Share this page!

Peter Ganong Profile picture
Twitter logo
6 Aug, 11 tweets, 4 min read
Tomorrow is jobs day and everyone wants to know how early cut off of pandemic benefits will affect employment

I can tell you the answer tonight... you won't learn anything tomorrow‼️

no state data are released ‼️

🧵 on evidence from 5 other data sources that *are* by state
TLDR: four data sources point to no significant/detectable effect, one data source finds a decrease in employment. issues with parallel trend assumptions and inference are plentiful though.
1) The BLS releases state-level employment estimates based on the establishment survey two weeks after the national numbers. For July data, have to wait until August 20. Analysis of the June state-level estimates is here. Looks like a noisy zero.
2) @arindube analyzes the Census Household Pulse Survey and also finds a noisy zero. arindube.com/2021/07/18/ear…
This is my second-favorite estimate because (a) representative sample and (b) has standard errors. My one quibble is that in HPS it looks like cutoff and non-cutoff states are close in terms of level and trend before cutoff. this is not borne out by more precise BLS data.
3) @aaronsojourner finds that regular UI claims (those not cut off by the end of pandemic programs) are similar in states with and without cutoff. So labor market doesn't seem stronger or weaker by this measure.
4) Payroll processor @GustoHQ has no net effect on employment and some puzzling results by age. Note that employment in this series is very volatile.

5) Covid-19 data all star @homebase_data finds that employment grew faster in states that did not cutoff benefits
An important caveat to estimates (2)-(5) is that they assume parallel trends between cutoff and non-cutoff states. This is probably the wrong assumption -- non-cutoff states grew faster than cutoff states from Jan to May.

An important caveat to estimates (3)-(5) is that they don't have standard errors! Traditional sampling uncertainty doesn't exist since these are population estimates, but visual inspection indicates that all three series are highly volatile.
Hard to know what is and is not significant. Maybe we can turn permutation test #econtwitter (looking at you @economeager!) on this problem to try to construct some intervals

If I've missed an estimate in my round up here, please post it in the comments.

• • •

Missing some Tweet in this thread? You can try to force a refresh
　

Keep Current with Peter Ganong

Peter Ganong Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @p_ganong

Peter Ganong Profile picture
29 Jul
New research on Pandemic Unemployment Assistance by JPMC Institute w @FionaGreigDC @Dan_M_Sullivan @pascaljnoel @JoeVavra

PUA is for people not eligible for regular UI

PUA meaningfully insures inc risk, but w/much longer wait for benefits!

jpmorganchase.com/content/dam/jp…
The plot above shows that for people who got regular UI in 2019, non-UI income falls at exactly the same time that UI kicks in (green line).

Regular UI in 2020 (orange line) income starts to fall four weeks before UI kicks in

PUA (blue) income starts to fall ten weeks before
There is also a smaller decline in income after UI receipt for the blue line. Two likely interpretations:
--PUA recipients account for smaller share of HH income
--some PUA recipients have already gone back to work by the time they finally get their benefits
Read 4 tweets
Peter Ganong Profile picture
29 Jul
🚨new results on the effect of UI supplements on job-finding 🚨

2 designs x 2 policy changes yield consistent pattern: small, precisely estimated disincentive effects

Disincentive remains small even after job openings up
An overarching theme of the pandemic has been to view the supplements as responsible for the biggest problems (slow employment recovery, usually conservatives) and the biggest successes (rising wages at the bottom, usually liberals).

Our results are inconsistent with both views.
Instead, it makes sense to think of the effects of pandemic UI primarily as an ambitious anti-poverty policy. I can’t think of a time before when a country gave *full* insurance to earnings losses (examples welcome in the comments)
Read 22 tweets
Peter Ganong Profile picture
28 Jul
Why is UI hard to access?

Most policy discussions (rightly) focus on issues of state capacity.

Amazing paper by Sorkin, Lachowska, and Woodbury focus on *firms* who make it hard to get UI and the seamy underbelly of experience rating.
I think of this as the academic paper which captures @IndivarD's political economy model of a major problem with the UI system
Experience rating is the system where firms that lay off more workers pay more in UI taxes.

Goal: discourage layoffs
Read 6 tweets
Peter Ganong Profile picture
16 Jul
Millions of people have had federal UI benefits cut off

Stated goal: speed the labor market recovery.

Is it working?

Tldr: It’s going to be really hard to use state employment data to do a good job of answering this question.

Looks like a noisy 0
So far, 26 governors have announced plans to cut off at least some federal benefits. 20 are cutting off all benefits by July 5. This is where we might expect to see the biggest effects.
In those states, over 1 million people had their benefits fully cut off and another 1+ million people lost the supplement by July 5.
Read 17 tweets
Peter Ganong Profile picture
15 Apr
“Should you coauthor your JMP?”

Motivated by @RyanReedHill and @carolyn_sms thread, @pascaljnoel and I decided to write our own coauthored JMP thread.

The “Hillstein” thread is excellent, but there are three additional *upsides* to doing this which we want to share!
1) Research is emotionally difficult, at least for me. I *hated* working alone. I was on leave when @pascaljnoel and I started working together in earnest. There’s a good chance I would not have finished my PhD if we hadn’t started working together.
Ask yourself “If you are going to work long hours and push your limits, would you rather do that alone or as part of a team?”
Read 15 tweets
Peter Ganong Profile picture
11 Feb
🚨 new paper on labor market & expanded UI benefits 🚨

Tldr: spending ⬆️ more than expected,
job search ⬇️ order of magnitude less than expected

1) rise of repeat unemployment
2) effect of UI on spending
3) effect of UI on job search
4) connections to current policy debate
“Spending and Job Search Impacts of Expanded Unemployment Benefits: Evidence from Administrative Micro Data”
With @FionaGreigDC @maxliebeskind @pascaljnoel @Dan_M_Sullivan @JoeVavra
bfi.uchicago.edu/working-paper/…
1) we can track workers’ experiences over the course of pandemic

Confirm well-known fact: long-term unemployment is high

New finding: *repeat* unemployment has been rising. (Estimates of long-term unemployment in the CPS miss this since they only ask about most recent spell)
Read 24 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!

Like this author's thread?

Get emails when @p_ganong has new unrolls!

Check out other threads from @p_ganong!

Read all threads by @p_ganong

:(