- Issue: The Affordable Care Act’s individual mandate requires most Americans to enroll in health insurance. In 2017, Congress eliminated financial penalties associated with failing to comply with the mandate, which becomes effective in 2019.
- Goal: To review the evidence for how individual mandates affect enrollment decisions, and to assess the effect of eliminating the penalty on enrollment, premiums, and the federal deficit.
- Methods: We reviewed the literature on health insurance mandates and conducted analysis using the RAND COMPARE microsimulation model.
- Findings and Conclusions: Consumers’ responses to mandates may be influenced by nonfinancial factors that are difficult to measure, including a desire to comply with the law, beliefs about enforcement, and inertia in decision-making. Under a range of scenarios that reflect alternative assumptions about responses to these factors, we find that enrollment falls by 2.8 million to 13 million people and premiums for bronze plans increase by 3 percent to 13 percent when the mandate penalty is removed. The impact on the federal budget deficit is more uncertain, with effects ranging from a reduction of $8 billion to an increase of $3.6 billion in 2020. The effect on the deficit depends on how enrollees who are eligible for tax credits and Medicaid — those who have little financial reason to drop coverage — respond to the penalty’s elimination.
The Affordable Care Act (ACA) includes a mandate for every person to obtain health insurance to guard against adverse selection in the markets. This occurs when enrollees are disproportionately older and sicker than the general population and can lead to high insurance premiums overall.1 Before the ACA, individual market insurers in most states could protect themselves against this kind of scenario by denying coverage to applicants at risk for high spending, charging sicker and older people higher premiums, excluding coverage for specific preexisting conditions, and not covering specific benefits such as mental health treatment and prescription drugs. These practices prohibited some individuals from getting coverage at all and left others with unaffordable premiums. The ACA required individual market insurers in every state to offer comprehensive coverage to all applicants at premiums that do not vary with health status and without restrictions on coverage for preexisting conditions. These changes aimed to expand access to health insurance for sick people who might previously have been denied coverage or priced out of the market. The goal of the individual mandate was to encourage young and healthy people to get or stay insured, which would help spread out the cost of sicker people who would enroll and use more services because of the ACA’s rule changes. The ACA further encouraged enrollment by offering tax credits to people who purchased insurance on the individual market and had low to moderate incomes (100% to 400% of the federal poverty level, or roughly between $25,000 and $98,000 for a family of four) and no other affordable source of coverage. The law also allowed states to expand Medicaid to all residents with incomes below 138 percent of poverty.
Although many consumers agree with insurance regulations that prohibit insurers from denying coverage to people who are sick or require high-cost care, the individual mandate was among the least popular provisions of the ACA.2 Soon after the ACA passed, the National Federation of Independent Businesses challenged the constitutionality of the individual mandate. The U.S. Supreme Court ruled that the mandate was constitutional in 2012, but in December 2017, Congress passed the Tax Cuts and Jobs Act, which eliminated the individual mandate penalty, effective January 1, 2019.
The Congressional Budget Office (CBO) estimated that eliminating the individual mandate penalty would reduce health insurance enrollment by 3 million to 6 million between 2019 and 2021, while increasing premiums on the individual market by around 10 percent.3 CBO made a point in its analysis of highlighting the inherent uncertainty of its results. The effect of eliminating the penalty depends on many issues: the cost of health insurance, the size of the mandate penalty, the availability of financial assistance like tax credits, and behavioral factors that are difficult to anticipate. These include consumers’ willingness to comply with laws, confusion surrounding mandate rules, perceptions regarding how strongly the mandate will be enforced, and inertia in decision-making, and could be affected by political beliefs, news reporting, and other factors.
The goal of this report is to analyze the potential effects of eliminating the individual mandate penalty, drawing from literature on early experiences with the mandate to guide assumptions. Because many of the factors that will influence consumer response are uncertain, we estimate effects under a range of assumptions. These results can help inform discussions at both the state and federal levels for policymakers who are considering state-specific mandates or devising policies to address the effect of the penalty’s elimination on enrollment and premiums.
How Is the Individual Mandate Penalty Calculated?
The individual mandate was phased-in over a three-year period from 2014 through 2016, and had two distinct components: a requirement to hold minimum essential health insurance coverage, and a “shared-responsibility” payment (i.e., penalty) for those who failed to comply with the requirement. Under the shared-responsibility payment, individuals who lacked qualifying coverage were required to pay the greater of two amounts: one based on a percentage of income and another based on an inflation-adjusted dollar value (Exhibit 1). Individual mandate penalties are assessed during the annual tax filing process; payments are made the year after the coverage lapse occurred. Per the Tax Cut and Jobs Act of 2017, the penalty will be eliminated beginning in 2019 although the act did not change the legal requirement to hold minimum essential health insurance coverage. It also left other components of the ACA, including regulations in the individual market, in place.
