Thank you for the clear explanation, seattlecyclone...this is the first time I've actually understood how it works. I ended up with a gold plan this year that was cheaper than the silver plans and my eyes glaze over when people try to explain that.
The gold plans cheaper than silver plans are themselves an interesting phenomenon.
For most people, bronze plans are designed to make the average customer pay 40% of their expenses out of pocket (insurance covers the other 60%), while insurance covers 70% for silver plans, 80% for gold plans, and 90% for platinum plans.
People under 250% of the poverty line qualify for "cost sharing subsidies" on silver plans (and only silver plans). These folks pay the same premiums as anyone else, but the out-of-pocket costs are reduced. Between 100-150% of the poverty line, insurance covers 94%. You get 87% coverage between 150-200% of the poverty line, and 73% coverage between 200-250%. Make a special note of this 100-200% zone: you pay silver premiums for coverage that is almost as good as (or better than) a platinum plan.
Of course, paying 94% of the costs while charging a premium designed to make a small profit at 70% of the cost is a money-losing venture for insurance companies. The ACA, as originally designed, would reimburse insurance companies for the extra costs here. A few years ago the Republicans in Congress decided they didn't want to pay for this anymore, so they stopped these reimbursements. The insurance companies needed to raise premiums to make up for that.
In most states they've settled on a practice called "silver loading" where the cost of these cost-sharing subsidies is baked into only the silver plans' premiums. Bronze/gold/platinum premiums aren't affected by this. If you buy a gold plan you're really paying for 80% coverage. If you buy a silver plan you're paying for a weighted average of 94/87/73/70% coverage. If the people getting 94% or 87% outnumber the people getting 73% and 70%, the weighted average could easily be higher than 80%, and the premium would therefore need to be higher than for a gold plan.
Once the balance starts to tip very far in this direction, people not qualifying for the cost-sharing subsidies might find gold or bronze plans to be a much better deal than silver, which tips the balance even farther the next year.
A side effect of this is that since the premium subsidies are based on silver plan prices, this imbalance from silver loading will make
everyone's premium subsidies higher than they would have been otherwise. Cancelling the cost-sharing subsidy reimbursements could easily cost the government more money than it saved.