CBO Director Issues Letter on Estimates of Enrollment in Short-Term, Limited-Duration Insurance

WASHINGTON, Sept. 26 -- The Congressional Budget Office issued the following letter from Director Phillip L. Swagel:

To: Honorable Tammy Baldwin, United States Senate, Washington, DC 20510

Re: CBO's Estimates of Enrollment in Short-Term, Limited-Duration Insurance

Dear Senator:

This letter responds to your request that the Congressional Budget Office examine recent research about short-term, limited-duration insurance (STLDI), reexamine its characterization of STLDI, and update its estimates of enrollment in STLDI accordingly./1

CBO reviews its projections of health insurance coverage annually. After performing its most recent review, which accounted for recent research on STLDI, the agency has concluded that the evidence available to date does not warrant changing its expectation that there will be two categories of STLDI plans--those that provide what CBO considers insurance coverage and those that do not. CBO anticipates that the majority of people who enroll in STLDI as a result of the most recent regulations will enroll in plans that do provide insurance coverage. CBO continues to monitor developments in the market for STLDI and will revisit that assessment in the future if new evidence supports doing so.

Background

Before 2014, the traditional role of STLDI was to provide temporary health insurance for individuals, filling short gaps in coverage that might occur in certain situations, such as moving from one job to another. STLDI is regulated by more than half of the states./2 But it is exempt from federal regulations governing other nongroup insurance--that is, health insurance sold to individuals, rather than provided to them through their employer.

In 2014, new regulations related to the Patient Protection and Affordable Care Act (ACA) changed the requirements for insurance sold in the nongroup market./3 Anecdotal evidence suggests that some consumers then began enrolling in longer-term STLDI in lieu of other types of health insurance because the premiums were lower--a result of the exemption from the ACA's nongroup insurance requirements./4 In response, the Administration issued new regulations in October 2016 that limited enrollment in STLDI to three months, starting in that December./5

In August 2018, after a change in Administrations, the new Administration issued another rule governing STLDI that reverted to the previous limit on the duration of such plans (364 days) and also allowed people to renew their policies for up to three years./6 That rule took effect in October 2018. The rule was challenged in federal court, but a district court upheld it in July 2019, and that ruling was affirmed on appeal. The insurance plans newly allowed by the rule started to become available in the fall of 2019.

* * *

CBO's 2019 Report

In January 2019, CBO published a report describing the estimated effects on enrollment of the Administration's 2018 rule./7 On the basis of interviews with insurers, state regulators, and other stakeholders, CBO identified two types of STLDI that were likely to emerge: traditional short-term plans (TSPs) and insured short-term plans (ISPs).

- TSPs would include short-term plans that were available before the 2018 rule took effect and similar plans that would provide coverage for longer periods. TSPs would offer limited financial protection against some high-cost events but would exclude coverage for many services provided by physicians and hospitals, and therefore they would not meet CBO's definition of insurance coverage.

- Coverage offered by ISPs, by contrast, would be similar to certain kinds of coverage that were offered before the ACA's requirements took effect, and they would meet CBO's definition of insurance coverage. For example, ISPs would cover medical expenses resulting from non-elective hospitalizations regardless of whether they originated with a visit to the emergency room./8

In the report, CBO projected that each year between 2019 and 2028, roughly 1.5 million people would be enrolled in STLDI as a result of the 2018 rule. About half of those people would have otherwise been enrolled in nongroup coverage, and about half would have otherwise been uninsured.

CBO also projected that after the rule's effects were fully realized in 2022, over three-quarters of people purchasing STLDI as a result of the rule would be enrolled in ISPs. Although it would take longer for insurers to develop and offer ISPs, interviews with insurers, state regulators, and other stakeholders repeatedly suggested that the increase in enrollment resulting from the rule was likely to be concentrated in ISPs./9 And historical enrollment data for the nongroup market before 2014 (when the ACA's regulations took effect) suggested that the number of people who purchased coverage similar to ISPs far exceeded the number who purchased coverage similar to TSPs.

Why Recent Evidence Does Not Warrant Changing CBO's Treatment of STLDI Plans

Some recent research has found that many of the STLDI plans currently being sold provide very limited benefits to enrollees, and that research raises questions about whether such plans constitute insurance. But the research differs from CBO's analysis in three important ways and therefore does not provide evidence to support changing the agency's projections.

