The Effect of Eliminating the Affordable Care Act's Tax Credits in Federally Facilitated Marketplaces

Key Findings

• Enrollment in the ACA-compliant individual market, including plans sold in the marketplaces and those sold outside of the marketplaces that comply with ACA regulations, would decline by 9.6 million, or 70 percent, in federally facilitated marketplace (FFM) states.
• Unsubsidized premiums in the ACA-compliant individual market would increase 47 percent in FFM states. This corresponds to a $1,610 annual increase for a 40-year-old nonsmoker purchasing a silver plan.

Since its passage in 2010, the Patient Protection and Affordable Care Act (ACA) (Pub. L. 111-148, 2010) has sustained numerous legal challenges. Most notably, in Nat. Fedn. of Indep. Business v. Sebelius (132 S. Ct. 2566, 2012), the U.S. Supreme Court upheld the individual-responsibility requirement to purchase health insurance under the federal government's taxing authority, but it made Medicaid expansion voluntary for states. The latest challenges to the ACA focus on whether residents of states that have not established their own insurance exchanges are eligible for subsidies under 26 U.S.C. § 36B. Although 16 states* and the District of Columbia have established their own exchanges, 34 states have not, instead deferring to the federal government to set up exchanges in their states. In its final rule, the Internal Revenue Service (IRS) interpreted the provision as allowing tax credits to be made available for eligible people purchasing health insurance in state-based marketplaces (SBMs) or federally facilitated marketplaces (FFMs) (45 C.F.R. § 155.20). Notwithstanding the statutory reference to “an Exchange established by the State,” supporters of the IRS's interpretation believe that federal lawmakers anticipated that each state would set up its own exchange and that they intended for eligible consumers in all states to have access to subsidies. However, opponents argue for a different interpretation of the provision, asserting that the IRS does not have the authority to allow subsidies in FFM states.

Legal challenges to the IRS's interpretation have worked their way through the federal court system. On July 22, 2014, two appellate courts reached conflicting decisions on the issue. In King v. Burwell (759 F. 3d 358, 2014), the Fourth Circuit ruled in favor of allowing subsidies in FFM states; in Halbig v. Burwell (758 F. 3d 390, 2014), a three-judge panel of the D.C. Circuit ruled against it. The full D.C. Circuit granted the federal government's petition for rehearing the case en banc on September 4, 2014. Prior to the case being heard before the full D.C. Circuit, the U.S. Supreme Court granted certiorari in King and will hear oral arguments in March 2015.

For two reasons, the framers of the ACA included subsidies for purchasing health insurance. Subsidies make health insurance affordable for a broad cross-section of low- and middle-income Americans. They also serve to reduce adverse selection by making health insurance more attractive to young and healthy people. Those who are young and healthy have a higher price elasticity for health insurance, meaning that they are more reactive to price and are more likely than older and less healthy people to forgo coverage if premiums are high. Without subsidies, the young and healthy are potentially exposed to higher premiums than they were before implementation of the ACA because the law prohibits insurers from using health history in developing premiums (except for an enrollee's smoking status) and restricts insurers' ability to charge lower premiums to younger enrollees.** The departure of young and healthy people from the health insurance market would cause premiums to rise further, leading to a cycle of additional departures and further premium increases. Such instability in the ACA-compliant individual market, which includes plans sold in the marketplaces and outside of the marketplaces that satisfy the minimum coverage and rating requirements of the ACA,*** would severely impair its viability and significantly reduce access to health insurance for millions of Americans under the current legislation.

A recent RAND report considered the implications of eliminating subsidies in all states (Eibner and Saltzman, 2014). The authors found that eliminating subsidies in all states would cause enrollment in the ACA-compliant individual market to drop by 13.5 million, or 68 percent, and result in a 43-percent spike in premiums in the individual health insurance market.