The ACA allowed several exemptions to the individual mandate penalty payments. People with incomes below the tax filing threshold ($10,400 for a single individual or $20,800 for a married couple in 2017) are exempt from the penalty, as are people who would have to pay more than 8 percent of income in 2014 (adjusted in subsequent years to account for rising health care costs) to enroll in the cheapest available plan. Following a 2012 Supreme Court decision that made Medicaid expansion optional for states, the U.S. Department of Health and Human Services clarified that people with incomes below 138 percent of poverty in states that did not expand their programs are also exempt.4 Other exemptions exist for members of federally recognized Indian tribes, people with religious conscience objections, incarcerated individuals, people with hardships like homelessness and bankruptcy, and several other groups.
Exhibit 2 shows how the penalty amounts varied across income levels in 2017, using a single individual and a family of four as examples. The penalty was zero for those with incomes below the tax-filing threshold; it then became a fixed amount (e.g., $695 per individual) for those with incomes immediately above the tax-filing threshold. For those with higher incomes, the penalty increases with income, eventually reaching a maximum level based on the cost of the national average bronze plan. For a single individual, the maximum was $3,264 in 2017 and applied to people with incomes above $140,000. For a family of four, the maximum penalty was $13,056, and applied to household income at or above $543,040.
Health Insurance Mandates: Evidence from the Literature
Massachusetts implemented a major health insurance reform in 2007, seven years before the ACA’s individual mandate took effect. The reform expanded Medicaid to people with incomes below 150 percent of poverty, offered tax subsidies to those with incomes between 150 percent and 300 percent of poverty without access to employer coverage, required employers to provide coverage or pay a penalty, and instituted an individual mandate. The individual mandate required people with incomes above 150 percent of poverty to enroll in insurance or pay a penalty based on half the cost of the cheapest plan available in the individual market.5 Massachusetts had significant regulations in its individual market both before and after the reform, including requirements that insurers must offer coverage to all applicants and that older adults can be charged no more than twice as much as younger adults. To assess whether the mandate prompted young and healthy people to enroll, Chandra, Gruber, and McKnight6 analyzed the health and spending profiles of individuals in Massachusetts who enrolled in the individual market before the mandate was effective, while the mandate was phasing in (during which penalties were lower than subsequent years), and after it was fully adopted. Average monthly claims among individual market enrollees decreased as the mandate was phased in, ultimately falling by 31 percent. Younger and healthier people tended to enroll later than older and sicker people, suggesting an inverse relationship between the size of the mandate penalty and the level of risk in the health insurance pool.
Hackmann, Kolstad, and Kowalski7 analyzed data from Massachusetts residents with incomes above 300 percent of poverty, and found that full implementation of the mandate was associated with a 23 percent decline in premiums and a 26.5 percentage-point increase in individual market enrollment among this group. The relatively large decline in premiums may reflect Massachusetts’ unique health insurance regulations, which may have led to disproportionate enrollment of individuals with high expected spending before the implementation of the mandate.
Affordable Care Act
Several recent papers look at the impact of the ACA’s individual mandate on enrollment and spending outcomes. Frean, Gruber, and Sommers8 used nationally representative data from the American Community Survey from 2012 to 2015 to analyze the relationship between ACA policies and coverage changes. They found that roughly 24 percent of the increase in coverage in 2014 and 2015 was because of marketplace tax credits and 36 percent was because of Medicaid enrollment (among newly and previously eligible people). Forty percent was unexplained by the policy variables included in the analysis. The analysis accounted for the size of the individual mandate penalty, suggesting that enrollees did not respond differentially to higher penalties. However, the analysis did not rule out the possibility that a “taste for compliance” — that is, a desire to comply with the law regardless of penalties or other enforcement mechanisms9 — led to a general increase in enrollment. Using consumer data from California and Washington, Saltzman10 found no evidence of a linear relationship between penalty amounts and demand for health insurance. However, he estimated a positive taste for compliance equal to roughly $67 per month, which was most pronounced among lower-income (i.e., <400% of poverty) populations.
Wettstein11 compared changes in the U.S. uninsurance rate before and after 2014 to changes in the Massachusetts uninsurance rate. He argued arguing that because Massachusetts implemented a health insurance mandate several years earlier, it was unaffected by ACA’s mandate. He limited his sample to people with incomes above 400 percent of poverty to avoid confounding because of changes in tax credit eligibility. The study estimated that the combination of insurance regulations and the individual mandate reduced the uninsurance rate by 19 percent in 2014. Moreover, the U.S. uninsurance rate continued to decline relative to Massachusetts’s in 2015. Because the individual mandate penalty increased from 2014 to 2015 while other policies remained constant, the author concluded that the size of the individual mandate penalty had a causal role in reducing the uninsurance rate. Wettstein further estimated that reductions in uninsurance were larger for younger relative to older people, suggesting that younger people were particularly responsive to the mandate.
