Evan Saltzman, Christine Eibner
E. Saltzman and C. Eibner, Donald Trump's Health Care Reform Proposals: Anticipated Effects on Insurance Coverage, Out-of-Pocket Costs, and the Federal Deficit, The Commonwealth Fund, September 2016.
Issue: Republican presidential candidate Donald Trump has proposed to repeal the Affordable Care Act (ACA) and replace it with a proposal titled “Healthcare Reform to Make America Great Again.” Proposed reforms include allowing individuals to deduct the full amount of premiums for individual health plans from their federal tax returns, providing block grants to finance state Medicaid programs, and allowing insurers to sell insurance across state lines. Goal: To assess how each of these reforms, when implemented individually, would affect insurance coverage, consumer out-of-pocket spending on health care, and the federal deficit in 2018. Methods: RAND’s COMPARE microsimulation model. Key findings and conclusions: The policies would increase the number of uninsured individuals by 16 million to 25 million relative to the ACA. Coverage losses disproportionately affect low-income individuals and those in poor health. Enrollees with individual market insurance would face higher out-of-pocket spending than under current law. Because the proposed reforms do not replace the ACA’s financing mechanisms, they would increase the federal deficit by $0.5 billion to $41 billion.
Since the Affordable Care Act (ACA) was enacted in 2010, critics have advocated that the law be repealed and replaced with an alternative set of reforms. Republican presidential candidate Donald J. Trump has offered a “repeal-and-replace” proposal titled “Healthcare Reform to Make America Great Again.”1 In this brief, we consider the impact of repealing the ACA and enacting three of the key policies proposed by Trump. The policies considered are only elements of Trump’s overall health care reform proposal, which includes several features we did not model, including increasing price transparency and removing barriers to entry in the prescription drug market.2 We analyzed each policy in conjunction with repeal of the ACA, rather than as a combined package. By considering each policy on its own, we can more easily understand each option’s effect on coverage, consumer out-of-pocket costs, and the federal deficit. The policies we consider include:
Because we analyzed only some of Trump’s proposed policies, we cannot conclude that a scenario that combined the effects of these reforms would be an accurate representation of the full impact of Trump’s health plan. As a result, we do not report a scenario combining these reforms in the main text of this brief, although it is available in the technical appendix.
We used the RAND COMPARE microsimulation model, which estimates the impact of health policy changes. Specifically, we analyzed how the proposed reforms would affect the distribution of health insurance coverage by income and health status, the federal deficit, and the level of out-of-pocket spending in the individual market. To quantify the impact on out-of-pocket spending, we focused on the individual market because many of the policies enacted by the ACA and proposed by Trump are targeted to this market. In particular, the Trump proposals would eliminate key ACA individual market reforms, including:
Trump's plan would remove these requirements and subsidies and introduce new policies that affect the individual market, including tax deductions and the ability to sell plans across state lines.
Modeling health reform proposals that have not yet been turned into legislation can be challenging because of lack of specificity. Further, Trump would implement several other proposals that could interact with the health policies, such as changes in tax rates. Consequently, we make several modeling assumptions, which we discuss briefly in the How This Study Was Conducted section at the end of this brief. A detailed description of the model and assumptions is provided in the technical appendix. In the technical appendix, we also compare our results to two previous studies that have estimated the impact of Trump’s proposals.7 (Also see a To the Point blog post and a digital feature on both candidates' proposals and their implications.)
Repealing the ACA would result in 19.7 million fewer people with health insurance in 2018 (Exhibit 1). This estimate assumes that individuals who newly enrolled in Medicaid under the ACA, but who were eligible under prior law,8 would remain enrolled even if the law were repealed.9 Repealing the ACA and adding a tax deduction for health insurance would result in 15.6 million fewer people with health insurance. The Medicaid block-grant program results in 25.1 million fewer people with health insurance, including approximately 5.5 million people who were eligible for Medicaid under pre-ACA rules who lose coverage because states may lack the funds to sustain enrollment among this population. Allowing insurers to sell across states lines reduces coverage by 17.5 million people.
