Economics & Marginalia: November 27, 2023

Hi all,

Today was the last day of my teaching for the year (and indeed, most of next year too), so with any luck the links won’t open with my complaining about perching my gigantic laptop across my knees in the most over-crowded bus known to man anymore for the next several months. But in the spirit of Thanksgiving in the states (which probably means you won’t be reading this till Monday at the earliest), I’ll be grateful for sitting next to an impeccably polite economist reading and correcting a draft paper, which is taking every ounce of my self-discipline not to read over his shoulder (he’s doing it with paper and pencil, too; I have a student who uses a ReMarkable and she let me have a go on it today and I regret to inform you I will soon be spending money that should go on food—or wine—on one). It almost doesn’t matter what sub-field he’s working in; it’ll likely be fascinating to me anyway. Economics is great, and in all seriousness, the message of most my teaching over the term has been that the world is a bit of a mess, but without economics it’s hard to fully understand why and how it’s a mess, and without economics it would likely be in even worse shape. So, on to the links:

1. Economics is great, but academia could definitely stand a few improvements. As ever, Andrew Gelman has something to say here. First, a piece which considers Dorothy Bishop’s [quick aside: I watch too much Marvel; I read the post all the way through mentally substituting ‘Kate Bishop’ for Dorothy Bishop without realizing it] proposal for an ‘academic fraud police force’, that reduces the need for unpaid detectives like the Data Colada team, and systematizes the search for fraud. At first thought it’s an attractive idea, but he very quickly homes in on the exact aspect that causes me discomfort: what kind of people would self-select into this ‘police force’? I genuinely don’t know but also feel strongly that without having a good sense of that we might do more damage than good (especially since the authority inherent in the ‘policing’ function can be misused more effectively than the amateur detective can manage, even though there are questions about who selects into that role too). And he also comments (favourably) on this article by Nina Strohminger and Olúfémi Táíwò on selection bias in what actually gets researched: another problem for science, one I can’t think of a good solution for. The point on the attention given to nudges crowding out research into harder, more expensive, but probably more effective solutions (I call them ‘shoves’) is well-taken.

2. I use ChatGPT a lot. It has dramatically improved my coding, and saved me many hours of trial and error; but I am neither filled with abject terror that AI is going to become some kind of dictator or cause our extinction (neither The Matrix or the Supreme Intelligence in Captain Marvel count as horror movies to me) nor am I giddy with the prospect of it completely revolutionizing the economy. Perhaps I’m being short-sighted or foolish (and given my recent record of cricket prediction, perhaps this should worry you). So I was no more than moderately interested in the Sam Altman saga. For those of you blissfully unaware, the OpenAI board, which oversees ChatGPT, ousted its CEO for… reasons. But through a combination of powerful friends, personal charisma and the sheer lack of transparency of that decision, he’s already returned to the role, with a new board in place. My interest in this is less on the personal dramas (Planet Money have the story, and transcript) and more in what this tells us about organizational behaviour. All bureaucracies need a governance structure, but this governance structure has limits to both power and legitimacy; it can only survive to the extent that it works within those limits, which means that organizational accountability always operates within bounds. This is true in both the public and private sectors, and has quite important implications for how much power any individual organization can responsibly have.

3. I’m glad to see ‘L’Affaire Gino’ has not stopped Data Colada publishing altogether: Uri Simonsohn has a post here on pre-registration.

4. This is one of the best things Tim Harford has written in ages, and not just for this line: “We already know that you can lose a string of British parliamentary elections while seeing your policies embraced by the political mainstream; Nigel Farage taught us that.” In it, he lays out his ‘loony’ policy platforms that might be embraced soon enough, and funnily enough one of them ‘abolish all planning permissions’ is very similar to an idea I proposed to a Director in the Cabinet Office once, in a discussion about how to implement ‘levelling up’. My proposal was that all planning permissions should change from default no to default yes, with cost and onus on opponents to new developments to make the case for them, and it should apply to everything. If the Council wants to turn Trafalgar Square into low-cost housing, it should start from ‘yes’, with a no campaign having to make the case for why this is a bad idea. If a developer wants to build a high-rise in the middle of a street of terraced houses, the same. And the bar for no should be high, just as it’s difficult to get a yes now. Suffice to say, my proposal did not make the shortlist.

5. This is a nice attempt from Ted Miguel and co to tease out how the benefits of deworming may be inter-generational. Pinning down why this might be the case is difficult, but they do a good job of laying out the evidence.

6. Ricardo Hausman cycles through metaphors for industrial policy before settling on the immune system; his explanation is long, but very good and worth reading. I particularly enjoyed his closing summation: “The fact that industrial policy can backfire does not imply that countries should eschew it. Learning how to deploy these interventions is as important for a well-functioning economy as developing sound education and health policy, and a failure to do so would likewise carry an unacceptable social cost.”

7. You know you’ve become old when your childhood memories belong in a museum. When I was very small, my parents used to work every other Saturday (as was standard at that time in Hong Kong); I would usually accompany my dad to work, since my mother’s work involved several layers of security that I had a habit of trying to circumvent. At that time, most of the desks in my dad’s office still had typewriters, and he’d plonk me in front of one, load it with paper and let me whack away at the keyboard (now I think of it, this may be where my incredibly loud typing, which once drove Nick Lea to switch my keyboard, comes from). Anyway—I’ve always loved typewriters, and even now think of novelists as people who sit behind them. LitHub have a series of photos of the most iconic typewriters in history (sadly they omit the ones from my father’s office, and the Wellington 4000—with feathertouch control, a reference my brother may be the only person on the planet to get). This gave me more joy than it is decent to admit.

Have a great week, everyone!

R