Eligibility for ACA Health Coverage Following Job Loss
|Methods and Definitions|
|This analysis uses our ACA eligibility model as applied to the 2018 American Community Survey as a baseline for all calculations. We rely on these calculations to assess the ACA eligibility of a cohort of workers prior to the pandemic (early 2020), during the pandemic (mid-2020), and in the following calendar year (early 2021). Assessing both insurance coverage changes and ACA eligibility at three time points might help policymakers understand both the immediate coverage needs of the population losing jobs due to the pandemic and the longer-term eligibility of the same population assuming they continue without wages into 2021.
In order to estimate the 2020 population within each state, we linearly extrapolated 2020 state population estimates based on 2018 and 2019 population estimates from the U.S. Census Bureau to determine a population increase factor between 2018 and 2020 within each state. We then applied this multiplier to the weight of each individual in the microdata to approximate state population sizes in mid-2020 rather than mid-2018. With the exception of this population multiplier, our baseline estimates (described in this brief as “May 2020”) align with other Kaiser Family Foundation products such as our ACA eligibility estimates of the uninsured population.
We summed initial unemployment insurance claims filed across the weeks ending March 7th, 2020 thru May 2nd, 2020 using the Department of Labor’s Employment & Training Administration state-specific statistics to arrive at a nationwide total job loss through early May of approximately 31 million workers. We also assumed unauthorized immigrants in the labor force lost employment proportionally without filing for unemployment. We did not make assumptions about other people losing jobs but not filing for unemployment insurance.
Within each state, we estimated who lost employment using sampling probabilities based on recent labor force changes by industry recorded by the March 2020 Current Population Survey. For example, leisure and hospitality workers appear more than five times as likely as agricultural workers to have lost a job in March 2020, and these relative probabilities guided sampling of who has become unemployed. We controlled to state unemployment totals (approximately 31 million nationally) for the citizen and legally-present immigrant population, and we separately controlled to state proportional unauthorized labor force unemployment estimate for the undocumented working population.
The American Community Survey does not distinguish between ESI policyholders and those covered as dependents. For all full-time workers losing jobs in our sample, we assumed a family-wide loss of ESI for all people who held ESI if there were no other workers present who both worked at least 30 hour weeks and earned at least $50,000 during the year. If a spouse with wage of at least $50,000 and weekly average hours over 30 were present within the family, we assumed the spouse held the policy (or another policy) and maintained ESI for the entire family. For part-time workers losing jobs, we assumed a family-wide loss of ESI only when no other workers were present within the family.
We calculated an industry-specific distribution of weekly state unemployment benefit payments from the 2019 Current Population Survey. We then applied a weekly state dollar amount onto most individuals who lost employment according to a random deviate sample using a gamma distribution. After adding unemployment benefit payments onto family income of those imputed to lose jobs, we then scaled each individual weekly payment to account for state-specific generosity using the Department of Labor’s 2018Q4 “Benefits Paid for Total Unemployment divided by Weeks Compensated for Total Unemployment” state-specific estimate divided by the nationwide average of $361.29. This nationwide weekly average amount matched our CPS-based calculated average. For any individuals imputed to receive a higher weekly state unemployment payment than the state maximum, we capped the imputed amount at the state maximum.
Medicaid eligibility is based on current monthly income. To calculate Medicaid eligibility immediately after job loss, we zeroed out wage and self-employment income for people who lost jobs and calculated monthly family income as a share of poverty based on other family income and the state weekly unemployment benefit. Following Medicaid eligibility policy, we did not include the Federal weekly supplemental unemployment payments of $600 in the Medicaid eligibility determination.
ACA marketplace subsidy eligibility is based on estimated annual income. To calculate ACA Marketplace subsidy eligibility immediately after job loss, we removed a share of annualized wages and self-employment income in proportion to the calendar week of job loss. For example, calendar year earned income for individuals imputed to lose jobs during the week of March 7th, 2020 were reduced by 75%. We also counted the receipt of Federal supplemental unemployment insurance payments of $600 for 17 weeks and multiplied the same imputed weekly state unemployment benefit by the maximum allowable weeks.
To re-calculate both Medicaid and ACA Marketplace subsidy eligibility for 2021, we assumed an exhaustion of both the state and Federal unemployment benefit amounts, no return to work among job losers, and counted only other income in the family.
Although our job loss imputation only edited the earned income and public assistance income of the individual worker, that worker’s income changes affect the Medicaid and marketplace tax credit eligibility of family members. Therefore, many statistics throughout this brief present the eligibility dynamics of Americans with any job loss in their family rather than solely the worker.
Medicaid/Other Public Eligible: Includes adults and children who were previously eligible for Medicaid and the Children’s Health Insurance Program (CHIP) but not enrolled as well as those newly eligible after job loss. Also includes some state-funded programs for immigrants otherwise ineligible for Medicaid.
Tax Credit Eligible: Includes individuals who are not eligible for other coverage, such as Medicaid or Employer-Sponsored Insurance (ESI), and who have incomes between 100% and 400% of the federal poverty level (FPL). This number also includes legally residing immigrants with incomes below the poverty level who do not qualify for Medicaid because they have lived in the U.S. for less than five years. Tax credit-eligible population in Minnesota and New York include uninsured adults who are eligible for coverage through the Basic Health Plan.
Ineligible for Financial Assistance due to Income, ESI Offer, or Citizenship: Includes individuals with incomes above 400% FPL and those with an offer of coverage from an employer (though we cannot determine whether the offer of ESI would be considered affordable under the ACA, which would make the individual ineligible for a premium tax credit). This number also includes undocumented immigrants who are barred from purchasing coverage through the Marketplace even without financial assistance.
In the Coverage Gap: Includes uninsured adults in states that have not expanded Medicaid and have incomes above the state’s Medicaid eligibility level (which, in many cases, is 0% FPL for adults without dependent children) but below the poverty, leading them to earn too much to qualify for Medicaid but not enough to qualify for tax credits. Adults in the coverage gap would be eligible for Medicaid if their state expanded under the ACA.