AN EVALUATION OF CAROLIANCE:
NORTH CAROLINA'S VOLUNTARY
HEALTH INSURANCE PURCHASING ALLIANCE PROGRAM FOR SMALL EMPLOYERS
 

Wake Forest University School of Medicine
Janice S. Lawlor, MPH
February 1999

(c) 1999 Wake Forest University School of Medicine. This publication may be downloaded, copied, or distributed free of charge only if (1) it is not edited or altered in any manner, (2) proper attribution is given for authorship and copyright, and (3) it is not sold and no fee is charged to recipients, even for reproduction and distribution costs.

I. EXECUTIVE SUMMARY

II. BACKGROUND AND METHODS

A. Background
B. Methods

V. TABLES
 
 

(c) 1999 Wake Forest University School of Medicine. This publication may be downloaded, copied, or distributed free of charge only if (1) it is not edited or altered in any manner, (2) proper attribution is given for authorship and copyright, and (3) it is not sold and no fee is charged to recipients, even for reproduction and distribution costs.

This report analyzes the purchasing cooperative component of North Carolina's small-group reform law. A separate report evaluates the success of the North Carolina's other small-group market reforms, which include guaranteed issue, renewability and portability, rating restrictions, restrictions on underwriting practices such as risk selection and preexisting condition exclusions, and reinsurance (1).

Caroliance has been a well-intentioned effort to improve competition in the small-group market so as to reduce premium prices and enhance product choices. The ultimate goal has been to increase the percentage of small groups with coverage in North Carolina. This study gathered information on the range of effects of the Caroliance program and on the reasons for these effects, from the perspective of the insurance industry and its government regulators. The findings are useful for illustrating the combined effect of discrete design features of a state-initiated health insurance purchasing alliance program. Unfortunately, the findings show that, on balance, Caroliance has not been a success.

Positive effects of Caroliance include:

• Improved access for higher-risk groups to more comprehensive benefit packages and to the statutory benefit products.
• Easier comparison shopping for purchasers as a result of standardized products.

• A small enrollment that is disproportionately high risk
• Premium rates for healthier groups that are not competitive with the outside market.
• Low carrier participation.
• Difficulties with the employee-choice feature.
• Failure to implement quality comparison measures.

 

A. Background

The North Carolina reforms, which took effect in 1992, included a system of regional alliances that small groups (2) could join on a voluntary basis for the purpose of purchasing health insurance. The idea was for the state to provide seed money and expertise to help the alliance system become operational, but eventually to bow out leaving the alliances to be self-supporting.

North Carolina's alliance program was named Caroliance. According to the program's leadership, its mission is "to increase the affordability, efficiency and fairness of health coverage for all small employers in the State of North Carolina" (3) and in so doing to decrease the level of uninsurance in the small-group market. The basic idea in creating Caroliance was to put in place an efficient mechanism for small groups to pool their purchasing power and reduce health insurance premiums. But the intended benefits of Caroliance were to go beyond price improvements, to also offer employees of small groups a wide choice of benefit plans and carriers. Furthermore, Caroliance was designed to make competition among carriers in the small-group market more equitable by requiring the sale of comparable benefit plans in the alliances, by monitoring clinical outcomes and customer satisfaction, and by utilizing a risk-adjustment mechanism to spread the risk evenly among participating carriers (3).

To oversee development of the alliance system, an 11-member board called the State Health Plan Purchasing Alliance Board was appointed by the North Carolina General Assembly in late 1993. Board members include the lieutenant governor, the state insurance commissioner and nine public appointments, six of whom must represent small business. An executive director was hired in January 1994 and three other staff members were hired over the next year. The State Board was charged with setting up from four to 12 noncompeting regional purchasing alliances. Six alliances were incorporated initially, each with its own governing board of local small-business owners and each with its own executive director and support staff. A third-party administrator (TPA), HealthPlan Services, was hired to assist the regions with marketing to small groups, enrollment, premium billing and collections, payment of agent commissions, and database maintenance. Participating carriers, called accountable health carriers, were recruited to offer a variety of standardized health insurance products through independent health insurance agents. Caroliance began selling its products in late 1995.

B. Methods

This study is part of an intensive study of seven states (4) that have enacted varying health insurance market reforms affecting small-group and individual health insurance. Funded by the Robert Wood Johnson Foundation, the larger study evaluates whether these reforms have achieved their intended objectives and avoided possible harms and failures.

In North Carolina, two rounds of interviews were conducted, the first in 1997 and the second in 1998. There were nine interviews in which Caroliance was discussed. These subjects included two representatives of the state Department of Insurance (DOI), four representatives of two insurance carriers, a Caroliance representative, two independent insurance agents, a benefits administrator, and an insurance industry consultant.

Data relative to Caroliance were analyzed using both qualitative and quantitative techniques. Interview notes and transcripts were coded and entered into a specialized computer program for content analysis. Quantitative data were analyzed with simple descriptive statistical techniques.

Criteria used to evaluate Caroliance included 1) whether the program has improved the availability of insurance products to small groups; 2) whether Caroliance has improved the affordability of products for small groups; 3) whether Caroliance has increased enrollment by small groups, i.e., whether availability and affordability together have translated into more insurance purchases; 4) whether Caroliance has improved the efficiency of the small-group market; and 5) whether Caroliance has improved fairness in the operation of the small-group market. Most of the findings reported here relate to the period prior to the enactment of the federal Health Insurance Portability and Accountability Act (HIPAA), which took effect July 1, 1997. However, the discussion also examines how interview subjects perceive the market in the short time since HIPAA has been in place.

