Katherine Swartz, Mark Hall, Timothy S. Jost
K. Swartz, M. A. Hall, and T. S. Jost, How Insurers Competed in the Affordable Care Act's First Year, The Commonwealth Fund, June 2015.
Prior to the Affordable Care Act (ACA), most states’ individual health insurance markets were dominated by one or two insurance carriers that had little incentive to compete by providing efficient services. Instead, they competed mainly by screening and selecting people based on their risk of incurring high medical costs. One of the ACA’s goals is to encourage carriers to participate in the health insurance marketplaces and to shift the focus from competing based on risk selection to processes that increase consumer value, like improving efficiency of services and quality of care. Focusing on six states—Arkansas, California, Connecticut, Maryland, Montana, and Texas—this brief looks at how carriers are competing in the new marketplaces, namely through cost-sharing and composition of provider networks.
Prior to the passage of the Affordable Care Act (ACA), most states’ markets for individual health insurance were dominated by one or two carriers that competed primarily on how well they were able to screen and select people based on their risk of incurring medical claims. They had little incentive to compete by providing efficient services; instead, their focus was on reducing their risk of covering people who might have very high medical costs. Consequently, many people perceived likely to have high medical costs were uninsured, because carriers either rejected them or offered policies at excessively high premiums.
A principal goal of the ACA is to shift carriers’ focus from risk selection to processes that increase consumer value, such as improving efficiency of services and quality of care. Carriers selling plans in the ACA marketplaces must accept all applicants and cover services for preexisting conditions. Further, the plans must cover the same minimum essential benefits and differ only in terms of actuarial value, which are identified by “metal” tier or level.1 A risk-adjustment program requires carriers that enroll a healthier-than-average population to then compensate other carriers that have less-healthy subscribers. In addition, premiums for individuals in the oldest age group cannot be more than three times the premiums for the youngest adult age group, and rating by gender and by how long a person has held a policy (durational rating) are prohibited.
Other researchers have documented the law’s basic success in attracting more carriers to the individual marketplaces and making them more price competitive.2,3,4 But research so far has not delved deeper into how this success might vary among different marketplaces, whether insurers still continue to compete on risk selection, and what forms of price competition are emerging.
In late 2013 and early 2014, we studied six states—Arkansas, California, Connecticut, Maryland, Montana, and Texas—and conducted interviews with a variety of public and private sector policymakers, academic analysts, and consumer advocates. These states were selected to showcase different demographics, economic conditions, and insurance markets. California, Connecticut, and Maryland established their own marketplace exchanges; Arkansas and Montana are cooperating or partnering with the federal government; and Texas defaulted entirely to a federally facilitated marketplace.5
This issue brief summarizes what we observed in the six states. We describe the marketplaces’ initial forms of competition and relate them to state-specific factors that might explain similarities or differences. Knowing more about carriers’ initial competitive strategies provides a baseline for assessing changes in competitive strategies as the marketplaces mature.
Before the ACA became law, most states had just one or two carriers that dominated the individual insurance market, even when a large number were licensed to sell policies in these markets.6 As a result, states and the federal government were concerned about attracting carriers to the marketplaces and acted accordingly.7 For example, although the ACA allows states to selectively contract with carriers, only California has so far opted to do so. In addition, the ACA created the Consumer Operated and Oriented Plan (CO-OP) program, which provided funding for new nonprofit insurance cooperatives. The co-ops are supposed to be managed with strong input from the people who enroll in the plans and will reinvest any profits as part of their efforts to offer affordable plans.
Efforts to increase the number of individual carriers are significant because the ACA’s premium subsidies are benchmarked to the second-lowest premium of silver plans in each market. If only two carriers compete, one will have the second-lowest premium so there is little incentive for either to drive costs down to lower premiums. But key informants noted that when there are at least three carriers present, all the carriers—unless they are colluding—have an incentive to reduce costs. Each wants to have at least one bronze plan and one silver plan premium that are at or below the benchmark silver plan premium in the marketplace so they can increase their market share.