While several additional countries, including Australia, Germany, Japan, the Netherlands, and Switzerland, have mandates to carry insurance, we found few studies of these countries that suggest clear lessons that can be applied to the United States. Switzerland and the Netherlands both adopted their mandates against a backdrop of near-universal coverage.12 In a review of three countries with mandates — Germany, Switzerland, and the Netherlands — van Ginneken and Rice13 report that uninsurance is rare (i.e., typically less than 2 percent of the population). Those who fail to comply with the mandate tend to be poor and often recent immigrants. Switzerland and the Netherlands take relatively aggressive steps to enforce the mandate, including autoenrolling individuals who are out of compliance and — in the Netherlands — garnishing wages. While Germany appears to have a less aggressive enforcement approach, nearly 90 percent of the population is automatically enrolled in public coverage.
In a study of the Australian health system, Stavrunova and Yerokhin14 found that a surcharge applied to higher-income people who did not enroll in private health insurance coverage had moderate effects, increasing enrollment rates by about 15 percent. A disproportionate amount of those who did not respond to the surcharge were younger than 30. However, it is difficult to generalize this experience to the United States because everyone in Australia was eligible for comprehensive public coverage. Private health insurance provided duplicative services, but with perks such as access to hospitals with more amenities, shorter waiting times, and more choice of physicians. Coupled with the framing of the penalty as a “surcharge,” the policy may have been viewed as a means-tested premium for public coverage, rather than a requirement to enroll in a private plan.
Behavioral Responses to the Mandate
Auerbach et al.15 posited that responses to health insurance mandates might be influenced not only by financial considerations such the magnitude of the penalty, but also by behavioral factors like awareness of the mandate, social norms, and consumers’ taste for compliance. More recent literature explores several of these issues. For example, using a sample of long-term uninsured people in South Carolina, Shi et al.16 analyzed whether the mandate prompted healthier people to enroll in the individual market and whether consumers’ awareness of the law affected responses. They found that individual market applicants who were aware of the mandate tended to have fewer long-term health problems than individual market applicants who were unaware. This could indicate that awareness of the mandate prompted healthy people to enroll, while less-healthy people enrolled regardless of whether they were aware of the mandate. Sixteen percent of those attempting to sign up for insurance were unaware of the mandate.
Ericson and Kessler17 used an experimental survey to assess whether individuals responded differently to a hypothetical requirement to obtain insurance described as a “mandate” versus a “tax.” They found that responses varied depending on how the requirement was described and changed over time because of current events. In early waves of the study — before publicity surrounding the 2012 Supreme Court case challenging the legality of the mandate — respondents reported a higher likelihood of purchasing insurance when the requirement was described as a mandate. However, following the political controversy around the Supreme Court case, responses were similar regardless of how the requirement was described. These results suggest that responses to the mandate requirement may be affected by framing by policymakers and the media. Further, highly publicized opposition to the mandate may have made some people ambivalent about complying.
The Kaiser Family Foundation also found that framing affected survey respondents’ perception of the policy. Support increased when people were told that mandate repeal could increase individual market premiums and reduce health insurance enrollment.18 Respondents’ support for the individual mandate also increased when they were informed that most people get insurance through an employer and that exemptions exist for certain groups, including those who may have difficulty affording coverage. Other evidence shows that people who self-identify as Republicans tend to have a less favorable view of the mandate (along with other ACA provisions), and may be less likely to respond to the mandate, than those who identify as Democrats.19
Responses also may depend on the costs that people face to enroll in coverage. People currently enrolled in Medicaid have no premiums, and hence limited reason to disrenoll in response to the removal of the mandate. However, by not enrolling in the first place, they avoid hassle costs associated with eligibility determination. Enrollees in employer-sponsored coverage and those who are eligible for tax credits on the ACA’s marketplaces also have limited out-of-pocket costs associated with obtaining insurance. The roughly 7.5 million people who pay full price for individual market coverage20 may be more responsive to mandate repeal than other groups.
Considerations for Modeling the Mandate
The literature, along with CBO and other analyses, suggests that people’s response to the removal of the individual mandate penalty depends on many factors: the value individuals place on being insured, out-of-pocket cost of insurance, the size of the mandate penalty, and nonfinancial considerations, such as a taste for compliance with the law. The following are key issues to account for when modeling the mandate.
