Abstract
- Issue: Republicans and Democrats agree on prioritizing choice in health insurance, but disagree on what it entails and how to achieve it. Choice and competition can create negative consequences, including adverse selection and consumer confusion.
- Goals: Examine the experiences of the Affordable Care Act’s marketplaces and recommend ways policymakers can harness choice and competition to improve coverage, satisfaction, and affordability.
- Methods: Review of existing evidence.
- Key Findings: There are multiple areas where insurance design could promote efficient competition and consumer choice. Experiences with the ACA have shown that health insurance marketplaces should include an urban area with adjacent rural and suburban communities to promote competition among insurers. Other recommendations include allowing smaller insurance carriers to base medical loss ratio rates on past years’ data; allowing insurers to bid against each other for contracts to serve a population; providing resources to allow consumers to make informed choices; and including features like essential health benefits to counteract adverse selection.
- Conclusion: Markets can deliver efficient premiums, access to care, and consumer satisfaction but only when they are carefully designed and actively managed through regulation.
Background
Even though the partisan divide over health policy will likely persist through the 2020 election, Republicans and Democrats agree on one priority: promoting consumer choice. That said, they disagree on what constitutes choice. On the left, some think it means choice of doctors or hospitals in a Medicare for All model, while some on the right view it as a choice of varying health plans. In the center, policymakers hope that competition and consumer choice at many levels can meet the health insurance and health care needs of American households and, at the same time, result in affordable insurance premiums.
Choice among competing health plans in a marketplace can have potential downsides. For example, if sicker people enroll in insurance while healthier ones opt out — a phenomenon known as adverse selection — insurers’ costs go up and they may choose to offer plans with only limited coverage. In addition, choice among plans may fail consumers if they select plans that do not fit their needs because they do not understand their choices. Regulation is needed along with competition — an insight offered long ago by Enthoven and informed by recent experience.1
This report recommends ways policymakers on both sides of the aisle could harness competition and choice in the individual health insurance market to improve coverage, satisfaction, and affordability. It first reviews lessons learned from the Affordable Care Act’s (ACA) individual marketplace and then applies those lessons to what will likely be the presidential candidates’ health plans.2 For Republicans, this plan may be the Graham–Cassidy–Heller–Johnson amendment: a state block grant, with few rules, that replaces the ACA’s coverage expansion. This idea was included in President Trump’s fiscal year 2019 budget. For Democrats, the campaign platform is likely to include some type of public plan similar to Medicare. Although Medicare for All is among the proposals under discussion, this report focuses on the models that offer a public plan as a choice alongside private plans. While disparate in their goals, the Republican and Democratic approaches support consumer choice and insurer competition and could both benefit from adopting the measures described in this report.
Creating Choice and Efficiency Through Competition
The individual health marketplaces would continue under the Republicans’ block grant proposal as well as the Democrats’ public plan choice proposals. In this report, I focus on four areas where meaningful lessons have emerged: geographic definition of markets; regulations that promote or inhibit market entry; the significance of adverse selection; and the impact of product complexity on the effectiveness of consumer choices.
Geographic and Population Size of Markets
In any market, for competition to function well, numerous firms must operate at an efficient scale. In addition, economies of scale suggest that “bigger is better” or suggest a tendency toward “natural monopolies” since larger insurers are typically better positioned to efficiently spread risk. This tends to undermine choice and limit the role of competition. Market design features that spread risk — like reinsurance and risk corridors — serve to reduce the importance of the “bigger is better” feature of health insurance.
The ACA gave states wide discretion in defining the marketplaces where consumers would shop and make their plan selections. Some states combined rural and urban areas into larger aggregates; others chose smaller, more focused markets. The result was considerable heterogeneity across states with regard to the definition of rating areas. This, in turn, creates wide variation in the number of firms competing in a rating area and ultimately the premiums for plans sold on the marketplaces. For example, in Florida rating areas were defined as counties, regardless of the size or density of the population. Texas took a very different approach to defining markets. The major cities in Texas anchored rating areas; those cities were linked to surrounding counties. As a result, 254 counties were divided into 26 rating regions.
