Best of #econtwitter - Week of September 19, 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
4/N temporary unconditional cash is very effective! impacts on HH consumption suggest these transfers pay for themselves in just 3 years. these impacts are remarkably stable across contexts (high external validity)
5/N great! so can we get larger impacts with larger transfers? yes! but, these impacts are less cost effective → larger aggregate consumption impacts by providing smaller transfers to more households
In 1969, the Federal Housing Administration (FHA) destroyed the records of its racially exclusionary🏡mortgage insurance from the middle third of the 20th Century.
We resurrect them.
nber.org/papers/w29244?…
NEW EVIDENCE ON REDLINING BY FEDERAL HOUSING PROGRAMS IN THE 1930s🧵
🚨New WP w/ Martti Kaila & @krista_ri_🚨 We show job loss causes much more damage for those from less wealthy backgrounds, w/ big implications for intergen mobility. Figure shows double unemployment post job loss for adult children born to bottom 20% relative to top 20% [1/9]
🚨New Working Paper🚨
The Impact of Financial Assistance Programs on Health Care Utilization
with @alyceadamsPhD, @ray_kluender, @JinglinWang4, @francisawong, @wesyin
NBER WP: nber.org/papers/w29227
Ungated version: stanford.io/2YSl8QX
Begin 🧵
My paper with the excellent Ian Fillmore, “Technological Change and Obsolete Skills: Evidence from Men’s Professional Tennis,” was just published in Labour Economics. It is a really fun paper (even R2 said so!) that I am excited to tell you about. 1/8
Such an important paper!
We have documented many belief/learning biases, e.g. confirmation bias, overreaction, etc
-How do biases affect learning in long run?
-When is standard model "robust" to a bias?
-What happens when ppl w/ diff biases interact? 1/3
Econometrica @ecmaEditors
More: survey on media and social capital; child maltreatment effects; perceptions of racial gaps; partisanship and life expectancy; followup on police calls and Floyd; program evaluation guidelines; Indian canals; women in tech self-reported skills; Polish secret police; sales tax breadth; fat tails and nowcasting
Public goods
Interested in international macroeconomics and finance? Find online the materials from the Stanford Initiative: videos, slides, data, and code.
stanford.io/3lKp3X5
Intended to be a starter kit for PhD students interested in large-scale empirical work in the field. 1/n
Interesting discussions
I just finished a 2 year stint at FB Core Data Science. Making the switch to industry left me with an appreciation for how academia often doesn't get the best out of people and mostly didn't got the best out of me. Here are some thoughts... 1/n
I once asked workers on MTurk to guess the hourly wage for a whole bunch of BLS occupations. People were pretty good for relatively low-wage occupations, but generally they way underestimated the returns to being a manager of some kind
My first reaction to this chart was: there's no way this is accurate. And it isn't! Food is not, in fact, "more expensive than almost anytime in the past 60 years".
Lisa Abramowicz @lisaabramowicz1
Re announcement today from World Bank
Perhaps it is more clear now why I encouraged the new World Bank President back in 2019 to outsource its entire research function.
"Diplomacy and science cannot both thrive under the same roof."
^more, more, joke, see media for more details / other side of the story
Profiling your code
Just talking to a coauthor about profiling code and realized this may not be known to many researchers who code a lot but don’t have much software background.
So here is a short thread on a tool that can speed up heavy computations by orders of magnitude.
Advice for students interviewing at the FMA and AFA meetings.
1. Do not drone on and on. The first thing you learn in a public speaking class is that your audience is not listening. So you have to stop and do a tiny recap every so often, so the non-listeners can catch up.
1/n
