Twenty-one million people in the United States were unemployed in May — more than 13 percent of the U.S. labor force.1 With more than half the U.S. adult population dependent on their employer or a family member’s employer for their health insurance, a key concern is that many of the newly unemployed have lost their health benefits along with their jobs.2
Federal survey data will not be available for some time to assess COVID-19’s impact on health insurance coverage. To provide an early look at how the pandemic has affected health coverage, the Commonwealth Fund and the survey research firm SSRS conducted a nationally representative telephone survey of 2,271 adults ages 18 and older from May 13 through June 2, 2020. We asked respondents whether they or their spouse or partner had lost a job or been furloughed since February, whether either or both had health insurance through that job, and what their current insurance status was. We also asked for their views on balancing health risks with opening the economy and whom they most trusted to make that determination.
- Among people who said they or a spouse or partner lost a job or were furloughed because of the pandemic, two of five had health coverage through the affected job.
- Among those who previously had coverage through an affected job, one of five said they or a spouse or partner were now uninsured.
One of five (21%) respondents who said they and/or their spouse or partner were working part-time or full-time on February 1, 2020, reported that they and/or their spouse or partner lost their job or were furloughed from their job as a result of the pandemic. Of this group, over half reported that either they and/or a spouse or partner had been furloughed (data not shown). About one of five reporting job loss were self-employed. Hispanics and people with incomes under $50,000 reported higher rates of job dislocation than whites and people with incomes of $50,000 or above. While Blacks experienced job dislocation at a higher rate than whites, the difference was not statistically significant.
Among people who reported job dislocation, 41 percent said they or their spouse or partner had insurance through the job in question.
Employer-based coverage remains an important source of health insurance for workers reporting coverage through a job affected by the pandemic. This is in large part because more than half of those experiencing job dislocation have not lost their job but are on furlough.
Among respondents reporting that they and/or their spouse or partner had coverage through an affected job, over half said they and/or their spouse or partner were still on an employer health plan during furlough. Fourteen percent said that at least one of them had gained coverage through an employer, 10 percent had coverage or were planning to get covered through COBRA, and 7 percent said at least one person had enrolled in Medicaid or an individual-market plan. But one of five adults reported they and/or a spouse or partner were now uninsured.3
The majority of adults (59%) who said either they or their spouse or partner lost their job or had been furloughed because of COVID-19 reported that no one had coverage through that job. Among these respondents, three in 10 were uninsured in May–June, and most had been uninsured prior to the pandemic. About one-quarter of respondents reported they had coverage through a job not affected by the pandemic, including a spouse or partner’s job. More than one-third either had coverage through Medicare, Medicaid, or through a plan purchased through the individual market.
While the Affordable Care Act provides coverage options for people who don’t have, or lose, insurance through an employer-based plan, switching plans after losing a job can mean losing access to preferred physicians and other providers. When health coverage is not connected to a job, coverage and access to care can be more stable over time. And given the impact of rising employer plan premiums on many with low and moderate incomes, the availability of federal subsidies through marketplace plans may make such plans a more affordable option for many.4
We asked survey participants whether they think people who get their health insurance through an employer should have the option of getting similar coverage, at similar cost, through government-regulated and -subsidized health plans. Three-quarters of respondents overall said they believed people should have such an option. Democrats, including those leaning Democratic, expressed the strongest support, but about two-thirds of Republicans also held this view.
A vaccine for COVID-19 is months, if not years, away. In light of this, federal and state officials are attempting to balance uncertainty over risks to people’s health and to health care system capacity if infections spike with the effects of a continued lockdown on the economy and jobs. Even as states have started to reopen, many have experienced a surge in cases and hospitalizations.
We asked respondents whom they most trust to make a decision to open the economy given the need to keep people safe. The top response was a public health official like National Institute of Allergy and Infectious Diseases Director Anthony Fauci, M.D. (37%), followed by state governors (26%). President Trump and mayors or local government officials garnered about equal support (15% and 14%, respectively). The president was the most trusted by Republicans (32%) and the least trusted by Democrats (5%) and Blacks (4%).
One of the most fraught issues during the pandemic is ensuring a safe work environment, both for people who never stopped working and for those who have returned to their workplaces or will soon return as states reopen their economies. When asked about potential ways to help prevent virus transmission in the workplace, majorities of respondents agreed that people should be required to wear masks in public, that people should be tested regularly, and that the government should have the ability to track people testing positive for the coronavirus. Though Democrats support these measures more strongly, large majorities of Republicans also are in favor of them.
Conclusion and Policy Implications
This survey, conducted in late May through early June, provides a snapshot of U.S. health insurance coverage in the midst of a severe recession. Twenty-one million people were unemployed in May, according to the U.S. Department of Labor.5 While this was a slight decline from April, the department noted that the improvement was driven by a reduction in the number of people who were temporarily unemployed; the number on permanent layoff actually increased.6
In our survey, the large share of respondents who said they or a spouse or partner were still getting coverage through a furloughed job accounted for the relative stability of insurance coverage in the findings. Whether those on furlough regain their jobs, or lose them permanently, will determine the longer-term effect of the pandemic on employer-based coverage.
