Sam Richardson recently joined the faculty at Boston College, where he is an Associate Professor of the Practice of Economics. He has also volunteered as a Certified Application Counselor each of the first two open enrollment periods. You can follow him on Twitter: @Prof_Richardson.
California and New York marketplace premiums for 2016 were announced last week. The reported premium increases affect the federal government’s bottom line, but are largely irrelevant for the 86% of marketplace consumers who receive subsidies.
To see why, we need to understand how subsidies are calculated. Families with incomes below 400% of the federal poverty level (FPL) receive subsidies such that the benchmark plan (the second cheapest silver plan) costs a certain percentage of their household income. This percentage ranges from 2% for those with incomes below 133% FPL to 9.5% for those with incomes above 300% FPL.
Suppose in a given market every plan’s premium went up by $100 for a particular consumer. Assuming the consumer’s income stayed constant, the subsidy would increase by $100, and the consumer could stay with the same plan for the same after-subsidy premium. Thus if all premiums move together (either up or down), the subsidized consumer is unaffected.
So what does matter for subsidized consumers? Possibly the most important statistic for these consumers is the difference in premium between the benchmark plan and the cheapest silver plan. Let’s look at two New York insurance markets: Buffalo and Utica [1]. Rounding to the nearest dollar, the premium for the cheapest silver plan for an individual for 2015 was $295 in Buffalo and $314 in Utica [2]. These numbers each increased by close to 20% for 2016, to $353 in Buffalo and $373 in Utica. Bad news for consumers, right?
Not necessarily. Let’s look at the benchmark second-cheapest silver plan in Buffalo and Utica. Buffalo’s benchmark plan cost $337 in 2015 and $361 in 2016, a 7% increase. Meanwhile, Utica’s benchmark plan increased 41%, from $356 to $503. This increase turns out to be good for the subsidized Utica consumer.
Consider a single person making $30,000 per year (257% FPL) [3], and assume she takes advice from Margot Sanger-Katz and Amanda Cox and shops each year for the cheapest silver plan. She is eligible for subsidies such that the benchmark plan costs her 8.31% of her income, or $208 per month. If she lived in Buffalo, she would have received a $129/month subsidy in 2015 and a $153/month subsidy in 2016. Thus the cheapest silver plan would have cost her $166 in 2015 and $200 in 2016, a 20% increase.
But if this same consumer lived in Utica, she would have received a $148/month subsidy in 2015 and a $295/month subsidy in 2016. Her cheapest silver plan would have cost $166 in 2015 and $78 in 2016, a 53% decrease.
From a policy perspective, premium increases play an important role in the government’s cost of subsidizing insurance. But low-to-middle-income consumers and their advocates should be looking at changes in premiums after subsidies, which are largely unrelated to the changes in pre-subsidy premiums that are typically reported.
Correction: While the overall point of the post is still accurate, I made an error in using the posted New York premiums to calculate subsidies. Because the same insurer can offer multiple plans in a tier, it is possible for the same insurer to offer the cheapest silver plan and the benchmark plan. In this case, the subsidy would not be based on the silver plan from the second-cheapest insurer, but rather the second-cheapest plan from the cheapest insurer. It is impossible to tell from the posted rates whether this is the case in 2016, but this was the case in 2015 in both Buffalo and Utica, meaning my calculated subsidies were too high. This highlights a drawback in New York’s approach to announcing premiums: it is impossible to calculate 2016 subsidies based on the state’s recent announcement. I regret the error.
[1] These markets are admittedly cherry-picked to illustrate a point.
[2] Note that premiums in New York do not vary by age.
[3] For simplicity, I will use the same FPL for both years; my numbers will be slightly off because the FPL increases each year.