A. Availability

1. Products for high-risk groups

The first question is whether Caroliance has made more insurance products available to potential subscribers in the North Carolina small-group market, particularly those groups and employees who were having trouble purchasing products because of existing health conditions. (Availability here only refers to the range of products offered for sale. Enrollment, or purchase of the products, is discussed later.) Guaranteed issue of the statutory basic and standard plans went into effect in 1992 for the entire small-group market. All carriers were required by law to offer those products to groups unable to pass medical underwriting requirements for the carriers' more popular "street" products. When Caroliance was created, it was restricted by law to selling only the basic and standard plans, the same statutory plans that were already available in the outside market. However, these plans are not viewed favorably in the market due to their low benefit level (particularly the basic plan) and because they are purchased primarily by high-risk groups. Caroliance felt at a disadvantage in only being able to sell these products, so petitioned for and received permission to also sell more generous benefit packages beginning January 1, 1996. These are called select plans. Caroliance allows the select products to be sold at preferred rates for groups that can pass medical underwriting and at higher guaranteed-issue rates for other groups. In this way, Caroliance actually made more appealing benefit packages available to groups with medical conditions, albeit at a higher rate. Such packages were generally not available to these groups in the outside market prior to HIPAA.

Caroliance also seems to have improved the perceived availability of the statutory plans, although technically these products were already available to groups before Caroliance came on the scene. From its beginning, Caroliance has made a conscious effort to market its products in a manner that emphasizes the availability of guaranteed-issue products for all who require them. This affirmative marketing of the guaranteed-issue products has been conducted through radio and print ads and in other print media. These products have not been actively marketed in the outside small-group market. Indeed, as our companion North Carolina report indicates, the phenomenon of "field underwriting" (agents steering the less-desirable business away from a carrier) appears to be occurring to some degree in North Carolina. As a result, there may have been some groups that would benefit from the statutory plans but failed to learn about them. Our information sources indicate that an undetermined number of agents also may have been unaware of the existence of the guaranteed-issue products in the outside market, prior to HIPAA. News reports and other published materials that we obtained pointed to several instances in which agents made comments characterizing the Caroliance guaranteed-issue products as being new to the market, when in fact these products predated Caroliance's existence. Furthermore, several of our interview subjects characterized Caroliance as "the insurer of last resort," a designation technically inaccurate and fraught with adverse selection implications, as we will discuss later. Nevertheless, the reality seems to be that Caroliance has improved the perceived availability of guaranteed-issue products for small groups (as well as the actual availability of more comprehensive benefit plans for groups with health problems).

2. Distribution of products across regions

Another consideration regarding availability is whether Caroliance has helped improve the distribution of carriers and products across regions. Although some parts of the state have had a good selection of plans, other areas have had fewer from which to choose. This has been particularly true in rural areas, where managed care plans with comprehensive benefits have been lacking. The question is whether the pooling of small groups in regional alliances might have attracted carriers that were already operating in the state to expand into new regions or offer more products, or whether new carriers might have been attracted to the state for the first time. Our information indicates that Caroliance has had little impact in this regard, since the carriers that have signed up to participate were already in business in these regions. However, we were told that Caroliance was instrumental in helping a local provider network in the largely rural Western region team up with a national carrier in the region to develop a new PPO product that sold quite successfully for about a year, until the carrier left the state as part of a corporate strategy unrelated to Caroliance.

3. Employee choice of plans

Caroliance has theoretically improved the access of small-group employees to a broader choice of health insurance products. Employers purchasing through the alliance program must offer employees a choice of at least two health plans and pay at least 50% of the employee-only rate for the lowest-cost plan. The two plans can be with the same or different carriers. All carriers must offer the basic and standard plans and at least one select plan. Offering plans with different carriers increases the likelihood that an employee will have access to a network including his preferred providers. Only if the employer pays at least 70% of the premium can employees be restricted to one plan.

In reality, however, the effect of the choice feature has been minimal, in large part because of disappointing carrier participation. Caroliance started selling its plans in late 1995. While a State Board report at the end of that year mentioned that 14 accountable health carriers had been certified by Caroliance, only six actually participated when the alliance program commenced operation. We were told that the other eight ultimately declined participation for a variety of reasons including amended business strategies, market consolidations, concerns about adverse selection, and failure of the Clinton health care reform plan. (Some sources reported that carrier participation was enhanced in the beginning because it was viewed as a wise political move since the lieutenant governor had sponsored the legislation for the alliance program.) Participation dwindled over the next year so that by the end of 1996, the number of carriers was down to five, only two of which were participating in all six regions. Two of the regions had only two carriers, three regions had three carriers, and one had four. By the end of June 1998, there were only three carriers. Only one is participating statewide. The number of regions has been consolidated to four. One region has only one carrier, and the other three have just two (Table 1). While the reasons for this apparent disenchantment of carriers with Caroliance will be explored later, we can conclude at this point that Caroliance was able to offer a fairly broad array of carrier choices for only a short time, and even then only in some regions.

We were told by Caroliance that this limited carrier selection led, in part, to fewer than 5% of employer groups enrolling with more than one carrier. Groups in the urban markets with more managed-care choices were more likely than groups in rural markets to split their purchases between carriers. It also was reported that among the smallest groups (particularly those with fewer than five employees), some employers pressured employees to all purchase the same plan. The Caroliance TPA does not handle claims problems, so in the smallest organizations this task often falls to the owner or head of the business. Caroliance reported complaints from managers of these smallest groups that trying to understand and negotiate the claims systems of multiple carriers is overly time consuming.