Exhibit 2 illustrates each policy’s effect on the number of people without insurance, by income level. All three policies would increase the ranks of the uninsured among those with incomes under 250 percent of the federal poverty level (i.e., $60,750 for a family of four). For those with incomes above 250 percent of poverty, the policies have mixed effects. Repealing the ACA would have little impact on insurance enrollment for people with higher incomes; the same is true of repealing the ACA in combination with the Medicaid block-grant program. However, repealing the ACA in combination with the tax deduction or allowing insurers to sell across state lines would increase the number of higher-income people with insurance. We estimate that 2.7 million more people with incomes over 250 percent of poverty would be insured with the tax deduction, and 1.4 million more higher-income people would be insured if insurers were allowed to sell across state lines.
We estimate that repealing the ACA would cause the number of uninsured individuals in fair or poor health to increase from 2.1 million to 5.8 million (Exhibit 3). Implementing Medicaid block grants or allowing insurance sales across states lines would further increase the number of uninsured in fair or poor health. Looking at the Medicaid block-grants option, the increase in the number of uninsured people in fair or poor health reflects the general decline in insurance, from 231.9 million to 226.5 million (Exhibit 1). However, in the sales-across-state-lines scenario, the number of uninsured individuals in fair or poor health increases relative to full repeal, despite the fact that more people are insured overall. This is because—as modeled—the sales-across-state-lines scenario leads to regulatory liberalization, making it easier for insurers to deny coverage to older and sicker people.
In Exhibit 4, we analyze how Trump’s plan would affect consumer spending for individuals who would have enrolled in individual market coverage (i.e., private coverage not obtained through an employer) under the ACA. We estimate that total out-of-pocket spending for individual market enrollees, including enrollee premium contributions and cost-sharing at the point of service, averages about $3,200 per year in the ACA scenario. Fully repealing the ACA would cause total out-of-pocket expenses to increase to $4,700. Repealing the ACA and replacing it with a tax deduction would result in average out-of-pocket spending of about $3,500 per year because the tax deduction is less generous on average than the ACA’s Advanced Premium Tax Credits (APTCs) and cost-sharing subsidies, particularly for the lower- and middle-income people who benefit from these policies. If insurers were allowed to sell insurance coverage across state lines, we estimate that average out-of-pocket spending would be approximately $5,700 annually. The large increase in out-of-pocket spending in the last scenario reflects several factors. First, it provides no premium or cost-sharing support for enrollees, leading to an increase in the proportion of health care costs they pay for. Second, it would lead to a proliferation of “bare-bones” plans, which have low premiums but high out-of-pocket cost-sharing at the point of service. Our analysis estimates effects for a standardized population of individuals who enrolled in individual market coverage under the ACA. Average out-of-pocket spending would be lower for the population that actually enroll in the sales-across-state-lines scenario because these enrollees are disproportionately young and healthy.10
Exhibit 4 shows that Trump’s proposed tax deduction would increase out-of-pocket spending for individuals enrolled in the individual market. In part, this is because the tax deduction tends to be less generous than the ACA’s APTCs for individuals with low and moderate incomes. Exhibit 5 shows how tax-related subsidies (e.g., deductions and APTCs) to individual market enrollees would differ between the ACA and Trump’s plan, by enrollees’ income and family size. While the results vary depending on enrollees’ age and family composition, individuals with incomes below 300 percent of poverty (approximately $35,640 for a single individual) tend to benefit more from the ACA’s means-tested tax credit structure, while individuals with incomes above that level benefit more from the Trump plan’s tax deduction.
Finally, we estimated the effect of the proposed reforms on the federal deficit (Exhibit 6). According to our analysis, repealing the ACA would increase the deficit by a net $33.1 billion in 2018. Although repealing the law would reduce federal outlays on Medicaid and tax credits, repeal would also eliminate the ACA’s revenue-generating provisions, such as changes to Medicare payment and taxes on health plans, medical devices, and other goods and services. For example, in the full repeal scenario, federal outlays are reduced by $35.9 billion relative to the ACA, while revenue is reduced by $69 billion, for a net increase to the federal deficit of $33.1 billion.
