Actuaries are guesers. They are systemic guessers, but they make guesses none the less. The larger the population and the richer the data history, the closer to reality their guesses tend to be as long as we can assume no massive step functions in the input-outcome matrixes. Actuaries really don’t like making guesses when they are working with mostly unknowns or at best extraordinarily wide parameters of plausible values as they know that their error bars will be significant.
The Exchange has been a source of night terrors for actuaries for eighteen months now. It was the ultimate in vaguely defined parameters. No single company knew what other companies were doing, no one knew exactly how people would choose plans, no one knew who exactly was in the eligible population and more importantly who would be in the sign-up generation. No one had a representative claims history.
Companies made massive efforts to get some structure to their guesses. For instance, my company participated in a consortium that interviewed 100,000 people with in-depth focus on several thousand people in order to get a feel for what potential Exchange customers wanted and what types of medical needs they would have. My company had a role play team that attempted to game out various strategies we could use as well as our competitors. Another company conducted over a hundred focus groups in a single market. Data miners created proxy consumers from claims data. But the actuaries and finance people priced 2014 Exchange products with at least severe glaucoma in their eyes.
Vision is improving now as the unknowns are shrinking and the known knowledge space is increasing. The 2015 product development round should see significant mistake correction on pricing. But let’s review a few common sources of error.
The biggest challenge for the 2014 Exchange product pricing was figuring out who was going to be buying any individual company’s Exchange product. Four million on-Exchange plans sold later, the actuaries are able to start to make solid assumptions about who will be on the Exchange next year. What metal bands were people buying and using? Did actuaries assume 35% Bronze or 62% Silver or 31% Gold, and how did that play out on the risk pool composition? Were plans getting the age mix that they anticipated. Right now, it seems that age mixtures are close to what people projected. Some firms are seeing younger/healthier mixes, other firms are seeing slightly older and sicker, but noise around central trends is expected from a variety of similar but not identical processes with different initial assumptions.
The strategic interaction of how plans and networks played against each other in a given regional market is the next critical piece of information. Some plans will find out that their super-narrow networks that excluded major regional academic medical centers were too narrow and attracted no one. Other narrow plans will find that they attracted the young and disconnected from the medical system, while broader networks attracted people with known medical issues.
Some plans will be discontinued, and other plans will be seriously tweaked in response to what other companies are doing in the same market. Business strategies will evolve as one insurer might choose to embark on a young and health strategy with concurrent risk corridor and risk readjustment payments going out, and others might choose to specialize in older or less healthy consumers with risk readjustment payments coming in.
Finally, people weren’t sure about claims scenarios. How would newly covered people use their insurance? Would it be a trickle of unmet need, or a deluge or just a regular marginal addition of utilization. Claims data will determine if an expensive specialist who is replacable is seeing enough members to keep in a network. Claims data and out of network authorizations can be used to determine if new providers need to be brought in.
2015 should see a more structured markets as some of the obvious sources of variance from projections to reality are disappearing.