In this article, we assess the expected change in enrollment and premiums in the ACA-compliant individual market in FFM states if the Supreme Court eliminates subsidies in those states (as opposed to our previous report, which considered the impact of eliminating subsidies in all states). Our analysis assumes that FFM states will not set up their own exchanges in response to a Supreme Court decision that eliminates subsidies in an FFM. Key findings of our analysis include the following:

1. Enrollment in the ACA-compliant individual market, including plans sold in the marketplaces and those sold outside of the marketplaces that comply with ACA regulations, would decline by 9.6 million, or 70 percent, in FFM states.****
2. Unsubsidized premiums in the ACA-compliant individual market would increase 47 percent in FFM states. This corresponds to a $1,610 annual increase for a 40-year-old nonsmoker purchasing a silver plan.*****

We have organized the remainder of this article into three sections. The next section discusses the approach we employed to leverage our previous analysis. The subsequent section presents the results of our analysis, and the final section discusses implications of our findings.

Methods

In October 2014, RAND released a report that considered several modifications to the ACA. Among the modifications considered were alternative subsidy mechanisms to the ACA's existing subsidy formula, the elimination of the individual mandate, and the elimination of premium subsidies in all states. To derive our estimates, we employed the Comprehensive Assessment of Reform Efforts (COMPARE) microsimulation model, an economic tool used to assess the economic impact of health policy reforms. COMPARE contains synthetic populations of people and firms using data from the Survey of Income and Program Participation (SIPP), the Medical Expenditure Panel Survey Household Component (MEPS-HC), and the Kaiser Family Foundation/Health Research and Educational Trust Annual Survey of Employer Benefits. In the model, an individual chooses an insurance plan from among the choices available to him or her by weighing the costs and benefits of available options, a process known in economics as utility maximization. Some might opt to forgo insurance. Firms in the model decide whether to offer insurance to their employees by considering the aggregate preferences of their workers—e.g., if the model predicts that many workers will prefer employer coverage to other types of insurance, the firm is more likely to offer. Before introducing any health reforms, we conduct a thorough calibration process to ensure that the model can accurately predict the pre-ACA insurance market (i.e., in terms of enrollment distributions, premiums, and firm offer rates). Once we validate the model's ability to accurately represent the pre-ACA insurance market, we introduce proposed health reforms and compare their effects on the insurance market. Our previous report compared the uptake (i.e., the decision to obtain insurance) and cost of health insurance under implementation of the ACA with subsidies being offered in all states and the alternative scenario in which subsidies are eliminated in all states, finding strikingly different outcomes in the insurance market between the two alternatives.

To assess the impact of eliminating subsidies in only the 34 FFM states, we developed an approach that builds on the simulation output from our previous report. For this analysis, we examined people residing in FFM states from the simulation output. In theory, the response to the elimination of subsidies in FFM states might differ from the response previously estimated for all states because uninsurance rates were higher in FFM states than in non-FFM states before the ACA (Henry J. Kaiser Family Foundation, 2014). As a result, people in FFM states might be more reliant on the exchange than people in other states are. In addition, FFM states disproportionately opted not to expand Medicaid, a decision that makes exchange subsidies available to people with incomes just above the poverty line (i.e., between 100 and 138 percent of the federal poverty line, or FPL). These low-income people might be particularly sensitive to the availability of subsidies.

In developing our analytic approach, we assume that FFM states would not reverse course and decide to set up their own exchanges should the Supreme Court side with the plaintiffs in King. Because the federal government completely funds subsidies provided to enrollees in the exchanges, some states might opt to establish SBMs, which might require an investment in new government infrastructure. By ignoring this possibility, we might overestimate the impact of eliminating subsidies in the FFM states. However, it is impossible to predict the extent to which state policymakers might change decisions about state exchange operations in response to the Supreme Court's ruling. Our objective is to shed light on how consumers might respond to an elimination of subsidies in FFM states, given states' current stances about exchange operations.