Repealing the ACA and replacing it with a tax deduction would increase the deficit by $41.0 billion relative to the ACA, mostly because of the federal cost of the tax deduction. Assuming the Medicaid block-grant amount is set to pre-ACA levels adjusted for inflation, we project that the block-grant program would increase the deficit by $0.5 billion. The block-grant scenario is less expensive than full repeal because, under full repeal, we assume that the federal government would continue to fund Medicaid costs for previously eligible individuals enrolled after the ACA was enacted. As modeled, the block-grant amounts are based on pre-ACA Medicaid spending, and do not account for this “woodwork” population (i.e., previously eligible individuals who enrolled in Medicaid following the ACA).
Assuming that Medicaid block grants are based on pre-ACA funding levels, they would be costless to the federal government relative to full repeal. However, states would face a conundrum regarding how to finance the population that is currently enrolled in Medicaid. We assume that states would eliminate eligibility for the ACA’s Medicaid expansion population if offered block grants based on pre-ACA funding levels. But more than half of those who newly enrolled in Medicaid in 2014 were eligible under previous rules.13 Unless the block grants covered this woodwork population, states would have to find alternative means to finance this group or would need to reduce enrollment to break even. In our analysis, we assumed that states would find ways to reduce enrollment, such as by cutting eligibility levels or reducing enrollment assistance and outreach. However, if states pursued a policy of financing the woodwork population, then both Medicaid enrollment and state budgetary costs would increase relative to our estimates.
Allowing insurers to sell across state lines would increase the deficit by $33.7 billion relative to the ACA. Like the other proposals considered, the sales-across-state-lines scenario reduces federal spending relative to the ACA, because of the elimination of ATPCs and cost-sharing reductions and reduced spending on Medicaid and CHIP. However, these savings are more than offset by the reductions in revenue caused by repealing the ACA, including the loss of revenue from individual and employer mandates and the elimination of ACA-related taxes and fees.
In this analysis, we considered three health policies proposed by Republican presidential candidate Donald Trump. Relative to the ACA, we found that all three policies would reduce health insurance enrollment and increase the federal deficit. While all of Trump’s policies reduce spending on health insurance programs and subsidies (e.g., Medicaid, tax credits) relative to the ACA, they also reduce federal revenue by repealing the ACA’s financing mechanisms, including changes to Medicare payment and taxes on medical devices, health plans, and branded prescription drugs. On net, the proposed reforms increase the deficit by $0.5 billion to $41 billion. We estimate that the tax deduction scenario would lead to the largest deficit increase, as a result of losses in tax revenue collected.
Because there were few details available about these policies, we made several key modeling assumptions regarding implementation. For example, we assumed that Medicaid block-grant funding would be based on pre-ACA spending. While different assumptions could lead to different results, it is generally true that features of Trump’s reform proposals are likely to lead to reduced insurance coverage for those with lower incomes and those with preexisting health conditions. First, the program does not replace the ACA’s subsidies to low- and middle-income individuals who were not eligible for Medicaid prior to the ACA and who lack affordable insurance offers through an employer. While Trump’s health insurance tax deduction acts as an implicit subsidy for health insurance, its effects disproportionately benefit those with higher incomes and higher marginal tax rates.
Second, none of Trump’s proposals guarantee that insurance will be available for individuals in poor or fair health who may have been denied coverage or charged higher premiums in the individual market under pre-ACA law. As a result, we estimate that the scenarios would increase the ranks of the uninsured in fair or poor health by 3.6 million to 5.0 million, with the highest numbers occurring in the Medicaid-block-grants scenario. The sales-across-state-lines scenario would lead to lower premiums on the individual market and result in about 2 million additional people being insured relative to the full-repeal scenario. However, because the policy does not require that insurers offer coverage to individuals with preexisting conditions, an additional 200,000 in fair or poor health would be uninsured relative to full repeal alone.
