Author: Carlton Harker, FSA, MAAA

Statutory Background

The COBRA law has special rules for determining COBRA premiums for self-funded health care plans. The premiums shall be a reasonable estimate of the cost of providing coverage for similarly situated beneficiaries determined on an actuarial basis taking into account factors that shall be prescribed by regulations. The COBRA law permits the plan administrator to choose between one of two methods of determining COBRA premiums for a self-funded plan.

Actuarially Determined Method

This is a method by which an actuary would perform such computation taking into account factors to be prescribed by regulations. By actuary, we would presume the law means a qualified actuary (minimally being a member of the American Academy of Actuaries). The regulations, directed by the law to be issued, have never been so issued. COBRA commentators are universally agreed that the law did not mean that a non-actuary could compute actuarially determined COBRA premiums by following what would be asserted to be actuarial methodology.

Past Cost Method

This method would project past claims forward to the upcoming plan year and spread such claims cost among plan participants. This past cost method must be modified where there are significant changes in benefits, eligibility, census, etc., between the old and the upcoming plan year. Inflation factors used in the projections are defined in COBRA law. Because this method is so simple, it is not further discussed. While not recognized by the COBRA law, the so-called Fully Insured Equivalent Method is further discussed along with the why the writer believes this Method carries with it some potential problems for the plan administrator.

Discussion of Actuarial Determination Method

Since actuaries are trained and qualified to project claims experience, based upon the past, into the future, by model building, the COBRA law permits such to be done by an actuary in computing COBRA premiums for the upcoming plan year. The actuary would, in the process of model building contemplate a large number of factors. Examples are as follows:

  • Benefit content
  • Benefit modifications
  • Monetary inflation
  • Claim reserves
  • Census trends, family content, etc.
  • Plan sponsor’s fixed costs
  • Direct: Stop-loss premiums, administration fees, etc.
  • Internal: Employer’s inside costs which are plan-related
  • Number/nature of shockers or lasered participants
  • Monte Carlo simulations to measure likelihood of stop-loss terms being too liberal or conservative
  • Slippage between promised and delivered stop-loss benefits
  • Complexities with benefit design
  • High or low plan options
  • Non-core benefits (dental, vision, e.g.)
  • Separate COBRA for Rx
  • Multiple tiers
  • Advantages of Actuarial Determination. A few of the advantages or by-products to the plan sponsor of having such determination are as follows:
  • Funding factors developed
  • Obtaining estimates of claim reserves as a by-product
  • Benefit content comparisons (high-low, e.g.)
  • Pricing ancillary or non-core benefits (dental, vision, disability)
  • Pricing of managed care programs
  • Stop-loss reviews, particularly with Monte Carlo simulations
  • Miscellaneous by-products (participant contributions, e.g.)
  • Avoidance of such premiums being challenged by regulators or attorneys
  • Being able to vary by age/geography (so long as plan is appropriately amended).

Discussion of Fully Insured Equivalent Method

 From the outset of COBRA, the practice of some has been to base COBRA premiums on the terms of stop-loss (funding factors and spec/agg premiums). Some, but far from all, of those who use this method contemplate the terms of stop-loss; examples include:

  • Aggregate factors may have a 15%-20% or 25% corridor
  • Terms may be 12/12, paid, etc. with some covered persons being lasered.

Often, COBRA premiums represent the worst case scenario of the plan sponsor whereby COBRA premiums are, as a consequence, significantly overstated. A very strong case may be made against using the fully insured equivalent method for these reasons:

  • The COBRA law does not recognize it as an acceptable method.
  • Stop-loss is not part of the plan in that the coverage has the plan sponsor (employer nearly always) as the applicant, owner, payer and beneficiary.
  • The method has been challenged successfully in the past by attorneys representing former employees in alleged wrongful employment termination case particularly as regards to calling of the aggregate corridor a plan expense.
  • Employers might well be concerned in those instances where the equivalency premiums are less than those which were actuarially-determined.
  • Total reliance that stop-loss paid benefits will match promised benefits is questionable. Often, stop-loss carriers pay less than what was expected. One simply needs to review some aggregate audits to understand how this slippage may occur.
  • Insurers do fall on bad financial times and fail; it is only a matter of time before an insurer offering stop-loss will do so. One stop-loss carrier, Legion Insurance Company, has gone into receivership. The statutory safety net known as the life and health guaranty association coverage will, in most instances, not be available to protect the employer where such failure occur.
  • Viewing the self-funded plan as a miniature insurance company with no capital or surplus, the employer is well advised to examine an actuarially-determined model for the upcoming plan year for basic financial prudency reasons.

