What did and didn’t work in unemployment insurance during the pandemic

The pandemic recession put enormous stress on the federal-state unemployment insurance (UI) system. A one-month increase in the unemployment rate from 4.4 percent to 14.8 percent, combined with legislation that dramatically expanded the reach of the UI system, sent caseloads to levels never seen before. As tens of millions of workers were submitting applications, policymakers directed UI agencies to fundamentally alter the way the system worked, setting up new programs and making a host of changes to typical operations.

From one perspective, the UI system met this unprecedented challenge. The federal government reports that states made more than $1 trillion in UI payments in the second quarter of 2020, in contrast to less than $50 billion in the prior quarter (both at annualized rates). This spending, combined with direct payments to households, caused personal income to rise instead of fall in the second quarter of 2020, even as workers lost their jobs. The UI payments constituted vital support for families, for the overall economy, and for the public health mission of enabling workers to avoid COVID-19 exposure during some of the worst months of the pandemic.

The creation of a new program called Pandemic Unemployment Assistance (PUA) represents another substantial achievement. State governments administer UI programs, and regular UI can only be paid to workers who meet state-specific eligibility requirements, such as having an acceptable reason for the layoff and a minimum level of previous earnings. PUA, which was created under the Coronavirus Aid, Relief, and Economic Security (CARES) Act, enabled UI to reach the self-employed (including workers earning income using online platforms) and those who otherwise would not meet regular eligibility standards, many of whom nonetheless experienced considerable financial hardship.

Another new program created under the CARES Act, Federal Pandemic Unemployment Compensation (FPUC), paid a flat dollar amount beyond the regular UI benefit, to further supplement the incomes of unemployed workers. (See the “Pandemic-prompted UI tools” sidebar for a summary of this and other CARES Act programs.)

But in several important respects, many states struggled to carry out the UI mission in its entirety. In the UI system, while both the federal government and state governments provide program funding, states serve as the program administrators. Some states were relatively slow in processing initial claims and setting up new UI programs introduced by the CARES Act. This delay was understandable: caseloads were unprecedented, the new programs were intricate, and the need to minimize fraud can slow payments. Another sign of struggle was states’ inability to fully utilize a longstanding UI program called short-time compensation (STC), which is designed to prevent layoffs by providing payments to employees so they can continue to work at temporarily reduced hours. Only around half of states have active STC programs, and even with supplemental federal funding, STC was underused during the pandemic.

FPUC was another program that wasn’t used to its full potential, with its design limited by state agencies’ capacity to administer the supplemental payments. Many economists, UI experts, and policymakers had hoped to pay out a bonus in the form of an increased wage-replacement rate—e.g., an additional pandemic benefit equal to some specified percent of previous wages—but found that this was impossible given the current state of the UI system infrastructure. Without the technology in place to adjust the wage-replacement rate, the only feasible approach was providing flat weekly increments to UI benefits (originally $600, and now $300 in participating states).

Below, we explore what did and didn’t work in state UI systems during the pandemic recession. Understanding the experience of 2020–2021 will be crucial for building UI programs that can effectively support workers in future downturns.

Pandemic provisions reach many more of the unemployed

The regular portion of the UI system is most successful in supporting full-time employees with continuous work histories who experience short spells of unemployment. But the program doesn’t cover many of those who would like to work but cannot find employment. In 2019, only around 28 percent of unemployed individuals at any given time received UI (U.S. Department of Labor, or DOL, 2019). Many of the unemployed are ineligible for regular UI, whether because they have insufficient recent earnings, voluntarily left employment, were self-employed, or were not available and searching for full-time employment. In addition, some of the unemployed have exhausted their regular benefits, which has become more common as states have reduced the maximum duration of benefits in recent years and the fraction of recipients who are long-term unemployed has risen (Nunn and Ratner 2019). Finally, some eligible unemployed individuals may not apply for benefits.¹

Recognizing the extraordinary burdens that the pandemic would put on workers and families, the federal government sought to increase eligibility for UI through the PUA program. This program is entirely funded by the federal government and open to applicants who are ineligible for regular UI benefits. This included workers who had too little earnings to qualify, those seeking part-time work, and those with self-employment earnings but insufficient wage- and salary-employment earnings. Individuals who quit their jobs for COVID-19-related reasons were also generally eligible for PUA, as were some individuals who had exhausted regular UI benefits.

