Coverage Losses by Congressional District Under the House ACA Repeal Bill

Waves of color surround the Capitol Dome in Washington, February 2017.

This column contains a correction.

For updated analysis, please see “Coverage Losses Under the ACA Repeal Bill for Congressional Districts in All States” by David Cutler and Emily Gee

Next week, the House is expected to vote on a bill that the Congressional Budget Office, or CBO, estimates would result in 24 million fewer Americans having health insurance coverage by 2026. The American Health Care Act, or AHCA, slashes funding for Medicaid, raises costs for consumers in the individual market, and reduces incentives for employers to offer insurance to their workers.

The CBO’s national coverage reduction estimate translates to an average of 55,000 people in each congressional district losing health insurance over the next decade. The coverage losses by district would vary depending on local demographics and state policy decisions.

In the table provided below, for each district, we have calculated the number of people who would be uninsured under the AHCA for several categories of coverage types. In general, we used the CBO’s national estimate and allocated it proportionally among districts; our detailed methodology is described in the appendix.

Most of the AHCA’s coverage reductions would come from the Medicaid program. By 2026, the CBO estimates that, under the AHCA, 5 million fewer adults would be covered by states opting into the Affordable Care Act’s Medicaid expansion in the future.* An additional 9 million people in current Medicaid expansion states and Medicaid’s traditional eligibility categories, including adults, children, and disabled individuals, would be uninsured.

Only the nonelderly, however, are included in the CBO’s projection of Medicaid coverage cuts. We estimate that an additional 914,000 seniors—or about 2,100 in each district—could be stripped of Medicaid coverage due to the AHCA’s caps on funding for that program. Although these seniors would still be insured by Medicare, they currently depend on Medicaid for financial assistance with premiums, out-of-pocket costs, and long-term care. The per capita caps on Medicaid spending could jeopardize nursing home access for seniors and their families.

The bill would also drive reductions in private insurance. By 2026, the CBO estimates 7 million fewer people would be enrolled in coverage through the workplace. The bill’s skimpy tax credits for the individual market, including the exchanges, and cuts to the Basic Health Program mean that 3 million people would no longer have coverage through the exchanges and other types of coverage.

Taken together, the coverage reductions would mean that 19 percent of nonelderly Americans—nearly 1 in 5 nationally—would lack insurance in 2026, compared with about 10 percent today. One has go back decades, to the 1960s, to find an uninsured rate that high. When it comes to health care coverage, the AHCA is truly a step backward.

David Cutler is the Otto Eckstein Professor of Applied Economics at Harvard University. Emily Gee is a Health Economist at the Center for American Progress.

To download the corrected table by congressional district, click here.

*Correction, March 21, 2017: This piece has been updated to correct the CBO’s estimates of changes in Medicaid coverage and to correct the numbers of North Dakotans and Louisianans enrolled in expanded Medicaid. The CBO projects that a total 14 million fewer people would have Medicaid coverage by 2026 under the House bill: 5 million fewer would be covered by additional Medicaid expansion in new states, and 9 million fewer would have Medicaid coverage in current expansion states and among pre-ACA eligibility groups in all states.

We have corrected our methodology for estimating Medicaid coverage reductions to reflect the CBO’s distinction between these groups. (Our methodology for all other coverage types remains the same as described below.) The CBO projects that under the ACA, additional Medicaid expansion would cover 5 million people and increase the proportion of the newly eligible population residing in expansion states from 50 percent to 80 percent by 2026. It projects that under the House bill, just 30 percent of the newly eligible population would be in expansion states. Extrapolating from the CBO’s numbers, we estimate the bill results in a Medicaid coverage reduction of 3.3 million enrollees in current expansion states by 2026. We then assume the remaining 5.7 million people who would lose Medicaid coverage are from the program’s pre-ACA eligibility categories. Applying the corresponding percentage reduction from the pre-ACA categories to elderly Medicaid enrollees implies that about 900,000 seniors could lose Medicaid coverage.

Because we do not know which individual states would participate in Medicaid expansion in 2026 in either scenario, our corrected estimates give nonexpansion states the average effect of forgone expansion and all expansion states the average effect of rolling back eligibility. We divided the 5 million enrollment reduction due to forgone expansion among nonexpansion states’ districts proportionally by the number of low-income uninsured. We made each expansion state’s share of that 3.3 million proportional to its Medicaid expansion enrollment in its most recent CMS report and then allocated state totals to districts proportional to the increase in nonelderly adult enrollment between 2013 and 2015. For Louisiana, which recently expanded Medicaid, we took our statewide total from state data and allocated to districts by the number of low-income uninsured adults.

Appendix: Methodology

Our estimates of coverage reductions by congressional district come from combining the CBO’s projected national net effects on coverage with state and local data from the Kaiser Family Foundation, the American Community Survey from the U.S. Census, and administrative data from the Centers for Medicare & Medicaid Services, or CMS. (Due to recent redistricting in Florida, North Carolina, and Virginia, the 114th congressional districts reported in Census data do not correspond to district boundaries for the 115th Congress. We therefore did not provide full coverage losses for those districts.)

Our calculation of reductions in employer-sponsored insurance were most straightforward. The CBO projects a net 7 million fewer people will obtain insurance through the workplace by 2026, and we assigned each district a share proportional to its number of nonelderly people with employer-sponsored coverage in the 2015 American Community Survey.

Estimating reductions in Medicaid by district required more assumptions. For coverage under the ACA’s Medicaid expansion, we had to decide which states would participate in expansion in future years.** We chose to freeze state expansion status as of March 2017. This essentially assumes that any current expansion state would partially cut back its program. We made each expansion state’s share of the 5 million proportional to its Medicaid expansion population in CMS’s 2016 enrollment report, which we supplemented with state data for Louisiana, and then allocated to districts by each’s increase in nonelderly adult enrollment between 2013 and 2015.

Another 9 million people would lose coverage under Medicaid’s traditional eligibility categories: low-income adults, low-income children, the aged, and disabled individuals. We used enrollment tables published by MACPAC to determine total state enrollment and each eligibility category’s share of the total, and we assumed that only some of the disabled were nonelderly. We then divided state totals among districts according to each’s Medicaid enrollment in the American Community Survey. Because each of the major nonexpansion categories is subject to per capita caps under the bill, we reduced enrollment in all by the same percentage.

Medicaid covers seniors who qualify as aged or disabled. Although the CBO did not specify the Medicaid coverage reduction that would occur among seniors under per capita caps, applying to elderly enrollees the same percentage reduction we calculated for nonexpansion Medicaid enrollees implies that 1.4 million could lose Medicaid.

Lastly, our estimates of the reduction in exchange, the Basic Health Plan, and other nongroup coverage are proportional to the Kaiser Family Foundation’s estimates of exchange enrollment by congressional district. The House bill reduces enrollment in nongroup coverage, including the exchanges, by 2 million relative to the ACA. To apportion this coverage loss among congressional districts, we assumed that the coverage losses would be largest in areas with higher ACA exchange enrollment and in states where we estimate the average cost per enrollee would increase most under the House plan. The CBO projects that 1 million people with “other coverage” types, including BHP, will be uninsured as of 2026. We assigned the 1 million reduction to the two states that currently offer BHP: Minnesota and New York.

**Authors’ note: The CBO predicts that additional states would elect to expand Medicaid coverage by 2026 under the ACA and some current expansion states would drop their programs under the AHCA, but it did not specify which states would change their expansion policies.