Different Periods Studied. Most of the available evidence about STLDI suggesting that it does not constitute health insurance coverage comes from the time before the 2018 rule took effect. Such STLDI typically provided limited coverage for three months or less.

For example, one study found that the five health insurers that earned the most premium revenues from STLDI spent only 39 percent of those premiums on enrollees' health care, compared with 73 percent for other coverage in the nongroup market--but that study used 2018 data./10 Similarly, the House Committee on Energy and Commerce gathered information about STLDI, including applications for enrollment in plans, marketing materials, and complaints from consumers--but most of those data were from 2018 and 2019, and insurers started selling the new types of STLDI only in the final months of that period./11

In its 2019 report, by contrast, CBO reported its projections of the average annual increase in STLDI enrollment that would result from the 2018 rule from 2019 to 2028, and those projections showed that the percentage of STLDI enrollment represented by ISPs would increase over the 10-year period. In estimating the effects of the rule on STLDI, CBO considered how long it would take insurers to develop and market the new types of coverage. The consensus among insurers, state regulators, and other stakeholders was that enrollment in TSPs would accelerate more quickly than enrollment in ISPs because TSPs were so similar to products already being sold, whereas insurers would need several years to create the necessary infrastructure to offer ISPs. CBO projected that increased enrollment in TSPs resulting from the 2018 rule would be observed starting in 2020 but that the effects on enrollment in both types of plans would not be fully evident until 2022. As a result, the percentage of people newly enrolled in STLDI who had TSPs would be larger from 2019 through 2021 than from 2022 through 2028.

Different Approaches to State Regulations. Because states have adopted different approaches to regulating STLDI, analyses of a sample of states may provide an unrepresentative picture of the market. For example, a recent study of STLDI in five states showed that those plans offered limited benefits, but those five states were among the minority of states that have not enacted regulations exceeding the federal requirements./12 Because coverage sold in those states is less regulated than coverage sold in most states, the STLDI plans considered in the study were likely to include fewer benefits and provide less extensive coverage than STLDI plans in a broader sample of states.

CBO's estimates, by contrast, accounted for the different kinds of state regulations affecting STLDI--increasing CBO's projections of enrollment in ISPs and reducing its projections of enrollment in TSPs. Specifically, some states entirely prohibit the sale of STLDI or limit the length of enrollment in STLDI to shorter periods than the federal rule allows, reducing enrollment in TSPs. Other states require STLDI to comply with the ACA's requirements for nongroup insurance, such as covering preexisting conditions and spending a defined percentage of premium revenues on enrollees' health care; those requirements increase enrollment in ISPs.

For example, Idaho has adopted regulations for two types of STLDI: plans that are similar to what CBO calls TSPs and "enhanced short-term plans," which are similar to ISPs. The enhanced short-term plans are required to cover preexisting conditions and the same categories of health benefits that other nongroup plans must cover. Although the requirements went into effect just this year, early evidence suggests that demand for the enhanced short-term plans has been considerable./13 Rhode Island also has regulations that could increase enrollment in ISPs: It requires STLDI to cover preexisting conditions, prohibits rescinding coverage when enrollees develop costly medical conditions, and requires plans to spend 80 percent of premium revenues on enrollees' health care.

Different Characterizations of Insurance Coverage. Other research frequently characterizes insurance coverage on the basis of its compliance with the ACA's requirements for nongroup health insurance. For example, under the ACA, insurers may not refuse to sell coverage to people with preexisting conditions or to renew their coverage. Insurers may not price premiums on the basis of people's health status or exclude the treatment of preexisting conditions from coverage. Insurance plans must cover 10 categories of health care services that federal law defines as essential and must spend at least 80 percent of premium revenues on payments for enrollees' health care./14

CBO, by contrast, defines health insurance coverage as a policy that covers high-cost medical events and includes coverage for services provided by physicians and hospitals. That definition includes plans that must comply with the ACA's regulations, but it also includes some coverage that is exempt from such regulations. For instance, it includes insurance coverage that was available before 2014, which may be exempt from many of the ACA's requirements.