By any measure, PUA was an important source of support for unemployed workers during the pandemic. Weekly continued claims in the program exceeded 10 million during the summer and early fall of 2020, and they remain at around 5.8 million nearly one year later. However, reported continued claims are an unreliable measure of unique PUA claimants at a given point in time because an individual can claim multiple weeks retroactively through PUA and system backlogs can lead to multiple weeks of unemployment being reported in a single week. Additionally, a sizeable share of the claims may be fraudulent, as PUA was particularly targeted by criminals. Nonetheless, the reported numbers suggest that PUA represented over one-third of claimants in summer 2020 and may be responsible for more than half of claimants in more recent months (Government Accountability Office, or GAO, 2020).

As Figure 1 shows, PUA continued claims increased quickly between April and June of 2020, reaching a peak in August and September 2020. They surpassed regular UI claims in the fall of 2020 and have remained about 2.5 million higher per week through June 2021. PUA has paid out over $85 billion in benefits to these newly eligible individuals.

A significant share of PUA recipients—roughly 40 percent of initial claimants—appear to be self-employed.² Although some states reported implausibly low or high self-employed fractions,³ worker groups other than the self-employed likely made up a large share of PUA claimants. Workers who lacked sufficient earnings histories, sought part-time (rather than full-time) employment, or quit jobs due to COVID-19 are all part of the remaining 60 percent of PUA claimants.⁴

Moving forward, policymakers may consider whether to make elements of the PUA program a permanent feature of the UI system. There have been proposals to provide benefits to specific self-employed individuals (including workers who earn income through an online platform) or new entrants to the labor market who have insufficient earnings history to be eligible for traditional UI (Dube 2021, Bennet and Wyden 2021). Alternatively, eligibility could be expanded on an automatic basis during economic downturns only. Any expansion will need to consider funding for the additional recipients’ benefits as well as the potential pitfalls in expanding eligibility to workers who have more control over their hours worked and employment status than traditional employees have. For example, UI generally is only available to individuals who lost employment through no fault of their own. That concept is difficult to apply to the self-employed or individuals who voluntary leave employment (Harris and Krueger 2015).

Limited administrative capacity constrains policy design

As the economy headed into uncharted territory in March 2020, Congress considered a range of policies to strengthen UI to support workers. However, certain features of the existing UI infrastructure severely constrained policymakers’ options. For example, states typically select a wage-replacement rate at which UI replaces lost earnings.⁵ As the pandemic recession began, many policy experts proposed that these rates be increased, but states found that it was difficult or impossible in the short run to make the necessary changes to UI systems, often due to inflexible or outdated information-technology systems (Dixon 2020). Instead, Congress decided to temporarily add a $600-per-week FPUC payment to every UI recipient’s benefit. This had the effect of raising some workers’ wage-replacement rates well above 100 percent, and other workers’ rates by lesser amounts. (See the “Why does it matter how UI benefits are increased?” sidebar for a discussion of the consequences of this policy design.)

Figure 2 shows the huge variation in replacement rates across workers that the $600 FPUC add-on introduced. The blue bars indicate replacement rates under regular UI, which are roughly 50 percent for the lowest 70 percent of earners and fall to 20 percent for the highest 10 percent of earners. The gold bars indicate replacement rates, including the FPUC add-on. With the supplement, replacement rates range from less than 50 percent (for the highest-income recipients) to over 300 percent (for the lowest-income recipients). While it is conceivable that the variation introduced by FPUC corresponded to the optimal replacement rate for each income group (in spring and summer 2020), it is more likely that it reflected the very limited options that were available to policymakers at that time. With more flexible tools, policymakers would likely have made different choices.

