Kosali Simon, PhD

An important feature of the Affordable Care Act is the creation of insurance exchanges, which are organized marketplaces through which individuals could begin to shop for and purchase coverage beginning in 2014. This study analyzes the decisions of new insurers and incumbent insurers already operating in a major market within a state to participate in state and federally facilitated exchanges.

Utilizing secondary data from the National Association of Insurance Commissioners* and government websites, we describe each state’s insurance market in 2012, summarizing the number of incumbent insurers by size and operations in the individual market segment. Next, we investigate the organizational, market, and policy-related factors associated with incumbent insurers’ participation decisions. Finally, we discuss the entry patterns of new insurers and briefly assess their potential impact on the market.

We use multivariate regression analysis to identify the organizational, market, and policy-related factors related to insurer participation in exchanges.

Only 10% of incumbent insurers are participating in exchanges in 2014, although considerable variation exists across states. Participation is more prevalent among larger insurers, local and regional insurers, and those with prior experience in other market segments in the same state. The entry of newly formed organizations, such as cooperatives (co-ops), and of existing insurers into new states is modest.

Robust participation of insurers is an important prerequisite to ensure competition in health insurance markets. Exchange administrators will need to better understand the strategic or operational reasons why insurers chose not to participate in the individual market exchanges in 2014.

*The NAIC does not endorse any analysis or conclusions based upon the use of its data.

Using data from the National Association of Insurance Commissioners and government websites, our analyses of insurer participation in individual market exchanges reveal:

		Ten percent of incumbent insurers are participating in exchanges; entry by newly formed organizations is modest.

		Participation is positively related to insurer size and prior experience in other market segments; it is not significantly lower in markets with a dominant insurer.

		States’ decisions related to exchange creation, Medicaid expansion, or active purchasing do not significantly affect participation.

		Robust participation is important for insurance market competition. Exchange administrators need to better understand both the facilitators of and barriers to insurer participation.

The creation of health insurance exchanges, through which individuals and small employers could begin to shop for and enroll in coverage in 2014, is a key provision of the Patient Protection and Affordable Care Act (ACA).

 While an underlying objective of exchanges is to strengthen competition in health insurance markets, exchanges may miss this goal if health plans do not participate. While much has been learned from the Massachusetts Health Connector regarding exchanges, the issue of plan participation remains unexplored. In the absence of empirical evidence, advice on exchange design abounds.

 The insurance industry supports unrestricted plan participation in ACA exchanges, while consumer groups and advocates recommend empowering exchanges to “keep out bad actors” and ensure value.

This study analyzes insurers’ decisions to enter state and federally facilitated individual market exchanges. We begin by describing the insurance market in each state in 2012, summarizing the number of insurers incumbent in the state by size and insurers’ presence specifically in the individual market segment. Next, we investigate the organizational, market, and policy-related factors associated with incumbent insurers’ participation decisions. Finally, we discuss the entry patterns of new insurers into exchanges and briefly assess their potential impact on the market.

Findings of this analysis provide a clearer understanding of the extent to which incumbent and new insurers found exchanges to be an attractive opportunity. Results also reveal how factors such as health insurance market structure, exchange governance related to qualified plan determination, and insurers’ size, group affiliation, and profit status affect participation. Findings could point to problems, but also to opportunities to improve the design of exchange governance and to encourage more robust insurer participation.

We used 3 sources of data from the National Association of Insurance Commissioners (NAIC) to identify insurers operating in each state in 2012 (the most recent year for data).

 The first source is the Supplemental Health Care Exhibit (SHCE). Starting in 2010, insurers selling coverage subject to new ACA regulations (eg, medical loss ratio regulation) filed this exhibit. The SHCE includes information on insurers’ enrollment, premiums, claims, and other information by state and market segment (eg, individual, small group, large group). The second source is the Annual Statements of Health Insurers’ Exhibit of Premiums, Enrollment and Utilization, also known as the “state page.” It complements the SHCE by documenting insurers’ activity in segments not clearly captured by the SHCE. Specifically, insurers with at least 95% of their business in health insurance file detailed information about their operations in Medicaid managed care, Medicare Advantage, Medigap, and the Federal Employees Health Benefit Program (FEHBP).