4. Group purchase without employer participation

One final note on availability is that Caroliance offers a novel feature for a group whose employer is not willing or able to contribute to the group's health insurance purchase. That is, a group may purchase Caroliance coverage without employer participation if 100% of eligible employees agree to participate. In such a case, the employer still sends in one payment to the TPA after payroll deduction for the employees, although no company contribution is included. Caroliance reports that only a handful of groups have utilized this option, presumably because the per-employee cost would be too steep without an employer subsidy. This practice is similar to list billing (where an employer facilitates billing individual employees' health insurance premiums by a carrier). However, list billing is not allowed in North Carolina because it segments the small-group market by allowing good risks to purchase in the individual market. If employees are allowed to purchase Caroliance coverage without employer contribution, the employees remain in the small-group market and thus do not weaken the risk pool.

B. Affordability

Research has shown that affordability, not availability, is the main deterrent to health insurance purchase by potential subscribers (5). Not surprisingly, "increased affordability" of small-group health insurance is the goal mentioned first in the mission statement of the Caroliance program. We interpret this to indicate a desire for reduced prices in the market as a result of competition (rather than through government subsidies to purchasers).

We tried to judge from two perspectives whether prices have improved. First, how have market prices behaved, pre- and post-Caroliance? Are prices lower in the alliance and perhaps even in the outside market due to some competitive effect of Caroliance, relative to prices before Caroliance began? Unfortunately we do not have data on comparable products across market segments (alliance and non-alliance) over time, so it is difficult to make meaningful statistical comparisons to answer this question. In a discussion of affordability in the non-alliance small-group market, our North Carolina report indicates that median average premiums among the top carriers selling to employer groups of size 2-49 decreased slightly from 1994 through 1996 and then rose sharply in 1997. Data on average premiums by carrier for Caroliance are unavailable for comparison. It is highly unlikely, however, that price movements in the outside market had anything to do with Caroliance, which was struggling to gain a foothold in the market. More likely, as the report speculates, the initial price stability following reform was due to the growing presence of managed care in North Carolina and to some carriers' recent practice of keeping rates low to increase market share. Rates appear to be rising now as carriers try to make up for low profits or losses in recent years.

Lacking the data to judge the rate effect of Caroliance over time, we can instead look at the relative position of Caroliance rates in the small-group market at a recent point in time. To this end we were provided the results of an April 1, 1997, rate comparison conducted by HealthPlan Services, the Caroliance TPA. The comparison listed annual composite rates charged by participating carriers in each region for the various standardized products offered in the regional alliances. In addition, comparison rates were shown for a sample of non-alliance carriers offering similar products in each region. All but two of the outside products were priced using medically-underwritten rates.

Analysis of the report reveals that in only four of the 23 comparisons (17%), in which similar products were offered by Caroliance and by a different carrier(s) in the region's outside market, was a Caroliance product the lowest-priced (Table 2). However, there were three additional cases in which the lowest Caroliance price was within 10% of the lowest price. So in approximately one-third of the cases, Caroliance had the lowest price or a competitive price. But even so, Caroliance products were priced at a competitive disadvantage in two-thirds of these comparisons. Furthermore, in nearly half of the comparisons, the lowest Caroliance price was at least 20% above the lowest price. And in a third of the comparisons, at least half of the Caroliance products were more than 50% above the lowest price. In contrast, there was only one comparison among the 23 (4%) in which the lowest outside price was more than 20% above the lowest price (a Caroliance price). There were no comparisons in which at least half of the outside products were 50% or more above the lowest price. In fact, only 2% of all the outside products were priced more than 50% above the lowest price, while 32% of the Caroliance products were at this level. We have no information on the market share of these products in the total small-group market so we cannot ascertain how representative they were of prices paid by most purchasers. Also, there is some question about the degree of comparability of some of the outside products to the Caroliance products. Furthermore, the rate comparison does not give examples of situations where carriers sell the same product both in and out of the alliance. Nevertheless, the data suggest that Caroliance products have not been generally competitively priced in the market and do not exert downward pressure on market prices.

Agents in North Carolina and around the country have told us that small groups are very price sensitive about health insurance premiums and will often switch carriers to save only a few dollars, particularly with the improved portability of coverage resulting from state and federal reforms. And even though agents say price is not the only factor influencing a buyer's decision to purchase (the provider network, claims service, and, to a lesser extent, quality measures are other factors), price is a crucial factor to small employers with tight budgets. In such an environment, Caroliance is clearly at a disadvantage.

Anecdotal evidence from our North Carolina interview subjects generally confirms the view that when agents obtain quotes for clients, Caroliance is not the lowest. Although there were reports of favorable prices in some instances (one subject told of a 17% discount on what was available outside the alliance), the prevailing opinion was that in more cases than not, groups can get better prices outside. One subject complained of a particular carrier that participated in the alliance but charged 8% more for the exact same product in the alliance than outside. While some states, such as California, prohibit carriers from charging more for the same coverage in the purchasing cooperative than in the outside market (6), North Carolina has no such explicit prohibition. However, North Carolina's reformed rating rules allow rate variation based only on a group's health status (+/- 20%), age, gender, location, family composition, and benefit level (1). The purchase venue for the coverage is not an explicit adjustment factor. Therefore, charging a group more for the same product inside Caroliance is a debatable practice. Carriers say that it is justified because selling through the alliances is actually more costly to them. Furthermore, as discussed below, Caroliance rating rules do not allow carriers to use the bottom end of the +/- 20% rating bands. This causes the healthiest groups to be charged more in the alliance system than outside.