Thus, the criteria used in recent research to characterize the comprehensiveness of coverage do not correspond to the criteria that CBO uses to determine whether coverage constitutes insurance or not. For example, the analysis of STLDI in five states explored whether STLDI plans were sold through an association; paid for health care services at substantially lower rates than other health plans did; lacked a provider network; failed to cover preexisting conditions, prescription drugs, maternity care, or mental health benefits; and denied enrollment on the basis of health status. Such characteristics were common in nongroup coverage in the years before the ACA took effect, and that coverage was generally considered insurance in CBO's cost estimates.15

It is difficult to determine whether the plans studied by the Energy and Commerce Committee would meet CBO's definition of health insurance, in part because the data describe insurers' practices but not the features of specific health plans. For example, many of the insurers that were analyzed offered "STLDI plans that include coverage for doctor's visits, hospitalization, urgent care visits, and emergency room visits subject to cost-sharing, including deductible and coinsurance." But it is not clear whether the plans included coverage of all of those services or whether some plans covered hospitalizations and others covered emergency care. The report also did not indicate how the coverage and cost sharing for such services would compare with coverage offered in the nongroup market./16

A related limitation of some analyses is that they draw from STLDI marketing materials and other plan documents but do not include enrollment in each plan or data describing payments for health care by specific STLDI plans. Such data would allow CBO to assess what share of STLDI products meet its definition of insurance coverage and how many people enroll in each type of coverage. Data about enrollees' health care spending--and the percentage of care that is paid for by STLDI plans--are critical to assessing whether or not the coverage is paying for enrollees' care when they have high-cost medical events. CBO understands that the National Association of Insurance Commissioners is conducting a comprehensive state survey of STLDI plans but has had difficulties, during the survey's first year, in collecting complete and consistent data from insurers.

New Data That Would Be Useful for Future Projections

CBO continues to monitor developments in the market for STLDI, and as new evidence becomes available, the agency will update its projections of enrollment in STLDI accordingly. Two types of data would be particularly useful as CBO projects future enrollment in STLDI plans.

- Claims Data for People Enrolled in STLDI and in Other Nongroup Plans for 2020 and Later. Such data would show what types of health care spending STLDI enrollees incur and how a plan's payments to providers for that health care compare between STLDI and other insurance products sold in the nongroup market.

- Enrollment in STLDI by State and by Year. Enrollment data would allow CBO to better understand how enrollment has changed since the 2018 rule took effect. State-level data are important because STLDI plans are likely to offer more comprehensive coverage in states that have required more comprehensive forms of STLDI.

State-level data would also allow CBO to observe whether STLDI is being sold in states that have prohibited such sales. Insurers often sell STLDI through associations, a practice that in some cases allows them to sell products that would otherwise be prohibited by state regulations. Although the practice is documented in several studies, including the recent report by the Energy and Commerce Committee, data about the extent of such sales nationally are incomplete. Similar challenges exist with other types of insurance coverage, such as farm bureau plans and health care sharing ministries, and CBO is very interested in whether there are enrollment data available that show the extent to which people are purchasing plans offered in another state when their own state prohibits such plans. CBO continues to monitor new research and new data sources in hopes of better understanding such issues when updating its projections of insurance coverage.

I hope you find this information useful, and I would be happy to discuss it with you.

Sincerely,

Phillip L. Swagel, Director

cc: Honorable Lamar Alexander, Chairman, Senate Committee on Health, Education, Labor, and Pensions

Honorable Patty Murray, Ranking Member, Senate Committee on Health, Education, Labor, and Pensions

Honorable Frank Pallone, Jr., Chairman, House Committee on Energy and Commerce

Honorable Greg Walden, Ranking Member, House Committee on Energy and Commerce

Identical letter sent to the Honorable Chris Murphy.

* * *

Footnotes:

1 The Honorable Tammy Baldwin and the Honorable Chris Murphy, United States Senate, letter to Phillip Swagel, Director, Congressional Budget Office (August 13, 2020), https://go.usa.gov/xGVY5 (PDF, 68 KB).

2 For a list of state regulations of STLDI, see Dania Palanker, Maanasa Kona, and Emily Curran, States Step Up to Protect Insurance Markets and Consumers From Short-Term Health Plans (Commonwealth Fund, May 2019), https://tinyurl.com/y56jep65.

3 For a description of the new regulations and how CBO expected they would affect premiums for nongroup coverage, see Congressional Budget Office, Private Health Insurance Premiums and Federal Policy (February 2016), pp. 17-26, www.cbo.gov/publication/51130.