Research has found that these choices had a considerable impact. Specifically, states that chose to combine counties into regional markets saw significantly more insurers participating in the marketplaces and significantly lower premiums than in states where markets were simply defined as individual counties.3 Research showed that if a state altered its market definition by enlarging its markets so that the number of people in a market increased substantially, the result would be an increase of between 27 percent and 37 percent in the number of insurers participating in the market, and premiums that are 3.3 percent to 5.4 percent lower.4
For some states there will be regions that will never be able to support more than one or a very few firms operating at an efficient scale. Those regions are characterized by small populations, large geographic areas, and a limited number of providers. Together these features limit the potential for market participation by insurers and interfere with robust premium and quality competition for consumers.
Policies That Affect Market Entry
Some regulations and policies — like medical loss ratios (MLRs), risk corridors, and reinsurance — have had notable impacts on the financial performance of marketplace insurers and will likely affect long-term insurer participation, consumer choice, and competition.
The MLR, as defined in the ACA, is the amount that insurers spend paying medical claims and on quality improvement activities relative to the premium, net of taxes and regulatory fees. Regulations require that insurers selling plans in the individual market pay out 80 percent of premium in benefits, which includes spending on quality improvement. MLR data are reported for each insurer for each market segment in which they sell plans (e.g., individual, large group) for a calendar year. Existing research suggests that the MLR regulations serve as a check on the exercise of market power and appear not to result in distorted administrative costs.5 Because smaller insurers tend to have more variable claims experiences, the MLR disadvantages them because they will be more likely to fail to meet the standard than larger issuers for reasons beyond their control.6 Early experiences with the marketplaces support that proposition. Insurers who did not enter the marketplaces in 2014 and 2015 tended to be smaller and have greater variability in their MLRs in prior years than those that did enter the marketplaces.7 Data also show that for the average rating area in 2015 and 2016 there were about 18 health insurers that sold health insurance in the geographic area but not in the marketplaces. Smaller insurers were overrepresented in that group. Research shows that smaller insurers were less likely to enter the marketplaces.8 The implication is that reducing impediments to smaller insurers from entering the marketplaces — such as reducing their risk of not complying with MLR rules — would increase the number of potential entrants.
The temporary reinsurance program in the ACA marketplaces, like other forms of reinsurance, was designed to allow insurers with insufficient resources to cover extreme losses to conduct normal insurance functions. It pays for coverage of very high cost cases (i.e., losses in the “right tail” of the distribution of costs per beneficiary). Reinsurance has a salutary effect on premiums because of this further spreading of risk beyond the individual insurer.9 Analysis by Jacobs and colleagues showed that reinsurance payments narrowed the claims deficits by nearly half for insurers in the top 10 percent of the claims cost distribution.10 That analysis also showed that the payments from reinsurance were especially important for the financial status of smaller insurers. The ending of the reinsurance program served to put upward pressure on premiums and disadvantaged smaller insurers because their size makes them less able to efficiently bear the risk of very-high-cost cases. As a consequence, the business case for participating in the marketplaces was weakened, resulting in less choice and less competition.
The risk corridors program also aims to reduce insurers’ risk. Its intent was to protect issuers from mispricing premiums in the early years of the marketplaces when experience was limited. While it is permanent for Medicare’s prescription drug plans, its continued value in this context has been questioned because numerous large insurers that participate in the Medicare Part D program have learned how to set premiums since the program’s beginning in 2006. Risk corridors and reinsurance serve some similar functions. Layton and colleagues demonstrate that the risk protection achieved by a combination of risk corridor and reinsurance that was initially used in the marketplaces can also be accomplished with a simpler reinsurance policy.11 A key lesson from the ACA is that it is critically important to follow through on regulatory promises, such as payments for the risk corridors. Failure to do so will undercut the risk-reducing features of those programs. In the case of the marketplaces, sufficient funds were not appropriated and the program could not be fully funded. This resulted in plans having to absorb those early pricing errors that in some cases were catastrophic.