Most of the comparisons reported above were of underwritten plans. Guaranteed-issue standard indemnity plans with a $500 deductible and 20% coinsurance were compared in two regions. In the Southeast region, the Caroliance prices ranged from 37% to 123% above the lowest price, which was outside the alliance. In the Western region, however, one Caroliance price was the second-lowest price at only 8% higher than the best price. But in the Western region, Caroliance also had the highest price, at 54% above the lowest. It would be useful to have a more thorough comparison of the statutory plan prices in and out of the alliances since Caroliance has been perceived as a safety net for groups having difficulty passing medical underwriting and finding affordable coverage. In the outside market, when guaranteed-issue plans are offered, the prices are often inappropriately high due to the use of actuarial techniques that are questionable in light of the intent of guaranteed issue and related reforms (1). Whether Caroliance has reduced prices for these statutory products is, unfortunately, a question that we cannot definitively answer given the limitations of our data. One agent reported to us that early quotes received from Caroliance for statutory plans were lower than the competition's. However, another subject reported that Caroliance's statutory product prices were at the same level as those in the outside market.

The problem of non-competitive rates is compounded by the addition of fees to the premiums paid by subscribers. Each employer group that joins Caroliance must pay a membership fee of $50/year. In addition, there are monthly administrative fees per employee covered, ranging from $3.25 for employee-only coverage to $9.75 for family coverage. The fees are a revenue source for the regional alliances. However, they also increase the effective market price to employers, who are already at a relative disadvantage in the market.

C. Enrollment

The most obvious indicator of the Caroliance program's success, or lack thereof, is enrollment: how many groups have found the available Caroliance products appealing enough to purchase through the program? And what types of groups have made this decision? Has Caroliance attracted a large, risk-diverse pool of subscribers that will encourage carriers to offer volume discounts, or has the alliance program attracted a disproportionate number of higher-risk groups? How does the Caroliance enrollment picture compare to that of the outside small-group market? Has Caroliance helped reduce the level of uninsurance in the small-group market?

Caroliance opened its doors to enrollment in November 1995. By the end of December 1996, Caroliance reported that enrollment stood at 453 groups and 1,616 lives. This compares to 63,933 small groups (and 689,297 lives) who purchased through the outside market in 1996, according to the North Carolina DOI. Thus, Caroliance had fewer than 1% of the small groups purchasing in the state that year. Since then, Caroliance enrollment in terms of groups has doubled, but the numbers are still modest. By December 1, 1997, enrollment topped at 1,020 groups. This was only 1.5% of the total small groups purchasing in the state that year. And that number dropped to 889 groups by April 30, 1998.

Who are these groups that have purchased through Caroliance? First, they are very small: 97% have fewer than 10 employees and 91% have fewer than five employees (as of 12/96). At the end of its first year, the average group size in Caroliance was 2.2 employees, compared to 5.7 in the outside market. Second, the groups represent a range of industries (manufacturing, retail, service, agriculture, professional, nonprofit, and "other"), but the service industry is predominant with 47% of the groups in late 1996 and 59% a year later (Table 3). Comparable industry data on the outside market were not available. Third, more than half of the Caroliance groups in 1996 and 1997 were previously uninsured (53% and 54%, respectively), compared to just a quarter of groups purchasing outside in 1996 and 19% in 1997. However, that does not necessarily mean that Caroliance has had great success in covering chronically uninsured groups. Since the DOI's definition of previously uninsured encompasses groups that were without group insurance for only 30 days before purchasing, this number may include new businesses and groups switching carriers after a relatively short lapse, or groups in which employees were switching from individual to group coverage. Unfortunately, from the data available, we have no way of determining with certainty how many groups that were without coverage for an appreciable period of time might have purchased Caroliance products. An analysis of the Census Bureau's March Current Population Survey indicates that the uninsurance rate among North Carolina workers in businesses with fewer than 25 employees has risen very slightly from 28% in 1994 to 29% in 1997 (7). This suggests, indirectly, that Caroliance has not reduced the level of uninsurance among small groups in the state. However, we were unable to resolve questions about the definition of "uninsurance" by the Census Bureau. Also, comparable statistics on uninsurance in the 25-50 market segment are not available. Nevertheless, it appears that Caroliance has had no impact on the overall level of uninsurance among small groups in the state.

The enrollment data do indicate, however, that Caroliance is providing products attractive to groups that have medical conditions which limit their purchase options to guaranteed-issue products and rates. Indeed, Caroliance is getting a disproportionate number of these subscribers. The latest data from the DOI show that at the end of 1997 less than 2% of covered lives in the outside market were in statutory guaranteed-issue policies, but Caroliance data for 1997 show about 70% of covered lives with guaranteed-issue policies (Table 4). Interestingly, 63% of the lives with guaranteed-issue rates were in the select plans, confirming the appeal of these enhanced products to groups that traditionally have had little access to them.

A final observation about Caroliance enrollment is that subscribers have overwhelmingly chosen fee-for-service plans (indemnity and discounted indemnity, i.e., preferred provider arrangements) over the more managed plans (health-maintenance or point-of-service plans). More than 80% of Caroliance lives are in indemnity or PPO options (Table 5). This may signal a relatively sicker health status of Caroliance groups. Although there is disagreement on this point, some research has shown that less restrictive fee-for-service plans attract sicker subscribers, who are more attached to their doctors and are less willing to switch to a managed care network without their providers (8) However, 80% of underwritten lives in Caroliance have chosen indemnity or PPO plans, a possible reflection on the weakness of HMO provider networks in the more rural areas, as mentioned earlier.

D. Efficiency

Has Caroliance contributed to efficiency in the small-group market, as its proponents hoped it would? This question actually involves three kinds of efficiency -- efficiency in gathering and providing information for purchase decisions (marketing efficiency), efficiency in administering policies once a purchase is made, and efficiency of care by providers.