Size of the penalty. Economic theory predicts that larger individual mandate penalties would lead to increased enrollment relative to smaller penalties. The evidence for this relationship is limited. While some studies find that compliance with the mandate increases with the size of the penalty,21 others have found no evidence that response to the mandate varies with size.22
Taste for compliance. Some studies suggest that individuals prefer to comply with the law and might opt to enroll simply because of the requirement, regardless of the size of the penalty.23 Auerbach et al.24 argue that the taste for compliance may vary depending on individual factors, and could increase with age. The Tax Cuts and Jobs Act of 2017 reduced the individual mandate penalty to zero, while keeping the requirement on the books, raising the question of whether the requirement itself has bearing on enrollment, even if the penalty is zero. In its November 2017 report, the CBO assumed that “with no penalty at all, only a small number of people who enroll in insurance because of the mandate under current law would continue to do so solely because of a willingness to comply with the law.”25
Knowledge of the penalty. Many studies have shown that people have limited health literacy,26 limited financial literacy,27 and are susceptible to cognitive biases that may impede rational decision-making.28 Data from the Commonwealth Fund Affordable Care Act Tracking Survey indicate that roughly 84 percent of the population is aware of the mandate; awareness is higher among people with incomes above 250 percent of poverty (90%) than among those with incomes below 250 percent (77%).29 Despite relatively high awareness, the complexity of the mandate formula, coupled with generally low health and financial literacy, may have affected consumers’ responses. If people underestimated the size of the penalty, this confusion may have reduced the mandate’s overall effect, while if people overestimated, it may increase the mandate’s overall effect. Income over the course of the year is also uncertain, which could make it hard for some individuals to estimate their payment. Difficulty in calculating the size of the penalty may explain studies such as Saltzman and Frean, Gruber, and Sommers,30 which both found lack of response to the mandate’s size. Some people may not have understood how the mandate was calculated, and behaved as if it were a lump-sum amount rather than based on a formula that varied with income, family size, and year.
Media coverage and political beliefs. People’s willingness to comply with the mandate may be influenced by media coverage and political beliefs and may change over time.31 People also view the individual mandate more positively when they are informed that eliminating the mandate penalty may reduce insurance enrollment and increase premiums.32
Exemptions. Several groups are exempt from the ACA’s mandate, including those who lack affordable coverage and who would have been eligible for Medicaid under the law but live in states that did not expand their programs. In general, the effect of the individual mandate will be weaker when more people are exempt, because fewer people face the penalty. However, confusion over exemption status could influence this effect. If consumers are unaware of exemptions, they may respond to the mandate even if it doesn’t apply to them. Alternatively, if people believe exemptions are commonplace and easy to obtain, they might anticipate being able to receive one even if this is not accurate. Perceptions about the availability of exemptions may depend on individual circumstances, such as whether friends and neighbors are exempt. Widespread exemptions also may interact with the taste for compliance. If many people are exempt, those subject to the penalty may feel less compelled to enroll to satisfy social norms.
Probability of paying the penalty. Those who expect the mandate will apply to them may have differing beliefs or expectations about whether they will pay it. On average, the IRS collects only about 82 percent of tax revenue owed.33 Federal policy also has led to relatively weak enforcement of the individual mandate penalty. For example, the IRS cannot take steps such as filing a notice of lien or criminally prosecuting those who evade the mandate.34 Further, in 2016, the IRS allowed people to file “silent returns” that did not include proof of health insurance coverage.35 Given these factors, some people may expect to avoid the penalty by failing to report health insurance status or by failing to pay all that they owe. Some people may expect to avoid the penalty and ultimately end up paying. For modeling purposes, we assume people expect to pay 80 percent of the penalty on average, but consider an alternative scenario where people expect to pay only 50 percent of the penalty.
Inertia in decision-making. Behavioral economics research shows that people tend to stick with decisions they have made in the past without reevaluating whether those choices continue to be optimal.36 Additional research shows that individuals place a higher value on a commodity once they have it compared to when they did not have it.37 This suggests that people who are newly enrolled in insurance because of the mandate may be reluctant to drop coverage, either because they don’t revisit the decision or because they value coverage more than they did before.
Welcome-mat effect. After the ACA’s coverage expansions took effect, Medicaid enrollment increased among individuals who had been eligible prior to the ACA.38 While the individual mandate may have motivated some additional enrollment among previously eligible individuals, an additional explanation is the so-called welcome-mat effect. Specifically, the ACA’s coverage expansions may have increased consumers’ awareness of the Medicaid program, outreach initiatives may have increased enrollment, sustained public focus on getting covered may have prompted people to apply, and other factors — such as the single streamlined application used to simplify enrollment and assistance from navigators — may have increased program uptake. It is difficult to disentangle the welcome-mat effect from other factors, such as the penalty (which applied only to the subset of Medicaid-eligible individuals with incomes above the tax filing threshold) and confusion over whether the penalty applied. It is also uncertain whether the welcome-mat effect is an enduring phenomenon.