Although the NAIC database is the most comprehensive source available to characterize insurers and insurance markets, it has some limitations, the most notable being incomplete representation of insurers in California. For example, several health plans in California do not file NAIC information (eg, Chinese Community Health Plan, Contra Costa Health Plan, Valley Health Plan, Western Health Advantage, Alameda Alliance for Health, Sharp Health Plan, Molina Healthcare, and L.A. Health Plan).

NAIC also provides company-level measures of insurers’ ownership status, national group affiliation (eg, Aetna, Humana), and whether the insurer is a health maintenance organization (HMO). From these databases, we calculated insurers’ enrollment and scope of operations across market segments within states (individual, group and/or FEHBP; Medicaid; multi-segment), as well as measures of state-level market competition.

To measure potential demand for subsidized coverage through exchanges, we augmented the NAIC data with the number of uninsured in each state earning less than 250% of the federal poverty line (FPL) estimated from the 2012 Current Population Survey Annual Social and Economic Supplement. Although premium subsidies continue until 400% FPL, the chosen threshold is the point at which cost-sharing subsidies phase out. To capture the policy environment, we utilized state-level information from the Kaiser Family Foundation on exchange type (eg, state, federal, or partnership), active purchaser status whereby state-based exchanges may contract with selected health plans and/ or negotiate premiums with plans, and states’ decisions to expand Medicaid in 2014.

We searched the websites of state agencies that oversee insurance regulation and state exchanges to identify insurers participating in individual market exchanges in 2014. These searches identified insurers offering plans in state-based exchanges (see the online appendix for the list of state websites). We obtained insurer participation in federally facilitated and partnership exchanges from the plan information provided by HealthCare.gov (2014), the official federal exchange website (

https://www.healthcare.gov/health-plan-information

). Next, we linked 2014 participants to the 2012 NAIC data using the insurer’s name, state, and, in some instances, NAIC company codes. We used Web searches to obtain information on 2014 exchange insurers that did not link to the 2012 NAIC data, including type (eg, co-op), and the year the organization was formed.

The unit of observation is an insurer-state (ie, if Insurer X operates in State 1 and State 2, this represents 2 observations: Insurer X-State 1 and Insurer X-State 2). Incumbent insurers are defined as insurers operating in the state in 2012 with enrollment in at least 1 of 3 market segments: individual, group/FEHBP, or Medicaid. We did not consider incumbent insurers to be insurers that operated only in Medicare Advantage, supplemental Medicare insurance (Medigap), dental/vision, or third-party administrative services in 2012. At the broadest level, there are 3199 insurer-state observations with activity in any insurance segment and of any size as documented in the NAIC filing data. Of these, 2439 have operations in at least 1 of the following segments that classify them as an incumbent insurer in that state: individual, group/FEHBP, or Medicaid. We consider new entrants to include insurers with existing operations in other states in 2012 that entered a state’s exchange in 2014, as well as organizations that were newly formed in 2012 or 2013.

We use a set of organizational, market, and policyrelated factors to model an incumbent insurer’s decision to participate in an exchange. The organizational attributes we examine are operations in different market segments in the same state (individual, group/FEHBP, or Medicaid), operations in multiple segments, and size. We measure size using the insurer’s combined enrollment across the aforementioned segments in a state and then recoded it into quartiles (1 = smallest enrollment quartile, 4 = largest enrollment quartile). Given significant economies of scale of offering insurance and building provider networks, we hypothesize that insurers with a larger presence in the state will be more likely than small insurers to participate in exchanges. We also consider whether for-profit insurers, HMOs, health insurers (ie, life, property/casualty, or fraternal), and multi-state insurers (operating in 2 to 10 states or more than 10 states, relative to single-state insurers) are more likely to participate. Finally, we include 5 binary indicators for insurers that are part of a larger, typically national, group of insurance companies (eg, UnitedHealth Group, WellPoint, Cigna, Humana, and Aetna) because insurers’ decisions to enter particular states may be influenced by the national company. In May 2013, Aetna bought Coventry Health Care. We treat these entities as separate groups.

