Charitable Discount Rates

Alternatively titled: It’s my charity, and I want it now!

Robert Wilbin at OvercomingBias recently wrote a short post arguing that when evaluating charitable options, we should apply discounts to costs and we should not apply discounts to benefits. This topic makes me uncomfortable, but I’d like to make three quick observations that might be helpful to the charitably-minded. [Edit: It appears as though the article has moved to Giving What We Can.]

First, people have a preference for helping others sooner rather than later. All else held equal, I would rather help a stranger today than a stranger a year from now. Assuming there’s countless children who constantly need saving, I’d rather save a million children now than a million and one different children a year from now. I suspect that’s the way the charitable feel, and I don’t think any amount of writing will change that. I treat preferences such as these as exogenous features to be maximized rather than parameters to be changed.

Robert acknowledges this point. He writes that:

Time preference appears similar to arbitrary prejudices regarding whose interests count that are generally rejected today, such as racism and sexism.

I’m uncomfortable with any reasoning of the form: “All good people admit that Y is bad. Y is like X. Therefore, X is bad.” Having a preference for your family is like racism. Most people like for their family to do well. Does that mean familial affection is bad or that racism is good? I reject both conclusions and instead reject making moral judgments based on surface similarities.

People will prefer charitable schemes that save lives now instead of later. If you want to produce a cost-benefit analysis useful to people, then you should take that preference into account.

Second, if you believe in the “pay it forward” principle, good deeds will multiply. The returns will not be realized by you, but that’s why it’s called charity instead of business. To make this warm and fuzzy concept more real, imagine you save a child from starving and he grows up to become a fisherman who keeps other children from starving.

Finally, I believe this conclusion will lead people to never give to charities. If a charitable fund can either donate $100 now or invest the money and donate $100*(1+g) in a year, they’ll always take the latter if they do not discount the benefits of charity. By recursion, this leads to a conclusion that is ridiculous to humans.


  1. “If a charitable fund can either donate $100 now or invest the money and donate $100*(1+g) in a year, they’ll always take the latter if they do not discount the benefits of charity.”

    No, there will be a time when economic growth is no longer possible and the probability of losses is so high that total wealth can only be maintained and redistributed, not increased.

    When this happens, paying the funds out to do maximum good at that time is maximally efficient.

    1. That’s a true disclaimer, I suppose. When the four horses of economic distress make ANY positive interest rate impossible, even he who never discounts will give money to charity.

      That doesn’t seem particularly likely or predictable, though.

  2. It is very likely and predictable, under the reasonable assumptions:

    1) that resource access is limited by light speed or some other limit and

    2) that the amount of economic output per resource unit is limited, e.g. by diminishing returns to efficiency innovations

    While this is clearly a while off and the world will look different when these limits are approached, they are still forseeable limits.

    Of course, along the way an investor may expect Black Swans and politicial or military disruptions, maybe technologies gone wrong, and so on.

    On the flip side, disruptive technologies (e.g. automation, fast reproduction, fast education, mind copying, new energy sources…) may make interest rates far more attractive, which actually makes the original argument stronger.

    Maybe you really face the question whether you want to make 1000 lives better now (in a not optimally productive way) or give free leisure lives to 10 million additional people in 200 years, using the resources from your R&D investments.

    Would you really trade 10 million lives for 1000 just because they are further away in time?

    I find my intuitions fluctuating.

Comments are closed.