Testimony on: Insurance Coverage and the COVID-19 Pandemic

U.S. House of Representatives, Committee on Energy and Commerce Subcommittee on Health

*The views expressed here are my own and not those of the American Action Forum. I am indebted to my colleague Christopher Holt for his assistance.


Chairwoman Eshoo, Ranking Member Burgess, and members of the committee, thank you for the privilege of appearing today to discuss insurance coverage and the COVID-19 pandemic. In this testimony, I hope to make three main points:

  • The enormous scale of job losses in the 2nd quarter of 2020 has raised the specter of widespread loss of employer-sponsored health insurance;
  • To date, however, there is no compelling evidence of a dramatic change in the fraction of the population without health insurance; and
  • To the extent that there is diminished employer-sponsored insurance, the marketplaces of the Affordable Care Act should be positioned to provide alternative coverage.

Let me discuss these issues in greater detail.

The Pre-Pandemic Insurance Baseline

Released last week, the Census Bureau’s report on health insurance coverage in 2019 is a timely look at insurance coverage prior to the onset of the COVID-19 pandemic. The report includes data from two different surveys: the Current Population Survey Annual Social and Economic Supplement (CPS ASEC) and the American Community Survey (ACS). According to the CPS ASEC survey, in 2019, 26.1 million (8 percent) of people residing in the United States were uninsured for the entire year. The ACS survey found similar but slightly higher uninsured numbers with 29.6 million (9.2 percent) of respondents reporting being uninsured at the time they were interviewed. Both surveys show that roughly 56 percent of the population was receiving health insurance coverage through employer-sponsored insurance (ESI) in 2019.[1]

The CPS ASEC survey provides characteristics of the uninsured, giving more perspective on those most likely to be uninsured prior to the pandemic. Unsurprisingly, those age 65 and older were least likely to be uninsured in 2019, with only 1.1 percent reporting having been uninsured that year. Adults age 19 to 64 were the most likely demographic to be uninsured at 11.1 percent, while only 5.2 percent of children were uninsured in 2019.

Race and ethnicity showed some of the largest disparities in rates of insurance coverage. Only 5.2 percent of White, non-Hispanics, and 6.2 percent of Asians were uninsured in 2019, while 9.6 percent of African Americans reported being uninsured. Most notably, 16.7 percent of Hispanics (inclusive of all races) reported being uninsured for all of 2019.

Poverty and employment also contribute to insurance disparities. Almost 16 percent of individuals below the federal poverty level (FPL) reported being uninsured in 2019. In contrast, 11.3 percent of those between 100 percent and 399 percent of FPL reported being uninsured, and only 3 percent of those at or above 400 percent of FPL. Similarly, 8.9 percent of full-time, year-round workers reported being uninsured in 2019, while 13.4 percent of those who worked less than full-time or were seasonally employed and 12.2 percent of those who did not work at all reported being uninsured in 2019.

Finally, marital status drove additional disparities, with 7.6 percent of married individuals and 15 percent of unmarried individuals reporting that they were uninsured in 2019. The report notes that this is likely indicative of the reality that many adults access health insurance through a spouse’s employer.

The Pandemic’s Impact on Insurance Coverage

According to the ACS and CPS ASEC surveys, between 178.9 million and 183 million U.S. residents received health insurance through an employer in 2019. In the second quarter of 2020, the spread of COVID-19 produced a sharp downturn, and furloughs and layoffs began to roil the American labor market. In total the U.S. economy lost 22.2 million jobs in March and April of 2020,[2] and according to the Bureau of Labor Statistics, as of August there were 13.8 million unemployed, as opposed to 4 million in February.[3] The substantial job loss, paired with the large number of Americans who receive ESI, has led to appropriate concerns about how many individuals have lost insurance coverage during the pandemic. The reality is that we do not yet have a full picture of the pandemic’s impact on insurance coverage, but new census data along with several recent studies do allow us to sketch out in broad strokes the range of impact we may be facing.