A UI system that could accommodate more complex policy designs would require additional modernization efforts and information-technology investments. Such efforts have had mixed success in the past, with initial decreases in program performance and increases in denial rates.⁶ UI modernization is difficult and disruptive, and previous experiences suggest that balancing automation with human involvement is important in maintaining accessibility and fairness. Around half of states have undertaken modernization efforts, and these efforts are likely to continue because the recent American Rescue Plan Act appropriated $2 billion for improved UI administration.

Some state UI systems struggled with payment timeliness

The benefits of UI—both for individual workers and for the overall economy—depend on how quickly payments can be delivered. Many families have limited ability to cope financially with an unexpected job loss, and UI payments must be timely in order to fully realize their fiscal stimulus during a downturn. Delay is costly: an additional week of delay between a last paycheck and first UI payment is associated with a 2.25 percent fall in household spending (Farrell et al. 2020). This can mean extra economic insecurity for families as well as less effective stimulus for restoring economic growth.

Prior to the pandemic, states were uniformly quick to deliver UI, making nearly all payments within 21 days. However, backlogs in the UI system have historically been an issue during economic downturns, and the pandemic recession was no exception. For example, in May 2020, two months into the recession, Mississippi was only making 20.7 percent of first payments promptly, well below the 87 percent that the DOL considers acceptable. Survey evidence suggests that during the early days of the pandemic, high claims volume across the country led to some unemployed workers being unable to apply for benefits.

Almost all states had difficulty maintaining pre-pandemic levels of payment timeliness, especially in the summer and fall of 2020, but the difference in performance was substantial throughout the pandemic. Figure 3 shows that in the least timely month over the previous year, August 2020, only half of states made at least 54 percent of payments within 21 days. However, by April 2021, the month of peak timeliness, half of all states made at least 80 percent of their payments within 21 days.

The early months of the pandemic were extraordinarily challenging times for UI agencies. It is therefore notable that differences across states persisted long into the pandemic: in May 2021, Rhode Island (the top performer) was making 96 percent of first payments promptly, compared to just 34 percent in Ohio (the bottom performer). Figure 4 shows that percentage for each state in May 2021.⁷

The pandemic recession presented massive challenges in addition to the unprecedented number of applications. State agencies were tasked with implementing new programs like PUA that enrolled tens of millions more workers and changed the nature of their benefits.

These new programs took time to set up and administer, and some programs took longer than others. In general, states were able to distribute FPUC payments more quickly than PUA payments. As Figure 5 shows, half of state agencies made their first FPUC payments within 17 days after the passage of the CARES Act, while half of states took more than 38 days to make their first PUA payments. Some states were also quicker than others to implement each program. For example (not shown in the figure), Texas made its first PUA payments in 16 days, while Kentucky took 77 days. There was variation in setting up FPUC payments as well, with Illinois making its first payments in 10 days and Wisconsin taking 33 days.

A full year into the pandemic, PUA and the still-extraordinary overall UI caseload continue to strain state agencies. Although some state governments have prioritized investments in UI modernization (Simon-Mishel et al. 2020), outdated information-technology systems have in many cases made it difficult for UI agencies to adapt to ongoing policy changes and accommodate new applicants. Continued investment in UI agency staff and computing, outside of the context of the pandemic, may allow states to improve the timeliness of payments in the next downturn.

Ineligible and fraudulent claims are a persistent challenge

The immense stress placed on the UI system coincided with a high volume of overpayments, resulting in the loss (sometimes only temporarily) of billions of taxpayer dollars. As of March 15, 2021, there had been $2.6 billion in regular UI overpayments, as well as $3.6 billion in PUA overpayments, over the course of the previous year (GAO 2021).

Overpayment dollars increased sharply during the pandemic, both due to errors and intentional fraud. In some cases, applicants did not know whether they qualified for a particular program. This confusion is understandable given the complexity of the new UI programs and the fact that many state UI agencies encouraged all potentially eligible claimants to apply, assuming that their states’ screening infrastructure would be able to discern whether an applicant was, in fact, eligible. However, that screening was not perfect. “Non-fraud” overpayments, though they always exist to some extent, ballooned during the pandemic as both unemployment and the complexity of the UI system increased.

