Silverloading game theory (revisited)

I’m working on an ACA/Silverloading manuscript this morning.  I think I will be able to write that sentence for at least another decade.  The big question that we’re poking at in this manuscript is how does Silverloading change the pricing dynamics from the insurers’ points of view.  My co-authors and I all agree that weird game theory happens.  I am trying to think my way through the problem and as I’m thinking about it, I wanted to pull up my first attempt at thinking through the problem from Spring 2017:

Anthem seems to be allocating all of the cost of the CSR non-payment only to Silver plans. Bronze, Gold and Platinum plans will be priced on medical trend only.

This is interesting as it opens up the potential for some very odd games. We need two cases.

The first case is if not all carriers use the same assumptions. I am assuming the prices of the carriers in a non-sabotage world would be converging as my baseline counterfactual. We’ll keep the world simple and just assume two carriers in a pricing region. I think the analysis logically expands to more than two carriers.

If one carrier elects to spread the cost of CSR throughout the entire product line and one carrier elects to concentrate the price increase in Silver, the spreader most likely captures the least expensive Silver plan. They are adding 10%-20% of premium to their non-sabotage Silver premiums. The concentrated Silver needs to add enough premium between 15% to 20% actuarial value which with induced demand factors probably means a premium increase of 25% to 30% over their non-sabotaged Silver plans. If each company only offers a single silver plan, there is a large silver gap and there are great deals for buyers of the spread Silver. Most likely the spread Silver company captures most of the CSR Silver market.

Odd things happen in other bands. The concentrated Silver has a massive price advantage on their Bronze plan as they add in no sabotage premium increase. The spread Silver plan has to add in a significant bump to their Bronze plans. And since there is a large Silver spread between the benchmark and the low cost Bronze plans, these plans will be very attractive to non-CSR eligible individuals. The same dynamic will occur at Gold. There is a decent chance the concentrated Silver Gold plan will cost less than their Gold plan and may be near the spread Silver’s premium.

I think in a situation where there is a split actuarial assumption on how to distribute the costs of non-payment of CSR, the company that spreads the cost over all metal bands will get almost all of the 100-150% Federal Poverty Line (FPL) Silver purchases, most of the 150-200% FPL CSR Silver purchases and very little else. The concentrated Silver carrier will get almost all of the people who would have otherwise bought the Silver 200-250% FPL CSR Silver as Gold buyers, and then pick up most of the Bronze, Gold and Platinum buyers.

I’m sort of identifying a Silverload vs. Broadload game here although those two terms were not invented for another few months. If there is a split in how costs for CSR benefits are attributed to different metal levels and somewhat similar pricing structures, there are points were different insurers are competing for very different market segments. I think this still makes at least directional sense.

The next scenario is if all carriers in a rating region adopt the Anthem assumption that their Silver premium contains the entire CSR sabotage price as an actuarial value adjustment….
Effectively Bronze plans in most counties will be low to no cost for most buyers who earn under 400% FPL. Weak gold plans (76%-80% AV) will be priced for subsidized individuals as if they are Bronze plans today. Platinum plans will have relative prices as if they are Gold plans today.

If every carrier in a region prices on the assumption that Wesley thinks Anthem is making, the average actuarial value of plans purchased on Exchange will increase significantly for everyone but for 100-200% FPL CSR Silver buyers. They will be getting the same (87% or 94%) actuarial value as before with perhaps some of the 150% to 200% FPL buyers choosing to buy down to a Gold to take a larger out of pocket in exchange for a lower monthly post-subsidy premium or buying up to Platinum for a small increase for less out of pocket.

I don’t think I quite got this one right as the Two Margin Problem suggests a much larger move to Bronze than I thought and I got the risk adjustment/competition incentives wrong on Silver pricing near Platinum instead of near Gold.