If anti-poverty programs can pay for themselves in two or three years rather than twenty, wouldn’t that make sense?
Today I have a post in the WashPo’s Monkey Cage on programs that give livestock or cash plus training and other services, such as supervision and advising. Some recent studies, including one of mine, say these are cost effective programs that pay for themselves many times over.
True. And this is a big deal. But my post shows it could take decades. Always read the small print:
- Two to three years after the livestock or cash, all but one of these programs are raising the incomes of the poorest households by $71 to $202 a year. Since a dollar goes much further in poor countries, that’s actually $250 to $500 a year in purchasing power. that’s big.
- Unfortunately several of the programs cost one, two or even three thousand dollars per person, mainly because of heavy supervision time and other staff expenses.
- This means that, even with the high payoffs every year, the average livestock-plus program will take 18 years or more to break even. That’s not a number I have seen in the news coverage or calls for scaling up these programs.
- But the lower cost programs (including livestock-plus in India and cash-plus in Uganda) are paying off in three to five years. And their impacts are still high.
- Half the expenses are for supervision. What if we dropped this paternalism? If benefits fall by less than half, then the program breaks even much sooner.
- We tried this in Uganda. Compared to cash and training with expensive supervision, cash and training alone had almost identical effects on consumption after a year. Some businesses were more likely to stay open, and profits were a tiny bit higher. But it’s hard to believe supervision passes a cost-benefit test.
The message is clear: charities need to shift the burden of proof to high cost components such as supervision and training. We need to be laser focused on how many years for a program to break even. If three years is possible, why accept 20?
Read the full post. If you’re interested in my numbers, here’s the table I calculated from the Science paper and my own work. If you want purchasing power parity figures use the multipliers (or just times things by three in your head.) Click to expand.