The Kaiser Family Foundation (KFF) estimated in a May 2020 report that 78 million individuals lived in a household that experienced job lost between March 1, 2020, and May 2, 2020. The report further estimated that 61 percent of those households were receiving ESI at the time they became unemployed, and that 9 percent were uninsured. Both figures are roughly consistent with the findings of the ACS and CPS ASEC surveys. KFF concludes that “nearly 27 million people could potentially” have lost insurance coverage because of pandemic-related job loss between March and May. The report indicates that as many as 19 million other individuals may have lost ESI initially but were able to switch to another family member’s ESI. KFF found that the vast majority of the 27 million people estimated to have entirely lost ESI had alternative insurance option available to them. Roughly 12.7 million were eligible for Medicaid, and 8.4 million qualified for subsidies through the Marketplace under the Affordable Care Act (ACA). KFF found that 5.7 million people who lost ESI coverage between March and May 2020 did not qualify for federal assistance. Of those, 3.7 million had incomes too high to qualify for ACA subsidies (though they may have accessed COBRA or non-subsidized individual market coverage), 150,000 fell into the coverage gap, and 530,000 individuals were ineligible because of their immigration status. Finally, 1.3 million had offers of ESI through another family member that precluded subsidies. By 2021, the report expects that even fewer of the uninsured will remain without options as more qualify for Medicaid and ACA subsidies the longer their unemployment continues.[4]

A second study, released by the Urban Institute in July, estimated—based on microsimulation modeling—that 10.1 million individuals will live in households that lose ESI coverage as a result of unemployment in the last three quarters of 2020. The modeling predicts, however, that of those 10.1 million, 32 percent will switch to another family member’s ESI, 28 percent will enroll in Medicaid, and 6 percent will purchase individual coverage, mostly using subsidies under the ACA. The modeling concludes that 3.5 million individuals will be left without insurance coverage as a result of pandemic-related job loss by the end of 2020.[5]

A third study, by Families USA, concluded that 5.4 million workers who experienced job loss between February and May 2020 became uninsured, but this study does not extrapolate those figures to account for additional household member who may have lost coverage.[6]

Finally, the Economic Policy Institute found in August that as many as 9 million workers have lost ESI coverage as a result of the pandemic but that recent job gains have erased some of this loss, for a net of 6.2 million workers losing ESI through June. This report also does not account for additional household members, and the authors note that not all those 6.2 million workers have become uninsured because some will have been eligible for Medicaid and subsidized exchange coverage under the ACA.[7]

More recently, there has been some direct surveying of the health insurance status of Americans in the Household Pulse Survey (“Pulse”) from the Census Bureau.[8] The first Pulse survey was conducted between April 23 and May 5, while the most recent survey (week 13) took place between August 19 and August 31. Before turning to the evolution of insurance status, two comments are of note. First, the Pulse reported a total of 23.2 million uninsured in week 1, substantially lower than the results from 2019 reported above. This is likely due to the differences in sampling techniques and the fact that Pulse is restricted to individuals 18 years of age and older. Second, the first week of the Pulse is toward the end of April; there were quite likely substantial job losses and potentially lost insurance coverage in late March and early April.

With those caveats in mind, Table 1 shows the change in health insurance status between week 1 and week 13. The table paints a striking picture quite different from the fears of large-scale insurance loss. Indeed, taken at face value, the first row shows that overall the uninsurance rate fell by 0.6 percent, attributable to a rise in both private and public insurance (respondents were permitted to respond with more than one source).

The remainder of Table 1 shows how insurance status changed for individuals of varying characteristics.  Rather than stressing individual data points, I would argue that the most important feature of the table is the lack of uniformity; the rate of uninsurance moves up and down across nearly all the characteristics. Since these are noisy survey data, probably the best way to interpret the results is that they suggest little dramatic change in insurance status over the time period covered.

In summary, microsimulation modeling, the number of job losses, and common sense all suggest that there may have been a large rise in the fraction of uninsured individuals. In contrast, direct survey data indicate little dramatic change in insurance relationships. Only time and more data will resolve this discrepancy.