Tax credits. People who receive tax credits to enroll on the ACA’s marketplaces may be less sensitive to eliminating the mandate penalty than individual market enrollees who do not receive credits because enrollees with tax credits pay only a portion of their premiums. Further, tax credits under the ACA reflect the cost of the second-lowest-price silver plan available to the enrollee, minus a contribution that scales with income. The design of the tax credit makes enrollees relatively insensitive to premium increases, because — when premiums rise — tax credits also increase. Awareness and understanding of the law likely influences the role of tax credits. Collins, Gunja, and Doty39 found that 40 percent of uninsured individuals were unaware of the ACA’s marketplaces and that roughly 35 percent of the uninsured have incomes in the range that makes them eligible for tax credits. These findings imply that some individuals may remain uninsured because they are unaware they are eligible for financial assistance.
Cost-sharing reductions. Along with tax credits, some enrollees are eligible for cost-sharing reductions (CSRs), which reduce out-of-pocket payments at the point of service (e.g., copayments, deductibles). By law, insurers must provide CSRs to tax-credit-eligible enrollees with incomes below 250 percent of poverty. However, Congress did not appropriate funding for CSRs and in late 2017, the Trump administration halted federal payment to insurers to cover these costs. In response, insurers in most states increased the premiums for silver plans40 resulting in higher tax credit amounts. The higher tax credits made coverage cheaper for many consumers, particularly for those who chose coverage outside of the silver tier (e.g., some consumers became eligible for free bronze plans). Response to eliminating the individual mandate penalty may change when CSRs are loaded onto silver plans. With higher tax credits, out-of-pocket premiums for tax-credit-eligible consumers will be lower.
COMPARE is a microsimulation model developed at RAND that is used to estimate responses to health reform policies, including the ACA. Modeled individuals in COMPARE decide whether to enroll in insurance and what type of insurance to choose by weighing the costs and benefits of available options, including the cost of the individual mandate penalty. However, the literature described above suggests that there are many noneconomic factors that could influence individuals’ response to the mandate and much uncertainty about their effects. To gauge sensitivity to these factors, we analyzed 10 scenarios that encompass alternative assumptions about how people respond to the mandate (Appendix 1). Most of the scenarios assess individual changes to our base modeling assumptions — such as replacing the linear penalty response with a taste for compliance. Combined scenarios A and B account for multiple nonfinancial factors simultaneously. By assuming that there is no inertia in decision-making and that the welcome-mat effect dissipates after mandate repeal, combined scenario A is designed so that individuals are relatively responsive to the mandate. In contrast, combined scenario B, which allows for inertia in decision-making and assumes the welcome-mat effect persists, is designed so that individuals are relatively unresponsive to the mandate. We estimate effects by comparing results from similar scenarios with and without the individual mandate penalty. A full description of the COMPARE model and the methods used to analyze each scenario can be found in Appendix 2.
Exhibit 3 shows the changes in enrollment that we estimate under each scenario. Declines in coverage range from 2.8 million in the scenario in which we assume there is inertia in decision-making, to 13 million in combined scenario A, which assumes the welcome-mat effect is tied to the individual mandate. In our base scenario we estimate that insurance coverage will decline by roughly 6.5 million.
We estimate that premiums will increase by 3 percent to 13 percent for bronze plans, and by –1 percent to 6.5 percent for silver plans, depending on the scenario (Exhibit 4). Premium changes for bronze and silver plans are equivalent in scenarios in which CSRs are paid by the federal government, because of the ACA’s risk-adjustment program, which transfers funding from plans with lower-than-average actuarial risk to plans with higher-than-average actuarial risk. However, when CSRs are loaded onto the silver plan, premium increases for silver plans are smaller than those for bronze plans, and sometimes silver premiums decrease when the mandate penalty is eliminated. For non-silver plans and in scenarios where CSRs are paid by the federal government, premium changes are driven by adverse selection only, which causes premiums to increase. When CSR costs are loaded onto the silver plan, adverse selection is partly offset by lower CSR spending, which occurs if the share of CSR-eligible individuals enrolled in silver plans falls. In two scenarios, the reduction in CSR-eligible enrollees more than offsets the adverse selection effect, leading to a net decline in silver (but not bronze) premiums.
Exhibit 5 shows the estimated effects on the federal deficit. In six of the 10 scenarios, eliminating the mandate penalty increases the deficit. However, this result is sensitive to modeling assumptions. We find deficit reductions in those scenarios that assume the welcome-mat effect is tied to the individual mandate, and — to a lesser extent — in scenarios that replace the response to the individual mandate penalty with a taste for compliance.
Across all scenarios, we estimate that the deficit impact ranges from a reduction of $8 billion to an increase of $3.6 billion in 2020.
