Conventional grant-making covers a project’s expenses, iff the project’s output seems to justify those expenses. In theory, this isn’t a great good idea. It’s not hard to find problems in practice either:
- Paying for inputs puts donors in the position of choosing what charities “should” spend. What salaries they should pay, how many staff they should have, how large their reserves should be, and so on.
- Paying for outputs instead makes price signals meaningful (both for ouptuts and inputs). For example, salaries play a huge role in resource allocation in the for-profit sector, and seem to help a lot, but they don’t seem very useful as a source of information in the non-profit sector. Paying for ouptuts would make it more realistic to offer services to non-profits (and provide useful information to anyone who wanted to offer such services, even if they were entirely motivated by altruism). And so on.
- Similarly, prices for ouptuts can help funders coordinate and allocate their resources effectively.
- Paying for outputs introduces material incentives for innovation.
Paying for outputs rather than inputs is harder: it’s easier to estimate “How much does a bednet cost?” than “How much good does a bednet do?”
But once you are making a serious effort to determine what interventions are most cost-effective, this problem becomes much less serious. It’s the difference between asking “Which of X and Y is larger?” and “How much larger is X than Y?”
Given that the cost-effectivenesss of X and Y changes continuously over time as the low-hanging fruit is exhausted, the harder comparison shouldn’t be that much harder. If you can tell that X is better than Y, but don’t know whether X is twice as good as Y, you are going to have a problem next year when X is twice as expensive.
That said, such quantitative estimates will be very noisy. In addition to the discomfort of committing to very uncertain estimates, this introduces risk into a charity’s income stream (via “how much are people willing to pay?”). It’s not clear if this risk is greater or smaller than the risk from “are people willing to pay?” under the status quo (since if you are paying for outputs, a charity can greatly increase its funding by lowering its prices).
In addition to the pragmatic benefits, the cultural effects of paying for results also seem significant and positive.
The impact purchase 