The standard story about Obamacare has two steps: (1) We need young people to join so that average costs are low. Prices will reflect average cost because of insurer competition and so healthy young people will cross subsidize less healthy older people. (2) This cross subsidy will only operate if the young get Obamacare. They may not because the price is greater than their payoff from going without insurance. Hence, the individual mandate is necessary to hold this together. If the tax is too small or the website failure too forbidding, young people will not join and the whole thing will collapse as adverse selection drives up prices and further reduces participation etc – the so-called “death spiral”.
But this story is persuasive if young and old people are in the same pool. Obamacare allows pricing based on age so young and old people are in different pools. The young do not subsidize the old. If the young do not get Obamacare the old still get their insurance and they can live happily ever after (or at least get statins and heart bypasses). At a second cut there is a bit of a cross subsidy because Obamacare imposes a 3-to-1 ratio on prices of older age groups versus new. If this constraint does not bind, no problem. Even if it binds, it is relaxed if the young do not participate in their pool so prices go up in that pool allowing higher prices in the older pools.
Still, within a fixed age based pool there can be adverse selection. How big is it? There is a working paper by Handel, Hendel and Whinston that gives us an idea. There is an impact of adverse selection because at least when you allow just two plans, one covering 90% of costs and the other 60%, only the latter trades. But what happens if you also drop the individual mandate? The last column of Table 12 on page 41 gives their forecast based on their model and the data. Participation is 87%-90% for those 50 and older. But is only 63-70% for those 25-40.
This is the death spiral, but only among the young. It does not affect the older population. First, age-based discrimination innoculates the old from the non-participation of the young. Second, the 50+ crowd (which I am fast approaching!) need health insurance so they all get it.
(Also, this analysis ignores subsidies which would increase participation further even in 25-40 age group…)