In this analysis, we set out to understand the likely effects of the impending elimination of the individual mandate penalty on health insurance enrollment, premiums, and the federal deficit, drawing from prior literature, behavioral economics, and microsimulation modeling. Given that the mandate is a relatively new policy, there is limited literature. However, the empirical studies that we identified found relatively consistent evidence that insurance mandates increase health insurance enrollment and that those who enroll because of the mandate tend to be younger and healthier than those who would enroll without the mandate.
Nevertheless, there is lack of consensus on the specific drivers of consumers’ response to the mandate. Economic theory suggests that individuals should be more responsive to a large penalty, compared to a smaller one. While there is some empirical evidence for this phenomenon,41 other studies have found a taste for compliance that does not vary with the penalty’s size.42 There is also evidence that behavioral factors such as awareness of the law, framing of the mandate as a penalty versus a tax, and political ideology may affect people’s response.43 Inertia in decision-making and limited health literacy also may affect people’s response to the mandate.44
Exhibit 6 summarizes the range of results we found in our analysis. When we used a microsimulation model to estimate response to removing the penalty under a variety of scenarios regarding consumer behavior, we found reductions in coverage ranging from 2.8 million to 13.0 million in 2020. The effects on enrollment were largest when we assumed the welcome-mat effect dissipated because of elimination of the mandate penalty and smallest when we assumed inertia in decision-making. We further estimated that premiums for bronze plans would increase by 3 percent to 13 percent, with the largest premium increases occurring in scenarios with the most substantial coverage reductions. While silver premiums also generally increased when the penalty was eliminated, these changes were smaller than changes for bronze plans because of reductions in CSR costs that can occur when the share of eligible individuals in the silver tier is reduced. In two scenarios, silver premiums fell slightly because of mandate repeal.
Our results suggest that removing the mandate may have uncertain effects on the federal deficit. CBO estimates a net deficit decrease of $14 billion in 2020.45 To the extent that people who receive federal financial assistance — either through marketplace tax credits or Medicaid — drop coverage in response to the penalty’s elimination, federal spending may fall, leading to decreases in the deficit. However, marketplace tax credits vary with premiums levels, and the federal government bears most of the extra cost associated with premium increases. When young and healthy people drop out of the individual market, premiums go up, increasing federal spending on marketplace tax credits. The deficit impact varies depending on whether the number of subsidized people who drop coverage is sufficient to offset the increase in marketplace tax credits. This result is very sensitive to assumptions. Those who are highly subsidized — such as Medicaid enrollees and people eligible for large marketplace tax credits — have little economic reason to disenroll from health insurance when penalties are eliminated because they pay little out-of-pocket for insurance. Hence, their response is likely driven predominately by noneconomic factors, such as awareness, inertia, and the welcome-mat effect. The empirical literature provides little guidance regarding the size of these effects, making it difficult to determine how they will affect enrollment. Our deficit results are most like CBO’s November 2017 estimates when we assume the welcome-mat effect dissipates when the individual mandate is repealed.
These findings present some important considerations for state policymakers contemplating state-specific mandates and for federal policymakers seeking to reduce individual market premiums despite the elimination of the individual mandate penalty. Notably, the effects of any state-based replacement for the individual mandate will depend on how the replacement is designed and publicized. States implementing their own mandates may be able to increase the impact of the policy by ensuring that affected individuals are aware of the requirement and that enforcement mechanisms are credible and effective. Further, opposition to a state-specific mandate could be tempered if states clearly communicate the rationale for the policy — that is, to reduce growth in premiums.46 Policymakers at both the state and the federal level may be able to reduce disenrollment by ensuring that people who are eligible for Medicaid and marketplace subsidies are aware that these programs remain in place. Keeping subsidized marketplace enrollees in the risk pool also may help to stabilize premiums.
We gratefully acknowledge Preethi Rao, Dylan Roby, and Chapin White, who provided thoughtful reviews of this analysis. We also thank Sara Collins and the Commonwealth Fund for their support.
1. David M. Cutler and Richard J. Zeckhauser, “Adverse Selection in Health Insurance,” in Frontiers in Health Policy Research, Vol. 1, ed. Alan M. Garber (MIT Press, 1998), http://www.nber.org/chapters/c9822.
2. Ashley Kirzinger, Elise Sugarman, and Mollyann Brodie, Kaiser Health Tracking Poll: November 2016 (Henry J. Kaiser Family Foundation, Nov. 2016), https://www.kff.org/health-costs/poll-finding/kaiser-health-tracking-poll-november-2016/.
3. Congressional Budget Office, Federal Subsidies for Health Insurance Coverage for People Under Age 65: 2018 to 2028 (CBO, May 2018), https://www.cbo.gov/system/files/115th-congress-2017-2018/reports/53826-healthinsurancecoverage.pdf.