Adverse Selection
The individual market is replete with incentives not to enroll the sickest and most costly segments of the population. This is well established in insurance markets with consumers that are heterogeneous with respect to health status, premiums that do not vary with health status, and health insurers that compete for enrollees. Adverse selection has traditionally been addressed through underwriting, preexisting conditions clauses, design of covered benefits, access to specialized provider services, and promotion and location of providers, among other strategies. For example, prior to the enactment of the ACA, 62 percent of enrollees in individual health insurance plans had no maternity coverage, 34 percent had no coverage for substance use disorder care, and 18 percent had no mental health care coverage.12
The ACA’s insurance reforms aimed to limit this type of adverse selection. It prohibited medical underwriting and annual and lifetime limits on coverage, and instituted minimum essential coverage standards. Its essential health benefit requirements and network adequacy standards also serve to reduce the incentives for insurers to engage in practices aimed at avoiding less healthy enrollees. The ACA’s essential health benefits require that 10 services categories, including maternity, substance use disorder, and mental health care, all be covered by all individual and small-group health plans. As a result, there has been an expansion in both coverage and treatment for those conditions.
Risk adjustment has been key to addressing incentives for distorted competition stemming from adverse selection.13 The marketplaces have a risk-adjustment system that is based on the current health profiles of enrollees rather than being set prospectively. It is based on a less detailed classification of the illnesses than in the system used in Medicare Advantage. This reflects the practical concerns of setting up a new program that serves previously uninsured people and a desire to balance the goal of eliminating incentives to avoid sicker enrollees with the possibility that insurers will have a new incentive to upcode. Insurers have an incentive to upcode because they can realize higher payments from coding cases as more severe.
Existing evidence shows that the risk-adjustment system used for the marketplaces generally worked as envisioned. The system resulted in a redistribution of payments that was consistent with reduced incentives for insurers to engage in actions that promote selection of the healthiest people.14 The risk-adjustment system shifted payments from insurers with low claims costs to insurers with high claims costs. Insurers with a moderate level of claims saw little changes in their net payments and receipts stemming from the risk adjustment. Research shows that reinsurance also serves as a complement to risk adjustment in weakening incentives to avoid enrollment of high-cost people.15
Complex Health Insurance Choices
The design of the ACA’s marketplaces relies on creating competition by including multiple insurers all vying for business.16 Price-linked subsidies on the marketplaces are based on the price of the second-lowest-cost silver plan.17 Insurers offering a premium that is the lowest or second-lowest-cost silver plan provide subsidized consumers relatively low out-of-pocket premiums. Thus, there is competition to be the lowest or second-lowest-cost silver plan. The fewer the number of insurers competing to be the least expensive plan, the less incentive insurers have to lower their premiums. The likelihood that an insurer will have the lowest or second-lowest-cost silver plan declines with the number of insurers competing and thus motivates more aggressive premium-setting as the number of plans rise.18 Evidence relating the number of insurers selling silver plans to premiums is strong. Multiple researchers have shown that premiums are driven lower as the number of insurers in the marketplaces increases.19 Analyses by Burke and Sheingold showed that for each additional insurer competing in a marketplace, premiums fell by between 2.8 percent and 4 percent. Estimates by Frank based on the marketplaces from 2014 through 2016 showed that premiums increased by an estimated 7.4 percent when the number of insurers fell below three, all other factors being equal.20
One downside of having choices in health insurance is that consumers are prone to making predictable errors when faced with numerous choices regarding complex products.21 These errors can undermine the benefits gained from competition. Insurance products are complex; benefit design involves copayments, deductibles, coinsurance rates, provider networks, prescription drug formularies, and a variety of specialized programs for specific illnesses. A large body of research has shown that few people buying health insurance have a complete grasp of even the most basic parameters of benefit design such as copays, coinsurance, deductibles, and out-of-pocket maximums.22 Marketplace design anticipates some of those difficulties. Health insurance products are presented in standardized groupings according to actuarial values — these are known as the metal tiers (bronze, silver, gold, platinum). The marketplaces feature calculators that estimate the expected costs of various health plans for individuals with specific characteristics and allow consumers to determine if their primary care physician is included in a health plan’s network. People also can get help from call centers and health plan navigators. These human interventions were associated with higher levels of coverage and enrollment in the marketplaces.23
Even with these forms of decision support in place there is evidence of errors in decision-making among marketplace consumers. In 2017, nearly 20 percent of consumers that returned to the marketplaces could have found a lower-cost option within the same metal tier. Marzilli Ericson and Starc24 showed that more standardization in the Massachusetts marketplace made a significant difference in the choices people made. Finally, studies have shown that as the number of health insurance choices increases, there is greater consumer “inertia” and less response to prices that results in reduced competition.25 Thus, choice can be a mixed blessing and needs to be carefully titrated to obtain the right balance of competition and clear-eyed consumerism.
