Thread by @ElizaForsythe

Many were surprised to PUA claims overtake standard UI claims this week in the UI data. Here is a thread with some data about the PUA program 1/

First, the data in the weekly claims reports from the DOL appear to include many claims that don't show up in the monthly ETA data on actual claims paid, so take the weekly news release with a grain of salt. 2/

If we look at the total number of weeks of claims actually paid by states, reported in the monthly ETA reports, we see that the number of PUA claims paid have been increasing each month, while standard UI claims have been falling (Aug data is incomplete) 3/

However, there is some funny business going on, because the dollars in PUA claims paid remain comparatively low and fell in July 4/

Which shows up if we look at the average weekly benefit paid per claim-- in July this feel to below the minimum PUA payment! 5/

So DOL data is inconsistent. To make progress, I look at predicted PUA claims in the CPS, using the machinery I developed with @CortesEcon in our paper: research.upjohn.org/up_workingpape… 6/

Here we see that the share of last year's employed workers who are now non-employed and eligible for standard UI and PUA have both been falling, and the PUA share of total UI eligibility is small 7/

Of course, this could be due to things that are missing in our model (such as if the employer is covered). But it does not look like the distribution of non-employed in July and August are tilting toward those who are more likely to be on PUA 8/

Finally, we can say what are the characteristics of the folks we think are eligible for PUA: we estimate about 20% are getting PUA due to too low of earnings in the past year, where most of the balance is self-employed. Note that this makes PUA very imp. for low-wage workers 9/

Conclusions? The data is still incomplete, so it is hard to know exactly what is going on with the rising share of PUA in the weekly DOL news release. Claims data and payment data tell very different stories. 10/

There have been reports of fraudulent claims, but given payments have been steady, hopefully this means that state UI systems are catching many of these. From the CPS data, it does not look like the non-employed in July and August should be more likely to be on PUA. /end

@EvermoreMichele @ENPancotti @keds_economist @ernietedeschi