Table 1

Change in Health Insurance Status, by Select Characteristics

Census Household Pulse Survey – Week 13 versus Week 1

Population Aged 18 and Older






Total Private Public
Total 0.6% 2.3% 0.7% -0.6%
    18 – 24 0.7% 3.2% -3.5% -0.7%
    25 – 34 -0.1% 2.3% -3.8% 0.1%
    35 – 44 0.4% 1.3% -1.0% -0.4%
    45 – 64 0.9% 1.7% 1.3% -0.9%
    65 and over -0.2% 4.4% 1.5% 0.2%
    Male 0.8% 2.2% 0.9% -0.8%
    Female 0.4% 2.5% 0.4% -0.4%
    Hispanic/Latino -1.9% -1.1% 1.4% 1.9%
    White 0.8% 2.6% 0.1% -0.8%
    Black 1.3% 4.4% 2.3% -1.3%
    Asian 2.0% 2.1% 1.5% -2.0%
    Less than HS -4.4% -3.2% 2.3% 4.4%
    HS / GED 2.5% 3.9% 1.2% -2.5%
    Some college 0.0% 2.9% -0.4% 0.0%
    BA or higher 0.2% 1.0% 1.4% -0.2%
Marital status
    Married 0.5% 1.5% 1.8% -0.5%
    Single -9.8% 3.9% -37.2% 9.8%
    Children 0.9% 2.1% 0.2% -0.9%
    No children 0.4% 2.4% 0.8% -0.4%
Lost employment income
    Yes 0.7% 1.9% 1.2% -0.7%
    No 0.2% 2.3% 0.1% -0.2%
    Yes -0.2% 0.0% 1.7% 0.2%
    No 0.9% 2.0% 4.2% -0.9%
Income (1000s)
    Less than $25 -0.9% 2.1% -2.0% 0.9%
    $25 – $34 -0.9% 0.4% 0.6% 0.9%
    $35 – $49 2.0% -0.2% 6.8% -2.0%
    $50 – $74 0.4% -0.4% 4.2% -0.4%
    $75 – $99 0.3% 1.6% 2.4% -0.3%
    $100 – $149 0.8% 1.5% 1.5% -0.8%
    $150 – $199 0.2% 0.8% 2.6% -0.2%
    $200,000 + -0.7% -0.2% 0.3% 0.7%


Addressing the Problem of Insurance Coverage Loss Due to the Pandemic

Several policies have been put forward since the onset of the pandemic, and corresponding job loss, aimed at addressing the problem of the newly uninsured. As we have seen, however, the problem may be less expansive than originally thought, and many of the proposed policies are not well targeted.

Medicaid and the Affordable Care Act

There was understandable concern about the potentially significant impact of job loss on insurance coverage during the early days of the pandemic. While a complete picture is not yet available—as discussed in this testimony—a variety of sources indicate the impact on insurance coverage may not have been as great as feared. While undoubtably many individuals have lost ESI coverage, most appear to have had other coverage option available—in particular Medicaid and subsidized individual market coverage under the ACA.

Existing federal programs paired with alternative sources of ESI should be filling much of the gap in coverage. Indeed, in many ways the pandemic and recession are the test of the ACA. Some have also suggested increasing ACA subsidies or expanding Medicaid eligibility in response to the pandemic, but the data to this point do not indicate that either course of action is necessary uniquely because of the pandemic. Social safety net programs exist for moments like these and appear to be meeting the need without additional expansion or increased federal spending.

Individual Market Special Enrollment Period

One proposal, included in the Health and Economic Recovery Omnibus Emergency Solutions (HEROES) Act, is to establish a special enrollment period (SEP) for individuals seeking coverage through the ACA’s insurance exchanges. SEPs allow individuals to enroll in subsidized coverage outside of normal open enrollment periods. There are several existing circumstances that would trigger a SEP for an individual, for example a life event such as giving birth. The value of a SEP specific to those who have lost work because of COVID-19, however, is likely negligible given that a loss of ESI already triggers an automatic SEP for the individual and their dependents. In other words, this remedy is already in place.