4. Centers for Medicare and Medicaid Services, HHS Final Rule and Treasury Notices on Individual Shared Responsibility Provision Exemptions, Minimum Essential Coverage, and Related Topics, Fact sheet (CMS, June 26, 2013), https://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2013-Fact-sheets-items/2013-06-26.html.
5. Massachusetts Department of Revenue, TIR 12-2: Individual Mandate Penalties for Tax Year 2012 (Mass. DOR, 2018), https://www.mass.gov/technical-information-release/tir-12-2-individual-mandate-penalties-for-tax-year-2012.
6. Amitabh Chandra, Jonathan Gruber, and Robin McKnight, “The Importance of the Individual Mandate — Evidence from Massachusetts,” New England Journal of Medicine 364, no. 4 (Jan. 27, 2011): 293–95, https://www.nejm.org/org/doi/full/10.1056/NEJMp1013067.
7. Martin B. Hackmann, Jonathan T. Kolstad, and Amanda E. Kowalski, “Adverse Selection and an Individual Mandate: When Theory Meets Practice,” American Economic Review 105, no. 3 (2015): 1030–66, https://pubs.aeaweb.org/doi/pdfplus/10.1257/aer.20130758.
8. Molly Frean, Jonathan Gruber, and Benjamin D. Sommers, “Premium Subsidies, the Mandate, and Medicaid Expansion: Coverage Effects of the Affordable Care Act,” Journal of Health Economics 53 (May 2017): 72–86, https://doi.org/10.1016/j.jhealeco.2017.02.004.
9. David Auerbach et al., “Will Health Insurance Mandates Increase Coverage? Synthesizing Perspectives from Health, Tax, and Behavioral Economics,” National Tax Journal 63, no. 4 (2010): 659–79, https://EconPapers.repec.org/RePEc:ntj:journl:v:63:y:2010:i:4:p:659-79.
11. Gal Wettstein, “The Affordable Care Act’s Insurance Market Regulations’ Effect on Coverage,” Health Economics 27, no. 3 (Mar. 2018): 454–64, https://onlinelibrary.wiley.com/doi/abs/10.1002/hec.3585.
12. Timothy Stoltzfus Jost, The Experience of Switzerland and the Netherlands with Individual Health Insurance Mandates: A Model for the United States? (2009), http://law2.wlu.edu/deptimages/Faculty/Jost%20The%20Experience%20of%20Switzerland%20and%20the%20Netherlands.pdf.
13. Ewout van Ginneken and Thomas Rice, “Enforcing Enrollment in Health Insurance Exchanges: Evidence From the Netherlands, Switzerland, and Germany,” Medical Care Research and Review 72, no. 4 (Aug. 2015): 496–509, http://journals.sagepub.com/doi/abs/10.1177/1077558715579867.
14. Olena Stavrunova and Oleg Yerokhin, “Tax Incentives and the Demand for Private Health Insurance,” Journal of Health Economics 34 (Mar. 2014): 121–30, https://doi.org/10.1016/j.jhealeco.2014.01.001.
15. Auerbach et al., “Will Health Insurance?,” 2010.
16. Lu Shi et al., “Does Awareness of the Affordable Care Act Reduce Adverse Selection? A Study of the Long-Term Uninsured in South Carolina,” Inquiry 54, (Jan.–Dec. 2017), https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5798667/.
17. Keith Marzilli Ericson and Judd B. Kessler, The Articulation Effect of Government Policy: Health Insurance Mandates Versus Taxes (National Bureau of Economic Research, Mar. 2013), http://www.nber.org/papers/w18913.
18. Ashley Kirzinger et al., Kaiser Health Tracking Poll — November 2017: The Role of Health Care in the Republican Tax Plan (Henry J. Kaiser Family Foundation, Nov. 2017), https://www.kff.org/health-reform/poll-finding/kaiser-health-tracking-poll-november-2017-the-role-of-health-care-in-the-republican-tax-plan/.
19. Kirzinger, Sugarman, and Brodie, Kaiser Health Tracking, 2016; and Anirban Basu et al., Political Polarization, Anticipated Health Insurance Uptake and Individual Mandate: A View from the Washington State (National Bureau of Economic Research, Nov. 2014), http://www.nber.org/papers/w20655.
20. Julie Rovner, “Overlooked by ACA: Many People Paying Full Price for Insurance ‘Getting Slammed,’” Kaiser Health News, Oct. 9, 2017, https://khn.org/news/overlooked-by-aca-many-people-paying-full-price-for-insurance-getting-slammed/.
21. Wettstein, “Affordable Care Act’s Insurance,” 2018; and Amitabh Chandra, Jonathan Gruber, and Robin McKnight, “The Impact of Patient Cost-Sharing on Low-Income Populations: Evidence from Massachusetts,” Journal of Health Economics 33 (Jan. 2014): 57–66, https://doi.org/10.1016/j.jhealeco.2013.10.008.