4 See Anna Wilde Mathews, "Sales of Short-Term Health Policies Surge," The Wall Street Journal (April 10, 2016), https://tinyurl.com/yxjbkw2s.

5 Excepted Benefits; Lifetime and Annual Limits; and Short-Term, Limited-Duration Insurance, 81 Fed. Reg. 75316 (October 31, 2016), https://go.usa.gov/xGYWb (codified as amended at various sections of 26, 29, and 45 C.F.R. (2020)).

6 Short-Term, Limited-Duration Insurance, 83 Fed. Reg. 38212 (August 3, 2018), https://go.usa.gov/xEcKs (codified at various sections of 26, 29, and 45 C.F.R. (2020)).

7 Congressional Budget Office, How CBO and JCT Analyzed Coverage Effects of New Rules for Association Health Plans and Short-Term Plans (January 2019), www.cbo.gov/publication/54915. The report also analyzed the effects of a new rule governing association health plans.

8 See Congressional Budget Office, Health Insurance Coverage for People Under Age 65: Definitions and Estimates for 2015 to 2018 (April 2019), www.cbo.gov/publication/55094.

9 CBO's projections in the 2019 report showed the increase in STLDI enrollment that was expected to result from the August 2018 rule, but those projections did not include all STLDI coverage. In CBO's assessment, analyses of the entire STLDI market would have shown a larger percentage of plans that did not provide what CBO defines as insurance coverage, because plans with durations of up to three months were already being sold before the August 2018 rule took effect. Such plans tended to provide coverage only for emergency care and not to provide coverage for preexisting conditions or preventive care.

10 Shelby Livingston, "Short-Term Health Plans Spend Little on Medical Care," Modern Healthcare (August 6, 2019), https://tinyurl.com/yyxzlfx2.

11 House Committee on Energy and Commerce, Shortchanged: How the Trump Administration's Expansion of Junk Short-Term Health Insurance Plans Is Putting Americans at Risk (June 2020), https://go.usa.gov/xGYWH.

12 Dania Palanker, Emily Curran, and Arreyellen Salyards, Limitations of Short-Term Health Plans Persist Despite Predictions That They'd Evolve (Commonwealth Fund, July 2020), https://tinyurl.com/y69f5ycw.

13 See John Tozzi, "Idaho's 'Enhanced Short-Term Plans' a Hit With Consumers," BenefitsPRO (January 14, 2020), https://tinyurl.com/y6z7gjpl; and Office of the Governor, Idaho, "Enhanced Short-Term Plans Available for Idaho Families in 2020" (press release, December 17, 2019), https://go.usa.gov/xGYWV.

14 Although most nongroup health insurance now meets the requirements established under the ACA, before 2014, more than half of people enrolled in nongroup coverage were in plans that would not have met those requirements. See Jon R. Gabel and others, "More Than Half of Individual Health Plans Offer Coverage That Falls Short of What Can Be Sold Through Exchanges as of 2014," Health Affairs, vol. 31, no. 6 (June 2012), pp. 1339-1348, https://dx.doi.org/10.1377/hlthaff.2011.1082.

15 Before the ACA was enacted, most plans sold in the nongroup market denied enrollment on the basis of health status or engaged in other forms of "underwriting," in which insurers evaluated people's health status before agreeing to issue them an insurance policy. In some cases, insurers denied enrollment to people with preexisting conditions, often providing comprehensive coverage to people who were healthier. In other cases, insurers issued coverage to people with preexisting conditions but priced that coverage to reflect those people's higher expected health care spending, imposed waiting periods before covering the preexisting conditions, or refused to cover health care claims for the preexisting conditions. Although such practices frequently prevented people from purchasing insurance, the policies sold generally did provide coverage (according to CBO's definition) to those who were able to get through the underwriting process.

Historically, STLDI plans did not engage in that type of underwriting, which is costly, because people were not enrolled in them for very long. Instead, the plans notified enrollees that they did not cover preexisting conditions, and when health care claims were incurred, the plans determined whether such claims were for preexisting conditions or for new or unexpected health care needs. The plans then denied claims that were for preexisting conditions. One of CBO's reasons for anticipating that it would take longer to develop ISPs was that insurers indicated that they would be likely to start using more traditional underwriting after the 2018 rule took effect.

16 A plan's cost-sharing requirements are the out-of-pocket expenses that enrollees pay.