Output metrics for this analysis included the change in enrollment and premiums in the ACA-compliant individual market. We did not consider other potential consequences of eliminating subsidies, such as effects on health outcomes, insurer profitability, or the federal budget. To calculate the change in ACA-compliant individual market enrollment, we examined people modeled as residing in FFM states from the output in our previous report for two scenarios: (1) the unmodified ACA scenario, in which subsidies are available in all states, and (2) the modified scenario, in which subsidies are eliminated in all states. A comparison of the enrollment decisions of people residing in FFM states in the two scenarios allowed us to compute the expected change in ACA-compliant individual market enrollment in FFM states should subsidies be eliminated in these states.

Estimating premiums was slightly more involved. In our previous report, we calculated premiums by creating a single risk pool using everyone in our simulation output enrolled in the ACA-compliant individual market nationwide. In reality, an individual rating area in each state has its own risk pool. The sample size of the SIPP does not allow us to reliably compute premiums using risk pools at the state level, let alone at the rating-area level. Hence, we formed a national-level risk pool, which provides us insight into how policy changes might affect the nationwide average premiums. (Premium changes at the state level might vary.) For this article, we extracted those from our previous results who resided in FFM states to form a new risk pool for estimating the average change in premiums associated with eliminating subsidies in all FFM states. We hypothesized that the underlying risk profiles of ACA-compliant individual market enrollees in FFM states differed from those in SBM states, primarily because of income, uninsurance rates, and Medicaid expansion decisions, and hence it was important to form a new risk pool of only FFM-state participants. The formula for calculating premiums for the FFM risk pool was the same as the one for the national risk pool in our previous report, with one small exception. As the ACA is being phased in between 2014 and 2016, the law provides funds to insurers to help offset the cost of covering very expensive enrollees through a temporary reinsurance program.****** Because the proportion of these very expensive enrollees is roughly the same as the percentage of ACA-compliant individual market enrollees residing in FFM states, we reduced the amount of reinsurance funds available to insurers in FFM states by this percentage when calculating premiums.

The approach we applied for this analysis does not obtain equilibrium estimates. As noted in our previous report, the COMPARE model is dynamic, allowing for repeated iterations of premium changes and individual responses to those premium changes until the simulation converges to equilibrium. To obtain an equilibrium estimate in each risk pool, we would need to continue the simulation, allowing people to respond to the new premiums, recalculating premiums based on revised insurance decisions, allowing people to respond once again to the new premiums, and so on. Because the sample of people residing in FFM states in our data is only a fraction of the national sample, we were concerned that rerunning the simulation with a much smaller sample would produce unreliable estimates. By not rerunning the model to equilibrium, we have probably biased downward the magnitude of the impact. For instance, the increase in premiums in FFM states is probably slightly greater than what we estimate because higher premiums could mean that some enrollees leave the risk pool, leading to premiums rising above our estimates. As indicated above, this downward bias could be mitigated if some FFM states elect to set up their own exchanges.

Results

In Table 1, we display the expected impact on the ACA-compliant individual market in FFM states if subsidies are eliminated. In our previous report, we projected that 19.8 million people would be enrolled in the ACA-compliant individual market nationwide; of these, approximately two-thirds, or 13.7 million, reside in FFM states. If subsidies are eliminated in FFM states, we project that enrollment in the ACA-compliant individual market will decline by 9.6 million to 4.1 million enrollees in FFM states, a 70-percent decrease. As noted in our previous report (Eibner and Saltzman, 2014), eliminating subsidies in all states would result in a slightly smaller decline of 68 percent. As stated, we assume that there is no relationship between the markets in different states, implying that the elimination of subsidies in FFM states has no impact on enrollment in SBM states.

Outcome Measure	Subsidies in FFM States	No Subsidies in FFM States	Absolute Change	% Change
Enrollment	13,700,000	4,100,000	–9,600,000	–70%
Premium^	$3,450	$5,060	$1,610	47%

^ Changes in premiums are shown for the specific case of a 40-year-old nonsmoker purchasing a silver plan. The 47-percent change in premiums is the same for all age groups.

