Sec L Anti-selection

Question 1

Approaches used for modeling antiselection

Answer 1

1. Partition models - the population is broken into subsets (e.g., healthy vs unhealthy), which are modeled separately. Individuals are ranked by cost, and a line is drawn to define the subsets
   a) Drawing the line in a Cumulative Antiselection Theory (CAST) model - the cutoff line is chosen so that the ratio of average claim costs between the unhealthy and healthy groups is a chosen multiple (such as 5 or 10)
   b) Drawing the line in a Minnesota Antiselection Model - this model defined boundary conditions on the antiselection which might occur in a specific situation, in order to estimate the antiselection expected under different scenarios
   c) Drawing the line in internal antiselection - a modified CAST model can be used, adding a decrement to reflect the potential of the insured to change benefit plans
2. Deterministic vs stochastic models - deterministic models typically project results based on expected values. But a true picture of the future requires distributions of potential values, which are provided by stochastic models
3. Markox processes - in most cases, a two-subset partition model is sufficient. But if more partitions are needed, a Markov chain can be created to determine the population distribution in future time periods

Question 2

Situations in which the CAST model does not work well

Answer 2

1. In the first 3-4 durations, when the impact of underwriting wear off overwhelms the CAST effects. The solution is to apply additional underwriting selection factors
2. In later durations, where only a fraction of the original populations remains. The solution is to choose a higher value of k₂, and recalibrate the model.
3. At all durations, when a rate spiral is sever and volatile. The projections formulas may need stronger terms to fit this type of situation, such as:

Shock lapse = [Rate increase - trend] / [(Rate increase - trend) + (1 + trend) / EF]

EF = elasticity factor, which measures the ability and willingness of the population to change coverage after a rate increase (e.g., may be 1.3 for healthy lives and 0.8 for unhealthy lives)

Question 3

Requirements that mitigate antiselection between exchange plans and off-exchange plans

Answer 3

1. Insurers must include all ACA-compliant policies in a single risk pool. So identical plans must have identical rates on and off the exchanges
2. Risk adjustment will be applied to even out risk between insurers and between the on- and off-exchange portions of the risk pool
3. Insurers must pay the same commissions to brokers and agents on and off exchanges
4. The exchange fee (3.5% of premium for each exchange policy) must be spread across the entire single risk pool, including off-exchange policies
5. Carriers participating in exchanges must offer at least one gold and one silver level plan on the exchange
6. Carriers are prohibited from marketing practices intended to discourage unhealthy individuals from signing up
7. Open enrollment periods are identical on and off the exchange

Question 4

Traditional techniques for controlling antiselection that are prohibited by the ACA

Answer 4

1. Underwriting, including offering alternative coverage or denying coverage
2. Health status rating
3. Pre-existing condition exclusions
4. Exclusionary riders
5. Lifetime or annual dollar limits
6. Limiting benefit coverage or imposing very high cost sharing designed to attract healthier risks
7. Rescissions, except in cases of fraud or intentional misrepresentation
8. Marketing practices that discourage unhealthy risks from signing up

Question 5

Assumptions needed for projecting values in antiselection models

Answer 5

1. Trends in claim costs (from aging, duration, secular trend, etc.)
2. Lapsation, separately for each sub-population (often varies by size of rate increase)
3. Movement between populations, often expressed as a net movement from healthy to unhealthy
4. The time value of money (interest)
5. Premium rate increases

Question 6

Mechanisms for controlling external antiselection

Answer 6

1. Individual underwriting before issue - includes initial screening of applicants by the agent
2. Pre-existing condition limitations
3. Requiring an enrollment mechanism that doesn’t permit antiselection (such as minimum participation percentages for associations)

Question 7

Types of antiselection

Answer 7

1. External antiselection - occurs as the person is first becoming insured. Those with expensive health conditions will seek insurance
2. Internal antiselection - occurs while the person is insured. When given the opportunity, healthy individuals will be more likely to decrease coverage while unhealthy individuals will tend to increase coverage
   a) Buy-down effect - upon receiving a rate increase, some policyholders switch to lower cost plans, so the actual premium increase will be less than expected
   b) Premium leakage - unhealthy individuals are less likely to buy down their benefits. So the claim cost reduction is less than the premium reduction and not enough premium is collected
3. Durational (cumulative) antiselection - occurs as people make decisions about whether to end coverage. Higher cost insureds tend to keep their coverage in force longer because they are:
   a) Less likely to be able to find coverage elsewhere (although this is no longer true in markets affected by the ACA)
   b) Less likely to be willing to become uninsured
   c) Emotionally less willing to change their insurance coverage