Economics & Marginalia: November 5, 2021

Hi all,

Here I was thinking I’d be able to start the links off with some feel-good Sri Lankan cricket highlights after we displayed some extremely rare competence in the Twenty20 World Cup. Instead it’s hard to lead with anything but how completely appalling every fresh revelation in the fallout from Yorkshire’s grubby and disgusting attempt to pretend systematic racism throughout the club was just ‘banter’ and nothing to apologise for. Racial slurs were used; an ex-England captain was alleged (by two people now) to have “jokingly” complained about there being too many of ‘you lot’ (that’s Asians, everyone) on the team; the chairman has resigned in fury that the club won’t do anything about it, saying that most of the coaching and management would need to be replaced to solve the problem. Michael Vaughan denies he made racist comments on the field, but you don’t need hearsay when you have his twitter account (here’s exhibit 2 for if that’s not enough for you). I’m pleased that the current England captain seems to be a genuinely good man who thinks this is as disgusting as the rest of us, but good god it’s been a depressing week to ben English cricket fan (even one who supports Sri Lanka).

1. Jason Kerwin wrote a very thoughtful, very interesting piece on why ‘nothing scales’ in economics, bemoaning the fact that so little of the research published in contemporary microeconomics works when you extend it beyond the small subgroup of people it was originally tested on, or when it’s implemented by different delivery agencies. I encourage everyone to read it—it is a helpful account of ‘heterogeneity of treatment effects’ (that is, that different kinds of people respond to the same policy or programme in different ways). But I also think there’s a deeper issue here: it reflects the poverty of ambition in microeconomics these days, driven by the very narrow focus of what gets published. The kinds of intervention Jason is complaining about the prospects for scaling are incredibly precisely-defined, sometimes just really small tweaks on top of big existing systems, like a new way of reminding people to pay their taxes. These small tweaks rarely scale. What does scale is building the system well in the first place. But building a good system shouldn’t be thought of as a single evaluable policy choice, but an accumulation of many choices, with path dependency and interactions… and consequently many ways of building something that works. So yes, specific, randomizable and publishable tweaks to a health system may not scale… but a functioning health system scales very well indeed. It’s just harder to define, and so economics doesn’t engage with that question much anymore.

2. Speaking of tax systems, a nice piece from Nishant Yonkan and co-authors (including Branko Milanovic) looks at the discrepancies between survey and tax data to investigate for which populations these data sources disagree and why this disagreement arises. It won’t surprise many to discover that it’s for the top 1% of earners that the discrepancy is biggest, or that it’s non-labour income that drives the divergence; it’s also quite neat how they show that changes in tax rules drive this effect. Documenting the ways in which the data are imperfect generates some of the most revealing information about inequality.

3. This is a really excellent note from Paddy Carter, looking at the economics of development finance. It’s an accessible (but not dumbed down) treatment of the economics of economic development, the specific role of development finance within that, and the circumstances under which it works best. My favourite parts are the sections on knowledge generation and spillovers in production networks, which do a lot of the lifting in terms of motivating the role of DFIs.

4. Hannah Ritchie on the importance of optimism and hope in driving change, starting from the perspective that constantly reinforcing that we are on the brink of catastrophe tends to disempower and disengage people, rather than rally them to fight for reform. It also reinforces my suspicion that 24/7 news is a net bad, and tends to reduce rather than increase engagement with news and issues (though it does increase engagement with news outlets, which is not the same thing).

5. This is a great, great episode of Planet Money, about the Federal Reserve’s “beige book” which provides a qualitative look behind the statistics on the economy it produces in such quantity. This humanizing approach to economics is important (and has a long history as anyone who has read much classical economics will immediately recognise). It takes a very emotional turn towards the end (transcript).

6. In a few recent papers I’ve used boxplots—indeed I was stopped from including an additional one in a recent paper by a kindly internal reviewer. I really like them, but this by Nick Desbarats (which I found via David McKenzie’s great links weekly) argues that they are too hard to interpret, sometimes misleading and should be discarded. I started with a strong feeling that I was going to disagree with it… but by the end I found myself convinced to use jittered strip plots instead almost every time I feel the urge to do a boxplot.

7. And finally, since I started with a massive rant, let me finish on something innocuous and amusing: the nominees for the year’s oddest book title are out and the early front-runner appears to be Is Superman Circumcised?, though I rather prefer Miss, I don’t Give a Sh*t, a book which evokes some of my early school days rather well. And on that revealing note…

Have a great weekend, everyone!

R