During the hectic early weeks of the pandemic lockdowns last year, the weekly unemployment numbers were truly shocking even to those among us who sadly saw it coming once the pandemic started.
In the week ending March 7, just under 200,000 workers across the United States filed for new unemployment benefits, while just under 2 million workers were on continuing benefits. Just four weeks later, new claims peaked at over 6 million the week ending April 3 and continuing claims one month later at 22.6 million. By the end of July, nearly 50 million applications had been processed – roughly one for every three workers!
The unemployment rate ballooned from 3.4% to 16.8% in just ten short weeks, and the county level map of the peak August unemployment levels show how widespread the pain was (the dark brown shading begins at the 30% level):
Fast forward past the three weeks to flatten the curve, and we are now 21 months into the pandemic, still very much in the middle of it all. Weekly new claims, while still higher than where they started, have averaged 247,000 per week. Total continuing claims lag the weekly claims and have been falling steadily and now averaging just over 1.75 million.
Unemployment levels are about 1% higher than where they started, largely because the defined labor force has shrunk substantially as people have abandoned the notion of going back to work.
During this time, we have tracked these numbers weekly, and using a sector-based model, produced block group level unemployment estimates over the entire period. Many of our readers subscribed to these free updates and utilized them, and we appreciate your comments and support for this project. But as the labor force situation has returned – not to normal – but to stability, the need for these weekly updates has diminished to the point where we feel confident that the program can be terminated, omicron notwithstanding. The release at the end of December will be the last of the series – barring renewed instability which we can all agree is not something any of us would welcome. These two maps – unemployment in March 2020 and November 2021, we hope, makes it clear why we are sunsetting the program:
We have two main takeaways –
First, and not a surprise to a company littered with geographers, effective spatial modeling is always worth the effort, because as we like to say, geography still matters. State and national unemployment numbers masked the true nature of the problem. In economically diversified areas, the unemployment effects were relatively minimal in comparison to highly tourist dependent areas where rates often exceeded 60%
Second, we were reminded that to remember that data is often collected for purposes other than that which it is used. The weekly numbers don’t measure unemployment, but rather the number of new, completed unemployment claims. Normally, these are so highly correlated that one forgets that we aren’t reporting exactly what is being measured. The collapse of the unemployment data systems made this perfectly clear – the numbers simply didn’t make any sense from week to week. If you didn’t read our article “Dashboards, Surrogates, and Wonky Gauges”, be sure to check it out.
So, is this the end of unemployment? No, of course not, but it is the end of our weekly coverage of it.
Recent Comments