Best of #econtwitter - Week of February 19, 2023: paper summaries
Welcome readers old and new to this week’s edition of Best of Econtwitter. Please submit suggestions — very much including your own work! — over email or on Twitter @just_economics.
Belated edition this week!
Paper summaries

How do beliefs on the morally appropriate way to treat different groups shape our policy views?
@BenjaminEnke (@HarvardEcon), Ricardo Rodríguez-Padilla (@Uber) & Florian Zimmermann (@UniBonn, @briq_institute) on moral universalism:
youtu.be/-wb8yB9hNX0


Are ethnic and partisan segregation related?
In France, electing a left-wing (instead of right-wing) mayor led to an increase in the local share of immigrants due to different public housing policies.
Impact was significant, lasting, and linked to higher reelection chances.



How much were election-denying Republican candidates punished in the 2022 midterms? @janetmalzahn and I put together the data and find they suffered roughly a 2.3 percentage-point penalty in the general election, on average.
gsb.stanford.edu/faculty-resear…


The penalty is also uneven. In some states (like Kansas), election-denying candidates severely underperformed other Rs. In other states (like Arizona), gaps were much smaller.

^Nate Silver: “2.3 points of vote *share* = 4-5 points of vote *margin* (how much you beat/lose to the other guy by) which is a pretty big effect. The view that "LOL, nothing matters" is sophomoric. Research is very robust that candidates pay an electoral price for extremism.”

**new** Using linked survey, census, admin and commercial data, we aim to produce the best possible estimates of income and poverty in the US.
We find median household income in 2018 is 6.3% higher than survey estimate, and poverty is 1.1 pp lower. (1/n)


New CEPR WP with @tlebarbanchon & @juliensauvagnat is out!
We study a topical subject 🔥: what is the role of hiring difficulties for firm growth?
The pandemic has turned the spotlight on this issue as the share of firms unable to fill jobs hit a record high globally🌎📈

CEPR @cepr_org

2/N More precisely, we show that when central banks keep the policy rate below the equilibrium/natural rate r* for an extended period, financial instability risks surge. We estimate r* following @marcodelnegro et al. The stance of monetary is the gap between policy rates and r*.


(1/n) A 🧵on over-parametrization and one of its applications in economics:
Over-parameterization and first order optimization methods (e.g., SGD) lead to "smooth" interpolating functions.
Smooth here means functions with small derivatives/gradients (think of Sobolev seminorms).


Maximilian Kasy @maxkasy

Law of Iterated... Quantiles??
🚨 New draft up 🚨, economic theory with many applications by @kaihaoyang2020 and Alex Zentefis.
papers.ssrn.com/sol3/papers.cf… #EconTwitter


Large language models (like ChatGPT and successors) will boost research productivity in economics and other disciplines, I predict in “𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗮𝗻𝗱 𝗖𝗼𝗴𝗻𝗶𝘁𝗶𝘃𝗲 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻 𝗳𝗼𝗿 𝗘𝗰𝗼𝗻𝗼𝗺𝗶𝗰 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵” ☛nber.org/papers/w30957

^“I estimate current models make me 5 to 20% more productive” 🥴
^“paper on late-life shifts in ethnoracial identification using linked Social Security applications”

Thanks for the shout out, @MargRev and @tylercowen!
I noticed several things come up in the comments that I would like to address here.
(Paper in question is here: nber.org/papers/w30934)

Marginal Revolution @MargRev

5 pages of coauthors!
164 teams: Same data, same hypotheses to test.
They call the variability of findings across teams "non-standard-errors."
A key finding:
"...uncertainty due to non-standard errors is similar in magnitude to that due to standard errors."

Cameron Pfiffer @cameron_pfiffer

What happens when you tell people they are morally obligated to donate to charity?
@SchoeneggerPhil & I address this question in our forthcoming paper in JBEE.
1/9


Donations for carbon offsets increased by 100% in 3 years. What does this market imply about people's valuations of environmental protection?
In a new study, I estimate implied willingness to pay for carbon mitigation using a field experiment with 250k consumers.
A summary!🧵


Under #affirmativeaction, Black and Hispanic students were twice as likely to be admitted to Berkeley/UCLA as white and Asian students with similar SATs and grades.
That advantage fell to 40% in 1998, and the policies implemented since then haven't moved the needle so much.


🚨New @nberpubs WP Alert!🚨 w/ @AndrabiTahir @BauNatalie Jishnu Das & @aikhwaja. We evaluate the effects of providing grants to public schools on learning outcomes in public & private schools in Pakistan. Spoiler: learning improves in both sectors! nber.org/papers/w30929


Here's my new working paper with @aaronrkaufman & @ChrisMCelaya. My first experiment!
tl;dr audit studies tend to underestimate discrimination because of respondent noncompliance
ungated: dropbox.com/s/0cnayx1hsbh0…
1/n


New paper claims that if you let parents raise their own children the children will become dumb and also criminals (in Finland).


And here's the evidence to support those claims...
oh hold on wait it looks like those confidence cross zero in every case?
you mean the big controversial finding is actually insignificant?
yes.



Ancient Babylon was a market economy. Using the clay tablets left by the priests of the god Marduk, which had daily astronomical information and monthly price data for six goods (allowing precise dating), Temin found prices followed a random walk, and were not administered.


More papers

New @nberpubs WP!📢
Multilaterals like @WorldBank @AfDB_Group fund 1000s of contracts each year to procure goods & services for infrastructure—but governments must follow donor contracting procedures. What do these do?
Wolfram @tedmiguel @e_o_hsu & I study Kenya electrification
