Economics & Marginalia: July 8, 2022

Hi all,

What are we supposed to do with a week like that? I usually poke (sometimes) gentle fun at the state of English politics during our now routine scandals, but satire seems hopelessly inadequate after a week of the most comical, incompetent, ridiculous political machinations in living memory. Boris Johnson has—finally!—agreed  to resign as Prime Minister (though if anyone believes a word he says after being caught out on [*deep breath*] Patterson, partygate, Pincher, peerages and (wall)paper, I have a bridge to sell to you), but let no-one say he did it with dignity. Instead, he clung on as the resignation tracker shot up like particularly virulent strain of Covid, going so far as to appear before a Parliamentary committee pretending that was openly mocking him, pretending there was nothing untoward happening even as the 44^th member of his Government had resigned in protest at his mendacity and incompetence. He appointed a Secretary of State for Education; she resigned within 24 hours, saying she couldn’t work for him. He appointed a Chancellor; he used official notepaper to tell him to quit (quite improperly, by the way, but I guess he thought it might be the only chance he would get to use it). Somehow even George Costanza behaved with more decorum in the same situation; and even John Belushi showed more self-awareness. Clownfall, indeed.

1. I’m going to be totally honest: so gripping was this week’s political omnishambles, I am seriously behind on my reading (nothing I could read was more entertaining than the Benny Hill music being constantly blasted right outside our office windows as Tory MPs were being interviewed). But one of the more thought-provoking pieces I did read was this, by Dani Rodrik in Project Syndicate: an argument that there is a gathering coalition around a new economic model, built on ‘production, work and localism’, rather than the globalist, consumerist finanancialism of the last few decades. Granted, one of the thoughts it provoked was “this is wrong, of course”, but nevertheless it’s worth reading. I do not see in the economy, in policy anyway, nor even in the tea leaves any real sign that localism is replacing globalism; perhaps there is more to the other two shifts, but I have my doubts there too.
2. A new VoxDev write-up finds that people take up preventative healthcare more when they are given an incentive that they lose if they don’t attend a healthcare appointment, rather than the promise of a guaranteed reward if they do attend (the size of the incentive doesn’t change, so you would expect there to be no difference between behaviours under the two systems). The authors (Emily Beam and others) ascribe this result to loss aversion—the phenomenon where people dislike losing things they already have more than they like gaining the same thing if they don’t have it—but I have two question marks. Firstly, the difference between the effects is pretty small: the loss framing gets 30% attendance and the gains one gets 28%, which matters, but isn’t exactly night and day. And secondly, a new paper, counting Colin Camerer among the co-authors, finds that most people aren’t actually loss-averse, at least in the US. Perhaps the small effect is driven by those people who are?
3. Branko on Chinese income and inequality during Covid-ravaged 2020, using the latest release of the Chinese national statistics.
4. Every week I could fill these links with stories about the many, many failures of the academic pursuit of science, and specifically the way its incentives reward poor practice, and create an equilibrium in which any given research finding has a probability of being wrong many multiples higher than appears at first blush. Two new entries into this sadly extensive genre: first, a new paper in which Vincent Arel-Bundock (creator of many of the greatest R packages in the known universe), Ryan Briggs finds that academic political science is *wildly* underpowered. Essentially this means that, given the best (based on multiple studies) guess of how big the effects most papers are studying are, the papers are using far too little information to investigate those effects. As a result they are unlikely to detect the true effect size when it shows up in their data, but will falsely find massive effect sizes when they randomly show up, as they are prone to do in small samples (think that it’s not unlikely to get only 6s rolling a fair die if you roll only three times, but it’s really unlikely if you roll it a thousand times times). And Felix Chopra and co-authors use a survey of academic economists to document the widespread beliefs that studies that finds no effects from an intervention or event are much less likely to be published.
5. And while still on the new research theme: a paper which finds that moderate-sized cash transfers had very limited effects on consumption and no, or negative, effects on well-being, with somewhat larger transfers doing no better. This is a paper that bucks a thousand priors—and they document that, too.
6. Via David McKenzie’s super links: Cardiff Garcia (still one of the greatest names ever) interviews Leonard Wantchekon about his life and career.
7. Lastly, James Caan died this week, which brought on intense nostalgia for his Sonny Corleone in The Godfather, to say nothing of Bottle Rocket, Thief, or a million other things he improved. The Ringer have a lovely obituary, and there are some fantastic Caan stories on twitter going around. Watch his movies, it will improve your weekend.

Have a great weekend, everyone!

R