John Minchillo/AP Photo
Lower-income individuals were more difficult to reach than before, meaning there was less comprehensive survey data on those more likely to have been unemployed during the pandemic.
Last week, the Census Bureau released new data showing that poverty spiked in 2022 following the expiration of prominent pandemic programs. The supplemental poverty rate—a measure that integrates taxes paid, cash and in-kind benefits received, and basic costs like medical expenses—rose from a record low of 7.8 percent in the prior year up to 12.4 percent. All the progress made in reducing poverty over the previous several years was officially confirmed undone.
In particular, the increase in child poverty due to the expired 2021 Child Tax Credit expansion drove media reports, leading to strong responses from lawmakers. Members of Congress and the Biden administration correctly identified the need to make the CTC fully refundable again (i.e., guarantee the full credit amount to children regardless of their household income level). But missing from the conversation, notably, was just how effective the pandemic unemployment benefit programs were at preventing poverty.
For much of 2021, unemployment insurance (UI) benefit eligibility was expanded to more adequately cover gig workers, self-employed individuals, and those refusing hazardous work conditions. Claimants also saw their weekly benefit sizes topped up by an additional $300, down from a $600 boost in 2020 at the height of the pandemic but still significant (and much more than the expanded CTC). Well over ten million claimants received these benefits in 2021.
One might expect the expanded UI in 2021 to have a greater impact on poverty than the expanded CTC, given that several hundred billion dollars in UI benefits were distributed to workers who would have otherwise seen dramatic drops in their income.
Data limitations in tracking poverty and income could have long-term consequences for workers, if the numbers lead federal policymakers to inadequately prioritize UI reforms.
But the Census numbers show that the pandemic unemployment insurance benefits produced a smaller reduction in poverty. The Census Bureau found that stimulus payments and the expanded CTC prevented 8.9 million and 5.3 million people, respectively, from falling into poverty in 2021. Unemployment insurance, in contrast, kept only 2.3 million people out of poverty, under half of the CTC’s determined impact.
These Census estimates may understate the full effect of unemployment insurance. The reason is technical: The Census Bureau may be undercounting the pandemic benefits sent to Americans.
Household surveys such as the Census Bureau’s Current Population Survey (CPS) have a history of underreporting cash benefits like unemployment insurance. For instance, survey respondents could forget the exact amount of money they received, or downplay the total because of the associated stigma.
The Census Bureau has also identified that survey non-response rates went up during the pandemic and were “more strongly associated with income than in prior years.” Lower-income individuals were more difficult to reach than before, meaning there was less comprehensive survey data on those more likely to have been unemployed.
A small but growing number of researchers have been able to address these sorts of survey issues over the past decade by incorporating Internal Revenue Service (IRS) tax records into their analyses. Some of the biggest revelations about social mobility and inequality—including research produced by Raj Chetty at Harvard, Emmanuel Saez and Gabriel Zucman at UC Berkeley, and Thomas Piketty at the Paris School of Economics—can be traced to newly gained access to previously inaccessible tax data. This tax data can similarly improve our understanding of the pandemic cash transfer programs, including the unemployment insurance expansions.
Last year, Jeff Larrimore, a section chief at the Federal Reserve Board, published a paper demonstrating how the Census survey data struggled to capture all the pandemic unemployment benefits sent out to American workers. According to Larrimore, the Census survey data doesn’t capture hundreds of billions of dollars in UI benefits sent out between 2020 and 2021 that are reflected by IRS records, which aligns closely with the numbers of other federal agencies like the Bureau of Economic Analysis. Not every dollar missed by the Census surveys went directly to those in or near poverty, but many of those funds likely went to lower-income workers.
When calibrating the Census estimates with the benefits reported in the IRS records, he found unemployment benefits kept an additional 3.9 million Americans out of poverty in 2021. Added to the 2.3 million that the Census estimated, that means that 6.2 million Americans were kept out of poverty thanks to expanded unemployment benefits, more than from the expanded CTC.
In a separate report released last year, Larrimore and his colleagues at the Joint Committee on Taxation, Jacob Mortenson and David Splinter, compared how many pandemic benefit dollars sent to workers in each income decile were reported in IRS records for 2020 but not reflected in the CPS data. They found that “the CPS underreported UI benefits in the bottom two deciles by $133 billion.” Larrimore applied the same approach in his sole-authored paper and found similar results for 2021. In total, CPS data missed over half of the benefits sent out and half of the ensuing poverty reduction.
According to Larrimore, Mortenson, and Splinter, dramatic benefit undercounts are not a new problem with the CPS. The Census also struggled to capture the unemployment benefits sent out to low-income groups during the Great Recession. The key difference from the recession 15 years ago is that millions more people received even more generous benefits during the pandemic, making the benefit underestimates especially pronounced.
These measurement issues may help explain why progressive policymakers and pundits lauded the CTC while overlooking UI when the new poverty numbers came out. Statements by the White House this past week have reaffirmed this view. The administration highlighted that “two ARP policies in particular substantially reduced poverty in 2021: the significant increase in the CTC, which included making the full credit available to the lowest-income children for the first time, and the Economic Impact Payments (EIPs) sent out in March 2021.”
Data limitations in tracking poverty and income could have long-term consequences for workers, if the numbers lead federal policymakers to inadequately prioritize UI reforms. Unemployed workers had much better financial security and bargaining positions during the pandemic recovery thanks to the benefit enhancements. Similar program expansions can help reinforce worker negotiating power moving forward, but fail to capture sufficient attention. The media, and policymakers, appear far more interested in the amount of fraud in pandemic-era unemployment programs in part because the GAO has continued to revise its estimates and publish new ones—an approach that would have strengthened the Census numbers and made the UI poverty impact more salient.
A more accurate understanding of the pandemic unemployment programs’ impact is critical to paving the way for such improvements.
