Recent studies of health insurance regulation all have concluded that state regulation of insurance issue, renewal and rating in general either reduces health insurance coverage or, on net, has no impact on coverage. Some of these studies have found that regulation may change the risk distribution of the insured population, raising coverage among high-risk groups and individuals but lowering coverage among low-risk groups and individuals, with no significant impact on overall coverage. All of this literature presumes that, by forcing insurers to accept and pool risk that they would otherwise deny or segment into high-risk rate classes, regulation raises insurer cost and, in turn, insurance prices. Low-risk groups and individuals are thus discouraged from buying coverage, and the rate of private insurance coverage declines. Implicitly, all earlier studies have assumed that insurance markets are competitive and, therefore, that higher price is an inevitable effect of regulation. Rarely are these studies able to observe price directly.10
This paper considers the impact of insurance regulation on the structure of health insurance markets. We do not presume that health insurance markets are competitive. Indeed, both the group and individual health insurance markets in every state are highly concentrated: typically, a few large insurers hold most of the market, and most insurers hold very little. Some segments of the market may be internally competitive, but the skewness of these markets in all states suggests that it is not competitive in the economist’s sense. Instead, the largest insurers (and even some small insurers) probably enjoy a degree of monopoly power, and therefore have some discretion about pricing. Smaller insurers (arguably operating with increasing returns to scale) may respond differently to regulation than larger insurers (with relatively constant returns to scale).
We test a number of regulatory variables against various measures of market structure, separately for the states’ group and the individual health insurance markets. Measures of market structure include the number of insurers selling coverage, the market share of different insurer types, and market concentration (measured as a Herfindahl index and large-firm market share). While we find occasionally strong year and state effects (that is, significant changes occurred in market structure between 1995 and 1997, and especially so in some states), much or all of the changes that occurred in market structure were independent of state regulation.
HIPAA requires guaranteed issue of all products in small-group health insurance markets, and we did find some effect of all-product guaranteed issue on market structure in states that had implemented these laws before HIPAA’s effective date, and the direction of the impact is of interest. Controlling for other factors that would influence the number of insurers in the market, states with all-product guaranteed issue (as HIPAA requires) had more insurers selling coverage — and less market concentration among the largest insurers — than states which required guaranteed issue of some products or did not require guaranteed issue at all. However, guaranteed renewal always accompanied all- product guaranteed issue. The combined effect of both all-product guaranteed issue and guaranteed renewal was a group market with both significantly more insurers and somewhat greater market concentration among the largest insurers.
In markets with no guaranteed issue or only guaranteed issue of some products, guaranteed renewal was associated with significantly greater market concentration. However, these more modest forms of regulation produced no significant change in the number of insurers in the group market.
In summary, it would appear that, if HIPAA’s all-product guaranteed issue provision has had any impact on the small group market, it probably has simplified the market, encouraging more insurers to remain in the market. Ironically, the complexity of markets with more modest reforms (some-product guaranteed issue or simply guaranteed renewal, such as many states had adopted prior to HIPAA) appears to have had dampened competition relative to HIPAA’s fuller reform by allowing insurance markets to remain much more complex.
While HIPAA’s guaranteed issue protections in the individual market extend only to the relatively few “HIPAA eligibles”, we also considered the impact of full guaranteed issue in the individual market, and found some significant effects among the few states that adopted guaranteed issue between 1995 and 1997. Specifically, our results suggest that guaranteed issue of only some products in the individual market appeared to favor greater commercial insurers, increasing their market share and driving less market concentration among the largest insurers. These results were highly significant. Conversely, guaranteed issue of all products significantly increased the largest insurers’ market share (and, with less significance, also drove lower commercial insurer market share). Whether these effects are seen as favorable or worrisome depends on the relative importance of economies of scale in production versus competition in determining the price of insurance. We return to this problem below.
HIPAA also limited the preexisting condition periods that group insurers may impose. We found that shortened waiting periods for coverage of preexisting conditions in the group market drove significantly greater concentration, causing some small group insurers to abandon the market (or to merge) and enabling the largest insurers to gain market share. However, the magnitude of these effects was small. In the individual health insurance market (despite insurers’ greater concern about adverse selection in this market), shortened waiting periods for coverage of preexisting conditions had no measurable impact on any measure of market structure.
HIPAA did not address insurers’ rating practices in either the group or individual health insurance markets. States that limit insurer rating – typically prohibiting insurers from using health status or age as a rating factor, sometimes establishing a composite rate band – have responded to their own policy priorities and political realities.
In the group market, the impacts of rate bands are complex: in states that constrain overall rate variation, the impact on the number of insurers writing coverage in the state was positive and, depending on the model’s specification, moderately to highly significant. However, nearly all states with a composite rate band also limit health rating, and limits on health rating fully offset the impact of the composite band – resulting in a negligible net impact on the number of insurers. States that limit health rating only (without constraints on age rating or on composite rates) had many fewer insurers in the group market, but we found no significant difference in market concentration in those states. Narrower constraints on age rating generally had no significant impact on any measure of market structure.
In the individual market, narrower limits on health rating appeared only to reduce commercial market share (favoring BCBS plans). As in the group market, narrower constraints on age rating had no significant impact on any measure of market structure.
Reviewing the impacts of regulation in markets prior to implementation of HIPAA, it is unclear whether the effects of regulation are in themselves or on balance positive or negative. Whether some forms of regulation produce lower prices and greater coverage depends on the relative strength of their intermediate effects on the number of insurers and on market concentration. In a declining-cost industry such as health insurance, fewer insurers would (all else equal) result in lower-cost production. In a competitive market, lower prices would result. However, in a monopolistic market, the loss of insurers may further dampen price competition and foster higher prices. While available research is not directly helpful in comparing these effects, the emerging literature on insurance coverage suggests that the upward price effect of greater monopoly power – absent regulatory constraints on monopoly pricing – may have outweighed the downward price effect of more efficient levels of production in response to market regulation.
Finally, it is important to understand that the empirical results presented in this paper, as well as in all other studies of state regulation to date, rest on very few observations of regulatory change during the study period. In effect, change in as few as one or two states drive all of the impacts that have been measured by any study to date. Despite the general strength of our analytic approach, the external validity of these results, based on change in only a few states over the study period, is suspect. It is essential that research of this type be extended to examine later years, after many states enacted new regulation or adjusted existing regulation to comply with HIPAA’s provisions, to confirm these results.