some extra money that I'd like to use to help these people out. Let's see, I have $100 to spend, I think I'll give $50 to one cause and $50 to another."

But I have training in economics and know better. I give all the $100 to fund doctors, because it's not like $50 by itself would do any good. Or maybe I give all $100 to fund research into the cure, because it's not like $50 would do any good. Or maybe I give nothing, because I've achieved nirvana and realize that I'm a rational selfish agent attempting to maximize my objectives, and what the hell do I care about other people anyway? But I know one thing for sure -- that providing to charity is a SINGLE OBJECTIVE, and therefore it would be ridiculous to give money to both causes.

You see, I am an economist.

Barbar, you're missing the point.

Let's look at your example. Two causes, one, a man who has a disease that will take $100 to cure it. Another, Children dying, it would take $50 each to save them. If you give half/half, you save one life. If you give all of it to the research, you save one life. If you give all of it to children, you save two lives. Which one is the best course of action? Unless you want to argue that saving one life is just as good as saving two lives, then the morally correct thing to do is to donate all your money to the children.

As to why people donate at all, I think there are two reasons for it. The first is that the donators want to maximize the benefit to humanity. (Things like cancer research, money for malaria, and political donations would fit in this category), secondly, they donate things that they feel they consume in part. Museums, PBS, and churches fall into this category. (Have you ever met anybody who donates to PBS that doesn't watch it?).

I think that, for the most part, a given persons donations will consist of multiple second-type donations, (to your church, various cultural things, and your alma mater) and pet cause of the first type.

Berber gives a perfectly good example but doesn't get it.
Maybe this has to do with some difference of opinion regarding what "should" means.
He's pretty horribly wrong with respect to what counts as realistic though. It's closer to correct that $100, spent rationally, should save either 1000 lives or 2000 than thatn $100,000 should save either one life or two.

Perhaps a better example would be giving $1 to prolong 10 lives for 4 years versus giving $1 to prolong 15 lives for 3 years. Maybe we could add some more details about the age, gender, economic productivity, living conditions, and generosity of the people being helped. Then the problem readily becomes solvable using calculus, and only a desire to feel good about ourselves prevents us from using the rational solution, right?

The problem is not that we want to feel good about ourselves more than we want to help people. The problem is that the human brain can recognize when it's about to plunge over a cliff, and people are smart enough to avoid this. Rather than enter an endless loop, we can make some arbitrary decisions and move on. Selfish, sure, but I don't see how you avoid it. Giving only to one charity rather than many is just the same sort of selfish solution (excluding considerations about transaction costs, of course).

I don't buy the analogy with the market portfolio, even if we assume zero transaction costs. It is much easier to argue that market prices efficiently reflect available information than it is to assume that MY normative values are equal to those of all other donors considered collectively.

If I am a small donor, then I probably cannot get the aggregate of all donations onto MY efficient frontier, no matter what I do. But I am likely to get it closer by concentrating than by not. So if I am rational and have my own values or sense of how the world should be, then there is no case for diversifying, regardless of transaction costs, information or whatever. It boils down to normative values, which vary across persons.

There would be an analogy to the market portfolio here only if I were a very large donor or if I cared not for others' welfare but only for how I impacted them personally.

"I don't buy the analogy with the market portfolio, even if we assume zero transaction costs. It is much easier to argue that market prices efficiently reflect available information than it is to assume that MY normative values are equal to those of all other donors considered collectively."

Yes, the argument that people should diversify donations would seem to be even less plausible that the argument that all investors should hold the exact market portfolio. My point was that diversification of small donations being optimal is an extreme case, and depends on very strong assumptions, just as the case where you should always hold the exact market portfolio does. You do not diversify donations for reasons similar to those that you partially diversify your investment portfolio for (even if you believe that more than one charity should recieve funding), as some people have argued.
That said, the conclusion that everyone will hold the same portfolio does (I think) depend on assumptions about subjective preferences (although they are preferences over things (i.e. probability distributions) that can be objective, which is probably not the case for charities), it does not follow directly from markets being informationally efficient. So I'm not sure the difference is as large as you suggest.

Logical or not, I will continue to diversify my charitable giving for a very simple reason: I can't decide.

According to Mr. Landsburg, if I were a truly generous person (and not merely interested in gratifying my own ego), I would simply pick the one charity that I consider, for whatever arbitrary reason, to be the best, figure out the maximum I can afford to give, write a check for that amount to that charity, and shut the door in the faces of all others. The problem is that if I say to myself, "Okay, I'm going to give away five thousand dollars this year, and it all has to go to just one place," I will develop analysis paralysis. With that sum of money riding on my decision, I'll look at all the worthy causes I normally donate to and think, "Yes, this is a good cause, but is it really the *best*? Am I really willing to deny my dollars to every other charity in the world, saying that I don't think they deserve the money as much as this one?" I'll delay and delay my contribution while I fret about which organization is truly the most worthy. And I'll end up giving nothing at all.

Giving $100 here and $200 there may be less efficient, but at least I end up giving *something* that way. And frankly, I think it's more important just to give something, anything, than to make sure all the money you give goes to the absolute best cause.

The bottomline word in your article title is "giving". In my opinion "diversify" reaches second place. The fact is, the more we educate ourselves to share our goods with people in need, the more sympathetic we become, the easier will be for us to understand their problems and to stay by their side with whatever we can. Human nobleness is the key. This way, we`ll get easily from cloth donation to automobile donation and other comparisons alike.

Comments for this post are closed