22. Frean, Gruber, and Sommers, “Premium Subsidies,” 2017; and Saltzman, Demand for Health, 2017.
23. Auerbach et al., “Will Health Insurance?,” 2010; and Saltzman, Demand for Health, 2017.
24. Auerbach et al., “Will Health Insurance?,” 2010.
26. Sylviana H. Barcellos et al., “Preparedness of Americans for the Affordable Care Act,” Proceedings of the National Academy of Sciences of the United States of America 111, no. 15 (Apr. 15, 2014): 5497–502, https://doi.org/10.1073/pnas.1320488111; and George Loewenstein et al., “Consumers’ Misunderstanding of Health Insurance,” Journal of Health Economics 32, no. 5 (Sept. 2013): 850–62, https://doi.org/10.1016/j.jhealeco.2013.04.004.
27. Ellen Peters, Louise Meilleur, and Mary Kate Tompkins, Numeracy and the Affordable Care Act: Opportunities and Challenges (National Academies Press, July 17, 2014), http://nationalacademies.org/hmd/~/media/files/activity%20files/publichealth/healthliteracy/commissioned%20papers%20-updated%202017/peters%20et%20al%202013%20numeracy%20and%20the%20aca.pdf.
28. Katherine Baicker, William J. Congdon, and Sendhil Mullainathan, “Health Insurance Coverage and Take-Up: Lessons from Behavioral Economics,” Milbank Quarterly 90, no. 1 (Mar. 2012): 107–34, https://doi.org/10.1111/j.1468-0009.2011.00656.x.
29. Sara R. Collins et al., Americans’ Views on Health Insurance at the End of a Turbulent Year (Commonwealth Fund, Mar. 2018), https://www.commonwealthfund.org/publications/issue-briefs/2018/mar/americans-views-health-insurance-end-turbulent-year.
30. Saltzman, Demand for Health, 2017; and Frean, Gruber, and Sommers, “Premium Subsidies,” 2017.
31. Ericson and Kessler, Articulation Effect, 2013; and Basu et al., Political Polarization, 2014.
32. Kirzinger et al., Kaiser Health Tracking, 2017.
35. Thomson Reuters, “IRS Won’t Reject Returns That Are Silent Regarding Compliance with ACA Individual Mandate,” Thomson Reuters Tax and Accounting News, Feb. 16, 2017, https://tax.thomsonreuters.com/media-resources/news-media-resources/checkpoint-news/daily-newsstand/irs-wont-reject-returns-that-are-silent-regarding-compliance-with-aca-individual-mandate/.
36. Miriam Krieger and Stefan Felder, “Can Decision Biases Improve Insurance Outcomes? An Experiment on Status Quo Bias in Health Insurance Choice,” International Journal of Environmental Research and Public Health 10, no. 6 (June 2013): 2560–77, https://doi.org/10.3390/ijerph10062560; and William Samuelson and Richard Zeckhauser, “Status Quo Bias in Decision Making,” Journal of Risk and Uncertainty 1 (1988): 7–59, https://sites.hks.harvard.edu/fs/rzeckhau/status%20quo%20bias.pdf.
37. Daniel Kahneman, Jack L. Knetsch, and Richard H. Thaler, “Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias,” Journal of Economic Perspectives 5, no. 1 (Winter 1991): 193–206, https://www.aeaweb.org/articles?id=10.1257/jep.5.1.193.
38. Frean, Gruber, and Sommers, “Premium Subsidies,” 2017.
39. Sara R. Collins, Munira Z. Gunja, and Michelle M. Doty, Following the ACA Repeal-and-Replace Effort, Where Does the U.S. Stand on Insurance Coverage? Findings from the Commonwealth Fund Affordable Care Act Tracking Survey, March–June 2017 (Commonwealth Fund, Sept. 2017), https://www.commonwealthfund.org/publications/issue-briefs/2017/sep/following-aca-repeal-and-replace-effort-where-does-us-stand.
40. Auerbach et al., “Will Health Insurance?,” 2010.
41. Chandra, Gruber, and McKnight, “Importance of the Individual,” 2011; and Wettstein, “Affordable Care Act’s Insurance,” 2018.
42. Saltzman, Demand for Health, 2017.
43. Shi et al., “Does Awareness,” 2017; Ericson and Kessler, Articulation Effect, 2013; and Basu et al., Political Polarization, 2014.
44. Loewenstein et al., “Consumers’ Misunderstanding,” 2013; and Samuelson and Zeckhauser, “Status Quo Bias,” 1988.
45. CBO, Repealing the Individual, 2017.
46. Kirzinger et al., Kaiser Health Tracking, 2017.