In April, a team of researchers from Brown, Toronto, Northwestern, and UCLA published a fascinating study on the long-term impacts of cash transfers in the United States, looking at impacts on kids whose mothers received transfers from a pension program in the 1910s-1930s. The paper isn’t experimental (the US government probably hadn’t seen the memo on randomized controlled trials in the 1930s), but the methods the authors use seem more likely to understate than overstate the impacts. The full paper is worth a look.
Quick takes:
-
- A study like this from the US is news. We still have plenty to learn about the long-term impacts of short-term transfers in developing countries, but there is no doubt they can be big – e.g. in Uganda, Sri Lanka, and (when combined with therapy) in Liberia (archive link). This is better evidence of long-term impact than we have for most things we spend bajillions of aid dollars on. In developing countries the question now is when and how cash has long-term impacts. In the US, on the other hand, we know next to nothing. The best previous paper we’ve read finds that transfers to Native American families increased schooling and reduced crime among their kids.
-
- It’s a significant piece of data work. This doesn’t entirely come across in the clinical language of the paper. The authors had to manually collect records on some 80,000 participants in the program they study (the Mother’s Pension), and then hand-match these to records from the Social Security Death Master File and to micro-data from the 1940 Census. Good on them for slogging through all that. Also good on the US for having data that can be matched to.
-
- The implied ROI on this program was substantial. By my back-of-envelope calculations, transferring $100 in 1935 dollars, or $1700 when adjusted for inflation, to a poor mother extended the lives of her sons by 0.1-0.3 years. In addition it raised earnings, increased schooling, and reduced malnutrition.
- What did the moms do? We don’t know why treated kids live longer, and there are probably a lot of reasons. The authors find that they had higher earnings, more schooling, and were less malnourished as young adults. Other stuff we can’t measure probably changed as well. And none of this tell us what moms did with the money that had these long-term benefits. Pure conjecture: I would guess some combination of (a) more schooling and (b) better early-childhood nutrition (also known pejoratively as “feeding them for a day”).