Policy Recommendations
What are the implications of these lessons for the individual health insurance market? Reform proposals have ranged from efforts to improve the existing marketplaces by adding a public option (e.g., “Medicare for More” proposals) to replacing the marketplaces with far less regulated state-based individual insurance markets (e.g., the Graham–Cassidy–Heller–Johnson amendment). In considering policy design, I appeal to three guiding principles: whenever possible, set conditions that will yield robust competition for consumers; protect against adverse selection; and offer consumers salient information and support in making choices. Using these principles and the key findings from this report, I offer five suggestions on the design of individual health insurance markets:
Setting Market Size to Promote Entry
Health insurance markets need sufficient population to support multiple issuers (preferably more than three) at scale if competition for consumers is to work. The evidence from the experiences of the marketplaces is strong on this account. The government entity overseeing the design of markets must structure them to maximize the likelihood that they will be able to sustain multiple insurers. In addition, using a uniform approach based on political divisions like counties is likely to frequently fall short of achieving the goal of promoting competition. The experiences of the ACA marketplaces highlight the fact that state decisions about market definitions have often been incorrect and have failed to promote competition. Therefore, the federal government should not be agnostic on the minimum size of local market designs. Marketplaces should encompass an urban area with adjacent rural and suburban communities to promote multiple entrants. However, planners must be cognizant of the fact that markets with larger land areas serve to dampen entry — this happens in many rural areas.26 These considerations can be included in Democrats’ proposals for a public plan option. They also could serve as criteria in the Graham–Cassidy–Heller–Johnson proposal for a state to receive a block grant. That is, federal requirements to qualify for a block grant would include defining markets, to promote competition among insurers at efficient scale.
Regulatory Flexibility to Promote Market Participation
Some existing regulations — including the MLR rules — aimed at consumer protection discourage smaller health plans from entering individual markets. This barrier to entry could be addressed by recognizing that smaller carriers experience greater claims variability and allowing them to calculate their MLRs based on the claims experience aggregated over several years. While this policy change alone would be unlikely to dramatically affect market entry, it would serve as a useful complement to reintroducing reinsurance into the individual health insurance market. Evidence suggests that the temporary reinsurance program in the ACA disproportionately benefited smaller plans and put downward pressure on premiums overall. The Graham–Cassidy–Heller–Johnson proposal contains some reinsurance provisions but these could be strengthened. Proposals aimed at adding a public option to the marketplaces must include a robust reinsurance program to properly balance risk-bearing between the public and private plans in addition to encouraging marketplace participation by smaller health insurers. Since the public option is backed by government, it has built-in protection — effectively a form of reinsurance. This means than small insurers also would need reinsurance in order to compete with the public plan.
Competition for Contracts as a Fallback Plan for Low-Competition Areas
Evidence indicates that for a number of markets across the country, economic circumstances will never generate sufficient numbers of issuers to create robust competition. In those circumstances, I propose that competition be reoriented. Instead of individual insurers competing to enroll consumers into their health plans, the competition would be among insurers to obtain a contract to serve a population. This would allow for competition among multiple insurers even when the markets are relatively small; insurers would be offering bids to obtain a franchise or partial franchise. There are many examples of successful uses of competition for contracts in health insurance. Employer-sponsored insurance commonly uses such approaches, such as requesting proposals and bids to serve their employees and dependents. State government health insurance plans have used competition for contracts to select pharmacy benefit managers to serve state employees, even those enrolled in different health plans. Similarly, state and local governments (in Massachusetts, Colorado, Arizona, Iowa, and others) have used competition for contracts to select managed behavioral health care organizations to serve their state Medicaid programs. These purchasing arrangements have resulted in multiple bids and strong price competition for markets that might otherwise not have generated robust competition for consumers. This approach might well serve as an alternative to a public plan or as a feature of the Graham–Cassidy–Heller–Johnson proposal.