In another move to engage carriers, most states declined to standardize benefit designs any more than the ACA required, thereby allowing carriers a wide variety of patient cost-sharing formulations (i.e., deductibles and copayments or coinsurance rates).8 Among the six states studied, California, Connecticut and Maryland limited the number of plans carriers could offer at each of the metal levels. In Arkansas, Montana, and Texas, informants explained that standardization of benefits was resisted, at least in part, in order to attract more carriers to the market. In these states, consumers confronted a relatively large choice of plans at each of the metal levels.9 Across all the rating areas of the 36 federally facilitated marketplaces, the average number of available plans (not including catastrophic plans) was 53 in 2014.10,11 This is far more than what is typically offered to people with employer-sponsored insurance and policy analysts are concerned that this could make selecting a health plan confusing.12 However, carriers and policymakers did not know what combinations of cost-sharing and premiums would be most attractive to the uninsured, especially those younger or healthier. These consumer choices will determine which combinations are offered in future years.
Finally, states were allowed to determine the number of premium-rating areas and their boundaries. Although the marketplaces are often referred to as if they are statewide, the premium-rating areas, consisting of counties or metropolitan areas, are the de facto marketplaces in most states.13 Among the six states studied, only Montana and Connecticut required carriers to offer plans in all rating areas in 2014. To increase their number of participating carriers, the other four states allowed carriers to choose where they would offer plans. As a result, in these states there are substantial differences in the numbers of carriers and plans available in each rating area.
Like most states, the six studied here have at least one carrier affiliated with Blue Cross–Blue Shield. In each case, that carrier had a large share of the state’s total and individual market health insurance business prior to the ACA and has retained the largest share in each of the six states’ marketplaces in 2014.
One reason Blues-affiliated carriers had a large first-year market share is that some of the biggest commercial carriers in the country—including Aetna, Cigna, and UnitedHealthcare—chose not to participate in the individual marketplaces in most of the study states.14 By staying out for at least the initial year, they avoided uncertainties about how they might fare under the new market rules. This initial absence provided opportunities for new or less-established carriers to compete. Each of the six states had at least one new entrant to the individual market and some states had several.15 There were three kinds of new entrants: co-ops (available in 23 states) Medicaid managed care plans, and smaller local or regional insurers that previously focused primarily on the group markets.
Connecticut, Maryland, and Montana have new co-ops. In Connecticut and Montana, the co-op was the critical third competitor needed to make the marketplace more price-competitive. Arkansas, Texas, and California have new entrants that previously provided managed care plans only to Medicaid beneficiaries or lower-income people in specific counties—for example, Chinese Community Health Plan and LA Care Health Plan in California. Connecticut and Texas had existing local or regional insurers in the group markets that saw the new marketplace as an opportunity to substantially expand their presence in the individual market. Many of these new entrants attracted substantial enrollment, especially when premiums were low.16 Interviewees told us that new entrants made other carriers more competitive in setting premiums, a point confirmed by other reports.17,18,19
The reform provisions of the ACA are intended to stop the practice of carriers selecting enrollees based on their likely use of costly health care. Instead, the ACA encourages carriers to compete in terms of how they provide value to consumers, where value is indicated by premium, expected out-of-pocket costs, and quality of providers. However, good information that consumers might use to compare quality of providers was scarce in the first year. Of the study states, California was the only one where even limited information about providers was available to people shopping for a plan in 2014.20
We detected no indication of any regulatory circumvention of the ACA’s basic provisions of guaranteed issue, age-adjusted community rating, and standardized benefits and cost-sharing. The law’s core requirements that prevent risk selection are being enforced in every state and carriers are complying.21
Key informants in each state commented that consumers appear to be price-sensitive and this is affecting the way carriers price plans. We observed that most insurance agents, despite some discontent, were constructively engaged with the ACA’s market reforms. In the states studied, numerous agents had received training to sell individual insurance in the marketplaces. Many agents see this as a business opportunity and understand that their role has shifted from medical underwriting to consumer assistance.
Although competition in the marketplaces appears to have shifted from risk selection and toward consumer value, the marketplaces in the six states do not fit neatly into a single, well-defined pattern. Various demographic and regulatory factors explain many of the differences observed among and within the six study states.
Exhibit 1 shows the number of plans being offered in the six states at each of the metal levels and indicates how many carriers are offering plans in each premium rating area. To see how people’s choices depend on where they live, we have interactive maps of each state that show the total number of metal level plans offered and number of carriers offering plans in each county within each premium rating area.
