New working paper on business death in history and during 2020, joint with @crane_leland , @Aaronflaaen , @Adrian_HaminsP , and Chris Kurz. https://www.federalreserve.gov/econres/feds/files/2020089pap.pdf

Folks are looking at a number of nontraditional indicators of business exit during 2020 (since official data lag). It's important to compare these to what we know historically; the question isn't "are lots of firms dying," the question is "is death higher than usual."

Historically, business death is common and, particularly recently, countercyclical. We summarize a few stylized facts that you should keep in your head.

We then look at a number of nontraditional measures of business death in 2020. For example, we describe a novel death measure in SafeGraph cell phone tracking data which suggests restaurant deaths are about 50% above normal.

Other indicators are elevated too, such as Homebase data. But ADP payroll data aren't so worrying, and particularly suggest that permanent death isn't currently accounting for a lot of employment losses.

Looking forward, though, small business owners in some sectors are pretty worried about the next six months, with death expectations above historical norms (from Pulse Survey)

Any conclusions about 2020 are tentative, and this is only a working paper, but we hope our historical discussion can inform the business death conversation broadly.

One other note: we were lucky the new Business Dynamics Statistics came out in time for this. In the appendix, we do a lot of comparisons between the redesign vintage and the legacy vintage of BDS. Our stylized facts are broadly robust to vintage, but there are some differences!

I think it goes without saying but I'll also add: nontraditional economic indicators have really shown their value this year. As always, though, their value is highest when coupled with things we know from official data. So, "why not do both?"