About 6 months ago we published a paper in PNAS that for many reasons went a bit under the hood. But given the growth of #EconSky, here's a short thread about it.
It's cool! It has:
- *Insane* microdata
- Simple approach
- Relevant policy result
It was likely also my last academic work. Here we go
It's cool! It has:
- *Insane* microdata
- Simple approach
- Relevant policy result
It was likely also my last academic work. Here we go
Comments
I recall seeing the paper although it is way beyond my sphere of competence
2
why on earth no link in this thread ?
https://www.pnas.org/doi/10.1073/pnas.2317589121
3
I will never understand the idea of threads for summarizing a paper; it is the worst sort of lazy science; I just do not get it, honestly
For the threads, I personally like them - I see them as a form of conference abstract served on a platter. You are interested? Great, read the paper. Not interested? Skim and discard.
But to each his own :)
Remember that magic threshold of 1 we all paid attention to? Now you know what it would give were you to close all supermarkets.
Now, the cool #data part
Plus background register data. Age, income, address...
Interested yet?
Then, we compare those who shopped at the supermarket within 5 minutes of them (treated)...
Neat? Yeah, we need to be *that* specific. Both shopping place and time is endogenous! Like, old people like to shop in the morning...
And we are not done yet!
Who tends to shop close to each other?
Your network.
Family, colleagues, friends...
Running our "experiment" on them would produce terribly biased results. They might have infected you elsewhere
So we identify your network from data, and remove those "false experiments" from our sample
Note: Masks and social distancing was imposed during the study period.