Exhibit 1. Range of Numbers of Health Plans by Metal Level
by State Offered in 2014
Focusing on the premium rating areas is crucial to analyzing the forms of competition in the states’ marketplaces because most states permit carriers to sell plans in just some premium rating areas rather than all of the state. California’s 19 rating areas have between two and six carriers each. Arkansas and Maryland allowed some managed care organizations to sell only in the areas where they have provider networks. In Texas, which has 254 counties organized into 26 rating areas, there are even differences within premium rating areas; some counties have only one or two carriers offering plans while other counties in the same rating area have four or more.
The variation in the number of plans and carrier options available in different parts of the states is easily seen by examining the interactive maps. They show that Connecticut and Montana required that carriers offer the same plans in all rating areas. In contrast, there are substantial differences in choices depending on where a person lives in Arkansas, Texas, and California. Further, the interactive maps suggest that in the states that did not require carriers to offer the same choices in all rating areas, choices differ based on the relative population size and per capita income in the premium rating area.
The types of products (i.e., HMO or PPO) offered are also important for assessing the extent of competition. When product choices are counted at the premium rating area level rather than aggregated to the state level, it is apparent that PPO and other non-HMO plans are more likely to be offered in areas that are more sparsely populated and have high numbers of poor and uninsured people (Appendix). Plans with limited provider networks are more likely to be PPO plans than HMO plans.
Finally, as other research shows, silver plan premiums are higher in rating areas with only one or two carriers.22,23 We also observed this effect, and saw that the competition to produce the lowest-priced silver plan did not mean that the insurer with that plan had the lowest premiums across all the metal levels.
The interactive maps and data show there are differences across the six states—and within four states—in how carriers are competing and the choices consumers have. In the six states studied here, carriers’ use of two approaches—relying on patient cost-sharing or using limited provider networks—created three distinct competitive strategies used in the states’ marketplaces. We observed four key state characteristics that aligned with the different competitive marketplace strategies.
In Exhibit 2, we illustrate each state’s unique combination of these characteristics. Distinguishing between characteristics that do not change quickly, such as population size, and characteristics that a governor or legislature could adjust is important for states that want to alter initial competitive strategies.
Exhibit 2. Key Characteristics Across Study States
|Arkansas||Population: 3 million; 56% of population in urban and suburban areas||Three carriers are competing; five of seven areas have two or three carriers offering plans. Prior to ACA the largest insurer had 78% of individual-market enrollees.||“Any willing provider” requirement for managed care plans (i.e., plans have to accept any provider who wants to be in their provider network)||Mixed views about reform; partnership state; innovative private Medicaid option|
|California||Population: 38.8 million; 95% of population in cities and suburban areas||Two to six carriers are competing in areas. Prior to ACA, four carriers had more than 5% of individual market enrollees; the largest had 48% of enrollees. Providers competitive in many urban areas, but less so in rural areas.||Active purchaser exchange; divides regulatory authority between managed care and indemnity plans||Long history of reform efforts; prior experience with failed insurance exchange|
|Connecticut||Population: 3.6 million; 88% of population in urban and suburban areas||Three carriers are competing in each area. Prior to ACA, the two largest carriers in individual market had 40% and 30% of enrollees. One of these chose not to compete in 2014. Some areas have dominant hospitals.||Strong network adequacy rules||Strongly supports reform|
|Maryland||Population: 6 million; 87% of population in urban and suburban areas||Six carriers competing but not in all areas. Prior to ACA, only two carriers offered plans in individual insurance market, largest had 72% of enrollees.||Tradition of strong regulation (e.g., hospital rates have been regulated since 1971)||Strongly supports reform|
|Montana||Population: 1 million; 56% of population in urban and suburban areas; one city with more than 100,000, two other cities with populations between 50,000 and 100,000||Three carriers competing in all areas. Prior to ACA, two carriers had 34% and 36% of individual insurance market enrollees and MT BCBS had 60% of large-group insurance market. Only two cities have two hospitals.||Active Department of Insurance||“Silent” partnership state (i.e., state conducts plan management activities to support certification of qualified health plans in the federally facilitated marketplace)|
|Texas||Population: 27 million; 85% of population in urban and suburban areas||Carriers and providers competitive in urban areas, less so in rural areas. Prior to ACA, five carriers each had more than 5% of individual insurance market enrollees; largest had 59% of enrollees.||Tradition of fairly light regulation||Hostile to reform|
With observations of only six states, it is impossible to say that particular characteristics are demonstrably more important than others in explaining the type of competition in specific marketplaces. Nevertheless, this analysis provides a starting point for examining competition in other states and a reference point for any future analysis of all 50 states.
