Economics & Marginalia: March 11, 2022

Hi all,

The news is just unrelentingly grim, isn’t it? One of the worst things about the social media world is how hard it is to escape as well; once you’ve read the news and whatever analyses you trust and just want to scroll through Jeffrey Woolridge dropping bombs on economics twitter, or Rachael Meager starting fights in empty rooms (to be fair, this particular tweet might be better characterised as starting a fight in an incredibly crowded room), but to get there you need to sift through the 30-piece threads on “Why my hitherto unsuspected expertise in military strategy, let me explain why Putin will fail. Hint: it’s all about the spark plugs used in their tank engines”; and even if you avoid those (and one of the reasons I follow relatively few people is that I want to have faith in the retweets that show up in my feed) you still have to process the news that India have accidentally fired a rocket at a potential nuclear antagonist. I mean, how?! Did someone sit on the control panel? The role of human error in history is probably larger than we ever realise with our post-hoc analyses, and right now that is making me very nervous indeed. Onto the economics, where error is sadly no less common.

1. So errors eh? I like the way Vincent Geloso just casually suggests that a chunk of the famous Piketty-Saez result about the reduction of inequality over the Second World War is down to a typo (happens to the best of us, guys); but do not let that distract you from how good and careful this paper is: he really digs into the data, really looks at all the code and as a result substantially revises the policy implications of an extremely influential piece of work. The thread is excellent, too. Make no mistake, though, this is not a story about science not working. This is the opposite: careful data work refining what we thought we knew, and ultimately reinforcing the unglamorous but important role of basic economic changes in understanding the world.
2. On a related note about how science proceeds, and its potential pitfalls: two very good pieces, one new and one newly in the news. First, Jessica Hullman on the various choices experimenters take in designing their experiments that can affect the results of their eventual research; this is one of those really hard grey areas to navigate. We design our experiments carefully, and test parts because we want to make sure we’re really testing the thing we care about; but you can never rule out we’re just narrowing down the path until we just the right combination of things that returns that magical ‘statistically significant result’. Even honest behaviour can undermine science. And relatedly, Nick Huntington-Klein and co-authors’ wonderful paper that shows that even innocuous data choices can have big effects on the result has just won an award. Science is hard, but so is knowing how to judge it, even when it’s carefully done.
3. Via Paddy Carter, I read this long but fascinating piece by Jean Dreze disputing the interpretation of the results of a famous and influential RCT. It’s difficult to summarise, and I highly recommend you read the whole thing. But the gist is really this: in analysing the results of the experiment, the authors focused on some matters to the exclusion of others, and this may provide a distorted assessment of the overall effect of the intervention; indeed it’s possible that the key result can be explained in other ways. I am not qualified to judge which reading of the results is right, but I will say this: almost all research is at risk of this kind of Rashomon narrative. We always have to select and arrange facts and analyses; the data never just speak for themselves. This kind of dialogue, between extremely smart and well informed people, is part of how our understanding advances.
4. I am a huge sucker for any paper that looks at how policymakers make choices and this Markus Goldstein write-up of a new paper by Mattie Toma and Elizabeth Bell is fantastic, and has sent me straight to the paper. They run experiments to show how even high-level policymakers are very susceptible to changing their decisions based on relatively minor changes in how the choices are presented which makes differences and simple calculations of impact and reach more obvious. This chimes with my own research and experience: policymaking is sophisticated in aggregate but it takes a lot of institutional scaffolding to make it so.
5. Everyone’s trying to become an expert on illicit financial flows now that the sanctions against Russian oligarchs have been announced, but here’s a tip: save yourself the time and just go direct to Matt Collin. His new paper, with Florian Hollenbach and David Szakonyi looks at recent high profile US Treasury attempt to eliminate anonymous ownership of US real estate (a popular way of stashing ill-gotten gains abroad) and finds… not much of an effect (excellent thread here). Illicit flows are complex, and hard to fully uncover; the kind of work that Matt and his co-authors do is important. Without it we simply make policy based on what sounds good, not on what works.
6. A really cool VoxDev write-up of a paper looking at how environmental regulations implemented in China caused negative productivity shocks, and how enforcement varied according to how easy it was for central government to monitor local governments—because local Governments subject to looser monitoring were more likely to enforce regulations laxly and thus incur fewer economic penalties from regulation.
7. And lastly, three pieces of marginalia to help arrest the doomscrolling. First, LitHub on the greatest opening lines in literature, featuring some absolute classics (the clock striking thirteen is particularly vivid, but my favourite is almost certainly the understated tone-setter from The Samurai: “It began to snow.”) Second, on a language-related trip, an analysis of 15 million wordle tweets that confirms what we already knew: people post their most impressive answers more than their failures. And lastly, The Ringer rank the top 25 Batman villains from the movies. They make many mistakes, but end up in the right place…

Have a great weekend, everyone!

R