For the last 6 months or more, I have been obsessed with this basic measurement problem: how many individuals who lost employment during the pandemic received UI benefits? Here is a short thread on measurement, with the hope to get new ideas. 1/

What can we measure? We can measure how many weekly UI claims each state reports paying. We can measure who is currently unemployed from the CPS. What's the issue? 2/

First, there is eligibility. You can be unemployed but be ineligible for UI, for instance if you are entering or re-entering the labor market or you left employment voluntarily. So unemployed may be an overcount. 3/

But you can also be eligible for UI but not be unemployed, if you are not on temp. layoff or actively searching. Many states waived search requirements for UI and there are differences between UI req. and the CPS measure. So the number unemployed may be an undercount. 4/

In addition, there are measurement issues on the UI side as well as known fraud. From my best estimates from the CPS on eligibility and DOL data on payments, the number and dollars of claims paid range from ~95 to 105% of estimated eligibility. 5/

http://publish.illinois.edu/elizaforsythe/files/2020/12/Cortes_Forsythe_Covid_Inequality.pdf

But these estimates give me pause because other sources show much lower receipt rates. For instance, only 3/4 of claimants in the @uscensusbureau Pulse survey report receiving benefits, however this could be due to delays in payment. 6/

But similarly, the @CAPolicyLab finds shares of unemployed that have received UI benefits to vary dramatically across geography, with some areas have less than half of unemployed receiving UI (but others with over 95%). 7/

What is absolutely clear is that the PUA program has been a fantastic success in expanding access to UI benefits. See this recent article for how bad UI access is without the CARES Act. 8/

https://www.nytimes.com/2021/01/21/business/economy/unemployment-insurance.html 8/

But I'm curious if others have ideas for data sources or how to measure this very important policy number, which tells us how well our expanded UI system is able to target folks who lost employment vs. the need for broader policy instruments, like checks. 9/9

Tagging some folks working in this space @p_ganong @TJ_Hedin @TillvonWachter @kevinrinz
@keds_economist @ernietedeschi @arindube @aaronsojourner @hshierholz @DanielBZhao @AbigailWozniak