Best of #econtwitter - Week of May 22, 2022 [3/3]

Welcome readers old and new to this week’s edition of Best of Econtwitter. Thanks to those sharing suggestions, over email or on Twitter @just_economics.

This is part three of three.

Editorial note: This week marks the two-year birthday of the newsletter 🥳. Always happy to see people getting value out of the weekly emails — we passed the 2,000-subscriber mark earlier this year — especially relative to the low marginal cost of production.

• A two-question survey / more general opportunity for feedback here! Only you can prevent selection bias, by filling it out :)

Paper summaries

Lee Crawfurd @leecrawfurd

Where did the industrial revolution start? Places with high skill but low wages journals.uchicago.edu/doi/10.1086/72…

^“this point about wages within England is technically not incompatible with Allen's high wage thesis which is about wage differences between England & rest of Europe”

Chris Blattman @cblatts

What if I told you that roughly $500 for therapy + a little cash helped the most troubled young men in West Africa drop their crime and violence by half. For at least 10 years. Almost 350 fewer crimes PER PARTICIPANT over 10y. That's ~$1.50 per crime averted. New paper & a 🧵

Mike Johanson @mikebjohanson

I'm excited today to announce a paper that I've been working on for the last four years, about RL agents discovering micro-economic behaviour: producing, consuming, and trading goods: 1/

DeepMind @DeepMind

Reinforcement learning agents can learn to cooperate, compete, and communicate in multi-agent settings. New research explores the emergence of micro-economic behaviour as agents learn to produce, consume, and trade resources in simulated worlds: https://t.co/wLNTdmUwza 1/ https://t.co/Afl60gnXVC

Mike Johanson @mikebjohanson

In this paper, using as little domain knowledge as possible, we found that state-of-the-art RL agents can indeed learn these behaviours: first producing goods for their own consumption, then producing to trade, then negotiating better prices... 12/

^DeepMind going after agent-based models 😬

Additional (1-tweet) commentary: “One person's cooperation is another's collusion...”

See also: the “generate a title from an abstract” and “AI writing assistant” tools linked in part 1 of this week’s edition…

Moshe Hoffman @Moshe_Hoffman

Hi y’all. I wanted to do a 🧵 summarizing our book that came out last month. (Thoughts, questions, more than welcome.) smile.amazon.com/Hidden-Games-S…

smile.amazon.comAmazon - Hidden Games: The Surprising Power of Game Theory to Explain Irrational Human Behavior: Yoeli, Erez, Hoffman, Moshe: 9781541619470…Buy Hidden Games: The Surprising Power of Game Theory to Explain Irrational Human Behavior on Amazon.com ✓ FREE SHIPPING on qualified orders

Ulysse Lojkine @ULojkine

New working paper : A general power index papers.ssrn.com/sol3/papers.cf… 1/

Susan Athey @Susan_Athey

In most econ applications, clustering std errors is a problem of design, not sampling. What does that mean? Clustering is needed if treatment assignment is correlated within clusters, not because some clusters were sampled and others were not. 1/8 @jmwooldridge

More: peer review in aid bureaucracy; emigration of indentured servants from colonial India; methods for predicting race; more RCTs for Medicaid, please; private health care for US vets; framing public policies as changing the status quo decreases public support for them

Interesting discussions

Tatyana Deryugina 🇺🇦 @TDeryugina

What benefits have you gotten out of #EconTwitter?

^recommended replies/QTs… even if the purpose of this newsletter is to disintermediate econtwitter (at the risk of Grossman-Stiglitzing ourselves)

Ben Grodeck @benleo_econ

@TDeryugina For me, it's had a number of benefits. - Learn about new & relevant papers - Information about conferences - Networked with amazing researchers. Now I am friends and coauthors with some of them. - Crowdsource information for questions I have - @KhoaVuUmn's memes

Nick HK @nickchk

I think one of the more useful things I do when I teach coding is to change the example I'm going to use on the fly to something I haven't done before so they can see me Google for example code and get errors and figure out how to fix them

Carly Urban @carly_urban

I'm officially promoted to "Professor." I want to scream loudly that I've done most of my academic career counter- "traditional academic econ" culture. A quick thread with some examples to maybe give some hope to others seeking a "different" path.

David Phillips @PhillipsEcon

On the occasion of detecting a correspondence/audit study in my inbox, apparently along with much of EconTwitter: a hopefully helpful rant on what to do if you or your grad student is thinking about running a study sending messages from fictional people to real people. (1/n)

Jimbo Brand @jamesbrandecon

I remember thinking in grad school “wow, I’ll never be able to learn this much this fast ever again once I finish the PhD,” and the magnitude by which I was wrong is hard to overstate. I don’t know if it’s the switch to industry, or if every PhD grad feels this way, but 1/

Jimbo Brand @jamesbrandecon

Sometimes I look back now and ask “what the hell did I even do all day?” I’m now somehow meeting more people, managing more projects, learning new institutional details, *and* researching/coding at a pace similar to my grad school rate. Mostly my takeaway from this is that 2/

Khoa Vu @KhoaVuUmn

Your Econ PhD application requires a short tweet from your dad/mom about you now. I don't make the rules.