Economics & Marginalia: June 24, 2022

Hi all,

There’s a mild statistical bent to all my favourite pastimes. Cricket is the most numbers-obsessed sport in the world; everything is counted, put into averages, investigated under all contingencies (sure, Mahela Jayawardene averages 59.72 at home, but shouldn’t we be concerned his average is a full 17 runs lower away?). Basketball is has evolved beyond the mere counting and joined the spatial analysis revolution—and hey, Amos Tversky has a paper on it! And electoral politics has spawned a ridiculous number of competing surveys, models and analyses. Still, sometimes the ‘eye test’ is undefeated. Statistics might tell you he was a tier below his illustrious teammates, but after fifteen minutes watching him you know that VVS Laxman is Very Very Special indeed; no-one ever needed to look at his RAPTOR score to know LeBron James is a genius; and there’s nothing a 30% swing to the Liberal Democrats can tell you that a Tory candidate locking herself into a dance studio and Boris Johnson giving interviews from Rwanda doesn’t say ten times as eloquently.

1. How should we define success in an intervention? Is it one that has large average effects? One that remains good when you scale it up nationally? One whose effects last a long time? One which makes progress against some specific difficult-to-achieve outcome? Or something that helps everyone it’s intended to target? It’s hard to get an intervention that satisfies any of these criteria, let alone all of them; this piece by Julie Buhl-Wiggers and co-authors looks at a specific education intervention—one that is pretty successful on many metrics, but has wildly different effects across the sample of people it was tested over. Most pressingly, it’s really hard to understand why the intervention has such different effects on different people. Increasingly, I’m struck by how important the noise around the signals we focus is, and how little—despite Fisher Black—we really understand it.
2. There’s something incredibly satisfying about the fact that Harrod (1936) accurately characterised the pricing behaviour of Harrods (est. 1849): specifically, fleece the rich. He predicted that as income increases, people become less sensitive to prices; a nice new paper by Kunal Sangani (found via Marginal Revolution) quantifies this: he finds that richer households systematically pay higher markups in the goods they buy, and that stores that specialise in higher income buyers have even higher markups than luxury brands in stores that have customers from across the income spectrum. This is a very cool paper, providing evidence for something we’ve often suspected, but also opening up new lines of inquiry.  
3. I found that paper via Tyler Cowen; in extremely tangentially related news, someone has used an AI to simulate a conversation between Tyler and Oscar the Grouch. This is a clear failure of AI: it reads much more like a conversation between Tyler and Tyler as Oscar the Grouch.
4. Andres Velasco takes aim at the deglobalization proponents, suggesting that neither the data nor the tea leaves suggest that globalization is on its last legs. Instead it is the deglobalizers who will be forced to find new work—a task that will be easier in a globalized world.
5. This is amazing: Scott Cunningham (whose book you should really be reading if you’re interested in econometrics) has a long podcast interview with Susan Athey; on the website it’s preceded by what is very humbly described a ‘Episode Details’, but is really more like an essay and then a transcript of the podcast. Athey is a properly impressive economist, not just in her outputs but in the way she thinks. Highly recommended.  
6. The FT’s summer reading list (from Martin Wolf) is typically superb, and features both Stefan Dercon’s Gambling on Development and Chris Blattman’s Why We Fight. Also from the FT: Tim Harford on the roots of altruism.
7. Thank you for all the recommendations of detective shows after my plea a couple of weeks ago: we are now about 7 episodes into Vera and have Inspector Montalbano and Endeavour queued up. My reciprocal recommendation is Ms. Marvel. I never expected American superheroes to debate their favourite SRK films (or to get it so wrong) over to feature an extended dance sequence, but it manages to do both while still being entertaining on its own terms. Growing up I cannot remember an equivalent movie or tv show for an Asian child. And lastly, the best thread of the week: about an Egyptian civil servant 4,000 years ago, and the contrasts he displays to his Odysseus and Achilles, whose stories are contemporary to him. As Dan Honig might say, Mission Driven Bureaucrats have always been around.

Have a great weekend, everyone!

R