Best of #econtwitter - Week of February 28, 2021
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.
Paper summary threads
Bottom line is the same: when the cross-country growth lit kicked off in the 1990s, rich and poor countries had been diverging *for hundreds of years*. Since then, that pattern has reversed. Who knew?
2/
Some striking stats on air pollution in China:
- PM2.5 levels declined by ~40 % 2013-2018
- Improvement was faster in cities that were
larger in population & wealthier
- They estimate that this decline in pollution will have raised avg life expectancy by 2.7 years per person
NBER @nberpubs
The tl;dr is the green line in this plot. We measure concentration in product markets - markets corresponding to things that are good substitutes for consumers - and we find that market concentration has actually fallen slightly on average, over the past 25 years!
Pot dealers’ comparative advantage is in black marketeering, not marijuana so marijuana legalization causes them to shift into selling new illegal goods not to become legal pot sellers.
Jennifer Doleac @jenniferdoleac
Other authors had recently found that the euro caused a very large increase in trade. Simply by plotting pre- and posttreatment trends relative to intuitive control groups, we see that isn't the case. Trade increased in non-EMU W. European countries by same amount since the 90s.
🦶🦶 New working paper 🦶🦶
I've been working on this one for a little while now.
andrewcbaker.netlify.app/publication/bl…
Excited to have this paper out! It was truly an interdisciplinary effort, both in terms of co-authors and methods (exp econ, finance, cog psych).
TL;DR: Ownership leads to overreaction to information, both pos+ and neg-, driven by `more-is-less' effect of attention.
🧵 1/N
QJE @QJEHarvard
We show that the longevity insurance provided via Social Security is worth 3.4% of ex-ante consumption when the riskiness of bequest income is taken into consideration. This is much higher than in a model where bequests are deterministic (and the welfare gain of SS is zero) 2/5
Parents' mental health is a MUCH stronger predictor of their children's overall health (measured in QALYs) than their physical health
3/n
Fabulous trial reduced loneliness, depression, and anxiety in Meals on Wheels participants who live alone over four weeks.
How? College volunteers w 2 hrs training simply called a roster of 6-9 folks 2-5X/week (elders' choice) to ask how they are.
New article from the #Seshat project on estimating agricultural productivity in past societies.
Methodology is still new and can be improved, so comments and suggestions welcome!
But the results are quite interesting.
journals.sagepub.com/eprint/NHF85PJ…
Paper by @jondr44 and @pedrohcgs on parallel trends & functional form (FF) must read for practitioners.
The functional-form-robust PT assumption really requires PT in *distribution* - unless you have good reasons for your particular functional form.
The age at which scientists or innovators achieve their moment of genius has been increasing. Half of all pioneering experimental contributions in science happen after age 40. The increase may be because there is so much more knowledge to master nowadays. nber.org/papers/w19866
More: learning by doing in cotton cultivation; laptops and education; Christian missions in Africa; stock ownership and brand loyalty
Jeffrey Wooldridge special section
Wooldridge becoming active on Twitter was the un(?)expected viral hit of the week. These are (short) threads:
Poisson regression can get one so far with so little trouble, why do so many still resist? Especially with panel data. It’s too bad we can’t give it another name to reflect the fact that its a fully robust estimator of conditional mean parameters.
Good reasons for using LPM by OLS over probit.
1. Simple.
2. Provides best linear approximation.
3. Seems to provide good APEs (but not always).
Bad reason: "The normality assumption for probit is too strong." Then I say, "The uniform distribution for LPM is too strong."
Check out his profile for more; “Is it possible -- over the course of a year -- to teach a kind of econometrics course on Twitter?”
Interesting discussions
Fantastic initiative by Review of Finance to help deal with the excruciatingly long publication process, in particular dealing with papers that were R&R elsewhere. Wish they do this in marketing! FYI @Moorman @ksudhir1 @DrGiesler @AndrewTStephen (& others)
aedmans @aedmans
Reviewing PhD applications, some candidates with 4-6 years (sic) of RA/predoc experience. The bar is (too) high
^provoked a lot of reactions, see replies/quote tweets
Economics seminar culture has many great aspects and some really bad aspects.
I love questions and interaction. I love being rigorous and getting to the heart of an issue.
AND it's been clear to me since I joined econ that there's a seminar culture problem
(Thread!)
Dr. Alicia Sasser Modestino @SasserModestino
^people are discussing seminar culture again (there was an NYT story): here, here, here, here, here, here. One story received a lot of attention
If you're in a top-10-ish US undergrad, there is a playbook which still gets you a good shot at top-10 econ PhD programs straight out of undergrad
(I think it's extremely unfair that this essentially only works for top US programs, but, info is info)
^provoked a lot of reactions, see replies/quote tweets; also next two tweets:
@AnthonyLeeZhang Seeing comments of the flavor 1) well I didn't follow this and have a great job/life or 2) criticizing thread because they don't like criteria. Neither diminishes value for writing this down. There are always exceptions to rules (e.g. me) and writing down is 1st step for change.
Suppose you've just received a slew of PhD admissions decisions and you're trying to decide what to do. You log onto twitter and get a mountain of confusing, rancorous discourse flying in 50 diff't directions. You hesitate but reaffirm--you've made it this far; no stopping now 1/
^see also here, and another thread by Zhang, remember this, once again among lots of replies
Interactive IV table?!? Fun to poke and prod at assumptions to see how things change and evaluate robustness. Very cool!
Maria Glymour @MariaGlymour