I recently took my college referee fitness test. The college fitness test has four events; a distance run, an agility box, and two sprint sets. On each event, the referee can score between zero and five points depending on time/distance. I passed.
My results could be described in the following ways:
- Piss Poor
These disparate descriptions are accurate.
They are accurate depending on what my results were compared against.
My results were piss-poor when compared against the other nine referees in my testing group. It was me, seven referees whose highest level matches last year were either NCAA semi-finals, or some version of an international match, and two 21 year olds who run the 800 in college. In all four events, I had or shared the lowest score. Against a narrow comparison group, I was slower than molasses.
My results were adequate in that I am fast and fit enough for the games that I work. I met or exceeded external absolute standards.
My results were great because I had personal bests in two events (the distance run and the first set of sprints) and matched my previous personal best scores on the other two events. Against an internal comparison, I improved.
Asking the simple question of what something is being compared against and therefore being evaluated against is a critical question in determining the value of an analysis.
Moving back to the healthcare world, this simple question — compared to what — is critical when looking at the Exchange age mixture.
Jonathon Cohn at the New Republic has a good summary:
And within those marketplaces that the federal government is managing directly, 28 percent of enrollees are ages 18 to 34…
As for the age mix, you may have heard that about 40 percent of the population eligible for coverage in the marketplaces is between the ages of 18 and 34. That’s true and, obviously, 28 percent is a lot less than 40 percent. The worry has always been that older and sicker people would sign up in unusually high numbers, forcing insurers to raise their prices next year and beyond.
But insurance companies didn’t expect young people to sign up in proportion to their numbers in the population. They knew participation would be a bit lower and they set premiums accordingly. Only company officials know exactly what they were projecting—that’s proprietary information—but one good metric is the signup rate in Massachusetts, in 2007, when that state had open enrollment for its version of the same reforms. According to information provided by Jonathan Gruber, the MIT economist and reform architect, 28.3 percent of Massachusetts enrollees were ages 19 to 34, a comparable age group.
Age is a good but not perfect proxy for health. The Exchanges need a population that is roughly similar in health composition to the general population to avoid premium spikes. An Exchange population of only Balloon-Juice blog hosts or 63 year olds with chronic conditions is an extremely expensive risk pool. A death spiral would be unlikely due to subsidy design but total federal costs would increase dramatically. An Exchange population with a good number of healthy people in it means premiums won’t spike.
So when you hear people comment on whether or not the Exchange risk pool composition is good or bad or adequate, ask — compared to what?
Against the most relevant comparison (Massachusetts in 2006), the risk pool composition is at least adequate if not good on a national scale. We don’t have the data to say what the Exchange risk pool looks like in any given state. My suspician is that states that embraced PPACA, the risk pools will be on average, better than the full resistance states. The states that embraced Obamacare have seen higher sign-ups and more effective outreach efforts. Young and healthy individuals have always been the hardest group to bring en masse into social insurance programs, so states that actively outreached to these groups probably would see higher enrollment than states like Georgia that actively ratfucked outreach efforts.
The first year risk pool age composition is not meeting ideals, but a reasonable expectation is that it would never reach ideal composition in the first year.