Week of November 5, 2018

1. Household Finance: One of the trips keeping me busy was to Mexico City for the PRONAFIM conference. Here's a video version of my current thinking on household finance, in Spanish.  
Of course, one of the key questions in household finance is to what extent a household is a household. I've had a hard time not thinking about this recent paper from Afzal et al, which through a series of "lab in the field" experiments, shows there are a lot of schisms in the household. Let me just quote from the abstract: "Subjects are often no better at guessing their spouse's preferences than those of a stranger, and many subjects disregard what they believe or know about others' preferences when assigning them a consumption bundle." Is there some explanation there for the puzzle in the Graphic of the Week (see below?). 
In the household finance realm I often pick on financial literacy--specifically as a bellweather for evidenced-based policy (if money is going into financial literacy, evidence isn't making a dent on policy). Here's some interesting new evidence on financial literacy and why it doesn't seem to work, from Carpena and Zia. They are looking for what parts of financial education might affect behavior, and find attitudes matter more than awareness or numeracy. I feel like that connects to this new paper from Gine and Goldberg documenting endowment effects in account choice in Malawi, and that the endowment effect can be overcome with experience, but maybe not.      

2. Inequality: Teaching a class on wealth inequality and policy makes anything on the topic grab my attention just a bit more. And there is a lot out there. On the downside, there's a lot out there and my attention is drawn to all of it. Here's a handy Twitter thread guide (and in a perhaps easier to follow/read format) to the global inequality literature that I found very helpful. Here's a new paper from Ayyagari, Demirguc-Kunt and Maksimovic calling into question the idea that a group of "star" firms are pulling away from others and boosting inequality. You probably already know about this, but the Chetty team has published their Opportunity Atlas. And here's a recent paper from Card et al. on the role of school quality in transmitting economic inequality in the US during the 20th century (in digest form here). 

3. Our Algorithmic Overlords: Nothing particular profound here but I couldn't resist pairing these two pieces together: a) "China’s brightest children are being recruited to develop AI ‘killer bots’" and b) A list of artificial intelligence programs that do "what their creators specify, not what they mean." I suppose since the actions of the AI programs sound a lot like children trying to annoy their parents, China's approach seems optimal? 

4. Migration: I mentioned too much travel as one of the reasons for the occasional nature of the faiV lately. One of those trips was to Vienna to present at the OSCE's Development and Migration meeting. It was an excuse to go back and re-read a paper I co-wrote with Michael Clemens on "Migration as a Household Finance Strategy." It's good, as you should expect from something that Michael led. I'm linking it here because it lays out a research agenda that's still valid. And so I can also link this video that came out of it. I'm pretty sure it's the video that got me the invitation to Vienna. 
Meanwhile, here's a new paper from Michael debunking brain drain (again; now read that like Eliza Doolittle) and a related article from Harvard Political Review. Or try the even shorter version, a tweet from Michael with this useful fact: "In the 71 countries that grew to middle-income or higher between 1960 and 2013, 67 had a concurrent rise in the emigrant share." 
Here's a recent story from the New York Times about how Uganda is integrating refugees successfully. And here's data on the crossing from Libya to Europe--1 in 5 die or go missing. hel Glennerster examines several cases where evidence led to scaling up n?

5. Evidence-Based Policy: Who isn't interested in making research more useful to policymakers and influencing policy with research? I mean, other than tenure committees. For everyone else, the Evidence in Practice project at Yale SOM has a report on their two years of work on how evidence can be better integrated into policy and practice. Check out particularly, page 32 and the "currency of exchange" for each of the groups involved in evidence-based policy development and implementation. Maybe print it and hang it on your office wall. (Full Disclosure: I participated in one of the Evidence in Practice workshops). The recommendations align pretty well with Oxfam's lessons for influencing policy, summarized in a post by David Evans at the Development Impact Blog. In fact, point 5 could reasonably be a summary of the Evidence in Practice report (I think that's a good thing. That's good right?). But you should still print and hang page 32 because you won't be able to do point 5 without it.    

 I  was chatting with Lore Vandewalle earlier this week (she's visiting at  NYU-Wagner this fall), talking about savings behavior and that led me to  look back at  an experiment she had run in India with Vincent Somville   where some people were paid cash and others were paid via deposit into a  savings account. While I had read the paper, it didn't really strike me  at the time how similar the results were to another savings  encouragement experiment-- Kast, Meier and Pomeranz in Chile .  So I show them side-by-side here. The interesting thing to me is that  balances for the treatment group rise quickly but plateau at a fairly  low level (e.g. in India it's about a week of food expenditures  according to Lore). It's consistent with using these accounts to manage  liquidity needs (buffering day-to-day or week-to-week volatility) but  not with precautionary or investment savings. But that doesn't explain  why the control groups also plateau, just at a lower level. Is there  another story I'm not thinking of? 

I was chatting with Lore Vandewalle earlier this week (she's visiting at NYU-Wagner this fall), talking about savings behavior and that led me to look back at an experiment she had run in India with Vincent Somville where some people were paid cash and others were paid via deposit into a savings account. While I had read the paper, it didn't really strike me at the time how similar the results were to another savings encouragement experiment--Kast, Meier and Pomeranz in Chile. So I show them side-by-side here. The interesting thing to me is that balances for the treatment group rise quickly but plateau at a fairly low level (e.g. in India it's about a week of food expenditures according to Lore). It's consistent with using these accounts to manage liquidity needs (buffering day-to-day or week-to-week volatility) but not with precautionary or investment savings. But that doesn't explain why the control groups also plateau, just at a lower level. Is there another story I'm not thinking of? 

Return to the Weekly faiV