1. Effective Altruism: It's the right time of year to be talking about charitable giving--most US-based charities take in about 50 percent of their annual revenue during the month of December. Here is GiveWell's list of recommended charities this year (NB: I'm on the board of GiveWell). Jennifer Rubinstein has a new essay about the "hidden curriculum" of effective altruism, as seen in Peter Singer's and Will MacAskill's books. There's always a hidden curriculum isn't there?
2. Evidence-Based Policy: Effective Altruism shares a curriculum, hidden or not, with evidence-based policy. At Stanford Social Innovation Review, Jennifer Brooks of the Gates Foundation has a post making the case for evidence-based decision-making. I suspect that prior to November 8th most readers of this newsletter wouldn't have thought the case needed to be made. One of Brooks' key points is the need for better data from rigorous evaluations so that there is evidence not just on effectiveness of a particular program, but information on how to improve other programs' performance. That just so happens to be one of the points in the conclusion to my shortly to be available book on the use of RCTs in development economics. You're running out of time to buy a copy for a holiday gift. It won't arrive until January regardless, but it's the thought that counts right? Oh wait--the whole point of effective altruism and evidence-based policy is that it's not the thought that counts.
3. African Bank Failures: It doesn't make the global news, but there have been a number of bank failures in sub-Saharan Africa in the last few months: Kenya, Mozambique, Zambia, and Uganda have all closed banks since the beginning of October. At FSDAfrica, Mark Napier looks at whether there's a trend to be concerned about. He forecasts a "rocky ride" for African banks, and lots of work for bank regulators, in 2017.
4. Cash, Cash, Cash (and Targeting): I've been trying to keep away from basic income and cash transfers for a few weeks. But some things are starting to build up. There's this new thing called the Economic Security Project (I'm not quite sure what it is) that seems to be organizing support for testing basic income in the US (but also making grants?). Here's their "Statement of Belief" with some notable signatories. Here's Rachel Schneider and Jennifer Tescher explaining their support, in part drawing on the US Financial Diaries research (yes, I'm biased to any argument from USFD research). And completely independently, here's Brown, Ravallion and de Walle looking at proxy-means testing approaches to target assistance to poor households. They find there are some methodological tweaks to standard approaches that would improve targeting, but that basic income performs as well at reducing poverty as any of the improved targeting approaches (but note that even still the best outcome is reducing poverty by 25%). And here's an Evansian review of Health CCTs if you haven't seen it yet--lots of good takeaways on program and study design.
5. Machine Learning: Back to evidence-based policy making. Well, perhaps I should say data-based policy implementation. Kleinberg, Ludwig, and the elusive Sendhil Mullainathan have a piece in Harvard Business Review--essentially a guide for policy makers on the possibilities and pitfalls of using machine learning for things like targeting. They discuss examples like setting bail and hiring police officers, but also how easy it is to be misguided by an algorithm if you don't understand it. I'm reminded of Matt Levine's phrase about algorithms being like genies, always taking instructions just a bit too literally. I've seen Sendhil present some of this work on how machine learning can be useful in improving targeting, but also dangerous by directing attention away from existing, undetected errors in the targeting process. Hopefully there will be papers available soon.
--------------------
Bonus Follow-Up: Last week I included some links to pieces that were widely circulating and getting a lot of attention with the idea that you didn't need such links from the faiV. I thought some of you might be interested in some data on that topic: David Roodman's blog on Worm Wars was 4th, Raj Chetty's data on mobility was 6th, and Gabriel Zucman's data on wages was 11th out of 20 links.
There's a crisis of opioid addiction in the US, mostly concentrated in rural areas, serious enough that it's showing up in life expectancy statistics. But it's also affecting newborns, with cases of addicted newborns increasing rapidly enough to strain hospital budgets. Of note, in the coverage of this issue there is no mention of "super predators." I'm not sure whether to be encouraged or discouraged by that. Source: JAMA Pediatrics