1. Marketing efficiency

When a small employer decides to purchase health insurance for its employees, there have traditionally been two options for gathering information on available products. The employer can contact several carriers directly, which is a time-consuming exercise, particularly if the group is too small to afford a benefits manager. The employer also can contact an independent agent, who will provide information on the products of various companies with which the agent is licensed. Both of these options existed before Caroliance. A feature that Caroliance added to the information process was the element of standardization of products to aid comparison. This is an efficiency improvement in the sense that the agent and the purchaser can more easily and quickly judge the relative merits of various carriers. Agents that we talked to, as well as some who were quoted in news articles, were complimentary of the program's responsiveness to information requests. A carrier subject also confirmed that Caroliance reduces insurers' marketing duties.

In regard to marketing efficiency, there is the sensitive issue of whether this goal is furthered by the current requirement that Caroliance products be sold through independent agents. One might argue that removing one operational layer (and the cost of agent commissions) from the system's structure could lower premiums. A counterargument might be that agents are already in place to efficiently perform a marketing/distribution function that would otherwise have to be developed in-house by the TPA and regions and would be reflected in the fees charged alliance members. When initial plans for the alliance system were being developed, it was proposed that Caroliance products be marketed without the involvement of independent agents. However, in the end, the decision was made to have agents serve as the sales vehicle. All direct requests from employers (14% of inquiries received since the program's start) are referred to agents in the appropriate region who have completed training on Caroliance products. Several of our interview subjects confirmed that this policy was made in response to complaints by agents who felt threatened by the possibility of direct sales of Caroliance products to small groups. Quantitative data are not readily available to compare the cost of the Caroliance agent-driven sales system with the potential cost of a direct marketing system. But it is instructive to note that some purchasing cooperatives in other states have found the involvement of agents to be an important factor in their success (9). In the case of North Carolina, we can conclude that trying to fit a new alliance system into the existing health insurance sales framework has been politically difficult.

Some of our interview subjects suggested that the alliance system should be modified to improve the efficiency and effectiveness of marketing efforts in the various regions. Problems cited included the division of the state into too many regions with duplicate offices and the need for increased emphasis on marketing, with more coherence between regions' sales efforts.

2. Administrative efficiency

Whether Caroliance has contributed administrative efficiencies also is debatable. Caroliance marketing materials emphasize the benefits of employee choice of plans and the ease with which employers can send one monthly check to the alliance to cover premiums for all carriers. The alliance disburses the appropriate amounts to carriers. As mentioned previously, some of the smallest groups in North Carolina have complained that the employee-choice feature causes claims hassles. We heard similar complaints from agents in Florida, which has a public purchasing cooperative system similar to North Carolina's and the same TPA firm supporting alliance operations. We also heard from agents in Florida that employers complained about the monthly premium payment period being reduced from 30 to 15 days in order to allow the TPA enough time to process the payment and distribute premiums to the carriers on time. In effect, adding another operational layer to the system caused the grace period for employers to be squeezed. Caroliance has a similar system; its business rules call for employers to be billed no later than the 5th of the month and payments from employers are due by the 15th, although coverage will not be terminated unless payment is not received by the 11th of the following month. We heard no complaints about the short grace period in North Carolina, although that may be an artifact of our sample of agents.

The message from carriers on the matter of administrative efficiency has uniformly been that Caroliance has not had a large enough enrollment to bring economies of scale to bear on the administrative component of rates. Even though Caroliance performs various functions for participating carriers (including marketing, group eligibility verification, underwriting, enrollment, premium collection, and agent commission payment), the carriers maintain that the Caroliance administrative system is incompatible with their own and a hassle rather than a help. Thus, rather than streamlining operations, Caroliance is viewed as an impediment to efficient operations. Subjects pointed out that the design of the Caroliance administrative system is based on the system at Healthsource, which ultimately declined participation in Caroliance. Subjects also noted that only if enrollment is sufficiently large will a carrier consider reconfiguring its system to fit Caroliance procedures and engender true economies of scale. Therein lies a dilemma for Caroliance. Carriers say there is not a sizable enough enrollment to justify system changes and premium discounts. But without better prices from its carriers, Caroliance cannot attract the enrollment numbers needed to earn these concessions. Caroliance had a goal at the outset of reaching 15,000 covered lives in three years, but attainment of that critical level has proven elusive.

Four design features of the Caroliance system are potential contributors to the alliances' inability to amass sufficient purchasing power with carriers. The first is the designation of Caroliance as a voluntary, rather than mandatory, alliance program. By allowing small-group purchasers in North Carolina a choice of the market in which to purchase (alliance vs. outside market), policymakers diluted the purchasing power of the alliances. Carriers no doubt would find an alliance of 63,000 groups more compelling than one of 450 or even 1,000 groups. But consensus for such strong government action is not easily achieved. Our sources tell us that in North Carolina, there was no discussion of a mandatory alliance program. A voluntary alliance was more politically viable than the proposals being discussed at the national level.

Some proponents of Caroliance have pointed to its lack of negotiating authority with carriers as another contributor to its generally noncompetitive rates. Caroliance can exclude carriers that do not meet minimum certification requirements. Beyond that, however, it must accept all willing insurers and the prices that they offer so long as the rates are set using the prescribed rating methods. In the parlance of health economists, Caroliance is a price-taker with no power to negotiate for reduced premiums or to forcefully encourage performance standards beyond a minimally acceptable baseline (10). Although the Caroliance business rules suggest that participating carriers offer a volume discount to larger groups, the alliances have no power to remove carriers from participation if price or outcome goals are not met. Several of our interview subjects commented that until Caroliance has a larger enrollment, the lack of negotiating power is a moot point.

