Premium Affordability, Competition, and Choice in the Health Insurance Marketplace, 2014. IV. Competition, Other Market Factors, and Different Measures of Marketplace Premiums

06/18/2014

Section IV Highlights

The Association Between Competition, Other Market Factors, and All Marketplace Premiums

  • The absolute number of issuers within a rating area did not, on average, have a significant association with the average premium for 27-year-olds for all plans by metal level (bronze, silver, gold, and platinum). In part, this difference may be due to markets with ore issuers exhibiting greater variability in premiums—that is, these markets had both higher and lower premiums within each metal level.
  • Consumers had a wider choice of plan types in areas with more issuers. The variability in premiums associated with a greater number of issuers was in part related to these rating areas being more likely to have the full range of plan types including; CO-OPs, HMOs and plans issued by insurers offering Medicaid plans in the market prior to Marketplace implementation.
  • CO-OPs and HMOs exhibited significantly lower premiums than other plan types.
  • Areas that had more concentrated hospital markets (higher HHI) exhibited higher average premiums.
  • A higher percentage of established issuers in a state was associated with lower premiums at each metal level.

In the preceding analyses, we examined factors that were associated with variation in one measure of market premiums (the second-lowest cost silver plan) across rating areas. In order to conduct a more complete analysis, we examined factors that affected the full range of Marketplace premiums in all four metal levels. In our initial models of the second-lowest cost silver plan premium, we also examined similar statistical models, replacing the second-lowest cost silver plan premium with both the average and median silver plan premiums for the rating area (not displayed in this Brief). In contrast to the results for the  second-lowest cost silver plan premium, the number of issuers did not have an effect on either the mean or median silver plan premium. 33

To further examine these findings, we conducted several other analyses. Specifically, we examine the extent to which a greater number of issuers results in greater variation in plan types being offered in the rating area. A greater dispersion of premiums could mean that the lower premiums offered are offset, on average, by higher premium plans offered for particular plan types.

First, we examined models at the individual plan level rather than for the rating area. Hierarchal linear modeling techniques were utilized to examine the effect of market factors and competition on all premiums across all four metal levels (bronze, silver, gold and platinum).34 We present in the discussion and Table 8 the results related to premiums for 27-year-old individuals.

In addition to the results for the number of issuers, there are several other results of interest. These models included indicators for whether the plan was a PPO, HMO, CO-OP or other type of plan.35 As displayed in Table 8, the results suggest that premiums for HMO plans were lower on average than PPO, point of service (POS) and exclusive provider organizations (EPOs). In addition, CO-OP plans tended to have lower premiums than non-CO-OP plans within areas, which is consistent with the intent of their creation— to offer competitive health plans. Established issuers were also associated with lower bronze, silver, and gold premiums, but higher platinum plan premiums. As discussed previously, such issuers may have greater knowledge of the area or have established provider markets that allow them to charge a lower premium; or, on the other hand, they may have a loyal customer base that is willing to accept higher premiums. These results indicate that both of these dynamics may be in play for plans at varying metal levels.

Another important factor was the measure of hospital market concentration (HHI). While this variable was not associated with premium levels for the second-lowest cost silver plan premium,  it does demonstrate a positive and statistically significant effect on the full range of premiums across all metal levels. This result supports the view that insurers likely have less price negotiating leverage in more concentrated hospital markets, resulting in higher premiums.

In the plan-level models, the number of issuers does not have a significant effect on premiums in any metal level. Thus, it appears that a greater number of issuers is associated with lower benchmark (second-lowest cost silver) plan premiums being offered, but is not related to the average of all premiums offered. Further statistical analyses offer a plausible explanation for this finding. We find that the variability in premiums increases with the number of issuers in a rating area. As displayed in Table 9, a common measure of variability—the coefficient of variation—increases with the number of issuers. This indicates that as the number of issuers increases, the number of plans offered also increases which leads to a greater dispersion of premiums. As the regression results in Table 10 demonstrate, the association between number of issuers and the coefficient of variation is statistically significant after removing the variation that might be attributable to other market factors. So, even while controlling for other factors that may contribute to the dispersion in premiums, the association between the number of issuers and premium dispersion remains.

The premium variation is in turn at least partly attributable to the plan types offered. The rating areas with more issuers are more likely to offer HMOs and CO-OP plans than those with only one issuer, while only areas that have four or more issuers offer the full range of plan types—corresponding to the dispersion of a full-range of premium rates (see Table 11). Consumers can expect more choice in plan types in markets with robust competition as measured by number of issuers participating.

TABLE 8 Hierarchical Linear Regression Model Results, Premiums for 27-Year-Olds by Plan and Metal Level, 2014 Health Insurance Marketplace