The table also provides changes in premiums, specifically for a silver-plan premium for a 40-year-old nonsmoker. Because of the ACA's community rating rules, the percentage increases in premiums are the same for all age groups, but the absolute increases will vary by age, smoking status, metal tier purchased, and rating area. When subsidies are available, average premiums in FFM states are $3,450 annually (though premiums might vary significantly within and between states). If subsidies are eliminated in FFM states, average premiums rise in FFM states by $1,610 to $5,060, a 47-percent increase. In our previous report, we noted that, were subsidies eliminated in all states, we would expect to see an increase of $1,490, or 43 percent, in premiums. This increase is slightly smaller than the projected increase in FFM states only, which is consistent with our observation that enrollment declines are more pronounced in FFM states.

Why are the consequences of eliminating subsidies greater in FFM states than in SBM states? One explanation is income. As a whole, the FFM states have higher proportions of low-income people, who tend to have a higher price elasticity of demand for insurance and who are thus more likely to drop insurance without subsidies. FFM states also tended to have higher uninsurance rates before passage of the ACA than other states did. Because of these two factors, FFM states are likely to receive more subsidy dollars per enrollee than other states and hence will lose more per capita subsidy funding if subsidies are eliminated. Furthermore, most of the FFM states did not pursue Medicaid expansion. In states that did not expand Medicaid, people with incomes between 100 percent and 138 percent of FPL become eligible for subsidies; without subsidies, it is highly unlikely that these very low-income people will remain in the ACA-compliant individual market. Therefore, socioeconomic distribution and decisions by policymakers are likely to have a more pronounced impact in FFM states if subsidies in the ACA-compliant individual market are eliminated.

Discussion

In this research report, we evaluated the potential consequences of a Supreme Court decision in the upcoming King case to eliminate subsidies in the ACA-compliant individual market in FFM states. Our analysis leveraged a previous study, in which we examined the impact of eliminating subsidies in all states and the District of Columbia. Similar to our previous findings, our study results demonstrate that the elimination of subsidies in FFM states would result in sharp premium increases and enrollment declines in the ACA-compliant individual market, causing significant instability and threatening the viability of the market. Furthermore, we find that the effects of eliminating subsidies are greater in FFM states than in SBM states, resulting in even larger percentage-wise premium growth and enrollment declines than if subsidies were eliminated in all states. The greater impact in FFM states is explained by three main factors:

1. Many FFM states did not expand Medicaid, a fact that created a group of very low-income people with incomes between 100 percent and 138 percent of FPL who would have been eligible for Medicaid had eligibility been expanded but who currently depend on subsidies to access health insurance in the marketplaces.
2. FFM states as a whole have higher proportions of low-income people, who receive larger subsidies and whose insurance decisions are more sensitive to price than higher-income people do.
3. FFM states had higher rates of uninsurance before passage of the ACA, further expanding the pool of potential enrollees who would benefit from subsidies.

As indicated above, our modeling approach established an estimate of the impact of eliminating subsidies in FFM states, assuming that policymakers in these states do not subsequently decide to establish their own exchanges. It is beyond the scope of our model to predict whether the political environment in these states might change in response to a Supreme Court decision eliminating subsidies. To the extent that governors and state legislatures opt to establish their own exchanges, the ultimate impact of eliminating subsidies in FFM states could be smaller than our estimates. In addition, our analysis did not consider other potential consequences of eliminating subsidies, such as the level of competition between insurers in the ACA-compliant individual market and the impact on health outcomes or the federal budget. Such fallout from a Supreme Court decision determining that the IRS did not have the power to permit subsidies in FFM states could be an important consideration for policymakers.

Our analysis indicates that eliminating subsidies in FFM states would have significant ramifications for the ACA-compliant individual market, potentially threatening its viability. In addition, eliminating subsidies in FFM states would hamper the ACA's ability to accomplish one of its key objectives: expanding access to health insurance coverage. This research reaffirms that subsidies are an essential component to the functioning of the ACA-compliant individual market.