The employee-choice feature also has been implicated as a factor in the failure of carriers to offer reduced prices. One subject was of the opinion that carriers were not inclined to offer volume discounts for larger groups (those with 20 employees or more) because the choice feature reduced the likelihood that a carrier would receive a whole group with the good, as well as the less desirable, risks. Another subject pointed out, too, that with only two rate tiers (guaranteed issue and preferred) in Caroliance, insurers are not able to offer a discount to the healthiest groups as they are in the outside market, where underwritten rates are divided into standard and low-risk tiers. Therefore, fewer healthier groups have been attracted to Caroliance. Caroliance maintains that it prefers only two rate tiers in order to streamline the underwriting process for the TPA and respond more quickly to agents requesting quotes.

3. Provider efficiency

Finally, regarding provider efficiency, Caroliance has not had the requisite bargaining power to compel carriers and their providers to reorganize and integrate systems of care to significantly lower medical costs and achieve better quality of care.
 

E. Fairness

1. Incomplete implementation of features to promote fair competition

Fairness for small groups in the context of the health insurance market no doubt means different things to different stakeholders. Caroliance was designed to bring a new structure to a portion of the market so that small groups could enjoy health benefits equivalent to those of larger groups in terms of price and choice. But the structure must have features to ensure a level playing field for participants. Was the resulting program able to deliver the desired results in a context of equity? If not, why not? How do the various stakeholders view Caroliance from the standpoint of fairness?

Prior to Caroliance and the other health insurance reforms of the early 1990s in North Carolina, commercial insurance carriers were increasingly competing using risk selection: carriers tried to encourage healthier groups to purchase by charging them lower premiums and discouraged higher-risk groups by charging extremely high prices or by refusing coverage. Increasing numbers of small groups were finding it difficult to purchase insurance if any members of the group had health problems. Purchasers and carriers with more liberal enrollment policies complained that competition was being increasingly driven by the use of risk selection tactics by carriers willing to use aggressive medical underwriting techniques to pluck the healthy risks from the market, leaving the less healthy risks to the more socially conscious carriers and health plans or without coverage. Marketplace reforms such as guaranteed issue and renewal, portability, rating restrictions, and risk adjustment mechanisms were attempts to reorient competition away from risk selection to other more socially productive goals, such as improved carrier efficiency and management of subscriber care (11).

When Caroliance was developed as a way to bring pooled purchasing power to bear on health coverage for small groups, several features were planned to discourage carriers from competing on risk selection and to encourage more equitable competition based on efficiency and quality. One feature -- having all participating carriers sell standardized benefit packages -- was designed to lead to easier comparison of products by purchasers and make it more difficult for carriers to select risks by varying benefits. For instance, without standardized benefits, some carriers might try to discourage groups with health problems by not offering prescription benefits. Another feature -- use of customer satisfaction and other quality measures -- would aid in evaluation of the carriers' performance in managing subscribers' care. A third feature -- the use of risk adjustment -- would help spread subscriber health risk evenly among carriers and thus encourage carriers to concentrate their energies on administrative efficiency and care management instead of on trying to avoid riskier subscribers.

As reported earlier, the use of standardized benefit packages was successfully implemented. However, the use of quality measures has been limited so far to ascertaining whether prices of Caroliance products are commensurate with the outside market and whether provider networks for HMO and PPO products are comprehensive enough to allow "excellent in-network health care," a relatively meaningless phrase in terms of judging carrier performance. Quality review within Caroliance has yet to reach the sophistication of comparing clinical outcomes and patient satisfaction by carrier. Caroliance is looking at developing an enrollee satisfaction survey, but has determined that enrollment currently is too small for survey results to be meaningful.

Also, no risk-adjustment mechanism has been put in place. Although Caroliance has attempted to identify an existing risk-adjustment methodology that will satisfy insurers' concerns, a suitable method has not been found. Reinsurance is one method of meeting the objective of risk adjustment. A voluntary public reinsurance pool is operating in North Carolina for small-group carriers, but many have been reluctant to participate (1). Caroliance has also explored the possibility of purchasing commercial reinsurance for policies sold through the alliances, but the idea was abandoned. It was reported that one carrier was willing to give a 6% discount on its premium price for policies covered by commercial reinsurance; however, the cost of the reinsurance and related commissions would have raised the cost to subscribers beyond the original premium level.

2. Carriers' concerns about adverse selection and rating policies

We were told that carriers in and out of the alliance system are generally uneasy about the smaller-size end of the small-group market, particularly about groups with fewer than 10 employees. The smallest groups are viewed as more risky because they are more prone to put off purchasing health insurance until an expensive health care need is imminent. As mentioned previously, Caroliance has a preponderance of very small groups (average group size of 2.2 employees). Our interview subjects indicated carriers' skepticism that the Caroliance system as now designed can ensure fair competition and protect them from the effects of adverse selection (i.e., receiving more than their fair share of sicker enrollees). Carriers are concerned that the Caroliance marketing program places too much emphasis on the availability of its guaranteed-issue products, thus attracting a disproportionate share of riskier groups. And even though products sold in the alliance are standardized, carriers worry about the failure to institute a viable form of risk adjustment. We were told that this led at least one carrier to scale back its participation.

Furthermore, carriers see the element of employee choice built into the system as a potential contributor to adverse selection, because a particular carrier might attract the sicker members of groups while the healthier members choose other carriers. For instance, an HMO plan might attract healthier members while those with more serious health conditions might be attracted to an indemnity plan so as to be able to continue to see a preferred physician. A Colorado purchasing cooperative has instituted "employer choice" to counteract this possibility. Under the employer-choice model, the employer chooses the type of plan (HMO, PPO, indemnity) to be offered employees, while employees choose from among the selections of that plan type (9).