Demographics. The six states’ populations in 2013 ranged from just over 1 million in Montana to more than 38 million in California. But as the maps clearly show, these totals mask the extent to which the populations are clustered around certain cities or counties while other areas are quite sparsely populated. Under the ACA, states were able to decide the number of premium rating areas they would have and their boundaries. Some states worked to create rating areas with roughly equal numbers of people but other states did not. The rating area boundaries follow county lines in most states, with areas consisting of more than one county. There are generally far fewer physicians and hospitals in rural areas so carriers have limited bargaining power for negotiating lower reimbursement rates or establishing exclusive provider networks in these areas. As a result, the premium rating areas have a direct influence on carriers’ options for how they might compete in a state or in specific parts of a state.
In some states or regions, we observed that carriers differentiated their plan offerings primarily in terms of patient cost-sharing (e.g., Montana) or by offering PPO plans with a combination of greater cost-sharing and restrictions on providers (e.g., rural counties in California and Texas). Carriers that rely primarily on limited provider networks are either not offering their products in sparsely populated areas or they are not competing in the state marketplace at all.24
Market structure. Arkansas, Connecticut, Maryland, and Montana each had a dominant carrier (i.e., Blue Cross Blue Shield) in their individual insurance markets before the marketplaces were established. In addition, Arkansas, Connecticut, and Maryland had managed care plans with smaller market shares. Arkansas and Maryland did not require these carriers to compete in rating areas where they lacked provider networks. In contrast, Connecticut required carriers to compete across the state; the one HMO offering plans in the marketplace is in all eight rating areas. Montana does not have any carriers offering managed care plans in the state.
In contrast, California and Texas had other carriers besides BCBS with substantial shares of the individual market before the ACA. Cigna, UnitedHealthcare, and Aetna offered PPO plans in the individual markets of these two states before 2014, but chose not to sell plans in Covered California in 2014, and although Cigna and Aetna sold plans in some areas of Texas in 2014, UnitedHealthcare did not. California and Texas also had strong regional HMOs (e.g., Kaiser Permanente in California, Scott and White in Texas) which chose to offer plans in some areas in 2014. California also has several Medicaid managed care plans and provider groups that have long focused on providing care to underserved and uninsured people.
The carriers in Arkansas, Connecticut, Maryland, and Montana continued to offer the same types of products that their residents had been used to purchasing in the individual and small-group markets before the ACA. Competition in Arkansas and Maryland reflects a combination of strategies while competition in Connecticut and Montana primarily focuses on cost-sharing differences. In some urban parts of California and Texas where there were large numbers of uninsured people, some carriers’ plans require enrollees to obtain care from a limited set of providers.25 Most of the competition in California’s rating areas is in terms of provider networks while carriers in Texas’s rating areas use a mix of strategies. The carriers in the marketplaces are not straying too far from what they offer outside the marketplaces.
Regulatory approach. States’ existing regulatory structures also influenced the forms of marketplace competition that initially emerged. How states regulate HMOs and other managed care networks is especially relevant, particularly the issue of whether states require carriers to include most providers in their networks or if they permit limited networks.
Many states regulate HMOs differently than PPOs, with a strong focus on HMO network adequacy and quality. Some states allow non-HMOs to sell closed network products, known as exclusive provider organizations (EPOs). Depending on the state and how a carrier is incorporated, a carrier might be restricted in the types of provider networks it can offer or it might be able to apply to a different state regulatory agency to offer the type of product it prefers.
