Emil O W Kirkegaard

These kinds of studies have been done before. I did one with @jfuerst0, and Borjas also did it in his recent book. Yes, origin still matters because origin = prior, and education etc. is not perfectly informative about relevant job skills.

Ah, gotcha. So something he briefly mentioned but didn't show a huge set of results for.

Results of the PIAAC show that immigrant status is very relevant.https://twitter.com/phl43/status/952673677904175104 …

I don't think the data explorer on their website allows you to look at this, but I had a look at the codebook of the dataset and it's very rich, so I'm sure there is a variable for country of birth. I suspect one could learn a lot analyzing this dataset.

(One thing I don't know, however, is whether the sample is large enough to reliably determine whether country of origin matters once you control for the usual suspects.)

Yeah, the 2016 ACS has more than 200,000 immigrants 18-54 and even then a lot of the countries have really small sample sizes.

Yeah, I suspect that sample size will be an issue with the PIAAC data, which is a pity because the dataset seems very rich from my skimming the codebook.