Economics & Marginalia: October, 22 2021

Hi all,

Normally, I complain of sleeplessness because my son has decided that sleep is for weaklings and that 3am is prime babbling, playing and general mischief-making time. But come autumn, I lose sleep for a different reason: living in the wrong time zone to watch LeBron James continue to lay the smack down on father time and Steph Curry make Paul George look an even bigger fool than he makes himself (no-one is ever going to forget Playoff P, Paul). Now that the NBA season is back, I can return to the time-honoured pursuit of fathers awake in the small hours: watching sports on mute while desperately trying to get a baby to listen to reason. If in the next few weeks, I resemble a zombie, you can split the blame between Luca and LeBron.

1. It’s hard not to lead with it when your colleagues suggest replacing a major arm of the international development architecture with three lines of Stata code, but that’s not far off what Justin Sandefur and Arvind Subramanian do in Project Syndicate. I get the strong sense that much of the outrage on Twitter has come from people who have read the headline and not the article, because while I don’t agree with every sentence, the final conclusion—that the Bank should focus on getting it’s most concessional resources to the poorest places that have fewest alternatives—is hard to argue with. I have a new paper coming out next week which extends a version of this argument to all donors. There are limitations (you should care about how your flows of money affect political equilibria; and how effectively they can be used), but as a starting point from which to deviate, it seems right. Where I think it’s more reasonable to disagree is whether the Bank’s data function has been fatally undermined by concerns with the Doing Business indicator. I would suggest not. Reasonable people can disagree about this. In any case, I think we can be pretty sure that Justin is off the Bank’s institutional Christmas card list for the time being.
2. Meanwhile, when my next paper is published, I might be off the OECD’s. In it I argue that a decade-plus of working on its ‘aid effectiveness’ agenda has achieved nearly nothing (link to follow next week). But it seems I may have been optimistic about the DAC’s impact—the true effect might be negative. Euan Ritchie, Anthony McDonnell and I have taken note of their proposals for doubling the ODA scoring of donated vaccines, some of which are so close to expiry as to be functionally useless (the equivalent of dumping some off chicken on a food bank), and concluded that it actually incentivises bad behaviour, not generosity. Add this to the way it scores loans and it’s ludicrously generous terms for ODA eligibility and they aren’t looking great right now.
3. I talk a lot about being a new father (and have had many excellent conversations about marginal vs. average effects of parenting choices, and why most of us worry far too much about them), but I’m extremely glad I’m not like the economist-parents covered in this Freakonomics podcast (which I discovered via David McKenzie’s great Friday links). I felt a lot of simpatico with Steven Levitt’s responses until he revealed he spent most of his younger childrens’ childhood golfing, which… I suppose he might derive a lot more utility from golf than I do (in fact, I’m certain of it), but it still seems like an odd choice.
4. A good grounding in statistics is important for almost everything. This excellent piece by Tom Chivers, about why the fact that almost all terrorist incidents in the UK are perpetrated by people already known to the Prevent programme is almost inevitable, is excellent. At root, this is all about Type 1 and Type 2 errors, and the relative costs thereof; and he correctly identifies that the best solution is to use something that has lower error rates, and Prevent ain’t it.
5. More coverage of last week’s econ Nobel: Tim Harford and Jorn-Steffen Pischke (of Angrist & Pischke) both weigh in, both worth reading. And while we’re talking metrics, here’s David McKenzie again, on the perils of leave-one-out instruments (I am a general IV sceptic, so this really speaks to me. There are a few exceptions of course).
6. I really liked this piece by Natalie Bau and co-authors on the effects of the pandemic and policies to respond to it had on women in India, with a particular focus on their mental health. Spoiler: it was not good. This kind of work is really important—there is such a dearth of good, well-collected data on these sorts of effect.
7. Lastly, it was the 40^th birthday of a close friend of mine recently, and I was called upon to stage a recreation of his favourite scene from his favourite movie: Jules and Vincent’s interrogation of Brett in Pulp Fiction (my son played Brett). It sent me down a Samuel L. Jackson rabbit hole (is there any better kind) and I (re)discovered two of the best things on the internet. First is new: a biography which asserts that his use of his favourite word (it starts maternally, but ends carnally) was a technique he developed to keep his chronic stutter under control. That might be why he says it with such immense relish, as this supercut of his every recorded use of the word demonstrates (I’ve linked to this one before, but—like his use of the word—it bears repetition).

Have a great weekend, everyone!

R