Both the demand for insurance and the extent of market competition are expected to affect an insurer’s decision to enter an exchange. We consider just one demand factor: the potential volume of new business in the market as measured by the estimated number of uninsured with incomes below 250% of the FPL. While individuals may be eligible for premium subsidies up to 400% of the FPL, recent evidence suggests that demand for individually purchased plans is price-sensitive.4 Those under the 250% threshold may be most likely to take up insurance since this is the only group to receive cost-sharing subsidiaries and therefore face lower out-of-pocket costs.

We use 2 market competition measures to investigate whether insurers in a more dominant position in 2012 are likely to enter exchanges: (1) an indicator for whether an insurer is a "dominant" player in the state (more than 50% market share); and (2) whether an insurer is a competitor of a dominant insurer. Insurers competing against a dominant firm may be at a disadvantage given economies of scale and bargaining power in provider contracting; therefore, they may be less likely to participate. On the other hand, nondominant insurers may perceive exchanges as opportunities to compete for new customers on a more even playing field, even if there is a “big dog” insurer in the market.

We also consider whether decisions by state policymakers affected insurer participation. Specifically, does the state run its own exchange, use a partnership exchange, or rely on the federally facilitated exchange? One might expect higher insurer participation in states that run their own exchanges because of better rapport between those developing exchanges, state regulators, and insurers.Also, in states with their own exchanges, administrators could choose to actively negotiate with insurers seeking to sell coverage. We hypothesize that states adopting this “active purchaser” model will have lower probability of insurer participation.

Finally, we compare insurer participation in states that chose to expand Medicaid prior to open enrollment for plan year 2014 compared with those that did not. The predicted effect is ambiguous: in states choosing not to expand Medicaid, individuals with incomes between 100% and 138% of the FPL are eligible for subsidized coverage in exchanges, leading to potentially greater demand. However, if these individuals have worse health status and use more healthcare services, insurers may be uncertain about their ability to predict premiums and less likely to participate.

We use bi-variate methods (ie, χ² tests, z tests for the difference of 2 proportions, and t tests for the difference of 2 population means) to evaluate differences in attributes between exchange participants and nonparticipants. Additionally, we estimate a binary logistic regression to examine the independent associations between each organizational, market, and policy-related factor and an incumbent insurer’s decision to participate in an exchange.

In 2012, there were 2439 insurer-state observations across the 50 states and the District of Columbia, reflecting 578 unique insurance companies. There are 141 unique groups with at least 2 insurers, and 201 insurers that are not part of any larger group. The 

shows the number of incumbent insurers by state and insurer size in 2012. The number of incumbent insurers ranged from 17 in Rhode Island to 91 in Texas. Large insurers (at least 1000 life-years of enrollment in a state) represented 43.8% of all insurers and 99.9% of all lives. Among incumbent insurers, 80% reported individual market activity in 2012, including 86% of small insurers and 73% of large insurers (data not shown).

Exchange Participation by Incumbent Insurers

summarizes 2014 exchange participation among incumbent insurers by state and insurer size. Overall, only 247 (10%) of incumbent insurers are exchange participants. Notably, 233 out of 1068 (22%) large incumbents are participating in exchanges, with the participation rate not appreciably higher among those with individual market presence in 2012 (24%). Only 14 small incumbent insurers (<1%) are participating in exchanges. Three states (West Virginia, New Hampshire, and Maine) have only 1 exchange participant, whereas 7 states have at least 10 participants.

compares the attributes of incumbent insurers that are participating in exchanges against those that are not. The table only includes insurers with at least 1000 life-years of enrollment because participation by small insurers is quite low. Among these large insurers, participants have higher enrollment than nonparticipants (225,000 life-years vs 55,000 life-years across all segments) and are more likely to operate in multiple segments in that state. Exchange participants are much more likely to be local or regional than national, and are more likely to be HMOs. For-profit insurers have lower representation among exchange participants (66%) than nonparticipants (84%).

Marginal effects and standard errors from the binary logistic regression are reported in 

. Each observation is a large incumbent insurer in a state (n = 1068). The dependent variable equals 1 if that insurer participates in the exchange, and 0 if not. To account for repeated observations among insurers operating in multiple states, we clustered the standard errors by NAIC company code.