As reported earlier, enrollment has been skewed toward products with guaranteed-issue rates, indicating that there has been adverse selection against the alliance system. Although we do not have information on loss ratios by carrier to verify this conclusion, it also is suggested by the fact that only 41% of enrollment in the select policies, which are comparable to popular products in the outside market, have sold at underwritten rates. Healthier groups are not being attracted to the alliances as frequently as less healthy groups, either because of rates or marketing practices.

Further analysis of the enrollment data indicates, too, that there may have been adverse selection within Caroliance, among its carriers. Table 6 indicates that Principal Health Care had more than its "fair share" of less healthy Caroliance lives. Eighty-six percent of its Caroliance lives had guaranteed-issue rates in 1997, while only 70% of total Caroliance lives were guaranteed issue. Blue Cross and Nationwide each had more "appropriate" shares of guaranteed-issue business at close to 70%. Mid-South, Kaiser and United Wisconsin Life had particularly low guaranteed-issue enrollment at 20%, 37% and 55%, respectively. Again, we do not have data to substantiate worse claims experience relative to premiums.

There also is some concern among carriers that Caroliance rating rules are not consistent with those in the outside market and thus give carriers less latitude in rating groups according to health risk than is allowed by law. North Carolina small-group reforms include rating restrictions that allow carriers to add or subtract 20% for a group's health status to the standard price of a given product for a group with particular case characteristics (age, gender, family composition, geographic location). Carriers also are allowed to increase renewal rates for each group by "trend" (a measure of marketwide cost increase) plus 15%. One carrier complained that Caroliance does not allow the 15% on top of trend for renewal increases. Caroliance says this characterization of the alliance policy is incorrect and results from a misunderstanding of its rules, which mirror those in the outside market. However, review of the 8/97 Caroliance Request for Participation, which was sent to prospective carriers, refers to "trend factors" with no mention made of an additional 15%. This may be simply a problem of definition; that is, Caroliance may assume that carriers know that trend can include the additional percentage. As discussed previously, however, there is an important difference in how Caroliance uses the 20% rating bands. Essentially it does not use the 20% discount off standard rates for the lowest-risk business. Interview subjects acknowledged that this policy probably has cost the alliances some healthier business because these groups can get lower rates on the outside from carriers taking advantage of the bottom end of the rating band.

A further difference noted between the Caroliance rating rules and outside practice is that Caroliance requires carriers to use blended family rates. The effect is that no distinctions are made for the number of children for which an employee wishes to buy coverage. Although this point was not brought out in our North Carolina interviews, agents in other states told us that such a policy can cause a hardship for financially strapped families that only have one child but must pay the same rate as those with multiple children.

3. Other perceived impediments to fair competition

While carriers have felt that they are vulnerable to adverse selection under the Caroliance system, Caroliance organizers have felt at a disadvantage under the rules of operation negotiated in the state legislature for the program. Research on the design of health insurance purchasing cooperatives has stressed the need for comparable operating rules inside and outside a cooperative so as not to introduce bias into the market (6). A key factor that made Caroliance stumble coming out of the blocks in 1995 was allowing it to sell only guaranteed-issue statutory plans, which reduced the alliances' appeal to healthier groups. While the problem of limited product selection was soon conceded and rectified by the addition of select products with underwritten rates, several subjects credited this initial mismatch between the products offered in Caroliance and the outside market with dampening early enrollment growth and contributing to the alliances' reputation as an insurer of high-risk business. However, even with a broader selection of products, there was still an imbalance in the two markets (prior to HIPAA) because all products sold through Caroliance were available on a guaranteed-issue basis, while only statutory products were required to be sold guaranteed issue in the outside market. Furthermore, there is no explicit legal prohibition in North Carolina to prevent carriers from charging more for the same product in the alliance system than outside.

Caroliance proponents have been concerned that the incentive structure for agents injects bias into the market against the alliances. Currently, carriers selling in the outside market are allowed to vary commissions by whether or not a group passes underwriting. As pointed out in our North Carolina report, commissions for medically-underwritten plans in the outside market are generally in the 8-10% range, while commissions for the statutory guaranteed-issue plans are about 5%. Caroliance rules suggest that commissions paid in the alliance system should be comparable to those paid outside. However, carriers are only given the options of varying commissions by benefit level (basic, standard or select) and product type (HMO, PPO, indemnity), not by rating type (underwritten or guaranteed issue). As a result, carriers have set flat commissions for all Caroliance business -- underwritten or not -- ranging from 5% to 7.5% in 1997 and 4% to 7.5% in 1996 (Table 7). The median Caroliance commission at the end of 1997 was 5%. One could argue that this is a reasonable carrier response because 70% of the covered lives in Caroliance are in guaranteed-issue plans. Agents with whom we spoke had the perception that commissions received for Caroliance sales are lower than for outside sales.