The survey findings also highlight the complexity of our health insurance system, and the ways in which families, both pre- and post-pandemic, piece together both their employment and coverage. It should also be noted that the survey only asked about spouses and partners; it did not take into account the coverage of children and other dependents. A full assessment of the pandemic’s effect on health coverage will need to also consider the implications for U.S. households.
Finally, the survey shows that about one in three people who lost jobs were uninsured before the pandemic hit. This points to the lingering gaps in our insurance system, driven by a lack of Medicaid expansion in 14 states; affordability barriers in buying coverage through the marketplaces; lack of awareness of coverage options, exacerbated by the Trump administration’s pull-back of advertising and general support for the ACA; and that undocumented immigrants are ineligible for subsidized coverage. Our findings suggest that the pandemic is likely increasing the number of people who are uninsured.
Federal and state policymakers have a host of options available to them, both short and long-term, to improve coverage. In the short run, they could mount an aggressive advertising and outreach effort to notify people who have lost employer-based coverage that they may be eligible for subsidized marketplace coverage or for Medicaid.7 Policymakers could also simplify documentation requirements for the marketplaces’ special enrollment periods.
In the long run, if the 14 states that have not yet expanded Medicaid eligibility continue to opt out, the federal government could design a subsidized insurance option, such as one offered through the marketplaces, for the more than 2 million very poor people who currently have no affordable coverage options.8 For those higher up the income scale, enhancing and extending the marketplace subsidies would address the affordability barriers that prevent many of these individuals from enrolling. And the survey shows majority, bipartisan support for allowing people with employer coverage the option to purchase a government-regulated and -subsidized health plan.
There is tremendous uncertainty over the future course of the pandemic and how long the accompanying economic crisis will last. Ensuring that everyone has access to affordable health care would at least relieve one major source of insecurity for millions of households.
How We Conducted This Survey
The Commonwealth Fund Health Care Poll: COVID-19, May–June 2020 was conducted by SSRS from May 13 through June 2, 2020. A total of 2,271 interviews were completed on the SSRS Omnibus. The survey was completed with the general population (adults 18 and older) and included an oversample of African Americans and Hispanics.* Approximately, 28 percent of the interviews were completed via landline (n=626) and 72 percent of the interviews were conducted via mobile phone (n=1,645).
This study asks respondents about changes in their employment and health insurance coverage, as well changes in the respondent’s partner/spouse’s employment and health insurance coverage, during the coronavirus pandemic. The unit of analysis is at the household level. “Household,” in this study, is defined as the respondent and the respondent’s partner/spouse. If both the respondent and partner/spouse had a change in employment, for example, we would consider this as one employment change for the household.
In total, 221 respondents reported they lost their job or were furloughed and their partner/spouse did not, 82 respondents said their partner/spouse lost their job or were furloughed and the respondent did not, and 30 respondents reported that both they and their partner/spouse lost their job or were furloughed. Of all respondents who reported they or their partner/spouse lost a job/furloughed, 60 respondents reported they had coverage through their own affected job and their partner/spouse (if married/living with a partner) did not, 25 respondents reported their partner/spouse had coverage through the partner/spouse’s affected job and the respondent did not, and 53 respondents reported both the respondent and partner/spouse had coverage through either their own or their partner/spouse’s affected job.
Each Omnibus survey consists of a minimum of 1,000 interviews, of which 700 interviews are completed with respondents on their cell phones and at least 35 are conducted in Spanish. Completes are representative of the United States population of adults 18 and older. The SSRS Omnibus uses a fully-replicated, stratified, single-stage, random-digit-dialing (RDD) sample of telephone households, and randomly generated cell phones.
Each SSRS Omnibus insert was weighted to provide nationally representative and projectable estimates of the adult population 18 years of age and older. The first stage weight or base weight took into account the disproportionate probabilities of household and respondent selection due to the number of separate telephone landlines and cell phones answered by respondents and their households, as well as the probability associated with the random selection of an individual household member. With the base weight applied, the sample was post-stratified and balanced by key demographics such as age, race, sex, region, and education. The sample was also weighted to reflect the distribution of phone usage in the general population. The study included an oversample of African American and Hispanic respondents. The weighting procedures adjusted for this oversampling.
The margin of error for this survey is +/– 2.4 percentage points. The landline portion of the survey achieved a 3.9 percent response rate and the cellular phone component achieved a 3.5 percent response rate. The overall response rate was 3.6 percent.
* This survey was included in the SSRS Omnibus for three weeks. To collect these 2,271 interviews, each week all African Americans and Hispanics were interviewed. Of the remaining respondents surveyed in the Omnibus, we completed interviews with seven out of every nine respondents age 18 and older for the first week. For the last two weeks, this number was decreased to ask six out of every nine respondents.
The authors thank Chris Hollander, Paul Frame, Jen Wilson, Jesse Baumgartner, David Blumenthal, Liz Fowler, and Eric Schneider, all of the Commonwealth Fund.