In addition to the three policies considered here, Trump also proposes to expand the use of health savings accounts, increase price transparency in health care, and remove barriers to entry in the pharmaceutical industry. Because we did not model these additional policies, we cannot comment on the full effect of Trump’s health reform proposals. However, in our technical appendix, we consider the combined effect of the three policies modeled in this brief. We estimate that the number of individuals with insurance under all three policies combined is similar to the number if the ACA were repealed without any replacement. Although enacting Medicaid block grants leads to a reduction in coverage relative to full repeal, adding the tax deduction and allowing sales across state lines brings enrollment back to the full-repeal level. However, relative to full repeal, the combined scenario has a lower impact on the federal deficit, and leads to reduced insurance coverage among the lowest-income groups. Both of these effects reflect the impact of the Medicaid block grants, which reduce insurance coverage for lower-income populations while also reducing federal spending.
We have not modeled how Trump’s plan would be financed but, if implemented, these policies would likely require new taxes or offsetting savings from other proposals to maintain deficit neutrality. Further, modifications or additions to Trump’s plan would be required if policymakers wish to avoid coverage losses, particularly for lower-income and less-healthy individuals. For example, refundable tax credits indexed by income instead of a regressive tax deduction could target subsidies to low-income individuals. High-risk pools also could provide a mechanism for those with preexisting conditions to obtain coverage. These reforms likely would expand coverage compared to the reforms considered in this brief; however, they also would increase the federal deficit relative to full repeal.
The RAND COMPARE model creates a synthetic population of individuals, families, and firms using national survey data. After calibrating the modeled behavior of people and firms to match actual outcomes, COMPARE introduces proposed reforms to assess how the choices of individuals and firms are affected. To evaluate the reforms proposed by Trump, we first eliminated all reforms introduced by the ACA, including the individual and employer mandates, premium tax credits and cost-sharing subsidies, and market rating reforms, such as community rating and guaranteed issue. We then added each of Trump’s reforms one by one to the full-repeal scenario using the following approach:
We allowed individuals in the model to deduct the full cost of employer coverage and individual market coverage from their tax returns in determining their federal adjusted gross income. We assumed that households could apply the deduction against their federal income tax obligation, but not against their state income or payroll tax obligations.
Medicaid block grants
To determine the amount of the federal grant, we estimated the level of Medicaid enrollment that would have existed in 2018 if the ACA had not been implemented, and calculated the federal contribution for this coverage. Because funding amounts would likely be based on pre-2014 rules, we assumed that states would first roll back their Medicaid eligibility limits to pre-ACA levels. However, the ACA increased enrollment among both those newly eligible and those previously eligible for Medicaid (i.e., the “woodwork effect”). We assumed that states would further reduce eligibility limits or otherwise discourage enrollment so that spending would not exceed the amount of the block grant.
Sales across state lines
We assumed that allowing insurers to sell across state lines would lead to regulatory liberalization relative to the rules in effect prior to the ACA. To model this effect, we adjusted the individual market to reflect insurance dynamics in the states with the least restrictive regulations prior to the ACA. In particular, we assumed that the premium difference between older and sicker individuals relative to younger and healthier individuals would widen. Further, we assumed that insurance denial rates (e.g., for preexisting conditions) would approach 30 percent—similar to what occurred in states with the highest denials prior to the ACA. Finally, we assumed that the elimination of benefit mandates and other consumer protections, such as risk adjustment, would exacerbate adverse selection and result in less generous, catastrophic-type plans becoming prevalent. As we discuss in detail in the technical appendix, there are several reasons why this degree of regulatory liberalization may not be realized. In addition, final legislation could include regulatory floors that seek to avert widespread insurer location in states with the least restrictive regulatory practices prior to the ACA. For example, Trump has publicly voiced support for protections to ensure that people with preexisting conditions have access to health insurance, although these protections are not discussed in his proposal. If the sales-across-state-lines policy were accompanied by restrictions on insurers’ ability to deny policies to those with preexisting conditions, and if other protections were put in place such as minimum benefit generosity levels, we would expect different results. Minimum benefit generosity requirements would likely cause insurers to offer more generous plans with higher premiums than we have estimated and lower out-of-pocket costs. If the individual market population has high price sensitivity, then it is likely that higher premiums would reduce enrollment relative to the predictions we make in our analysis. This reduction in enrollment could occur even if restrictions on denial rates allowed a larger share of the population to purchase individual insurance than we assume.
