Economics & Marginalia: May, 21

Hi all,

Brood X cicadas are remarkable creatures. They hibernate underground for 17 years, foraging on the roots of trees under cover of darkness – sometimes killing them in the process – oblivious to the changes wrought on the world above them. After nearly two decades they build little mud tunnels to the surface to moult into their adult form in huge numbers – estimates suggest billions emerged in the US this week. They emerged to a changed world. Apple’s portable music product went from looking like this to this (hands up; how many of you had forgotten that we used to have to press actual buttons on iPods back in the day) and a Senator from Georgia went from looking like this to looking like this. But not everything changed. In 2004, you would have been forgiven for identifying this man as the greatest basketball player in the world  - at worst you’d have been two years early. In 2021, asked to pick one player to play for your life, if LeBron James isn’t your pick, you may as well head to the crossroads with your guitar and hope you’ll at least get some lessons in return. Every year when the playoffs roll around I run short of superlatives for him; and by the way things have started, I should look for my thesaurus again now.

1. I am an evangelist for Cost-Benefit Analyses. I think, done well, they’re an immensely powerful tool for public policy, and done badly they’re still better than most alternatives, at least providing a transparent structure against which to understand and question the choices made. There is a real art to doing them well, though. You need to pay enormous attention to the details, but also need to confront your own uncertainties head on. The best cost-benefit analyses tell you under which circumstances the decision would be reversed. Dietrich Vollrath gives a masterclass in both the attention to detail needed for a difficult CBA and in how to communicate the limits of one’s own certainty in the analysis. He examines the Biden administration’s infrastructure plan and sets out why he supports it, and why he might be wrong, all using the simple framework of a net present value calculation. Everyone who does these in public policy should read this.  

2. Doing a CBA really well requires knowing when the data you have are bad, and how to work around their weaknesses (and indeed, when to avoid them altogether and take a different approach or use extreme values to try and show the range of plausible outcomes). There’s no shortage of examples of bad data in development, but I really liked this pair of deep dives into the construction of a completely different data set: metrics for the quality of defence played by individual players and teams in basketball. Defence is notoriously difficult to measure, but the alternative of going with ‘the eye test’ is even worse than using no data at all. Better to understand the ways the data is weak and to use it with your eyes wide open.

3. In the grand scheme of things, data on basketball players doesn’t matter that much (well, it does to me, but I can see how it’s perhaps a little less important to normal people). Tim Harford looks at a rather more pressing case: data collected on Covid treatments in the UK to make the case for much more systematic investment in the data infrastructure in both rich and poor countries alike. He points out virtually every great global crisis has demonstrated that we just don’t have the data we need to make public policy well, and to do it quickly. I am fully on this bandwagon, and you should be too.

4. I really loved this symposium of education essays organised by my colleagues at CGD, in response to Girin Beeharry’s Manifesto for Global Education. It brings together a huge range of thinkers; if you’re not sure which ones to read, Susannah Hares’ thread is a good place to start.

5. NPR on how badly people calculate risk, taking the example of the Johnson & Johnson Covid vaccine as their starting point. They’re careful not to overegg the point: if we were really that terrible at calculating risk, we probably wouldn’t have survived this long. Nevertheless, as the kind of things that used to kill us most often become more and more manageable, the ones that remain become relatively more complex to respond to; and how we understand the risks of different actions is a big part of that (transcript).

6. Branko Milanovic’s layman’s guide to inequality is – as you would expect – largely brilliant. However, there are two points on which I’d disagree with him, both in the section considering the implications of trends in inequality. The first is the statement that increasing global inequality will increase migration; this may possibly be true, but it’s also the case that – as Michael Clemens and Mariapia Mendola suggest – increasing incomes among the poor will also increase migration (my bet is that this will be true even when such increases decrease global inequality). Why? The cost of migrating is, for most, the more salient concern in migration choices than the exact magnitude of the gap in incomes at destination and home countries. And secondly, he suggests that African convergence has stalled; my colleague Justin Sandefur may beg to differ.

7. “Proceedings of the Second International Workshop on Nude Mice” may be the greatest book title of all time, but there are some phenomenal contenders in this list compiled by LitHub; I note that one David Evans has written “Does God Ever Speak through Cats?” Such is his productivity and the breadth of his interests that I wouldn’t be wholly surprised if it was my erstwhile colleague and development economist extraordinaire… The only discovery of the week that made me happier was that there is now an R package to make all your graphs look like Taylor Swift albums.

Have a great weekend, everyone!

R