Economics & Marginalia: March 8, 2024

Hi all,

Sorry for the lack of links last week: childcare responsibilities trumped reading and musing on the assorted economic and research detritus that was settling on my brain. But what you lost last Friday, you gain today: though I’m going to try and stick to 7 items, as usual, there will be a little more clickthrough content for you to ignore or peruse as is your norm. Some of this is informed by my two days at a conference (from which I’m travelling home right now—if these appear late blame the unpredictable world of the UK train system). It was excellent, though as is often the case, the side-discussions were often the best, mainly because we could follow our interests even away from the agenda. The good thing is that it was organized to provide a lot of space for that sort of side-conversation. Anyway: to the links.

1. It didn’t come up at the conference, but this piece in The Economist turned up virtually everywhere else: a long, critical look at the ethics of randomised control trials (paywalled, but worth it, or worth borrowing). Now, stop groaning: I know we’re all exhausted by the RCT-wars, and yes, there are some missteps in the article. For one thing, while it’s correct to say RCTs are better at answering some questions than others, I don’t like the description of the questions RCTs answer as ‘small’. There is nothing small about understanding how to support women’s political participation; or about how to increase vaccination rates. But I do find people too quick to dismiss questioning of the ethics of RCTs by focusing on the randomization process, which is often defensible as at least as fair as any alternative method of allocating scarce support. The issues are more complex than that. Sometimes we have strong knowledge about the main effect of the intervention being randomized, but want to study secondary effects. Is randomization then the fairest way of allocating it? Perhaps we should allocate based on need for the primary purpose, rather than randomize to learn about a secondary purpose? I’m not suggesting the answer is always to do that, but just that there are many more ethical dilemmas beyond that first filter. Let the record state I like RCTs: I cite them regularly, and have learnt very much from them. But every form of research requires critical evaluation every once in a while, even if many of the criticisms aren’t quite right.
2. Related: Ken Opalo discusses the difference between academic and policy research. This is excellent, and something that I have written about in the past and expect to write about again in the near future. The nature of the policymaking problem is almost always very different to the nature of the research questions academics study, and the process towards ‘better’ looks different for both. Policymaking is often about starting in the wrong place and trying to scramble to a better one in real time, and thus path-dependent; and the path is hemmed in by all sorts of obstacles and difficulties. And Ken is absolutely right to say both that most researchers are bad policymakers, and that much policy in many developing (and developed: look at the UK!) countries is very bad indeed. Highly recommended.
3. It is International Women’s Day, and many outlets have clearly been saving up for it, and there is a lot of good stuff out now. I found this, on Development Impact completely fascinating: Kathleen Beegle and co summarise  a pair of great experiments into how women function in teams, and as leaders. The first experiment is really brilliant: it shows how women have less influence in male-dominated teams (and this is true with both women and men in those teams), but that this effect disappears when women lead those teams. Also excellent is Ashwini Deshpande’s piece on VoxDev about women’s labour force participation in India, and the need to dig very carefully into the data before drawing conclusions. The data is, unsurprisingly, often mismeasuring women’s actual economic activity. And in a similar vein is Tanushree Goyal’s work on the mechanisms by which female political leaders help close gaps in women’s political knowledge and participation.
4. One of the more interesting discussions I had at this conference was about AI and how it is and will shape our work (I already use AI a lot in coding, for example). Three good things on this general topic: first, two pieces by Tim Harford. The first looks at how assistive technologies can both support and sometimes undermine human decision-making, using the terrifying example of Flight 447. The second looks at what we can learn about AI from the development of the spreadsheet (though note Ethan Mollick’s good warning about taking too much succour from this example, which has only limited applicability). And on the first of these two topics, I thought Jessica Hullman’s piece on Stat Modelling was great. Requires some concentration, but she beautifully explains how our most common assessments of AI-supported decision-making may be misleading. (and as a bonus, here is Rohit Krishnan’s piece on Google’s very bad week, which is ironically very good indeed).
5. This is short and excellent: Hannah Ritchie explaining why our fossil fuel reliance is more of a geostrategic risk than our reliance on concentrated sources of the minerals used to in the production of our renewable energy infrastructure.
6. Two good pieces on trade liberalisation and its effects. In the first, Dani Rodrik argues that the backlash against globalisation is not simply irrational anger but at root owes something to our failure to truly understand its distributional effects and remedy then (which I think is more-or-less correct; and worse, our understanding has outstripped our remedial attempts). That is in ProSyn and may need you to register. And secondly, a write-up of research that suggests that trade liberalisation in Brazil led to reduced support for left-wing political candidates by weakening the trade unions.
7. I used to think I had a good memory (and in all seriousness I did and mainly still do). But I have in the last few years met two people whose powers of recall so leave mine in the dust that I was stunned. This, from LitHub, might help me explain them. It’s about Solomon Shereshevsky, a man who effectively remembered everything, in large part because of his synesthesia. Synesthesia is a condition where stimuli affect all of your senses: you smell sounds, taste sights and so on. These multiple stimuli gave him many ways of ‘locating’ and fixing things in his memory; and the article takes this starting point to look further at memory formation. It’s even better than it sounds, not least for the image of cooking pasta with Marvin Gaye (my own association with cooking pasta is always the beginning of The Wind-Up Bird Chronicle, in which the main character opens the book by cooking spaghetti to the sounds of Rossini’s The Thieving Magpie). And on that note…

Have a great weekend, everyone!

R