We report all results for the calendar year 2018.
1 See: https://www.donaldjtrump.com/positions/healthcare-reform.
2 Trump proposes three other reforms that we cannot model and, therefore, do not include in our analysis. First, Trump proposes allowing individuals to use health savings accounts (HSAs). HSAs exist under current law and it is unclear if and how Trump would modify HSAs. Second, Trump’s plan requires “price transparency from all healthcare providers.” As discussed in detail by Cutler and Dafny (2011), the impact of price transparency is theoretically ambiguous. Empirical evidence for price transparency is limited, but evidence from recent initiatives in California and New Hampshire suggests that price transparency had little impact on prices (Cutler and Dafny, 2011). Finally, the plan advocates removing “barriers to entry into free markets for drug providers that offer safe, reliable, and cheaper products.” Two key barriers to entry in the pharmaceutical industry include patent protection and Food and Drug Administration (FDA) regulation. Trump’s plan provides no detail on how these barriers would be modified, and very little evidence exists to evaluate the impact of such modifications. (Notably, an analysis of Trump’s proposed reforms by the Center for Health and Economy also omits consideration of these three reforms.)
3 Examples of new taxes include the section 9010 tax, the Patient-Centered Outcomes Research Institute tax, taxes on medical devices and branded prescription drugs, the tanning tax, and an increase in the hospital insurance tax for high-income individuals.
4 In addition to the tax exclusion for ESI, under current law, self-employed individuals can deduct individual-market health insurance premiums, taxpayers can deduct medical expenses exceeding 10 percent of adjusted gross income if itemizing, and qualified individuals can claim the Health Coverage Tax Credit (HCTC) under the Trade Preferences Extension Act of 2015.
5 Many states expanded Medicaid beyond categorical eligibility limits prior to the ACA. We assume the block grant would be sufficient to fund these expansions at pre-ACA enrollment levels, but would not cover early state Medicaid expansions (prior to 2014) that occurred under the ACA.
6 Age rating is limited to a 3-to-1 ratio (i.e., 64-year-olds cannot be charged more than three times as much as what 21-year-olds are charged), smoking rating is limited to a 1.5-to-1 ratio, and geographic rating areas must be based on counties, MSAs, or three-digit zip codes with limited exceptions.
7 Center for Health and Economy, Healthcare Reform to Make America Great Again (H&E, July 7, 2016); and Committee for a Responsible Federal Budget, “Analysis of Donald Trump’s Health Care Plan” (CRFB, May 9, 2016).
8 “Previously eligible” in this context refers to people who were eligible for Medicaid prior to the ACA’s coverage expansion.
9 The woodwork effect leads to increased Medicaid enrollment under repeal compared to the pre-ACA environment. Without the woodwork effect, we estimate that 25 million fewer people would have health insurance if the ACA were repealed.
10 Actual enrollees are younger and healthier because older and sicker people are more likely to be denied coverage in the sales-across-state-lines scenario.
11 Congressional Budge Office, Budgetary and Economic Effects of Repealing the Affordable Care Act (CBO, June 2015).
12 G. Claxton, M. Rae, N. Panchal et al., “Health Benefits in 2015: Stable Trends in the Employer Market,” Health Affairs, Oct. 2015 34(10):1779–88.
13 M. Frean, J. Gruber, and B. D. Sommers, Premium Subsidies, the Mandate, and Medicaid Expansion: Coverage Effects of the Affordable Care Act, NBER Working Paper #22213 (National Bureau of Economic Research, April 2016).