We see clear evidence of a size gradient in participation: insurers in the second, third, and fourth enrollment quartiles are 7.5, 19.8, and 31.4 percentage points, respectively, more likely than those in the first quartile to participate in the exchange. Insurers with prior experience in the group/FEHBP market are 15.7 percentage points more likely to be participants; insurers that operated in the individual market are not any more likely to be participants. We also observe lower probability of participation among insurers affiliated with UnitedHealth Group (-30.8 pp), Aetna (-13.2 pp), and WellPoint (-9.5 pp). The large magnitude of the United- Health Group coefficient is unsurprising given that they announced their intention to remain absent from several exchanges.

Among incumbent insurers, 24.3% are operating in states with a dominant insurer. Table 3 clearly shows that participation is higher among dominant insurers in a state. However, we do not find any evidence that competitors of a dominant insurer are less likely to participate. The competitors may perceive exchanges as a strategic opportunity to compete for new customers receiving subsidies under new “rules of the game.”

Insurers’ decisions to enter exchanges are not significantly associated with the number of low-income uninsured in the state. Also, there is little evidence that policy-related factors, such as exchange type or activepurchaser status, influenced participation. Insurers in states that did not expand Medicaid are less likely to participate in exchanges, but this effect is not precisely estimated.

In 2014, 30 of the 310 insurers participating in exchanges are new, while 247 are incumbents with NAIC filing information. Thirty-three participants existed before 2012 but did not file with the NAIC or were present in a state but did not operate in the individual, group/FEHBP, or Medicaid segments in 2012. Twenty-four states and the District of Columbia have a least 1 new insurer participating in the exchange, the majority of which are co-ops.

Only 10% of incumbent insurers from 2012 are participating in exchanges in 2014, although considerable variation exists across states. Even incumbent insurers with experience operating in the individual market are no more likely to participate than other insurers. Exchange participation is more prevalent among larger insurers, local and regional insurers, and those with prior experience in other market segments, including group/FEHBP. Insurers in states with a dominant competitor are not less likely to participate despite potential competitive disadvantages related to economies of scale and lack of negotiating power in provider contracting. Finally, entry is modest for newly formed organizations (eg, co-ops) and existing insurers entering new states .

Of course, other factors not accounted for in our analyses may have influenced insurers’ participation decisions. For example, we were not able to account for the product composition or past performances of incumbent insurers in 2012. Insurers specializing in limited benefit plans may have found it too costly to adapt their products to comply with ACA regulations, and thus chose not to enter.

Low participation rates during the initial year suggest that many insurers adopted a “wait and see” strategy. Notably, several large national insurers decided to participate in only a few states in the initial year compared with the total number of states where they operate.

 Anecdotal evidence suggests that insurers’ uncertainty about the composition of the risk pool and its implications for premiums and profitability may have affected their willingness to participate. In an attempt to address such concerns, the ACA includes 3 different types of risk mitigation strategies: risk corridors, reinsurance, and risk adjustment.6 Additionally, to bolster the number of insurance options available to consumers, the ACA created the Multi-State Plan Program. This program charges the Office of Personnel Management with negotiating contracts with at least 2 national insurers, of which at least 1 is nonprofit, to offer plans in all exchanges within the next 4 years.

Policy makers clearly need to monitor insurer participation and the composition of participants in subsequent years. Efforts should be made to better understand the strategic or operational reasons for nonparticipants’ decisions. Exchange administrators may also consider targeting Medicaid plans for participation, given their experience with populations most likely to gain subsidized coverage through the ACA.

Future work is needed to understand the extent of competition in exchanges, including plan choice and premium variation at the state level and in local geographic markets. Longitudinal research is also needed to understand insurers’ participation over time as exchanges mature and ACA provisions are fully implemented.

From Division of Health Policy and Management, University of Minnesota, Minneapolis, MN (JMA, RF); School of Public and Environmental Affairs, Indiana University, Bloomington, IN (KS).

The study received financial support from the State Health Reform Assistance Network, a program of the Robert Wood Johnson Foundation.

The authors report no relationship or financial interest in any entity that would pose a conflict of interest with the content of this paper.

Articles

This study analyzes the factors associated with insurer participation in the initial year of individual market exchanges created under the Affordable Care Act.

Kosali Simon, PhD

Articles

Did They Come to the Dance? Insurer Participation in Exchanges