A variation on this theme would be to make use of a public plan option only in places that do not generate sufficient market participation to support meaningful competition for consumers. Establishing a public option triggered by a specified level of market concentration is one way to address markets that fail to support competition. For example, if a market had a Hirschman-Herfindahl index (HHI) value of 3000 (i.e., a high level of market concentration) or more for three consecutive years, it would trigger the establishment of a public plan for a period of five years, at which point the competitive conditions would be reassessed. (The HHI, a standard measure of market concentration, is the sum of the squares of the market share of the firms. An HHI of 2500 means that the market has four equal-sized firms, each with a 25% share.)
Consumer Decision Supports
Research shows that for consumers to be effective shoppers, they need to be able to easily sort through health insurance options. Benefit standardization has been shown to help on this front. It also helps to mitigate harmful competition stemming from adverse selection incentives. Providing consumers with clear, accessible information on product characteristics they value is important. Some evidence shows that provision of information by letter and email increased shopping but had little effect on plan-switching behavior.27 Other evidence highlights the impact that human assistance has in getting people to enroll in marketplace plans. Reducing the number of health plan features under consideration and targeting those most important to consumers can serve to reduce choice overload and make consumers more responsive to key plan differences, thereby focusing competition on parameters that matter most to consumers. The evidence also suggests that making help easily accessible from human assisters via telephone or in person would promote more effective shopping. For the public option proposals, this would mean making the information on all plans equally available to consumers. For the Graham–Cassidy–Heller–Johnson proposal, states should be required to provide such information and consumer supports to qualify for a block grant.
Countering Selection Incentives
Finally, all proposals must address selection incentives. For proposals that build on the ACA, such measures could be enhanced. Essential health benefits and risk adjustments are the two most powerful mechanisms for addressing the inefficiencies that result from adverse selection. Evidence from a range of insurance markets implies that if competition is to focus on price and not on selection of healthy enrollees, essential health benefits provisions, standardized benefits, and risk adjustment must be part of market design. For the Graham–Cassidy–Heller–Johnson proposal, this means that those provisions must be key criteria for receipt of a block grant. For the public option proposals, the public plan must provide the same benefits (i.e., essential health benefits, metal tiers) and be subject to risk adjustment in the same fashion as private plans, otherwise competition would lead to adverse selection. Modern risk-adjustment systems have accomplished a great deal in limiting selection. This has been studied extensively within the Medicare Advantage programs. There is, however, a tension between the benefits of reducing adverse selection and the potential costs of upcoding. More detailed illness classifications in risk-adjustment programs make it easier for insurers to upcode. The evidence suggests that it is critically important to use designs for individual health insurance markets that allow for both essential health benefits and risk adjustment to be incorporated. Without such mechanisms an efficiently functioning market is unlikely.
Discussion
Promoting competition that results in efficient health insurance markets cuts against the grain of some strong beliefs on both sides of the American political divide. Republican members of Congress shy away from regulations that serve to standardize products and affect the premiums paid to insurers. Democrats, in contrast, are frequently mistrustful of the profit motive in health care and suspect that there is little ability to rein in the tendency of insurers to compete for good risks and shun people with preexisting conditions. Lessons from Medicare Advantage and the marketplaces suggest that health insurance markets can deliver efficient premiums, access to care, and consumer satisfaction, but only when markets are carefully designed and actively managed through regulation.
Successful markets require more regulation than many Republicans would prefer; on the other hand, Democrats should recognize that markets can work efficiently, albeit with the government playing an active role. Designing health insurance markets requires a regulatory platform that equips consumers and sellers with information, supports risk protection, and offers incentives for efficient choice and supply. Regulations must be flexible and subject to modification as conditions in individual markets change. Realizing the promise of markets for health insurance will require greater unity of purpose from our political leadership and administrative agencies.
Acknowledgments
The author is grateful to Sara Collins, Jeanne Lambrew, and Tom McGuire for comments on an earlier draft of this report.