References

Code of Federal Regulations, Title 45, Public Welfare, Subtitle A, Subchapter B, Part 155, Exchange Establishment Standards and Other Related Standards Under the Affordable Care Act, Subpart A, General Provisions, Section 155.20, Definitions. As of January 6, 2015:
http://www.ecfr.gov/cgi-bin/text-idx?SID=5122599e459bcb5068beb07185692403&node=se45.1.155_120&rgn=div8

Eibner, Christine, and Evan Saltzman, Assessing Alternative Modifications to the Affordable Care Act: Impact on Individual Market Premiums and Insurance Coverage, Santa Monica, Calif.: RAND Corporation, RR-708-DHHS, 2014. As of January 6, 2015:
http://www.rand.org/pubs/research_reports/RR708.html

Halbig v. Burwell, 758 F. 3d 390 (D.C. Cir. 2014).

Henry J. Kaiser Family Foundation, “Uninsured Rates for the Nonelderly by Age,” c. 2014. As of January 7, 2015:
http://kff.org/uninsured/state-indicator/rate-by-age/

King v. Burwell, 759 F. 3d 358 (4th Cir. 2014).

Nat. Fedn. of Indep. Business v. Sebelius, 132 S. Ct. 2566 (2012).

Public Law 111-148, Patient Protection and Affordable Care Act, March 23, 2010. As of January 6, 2015:
http://www.gpo.gov/fdsys/granule/PLAW-111publ148/PLAW-111publ148/content-detail.html

RAND Corporation, “Comprehensive Assessment of Reform Efforts (COMPARE),” undated. As of January 6, 2015:
http://www.rand.org/health/projects/compare.html

U.S. Code, Title 26, Internal Revenue Code, Subtitle A, Income Taxes, Chapter 1, Normal Taxes and Surtaxes, Subchapter A, Determination of Tax Liability, Part IV, Credits Against Tax, Subpart C, Refundable Credits, Section 36B, Refundable Credit for Coverage Under a Qualified Health Plan. As of January 6, 2015:
http://www.gpo.gov/fdsys/granule/USCODE-2010-title26/USCODE-2010-title26-subtitleA-chap1-subchapA-partIV-subpartC-sec36B

Notes

* California, Colorado, Connecticut, Hawaii, Idaho, Kentucky, Maryland, Massachusetts, Minnesota, Nevada, New Mexico, New York, Oregon, Rhode Island, Vermont, and Washington.

** Specifically, under the ACA, a plan can charge a 64-year-old no more than three times what it would charge a 21-year-old in the same rating area with the same plan.

*** The Obama administration has allowed individual market plans that do not comply with the ACA's requirements to continue being offered through October 1, 2017.

**** Of these 9.6 million people, 8.0 million become uninsured. The remaining 1.6 million find coverage through another source.

***** Under the ACA, an enrollee in the ACA-compliant individual market can choose from one of four tiers of cost sharing: bronze, silver, gold, and platinum. Someone enrolled in a bronze, silver, gold, or platinum plan would be expected to pay 40, 30, 20, or 10 percent, respectively, of his or her covered medical costs. The insurer would be expected to pay the remaining 60, 70, 80, or 90 percent of the expected costs for a bronze, silver, gold, or platinum plan, respectively. In addition to the four main tiers of cost sharing, a catastrophic plan is available for people under the age of 30.

****** The ACA includes two temporary programs to keep premiums in the ACA-compliant individual market affordable and stable during the first three years of the market's existence. The reinsurance program makes funds available to insurers that have very expensive enrollees. The risk-corridor program limits both insurer losses and insurer gains, protecting against inaccurate insurer premium-setting in the initial three years. A third provision, risk adjustment, redistributes funding from low-risk to high-risk enrollees. Unlike risk corridors and reinsurance, the risk-adjustment program is permanent.

This research results from the RAND Corporation's Investment in People and Ideas program. Support for this program is provided, in part, by the generosity of RAND's donors and by the fees earned on client-funded research. The research was conducted within RAND Health, a division of the RAND Corporation.