Read on Twitter
Lots of focus recently on the weekly UI claims report from @USDOL, and rightfully so. UI claims data give us real-time measures of employment status changes amid the pandemic. They were our first sign of pain in March and have remained a valuable tool for monitoring the economy.
In a report released yesterday, @EvanNewell and I show that the initial claims numbers and, to some extent, continuing claims numbers are not as accurate as we might have hoped, challenging our ability to compare UI claims between states and over time. https://haslam.utk.edu/whitepapers/boyd-center-business-and-economic-research/understanding-accuracy-unemployment-insurance
In GA, the number of initial claims in leisure and hospitality since the pandemic began represents 163% of 2019 employment in that industry.
In PA, continuing claims data report more people receiving PUA for the week of June 20 than reportedly applied for PUA in all prior weeks.
In PA, continuing claims data report more people receiving PUA for the week of June 20 than reportedly applied for PUA in all prior weeks.
These errors appear to be due to challenges states faced to stand up a PUA claim reporting system on limited notice, to duplicate claims from legitimate claimants, and to applicant fraud.
We close by noting that these data have the potential to be an extremely useful tool for policymaking, but they have to be accurate to be useful. Federal investments in upgrading the UI data infrastructure at the state level would likely have large public returns.