  Log of the Average Premiums for 27-Year-Olds By Metal Level
Bronze (n=5,721) Silver (n=6,896) Gold (n=5,221) Platinum (n=1,523)
Market Characteristics by Plan Coefficient(P-Value) Coefficient (P-Value) Coefficient (P-Value) Coefficient (P-Value)
Number of Health Insurance Issuers 0.00 (0.97) -0.00 (0.36) -0.00 (0.27) 0.00 (0.46)
PPO Plan (1,0) 0.11 (<0.001) 0.09 (<0.001) 0.10 (<0.001) 0.05 (0.003)
HMO Plan (1,0) -0.04 (<0.001) -0.05 (<0.001) -0.06 (<0.001) -0.04 (0.001)
FFM State (1,0) 0.02 (0.76) 0.01 (0.93) -0.02 (0.70) 0.01 (0.95)
Medicaid Expansion State (1,0) -0.00 (0.96) -0.00 (0.93) -0.02 (0.73) -0.00 (0.96)
Full Community Rating State (1,0) 0.52 (<0.001) 0.50 (<0.001) 0.49 (<0.001) 0.58 (<0.001)
Established Issuers (1,0) -0.08 (<0.001) -0.08 (<0.001) -0.07 (<0.001) 0.06 (0.001)
Issuers Offering Medicaid Plans (1,0) 0.06 (0.001) 0.02 (0.17) 0.02 (0.27) -0.00 (0.95)
Issuers that are CO-OPs (1,0) -0.09 (<0.001) -0.05 (0.001) -0.07 (<0.001) -0.08 (0.02)
Log of Hospital Market Concentration (HHI) 0.02 (0.04) 0.02 (0.05) 0.02 (0.01) 0.05 (<0.001)
Constant -0.19 (0.90) 0.75 (0.59) 0.43 (0.76) -1.38 (0.58)
Wald (Х2) 776.77 (<0.001) 901.93 (<0.001) 991.61 (<0.001) 118.94 (<0.001)
Log Likelihood 5,082.93 6,674.71 4,714.07 1,247.09

Source: ASPE computations of plan and premium data from the following publicly available sources: Healthcare.gov, state rate filings (where available), and State-based Marketplace websites.
*Other market characteristics for a rating area include the percent of the population that is uninsured, log of state health care expenditures, and the log of the population density.
NOTE: Other model specifications included: 1) excluding the log of state health care expenditures and 2) excluding the log of state health care expenditures and replacing it with the log of the average state small group premium. Results were consistent across specifications.

 

TABLE 9 Coefficient of Variation of Silver Plan Premiums for 27-Year-Olds, by Rating Area and Number of Issuers, 2014 Health Insurance Marketplace

Number of Issuers Coefficient of Variation
Only 1 Issuer 0.06
2 or 3 Issuers 0.09
4 to 6 Issuers 0.12
7 to 11 Issuers 0.15

Source: ASPE computations of plan and premium data from the following publicly available sources: Healthcare.gov, state rate filings (where available), and State-based Marketplace websites.

 

TABLE 10 Coefficient of Variation (CV) Regression Analysis Including the Number of Issuers and Other Market Characteristics,* Premiums for 27-Year-Olds by Rating Area and Metal Level, 2014 Health Insurance Marketplace

  CV Based on Premiums for 27-Year-Olds by Metal Level
Bronze Silver Gold Platinum
Market Characteristics by Rating Area Coefficient (P-Value) Coefficient (P-Value) Coefficient (P-Value) Coefficient (P-Value)
Number of Health Insurance Issuers 0.01 (<0.001) 0.01 (<0.001) 0.01 (<0.001) 0.01 (<0.001)
Number of Observations (Rating Areas) 483 494 494 205
F-Statistic 12.83 (<0.001) 9.42 (<0.001) 9.22 (<0.001) 26.11 (<0.001)
R2 0.32 0.31 0.32 0.49

Source: ASPE computations of plan and premium data from the following publicly available sources: Healthcare.gov, state rate filings (where available), and State-based Marketplace websites.
*Other market characteristics for a rating area include established issuers as a proportion of all issuers, issuers offering Medicaid plans in the rating area prior to the implementation of the Marketplace as a proportion of all issuers, indicator that a CO-OP has plans available in the rating area, indicator of a Federally-facilitated Marketplace state, indicator of a Medicaid expansion state, indicator of a full-community rating state, log of the hospital HHI, the percent of the population that is uninsured, log of state health care expenditures, and the log of the population density.

 

TABLE 11 The Percent of Rating Areas with at Least One Silver Plan of Selected Types, by Number of Issuers, 2014 Health Insurance Marketplace

  The percent of rating areas with at least one silver plan that is one of the following types:
HMO CO-OP Medicaid* HMO, CO-OP, Medicaid
Any Number of Issuers 55% 32% 22% 5%
Only 1 Issuer 21% 0% 2% 0%
2 or 3 Issuers 46% 31% 13% 0%
4 to 6 Issuers 75% 40% 26% 8%
7 to 11 Issuers 62% 45% 77% 21%

Source: ASPE computations of plan and premium data from the following publicly available sources: Healthcare.gov, state rate filings (where available), and State-based Marketplace websites.
*These are plans provided by issuers that were offering only Medicaid plans in the market prior to the implementation of the Marketplace.


33 Our results are comparable to those from Dafny, Gruber and Ody (2014) who found a relatively consistent relationship between their measure of competition and the second-lowest cost, mean, and median silver plan premiums in a rating area. While Dafny et. al. use the change in issuer HHI if United Healthcare had entered the 2014 Marketplace to test the effect of competition based on pre-Affordable Care Act shares of the individual market, in our analysis, we incorporate both the number of issuers and the proportion of issuers that had been established in the individual market prior to the implementation of the Marketplace. While we did not find a statistically significant association between the number of issuers and the mean and median silver plan premiums, it is notable that we find that established issuers generally offered lower premiums.

34  Hierarchical linear modeling (HLM) regression techniques are designed to deal with clustered or grouped data in which analytic units are naturally nested or grouped within other units of interest. For example, the number of Marketplace plans nested within issuer nested within rating area nested within state. Hierarchal linear models recognize the existence of such data hierarchies by allowing for residual components at each level in the hierarchy. In analyzing premiums of Marketplace plans being offered by issuers within a rating area, interest centers on the effects of plans, issuers, rating area, and state characteristics.

35  The reference group for the PPO and HMO variables consists of plans that are POS or EPO. For the CO-OP variable, the reference group is established issuers, including all commercial and Medicare plans.

View full report

Preview
Download

"2014MktPlacePremBrf.pdf" (pdf, 346.96Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®