COBRA Subsidies

Another proposal has been to provide a federal subsidy for COBRA premiums. COBRA is a transitional insurance program dating back to 1985 that allows employees to continue with their existing ESI plan for between 18 and 36 months in most cases, provided they pay both the employee and employer shares of the premium. COBRA provides continuity, but it also can be prohibitively expensive for the individual.

Subsidizing COBRA premiums would help recipients, but it’s not clear this would be the best use of federal dollars. As evidenced above, most people who have lost ESI already have access to either other ESI coverage through a family member, Medicaid, or subsidized coverage on the exchange. HEROES included a proposed 100 percent federal financing of COBRA premiums between March 1, 2020, and January 31, 2021.[9] Such a proposal would almost certainly be more costly to taxpayers than covering these individuals through Medicaid or subsidized exchange coverage.

There is some precedent for subsidizing COBRA; the American Reinvestment and Recovery Act—passed in response to the financial crisis—provided a tax credit covering 65 percent of the cost and an unemployed worker’s COBRA premiums. The subsidy did not prove particularly popular, however, with only an estimated 38 percent of eligible enrollees taking advantage of it.[10]

At the same time, there may be value in keeping people in the insurance plan and provider network they are already familiar with for the remainder of the plan year. Amid all the other disruptions buffeting the country, having millions of newly unemployed Americans trying to acclimate to a new insurance plan and provider network in the middle of a public health crisis could be counterproductive. To preserve continuity, the newly unemployed could be granted a COBRA subsidy equivalent to the ACA premium subsidy they would receive on the exchange through the end of 2020, at which time they could transition to exchange or Medicaid coverage for the start of the new plan year.

Short-Term, Limited-Duration Insurance

Short-term, limited-duration health insurance plans (STLDIs) have been around for years. Traditionally they have functioned as a bridge to longer-term health insurance options. For example, an individual moving between employers who expects to have a two-month gap in their health insurance might consider enrolling in one of these plans for the gap period. The plans became controversial because they exist outside of the ACA’s insurance market reforms, which gives insurers a lot of flexibility. Insurers can pick and choose what they cover and who they sell to, and they can set charges based on health status. These plans can be dramatically cheaper than what’s available in the ACA exchanges—in part because they can offer dramatically less coverage—but this can make them an attractive option for healthy people with incomes too high to qualify for federal subsidies. The Obama Administration issued a rule in October 2016 limiting the duration of these plans to less than three months, significantly shorter than their traditional duration of less than 12 months. The Trump Administration has reversed course on that rulemaking and allowed STLDIs to be issued for up to 364 days. Additionally, STLDIs are now renewable—at the discretion of the issuer—for up to 36 months.[11]

The reality is that these plans are not ideal for everyone, or even most people, long-term. They do, however, already exist, and provide an option for those who are recently unemployed and seeking low cost, catastrophic coverage for a limited period.

Thank you. I look forward to answering your questons.




[1] https://www.census.gov/content/dam/Census/library/publications/2020/demo/p60-271.pdf

[2] https://www.whitehouse.gov/articles/june-jobs-numbers-shatter-expectations/#:~:text=From%20February%20to%20April%2C%20nearly,of%20those%20jobs%20have%20returned.&text=According%20to%20the%20BLS%2C%20the,percentage%20points%20for%20Black%20Americans.

[3] https://www.bls.gov/news.release/pdf/empsit.pdf

[4] https://www.kff.org/coronavirus-covid-19/issue-brief/eligibility-for-aca-health-coverage-following-job-loss/

[5] https://www.urban.org/sites/default/files/publication/102552/changes-in-health-insurance-coverage-due-to-the-covid-19-recession_4.pdf

[6] https://www.familiesusa.org/wp-content/uploads/2020/07/COV-254_Coverage-Loss_Report_7-17-20.pdf

[7] https://www.epi.org/publication/health-insurance-and-the-covid-19-shock/

[8] https://www.census.gov/data/experimental-data-products/household-pulse-survey.html

[9] https://www.americanactionforum.org/insight/health-policy-provisions-in-the-heroes-act/

[10] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3513614/

[11] https://www.americanactionforum.org/weekly-checkup/sizing-up-short-term-insurance/