Overpayments due to fraudulent claims also spiked during the pandemic. However, currently available data may understate the true levels of overpayments and may not include all uncovered fraud (DOL Office of Inspector General 2021). For example, the state of California announced that $30 billion may have been lost to fraudulent claims between March 2020 and mid-January 2021. In contrast, the DOL’s data report only $946 million⁸ in fraudulent overpayments across the entire country in the last three quarters of 2020—almost certainly a considerable underestimate.⁹

Any program, particularly one with complicated eligibility requirements, will have some degree of overpayment. While non-fraudulent overpayments may have gone to ineligible individuals, fraudulent overpayments at times went to criminal enterprises during the pandemic. In an effort to reduce fraud, the federal government has created tools and resources for states through the DOL and provided funding to help them reduce fraud.¹⁰ State policymakers will have to balance cost, accuracy, timeliness, and accessibility to their UI programs in determining the level of overpayments that are acceptable. Decreasing overpayments could improve state trust-fund balances, but if the method is to spend additional time verifying eligibility, it could reduce timeliness or discourage applications for UI. UI system enhancements (e.g., improvements in processes and technology) that mitigate overpayments while minimally affecting legitimate recipients will be necessary to help UI better serve workers and taxpayers.

UI can be leveraged to prevent layoffs

UI during the pandemic recession was broadly effective in supporting workers who were laid off. However, it did comparatively little to help workers who retained their jobs but had their work hours reduced, and similarly did little to discourage layoffs in the first place. This was a missed opportunity given that a little-used program—STC (short-time compensation)—already exists to serve these purposes.

With STC, employers can temporarily reduce workers’ hours—between 10 and 60 percent—as an alternative to layoffs, and STC replaces some of the workers’ lost income.¹¹ For example, if an employer reduces a group of employees’ hours by 25 percent, those employees would be eligible to receive 25 percent of the UI that they would have received if they were laid off. This is in addition to the 75 percent of wages that the employee still receives for their continued employment.

Despite the benefits described in our “How can STC help during a recession?” sidebar, STC makes up a small fraction of overall UI and, as Figure 6 shows, STC programs exist in only 26 states. Each state sets its own requirements for employer eligibility, in terms of the allowable hours reductions and how those reductions can be applied across workers. Generally, employers must submit a plan to the UI agency outlining the expected duration of the hours reductions and the specific percentage of hours reduced for each affected group of workers.

As shown in Figure 7, STC historically makes up less than 1 percent of total UI continued claims. The proportion of STC usage peaks during severe economic downturns, such as the Great Recession and the pandemic recession, although it remains a small share of overall UI claims. Increased usage during recessions reflects the fact that STC is particularly well-suited to help firms during temporary declines in demand. Additionally, the federal government tends to provide states with funding for STC programs during recessions, as it did during the pandemic recession, when 100 percent of STC expenditures were federally funded (in states with STC programs).

Given the advantages of STC and its federal support during recessions, why does take-up remain low? Lack of employer awareness could be one possible hindrance to STC usage, and informational campaigns may increase STC utilization (Balducchi et al. 2015, Houseman et al. 2017). Another issue is that STC is most likely to be useful during times of economic distress for a firm, which is a difficult time for employers to research, enroll in, and administer a new program. A related concern is that administrative burden and lack of flexibility within the program could also contribute to low levels of use. In some states, firms are not allowed to hire workers into a unit that is currently under an STC agreement. (For example, New York allows hiring to replace workers who have left a unit that is under a work-sharing agreement, but no other hires in the unit are allowed.) Another administrative issue is that firms operating in multiple states may find it difficult to create plans that comply with each state’s STC requirements, because states vary in the percentage reduction of hours that is required for eligibility and in whether individuals within a plan can have different hours reductions. (For example, Texas allows workers to have varied percentage reductions within the same plan, provided they all fall within the eligible range.)