Special Problems

Typically there are significant challenges involved with COBRA calculations:

  1. The experience is composite but the plan has high-low benefit structure. In this case, a benefit content analysis is needed.
  2. Fixed costs are provided two-tier but the COBRA premiums are to be four-tier
  3. Data is composite as regards core and non-core benefits but such COBRA premiums are to be shown separately.
  4. Plan sponsor wishes COBRA premiums to vary by age or geography area.
  5. Plan sponsor wishes to use COBRA for reasons beyond COBRA such as preparing 1099s to highly compensated where benefits are discriminatory; or as a basis for determining participant contributions; or for funding purposes; or for intercorporate expense transfers.
  6. Stop-loss is specific-only and experience data is limited or the plan is a new plan with no prior claims experience.
  7. Managed care arrangements. What are out-of network COBRA premiums, e.g.?
  8. May COBRA premiums vary by the financial experience of sub-groups? Unless such sub-groups are separate plans, the response is no.

High-low Plans. Consider a situation where the claims experience is composite of 300 participants but there is a Plan Option A (100 participants) and a Plan Option B (200 participants). Such situation requires a benefit content analysis by where the economic value between A and B are measured; A has a cost index of 100 while B has a cost index of .88, e.g.

Premiums Involving Multiple Tiers. Experience is typically maintained composite; COBRA premiums are determined by tier (such tiers usually follow the bases of participant contributions). What is needed is a cost index of each tier. For example:

·        Two-Tier I; 2.4 F
·        Three-Tier I; 1.8 P+1; 2.5 F
·        Four-Tier I; 1.6 P/1; 1.8 P/S; 2.6 F
·        Five-Tier I; 1.6 P/C; 1.8 P/C; 2.1 P/Children; 2.7 F

There are several observations which are relevant to these COBRA premium tiering challenges:

  • Such tier as P/Child or Children is illogical in that the initial reason for P/C tier was to help the single mother; since, the number of P/Children family units has grown dramatically. A P/Children group is a family and should be so treated.
  • A tier such as 1P; 2.2 P/S is illogical in that the participant and spouse may each elect COBRA as individual beneficiaries making the 2.2 premium pointless or ineffectual.
  • Having the I premium as high as practical relative the family, e.g., is logical from a risk management standpoint when considering the ability of the COBRA participant to select against the plan by electing COBRA for sick child only but not on the other family members in good health.

Data is Composite but Core and Non-Core Premiums Needed. In such instances it is mandatory that some serious attempt to split out the core and non-core claims be made. When not possible, an estimate is the only option. Dental, e.g., is typically 8-12% of medical.

COBRA Premiums by Age and Geography. Nothing prevents a plan from considering age and geography as a factor in funding policy so long as such practice is formalized in the plan document. Once the plan document establishes age and/or geographic variations, such must, under the similarly situated rules, apply such to COBRA premiums.

Other USES for COBRA Premiums. For various functions of the plan sponsor. Having COBRA premiums actuarially-determined may prove useful. For example:

  • Where IRS Form 1099s are to be given to certain employees representing the economic value of their health care benefits, such COBRA premiums are used in the completion thereof.
  • COBRA’s are useful in setting participant contributions.
  • The claims-only portion of the COBRA premium should be used when establishing funding levels.

Limited Claims Experience. For a new plan or a plan for which claims experience is not available or inappropriate, such COBRA premiums must be estimated on best evidence basis.

Managed Care Arrangements. Where plan benefits cost indices are 100 for in-network and 80 for out-of-network, how should COBRA premiums be shown. The preferred way is to have the COBRA premium calculation assume that the COBRA beneficiary has such network option whether residing in the plan’s geographic area or out of such area as a move away.

Financial Experience of Group.   Where two divisions share a common plan (defined as one plan name, sponsor DOL number) COBRA premiums for both divisions must be the same regardless of each division’s different claims experience. See Draper V. Baker Hughes. Inc. 892 F.Supp.1287 (E.D. Calif. 1996).

Adjustments Often omitted.

There are four COBRA premium determination adjustments which are usually omitted in error. Each of the four adjustments should be made but only if such adjustments are based upon some actuarially-supportable bases or adjustments. Such adjustments are these:

  • Employer Cost of Plan Administration.
  • Seasonal Variation of Claims.
  • Aggregating Specific.
  • Lasered Participants.

Requisite Features of the Actuarial Methodology

 These are the features of the actuarial methodology which set it apart from any other methodology:

  1. Where the stop-loss has the employer as the applicant-ower-payer-beneficiary no reference and/or reliance thereto should be made except as regards fixed costs of such stop-loss.
  2. Modeling of the future is the proper methodology to be followed.
  3. Where multiple plan options are covered by the same aggregate stop-loss terms, a benefit content analysis, actuarially made, should measure the option cost variations. Such variations are reflected in the COBRA premiums.
  4. Special analyses, based upon actuarial experience are needed for (a) multiple tiers and none-core benefits (dental, vision, etc.)
  5. Special analyses are increasingly being neede for the Rx benefits which may be offered on a pick-choose basis.
  6. Where past experience is available and meaningful, COBRA premiums should measure the cost of these three items:
  • Lasered participants.
    • Aggregating specific.
    • Employer’s internal plan costs.
  1. When partial-year projections are made, the projections should adjust for seasonal variations in the paid claims.