Huge puzzle about the role of UI in the economy now

1) $600 UI supplement expired, removing $30 billion/week from the economy
2) Spending trackers for proprietary data look *flat*!

Answer: 1) does matter, but 2) unlikely to capture 1) b/c of wonky data issues.

Deets in

For Fact 2, on real-time spending trackers, see @OppInsights. Also, the Chase card spending tracker has similarly flat-lined in first half of August.

Sample article pointing out the apparent puzzle by @jimtankersley @jeannasmialek https://www.nytimes.com/2020/08/21/business/economy/coronavirus-economic-recovery.html

What is going on? Some economists speculate: maybe people have saved up b/c UI benefits exceeded pre-job loss wages?

Maybe, but we don't know yet b/c of a wonky data problem which makes fact 2 pretty uninformative about the role of 1). https://twitter.com/SHamiltonian/status/1296944018035576832?s=20

It comes down to payment method.

* About ¾ of UI benefits are issued via prepaid debit card.

* A spending series built off regular debit and credit cards won’t catch the effects of these UI benefits.

https://www.federalreserve.gov/paymentsystems/2019-December-The-Federal-Reserve-Payments-Study.htm

In fact, you might even see households spending *more* on their regular debit and credit cards once they get less in benefits on their UI cards!

best evidence so far is from @facteus. they have a sample of prepaid cards for UI benefits and shows a big drop in spending at the start of August specifically on those cards
source: @jeffsparshott
https://blogs.wsj.com/economics/2020/08/21/newsletter-u-s-labor-market-cools-u-k-output-jumps/
(h/t @jialanw for link)

When will we figure out the immediate impact of the supplement expiring?

It's hard for the Facteus data to be decisive since (a) it greatly overrepresents low-income households and (b) we don’t see household spending, just spending on the prepaid card.

We will get the first Census estimate Retail Sales for August until mid-Sep.

We might learn more sooner is if there’s a data series of spending that uses data collected on the merchant side, rather than the household side. But I don't know of any such series.

Also, obvi, #JPMCInstitute ( @Dan_M_Sullivan @Farrell_Diana @FionaGreigDC @maxliebeskind @pascaljnoel @JoeVavra) is interested in understanding effect of $600 supplement.

Our August data will arrive more slowly than Census, so it will be a while before we can analyze this (and in any case, the decision about when to release analysis is not mine)

Finally, I’m sure I’ve missed some of the evidence/analysis in press and by economists in this thread. Please add links to this thread!

{end mega thread}