Looking ahead to post-pandemic UI

UI was arguably the most important single element of the fiscal response to the pandemic recession. State UI systems, working under unprecedentedly difficult conditions, provided vital support both for the overall economy and for workers who bore the economic brunt of the pandemic. However, the recession also exposed weaknesses in UI that policymakers should address prior to the next downturn:

• Incomplete coverage of workers. PUA dramatically expanded—on a temporary basis—who was eligible to receive UI benefits, but the choices made during a devastating pandemic are likely not the choices that make sense for a permanent expansion in UI coverage.
• Limited flexibility. The flat benefit add-on provided through FPUC was likely not the optimal UI policy, but it was infeasible for state agencies to quickly adjust their benefits formulas.
• Timeliness of payments. States varied in how prepared they were to set up new UI programs and accommodate huge surges in applications.
• Fraudulent claims. The imperative to quickly make payments and set up relatively generous new UI programs likely contributed to overpayments and fraud on a large scale.
• Limited ability to prevent layoffs. States that do not have STC programs should consider adding them, and states with STC programs should explore options for increasing employer take-up, thereby bolstering the ability of UI systems to prevent layoffs.

As the labor market recovers and demands on UI subside, federal and state policymakers should seize the opportunity to address the challenges highlighted during the pandemic recession. Researchers can contribute by studying the effects of CARES Act programs and illuminating the choices that policymakers face. As the pandemic demonstrated, it can be difficult or impossible to predict the next recession, making it all the more urgent to bolster the resilience and capacity of UI systems before they are tested again.

The authors thank Michele Evermore, Laura Kawano, and Libby Starling for insightful feedback.

Endnotes

1 See Table 3 from the Characteristics of Unemployment Insurance Applicants and Benefit Recipients Summary from the Bureau of Labor Statistics. The most common reasons the unemployed do not apply is because they believe they are not eligible, but significant shares of the unemployed decide not to apply for other reasons, including a negative attitude about the program or because they expect to start working again soon.

2 Authors’ calculations from DOL Employment and Training Administration (ETA) Table 902P data.

3 The overall figure is 42 percent using claims from all states and 40 percent when we restrict it to states with plausible reports of the fraction of claims coming from the self-employed.

4 Each state has its own eligibility rules for UI, but a number of states have added flexibility on what earnings can count toward eligibility. We would expect states with more flexible traditional UI requirements to have a smaller share of individuals on PUA, specifically non-self-employed PUA, than states with stricter requirements. When we test the relationship between the 2019 state-recipiency rates with the share of total weeks compensated under PUA relative to traditional UI, we find no clear evidence of a relationship.

5 These replacement rates are subject to a minimum and maximum weekly benefit amount. A further complication is that not all states determine weekly benefit levels based on a replacement rate.

6 See Simon-Mishel et al. (2020) for an in-depth report on UI-modernization efforts.

7 Delays also arose due to the timing of extensions of UI programs like PUA. Funding for PUA was set to expire on December 26 in accordance with the CARES Act; however, the next pandemic relief bill was not signed into law until December 27. Although the gap did not mean that unemployed individuals ultimately missed out on any payments, it did result in some delayed payments.

8 Includes regular UI, FPUC, PUA, PEUC, and PUA overpayments. Authors’ calculations of DOL ETA Table 227.

9 Overall reported overpayments during the pandemic as a percentage of total program payments actually fell relative to previous years. For example, state-reported regular UI overpayments were around 2 percent of total payments between the second and fourth quarters of 2020, compared to 3.4 percent in 2019 and 3.5 percent in 2018. Reported PUA overpayments were around 4.5 percent of benefits, with some states not reporting any detected overpayments. Future analysis of Benefit Accuracy Measurement (BAM) data may shed light on whether and to what extent the percentage of overpayments increased during the pandemic. The BAM program is the audit program that estimates overpayments in each state’s UI system using a selected sample, but many states requested a suspension of the program at the beginning of the pandemic due to high workloads. Our analysis uses state-reported ETA Table 227 information combined with DOL ETA Table 5159 information to compare benefit payments to state-reported actual overpayments uncovered by the state.

10 See this August 31, 2020, directive from the DOL describing tools available to identify fraud, and this February 12, 2021, DOL news release about funding to address fraud.

11 Actual eligible-hours percentage reductions vary by state, but they must be between 10 and 60 percent to meet DOL requirements for the program.