The Caroliance policy stems from the belief that carriers' paying lower commissions for guaranteed-issue products violates the spirit of the fair-marketing requirements of the small-group reform laws. However, some of our interview subjects asserted the right of carriers to vary commissions to reflect expected profitability. Also, the fact that guaranteed-issue premiums are higher means lower commission rates produce similar commission totals for agents. And lower commissions help to keep the cost down for consumers. The Department of Insurance believes it has no authority to regulate commissions and has declined to interpret fair marketing as requiring equal commissions for underwritten and guaranteed-issue sales. Nevertheless, this difference in commissions between the outside and alliance markets increases the incentive for agents to steer underwritten business to the outside market, to the possible detriment of the alliance risk pool. Such an observation is not meant to impugn the ethics of agents involved in the marketing of health insurance. The agents we talked with seemed sincere in putting their clients' interests first. And our data indicate that the best prices for clients have generally been found outside the alliance. However, it would be naive to think that commission levels do not affect to some undetermined degree where agents focus their efforts and attention. It is interesting to note that Nationwide, which paid the highest Caroliance commissions in 1996 and 1997, at a rate near the outside market's rate for underwritten sales, had by far the highest percentage of enrolled lives in the alliance program in mid-1997 -- 73% overall, 71% underwritten, 75% guaranteed-issue (Table 8). These numbers suggest a relationship between agent commissions and health insurance sales that should be studied further by policymakers seeking to improve market competition for the benefit of insurance purchasers. Whether or not North Carolina policy should be changed to mandate equal commissions for guaranteed-issue and underwritten business remains to be decided.

Beyond the matter of commission levels, some interview subjects reported that field underwriting is a problem for Caroliance. As noted in our North Carolina report, field underwriting involves agents' steering higher-risk applicants away from particular carriers in response to subtle or sometimes not-so-subtle pressure from carriers. Field underwriting violates fair-marketing laws. Although all carriers in the pre-HIPAA small-group market were required by law to offer the guaranteed-issue statutory plans to groups that could not pass medical underwriting, some carriers discouraged agents from submitting these applications. The agents, instead, referred the groups to other carriers who were meeting their responsibilities under the guaranteed-issue reforms. Both insurers and administrators are concerned that Caroliance has received a disproportionate number of guaranteed-issue groups, in part because some carriers in the outside market have not acted responsibly in regard to this higher-risk business.

One final point that our subjects made about market fairness is that many independent agents in North Carolina view the alliance system as an inappropriate competitor with them in health insurance sales. Many agents are wary of any government involvement in the health insurance market and feel that Caroliance is competing with them, using public funds, as an intermediary between carriers and purchasers. In addition, some agents are suspicious that if Caroliance were to reach a large enough enrollment, agents would be cut from the sales loop and replaced with direct marketing by the alliances. The idea of direct marketing was considered when Caroliance was being developed, but was dropped in response to a hostile reaction from agents' trade groups. So even though Caroliance is now required by law to sell its products through agents, we were told that bad feelings still permeate the message about the alliance program that some agents give purchasers.
 

F. HIPAA

The Health Insurance Portability and Accountability Act passed by the federal government in 1996 took effect July 1, 1997. As a result, all products sold by carriers in the small-group market must be made available on a guaranteed-issue basis for groups size 2-50. In our second round of North Carolina interviews, we asked what effect HIPAA has had in the short time it has been in force. Some subjects in the first round of interviews had predicted that HIPAA would help level the playing field for Caroliance by making all products guaranteed-issue. Subjects report that HIPAA has not benefited Caroliance but has exacerbated the problem of field underwriting and increased the resulting flow of business to Caroliance that carriers in the outside market deem unappealing. Some carriers that have traditionally been wary of the smallest groups but were only required to sell them the statutory plans, resist having to sell any product to groups that they perceive to be of higher risk. Thus, we were told that some carriers are discouraging agents from sending them these smallest groups by lowering commissions for these sales (one carrier lowered commissions to 3% for groups under 15) and by encouraging agents to send the groups elsewhere, even though such instructions are prohibited under the law. Caroliance feels that it is receiving more of this "last resort" business and points to a reduction in its average group size since HIPAA went into effect.

Some subjects commented that with HIPAA, there is little need for Caroliance because all products in the market are now available to any small group of 2-50 on a guaranteed-issue basis, and self-employed persons still have access to the statutory plans from any carrier. The observation also was made that the small number of Caroliance carriers effectively removes employee choice as a reason for its existence. Nevertheless, Caroliance supporters still point to its pooling power as an important attribute to justify its existence even under HIPAA. Caroliance also is said to be one of only two remaining sources of underwritten coverage for groups of one, the other being Blue Cross in the outside market.
 
 

1. Hall, Mark A. An Evaluation of North Carolina's Small-Group Health Insurance Reform Laws, Wake Forest University School of Medicine, Winston-Salem, NC, February 1999.

2. A small group was defined in 1992 and 1993 as having 3-25 employees, in 1994 as having 2-49 employees, in 1995- June 1997 as having 1-49 employees, and since July 1997 as having 1-50 employees.

3. State Health Plan Purchasing Alliance Board. Report to the Joint Legislative Commission on Government Operations, Raleigh, NC, December 31, 1995.

4. The study states are Colorado, Florida, Iowa, New York, North Carolina, Ohio and Vermont.

5. Kaiser Family Foundation and the Commonwealth Fund. The Kaiser/Commonwealth 1997 National Survey of Health Insurance, December 1997.

6. Buchmueller, Thomas C. Managed Competition in California's Small-Group Insurance Market, Health Affairs, Volume 16, Number 2, March/April 1997, 218-228.

7. Wagner, Reenie. Analysis of March Current Population Survey, 1992-1997, October 1998.

8. Wrightson, Charles W. HMO Rate Setting and Financial Strategy, Ann Arbor: Health Administration Press, 1990, 245-294.

9. Curtis, Rick and Kevin Haugh. Small-Employer Co-ops Pick Up Speed, Business and Health, July 1996, 29-36.

10. Wicks, Elliot K. Aggressive Regulator or Passive Price-Taker: What Role Should HIPCs Play?, Journal of American Health Policy, July/August 1993, 21-25.

11. Hall, Mark A. Reforming Private Health Insurance, Washington, D.C.: The American Enterprise Institute Press, 1994, 53-54.