Officials cast doubt on claim that 73% of Vermont’s unemployment claims came from women

Published: 5/6/2021 9:56:48 PM
Modified: 5/6/2021 9:56:46 PM

In January, Vermont media outlets — including VtDigger — reported on the impact of the COVID-19 pandemic on women’s employment, focusing on the disproportionate number of women across the state who lost their jobs or were forced to leave the workforce to care for loved ones or themselves.

Those stories hinged on a statistic that appeared to demonstrate that disparity with simple, direct clarity: 73% of recent unemployment claims in Vermont came from women.

The 73% figure has since been cited in legislative debates, including over whether to provide additional unemployment benefits to those with dependent children. But the reality behind that figure — and the perils of studying employment data in Vermont — is far more complicated, as Seven Days reported earlier this week.

“People just glommed onto this statistic and just applied it universally,” said Mathew Barewicz, an analyst at the Vermont Department of Labor. “And anytime we’ve been in testimony and discussing this, we try and turn the attention away from that number because there’s uncertainty around it.”

There’s no doubt that women have struggled during the pandemic. Reams of evidence, including the department’s own figures, show that women lost ground in the workforce in 2020, though it’s unclear whether that has changed in 2021.

According to national data, women tended to work in sectors that lost the most jobs during the pandemic, such as tipped positions, and were also more likely to take on child care duties that affected their employment.

But when it comes to the exact percentage of Vermont unemployment claims that came from women, “we don’t know at this time,” Barewicz said.

“The data we have confidence in at this time indicates that it’s not, you know, a 70/30 split. It’s probably more like 60/40,” he said. “I think sometimes we get hung up on data, and I guess my point is, the data is less important than what the takeaway is: that women are being disproportionately impacted. End of sentence. The public policy discussions go from there.”

The origin of the 73% statistic is the U.S. Department of Labor’s “Characteristics of the Unemployment Insurance Claimants” dataset, published regularly since 2007.

Midway through each month, the federal government obtains data from the states for one particular week, to get a snapshot of who is requesting unemployment from the state. In November, Vermont’s records showed that women made up 73% of that sample.

Joyce Manchester, senior economist for the Legislative Joint Fiscal Office, first publicized the 73% figure in a fiscal note to lawmakers. But she has since backed away from it, citing Barewicz’s misgivings over it.

Cary Brown, executive director of the Vermont Commission on Women, has been tracking that claims data since the start of the pandemic. She keeps a particular eye on how great the disparity is in Vermont, as compared with the rest of the nation.

There’s no reason to conclude those numbers are misleading, she said.

“When you look at the past year, you see that the share of people who filed for unemployment claims in Vermont, it’s much higher for women than men in Vermont, which was true across the country,” Brown said. “But the difference between those claims and men’s claims in the country as a whole was not as great as it was in Vermont. I just don’t know how you can dispute that. That is true.”

While the data may not be false, Barewicz said it may not show the whole picture. The state Department of Labor and the federal Department of Labor have been pushed to the breaking point processing unemployment claims since they began to spike in April 2020, meaning some could be lost in the fray.

“As a result, systems were stood up that did not talk to all other systems,” he said. “When you look at the data in that federal website, (it’s) observations for only 10,000 people. And at those times, the Department of Labor was servicing 30,000 to 50,000 Vermonters, so there’s clearly a gap in information.”

He said the data could be right for the “snapshot” of the week it was taken. But that doesn’t mean that it could be projected to paint a picture of the broader impact of the pandemic on women.

Barewicz also pointed out a well-known flaw in using unemployment claims to measure the impact of job losses: It doesn’t account for people who aren’t collecting unemployment insurance. Brown echoed the point.

“What unemployment numbers don’t capture are the people who are no longer looking for work, for the people who’ve just left the workforce entirely,” she said. “And we know that women did that at a much much higher rate than men did in response to pandemic conditions.”

The Vermont Department of Labor has tried to collect its own data to demonstrate the pandemic’s impact on women. Its review of earnings by job and gender shows some disparities: The female share of the Vermont workforce dropped by one percentage point, from 51% to 50%.

Female employment also fell by 13% in the second quarter of 2020, compared to a 10% drop for men.

The latest data from the department is only available through the third quarter of 2020, compared to March 2021 for the unemployment claims data. Barewicz said it can be challenging to get detailed, timely data at the speed people want it. “In the data shop, people always want tomorrow’s data today,” he said.

“There’s a lot of conflicting information out there and a lot of things pointing in different directions, and putting it all together in a comprehensive, coherent story can be difficult at times,” he said.

As an example, he noted that the unemployment rate in Vermont, currently at 3%, can be difficult to put in context. That rate, like claims data, only reflects the status of those who are considered to be in the workforce in the first place.

“One person can either conclude the economy’s fine, or another person can conclude that the data is garbage,” he said. “And the reality is neither of those are true, because you need to understand how to read that information and understand how that information is tracking.”

That nuance can get lost in conveying numbers to a layperson, or to the media. Sometimes when he talks about data sources, definitions and caveats, Barewicz said, he can “see people’s ears close.”

Barewicz is working on a data report right now that will hopefully provide clearer answers. “That’s the work we’re doing right now,” he said, “to try and create a more comprehensive picture about this.”

Valley News

24 Interchange Drive
West Lebanon, NH 03784


© 2021 Valley News
Terms & Conditions - Privacy Policy