Texas, for instance, has a mix of HMOs, PPOs, and EPOs that is based partly on differences in existing state regulations of each type of managed care plan. Texas has stricter regulation of network adequacy for HMOs than for conventional insurance, with specific limits regarding how far a person must travel to reach a primary care provider or a hospital emergency room. Accordingly, the HMO plans offered in the premium rating areas are concentrated in urban areas where they constitute a majority of the marketplace products. In Texas’ more rural rating areas, the majority of products offered are PPOs. Similarly, the mix of HMO, PPO, and EPO plans in California is driven by different provider markets and the varying ways California regulates these plan types.26
In Connecticut, existing regulatory policy discouraged carriers from offering different provider networks in different plans or market segments because of concerns about creating consumer confusion. Carriers must use networks that are “substantially similar” to those they offer in the large-group market, which is defined as including 85 percent of the same providers.27 Not surprisingly, Connecticut required carriers to offer the same plans in all areas of the state. Montana, where only two cities have more than one hospital, also required carriers to offer the same plans in all areas of the state. In Arkansas, sources said that a state regulation limiting how much insurers can penalize patients for going out of network made it difficult to create plans with limited provider networks.28
Attitudes about and history of reform. The extent to which a state was actively involved in setting up its marketplace influenced its competitive strategies, as did the state’s history of health reform. Among the six states studied, California and Texas are at the bookends of the spectrum of reform efforts and involvement in setting up a marketplace. Texas chose not to run its own marketplace and the state’s Department of Insurance has essentially stayed out of the marketplace operations.
The board of Covered California—California’s state marketplace—used selective contracting authority to require carriers to compete for approval to sell in each of the 19 rating areas in the state. Although the Covered California board did not use its selective contracting authority to exclude qualified carriers, it defined more specifically than other states what qualifications were necessary for a carrier to participate. Moreover, the fact that California could use its authority to reject any bids on the basis of proposed rates encouraged some carriers to offer plans with narrow provider networks. Several sources told us that the decision to use selective contracting grew out of the state’s history of managed competition in the California Public Employees’ Retirement System—also known as CalPERS, the state’s public employee program for pension and health benefits—and in the Health Insurance Plan of California for small employers, as well as California’s approach to regulating managed care plans since 1975.29 This helps explain California’s limits on the plan designs that carriers can offer at each metal level. This, in turn, explains why competition in most of California’s rating areas is focused on provider networks.
Maryland chose not to conduct selective contracting but required marketplace participation by all insurers that had pre-ACA health insurance market shares above $10 million in the individual market and above $20 million in the small-group market. The result is that competition among carriers exhibits a mix of cost-containment approaches.
The other states neither required nor restricted participation by carriers. Instead, they accepted all qualified carriers that chose to participate.30 However, these states were far from passive in their dealing with carriers. Connecticut and Maryland required carriers to offer standardized benefit designs and limited the number of nonstandard products carriers may offer. In contrast, Arkansas and Montana, despite having only three carriers each, have a very large number of plans available.
Although the ACA does not require states to regulate the premiums of plans offered in the marketplaces, all the states except Texas chose to actively review carriers’ proposed 2014 rates as a way to encourage low premiums. As part of their review, insurance regulators or marketplace officials questioned some of the assumptions carriers used in developing their proposed rates. This resulted in substantial reductions of 10 percent to 20 percent by one or more carriers. In Connecticut and Montana, at least two carriers reevaluated their proposed rates; in Arkansas and California, some voluntary rate revision occurred even before proposed rates became public because officials quietly advised higher-priced carriers to reevaluate their initial filings.
This dynamic caused the marketplaces to be highly price-competitive, at least for the bronze and silver plans. However, it was critical to this dynamic that there were at least three competing carriers in most of the premium rating areas in each of the six states. (There were exceptions in rural parts of California, Arkansas, and Texas).
Our study of six states’ experiences during the first year of the ACA marketplaces shows that when carriers are not allowed to compete by risk-selecting who they insure, they shifted to competing with different patient cost-sharing requirements and the composition of provider networks. But these competitive strategies are not the same across the country. Carriers compete differently in different states and rating areas, and the choices people have depend on where they live. The four factors we identified provide a framework for understanding the differences in competitive strategies among the marketplaces. But the fact that varying competitive strategies characterize different regions within a state raises a concern about carriers’ potential ability to avoid covering people who may have higher risks of costly medical conditions. If states and the federal government do not implement a statewide risk-adjustment mechanism, then the goal of shifting competition away from risk selection may be jeopardized.
This study also provides a reference point for identifying changes in the forms of future competition, as it reflects only the initial year of the ACA marketplaces. The ACA reforms will surely stimulate continuing adaptations by carriers, providers, and policymakers, and we expect the competitive strategies in the marketplaces to evolve as consumers and carriers gain more experience with marketplace competition.
Appendix. Types of Health Plans by Premium Rating Area–Region by State
|Types of||Health Plans|
|State & Rating
|Rating Area 1||12||29||0||0|
|Rating Area 2||6||11||0||0|
|Rating Area 3||12||29||0||0|
|Rating Area 4||6–12||11–29||0||0|
|Rating Area 5||0||11||0||0|
|Rating Area 6||0||11||0||0|
|Rating Area 7||6||29||0||0|
|Rating Area 1||0||12||6||0|
|Rating Area 2||0||12||6||0|
|Rating Area 3||0||12||15||0|
|Rating Area 4||0||6||17||6|
|Rating Area 5||0||12||14||0|
|Rating Area 6||0||12||11||0|
|Rating Area 7||0||12||14||0|
|Rating Area 8||0||12||11||0|
|Rating Area 9||0||12||6||0|
|Rating Area 10||0||12||6||0|
|Rating Area 11||0||12||9||0|
|Rating Area 12||0||12||11||0|
|Rating Area 13||0||12||6||0|
|Rating Area 14||0||12||6||0|
|Rating Area 15||0||6||22||6|
|Rating Area 16||0||6||22||6|
|Rating Area 17||0||12||17||0|
|Rating Area 18||0||6||12||6|
|Rating Area 19||0||6||27||6|
|Rating Area 1||0||16||0||0|
|Rating Area 2||0||16||0||0|
|Rating Area 3||0||16||0||0|
|Rating Area 4||0||16||0||0|
|Rating Area 5||0||16||0||0|
|Rating Area 6||0||16||0||0|
|Rating Area 7||0||16||0||0|
|Rating Area 8||0||16||0||0|
|Rating Area 1||9||4||18||6|
|Rating Area 2||9||4||14||6|
|Rating Area 3||9||4||14||6|
|Rating Area 4||8||4||15||6|
|Rating Area 1||4||25||0||0|
|Rating Area 2||4||25||0||0|
|Rating Area 3||4||25||0||0|
|Rating Area 4||4||25||0||0|
|Rating Area 1||0||12||11||0|
|Rating Area 2||0||12||11||0|
|Rating Area 3||0||19–30||44–50||0|
|Rating Area 4||0||19||12–17||0|
|Rating Area 5||0||12||35||0|
|Rating Area 6||0||12–18||19||0|
|Rating Area 7||0||12||12||0|
|Rating Area 8||0||12–30||6–14||0|
|Rating Area 9||0||12||35||0|
|Rating Area 10||0||25||6–17||0|
|Rating Area 11||0||19||19–46||0|
|Rating Area 12||0||12||8||0|
|Rating Area 13||0||12||6||0|
|Rating Area 14||0||12||11||0|
|Rating Area 15||0||12||8||0|
|Rating Area 16||0||12||11||0|
|Rating Area 17||0||12||11||0|
|Rating Area 18||0||12||14||0|
|Rating Area 19||0||19||6–42||0|
|Rating Area 20||0||23||6||0|
|Rating Area 21||0||12||6||0|
|Rating Area 22||0||12||6||0|
|Rating Area 23||0||12||6||0|
|Rating Area 24||0||19||52||0|
|Rating Area 25||0||12||6||0|
| Rating Area 26
The authors are grateful to all the many people who were interviewed in Arkansas, California, Connecticut, Maryland, Montana, and Texas in connection with the research for this project. Without their generosity in spending time with us, this project would not have been possible. The authors also are grateful to comments from Frank Levy, Nancy Turnbull, and participants in the RWJF Scholars in Health Policy Research Seminar, and for funding support from The Commonwealth Fund and to Sara Collins at The Commonwealth Fund for her comments and support. Finally, we thank Giovanni Zambotti of the Center for Geographic Analysis at Harvard University for creating the interactive maps.