A while back, ACA Death Spiral had an interesting post in which he tried to make sense of CBO’s claim that the Obamacare-related
health-insurance synthetic CDOs “risk corridors” your Congressmen have committed your tax money to would actually (a) make the government a net $8 billion over three years, in the process (b) receiving from 2x as many insurers as it pays to. You should go read that post, but his method seems sound enough: he just tries to calibrate a reasonable loss-distribution to that payoff-outcome. When he does he finds that to realize such an outcome, insurers will have to be making 7+% profits on average.
I didn’t follow his Black Scholesy calculus to the letter but I did try to replicate the basic finding using his same assumptions (namely that the average premiums are $3962, & the total count of assumed signups) and some poor-man’s spreadsheet mumbo jumbo. Wanna see? Well I’m gonna show you anyway.
Here’s a simplistic model then: pretend all Obamacare risk-corridor-eligible insurers will fall into three ‘buckets’. Gainers (who make X% profit), Breakevens (0% profit), and Losers (who lose X% on each policy). How many Gainers vs. Breakevens vs. Losers are there gonna be? And how much Gaining/Losing do they do? Lay out the 3 buckets in a spreadsheet, use ‘solver’, and try to get USG Profit to =$8 billion and Gainers/Losers = 2.
One thing here is obvious, if you get 2x as much from Gainers as you pay to Losers then (in this setup) the former bucket has to be 2x as big as the latter. Just making them 67% vs 33% is one way to get there.
The end result is that if we live in this 3-bucket world, the CBO is saying that insurers will make 6.9% profits on average (the number in green).
“But your breakevens bucket is empty”, you protest. You’re right. That’s weird. Let’s look for a solution in which the plurality of insurers just break-even, since that’s more realistic. Obviously, to still give USG the same profit, we’ll need to make the Gainers/Losers a bit more Gainy/Losey. We could have a world where Gainers make ~32% profit:
This means overall insurer profits average 6.2%. Either way, we’re ending up with a number in the 6-7% range.
“But we don’t live in a 3-bucket world” you protest. And indeed there’s something very artificial about a 3-bucket world with only ‘gainers’ and ‘losers’.
So let’s make more buckets. Imagine some insurers make -100% (i.e. lose $1 for every $1 in premiums they collect), some make -90%, -80%, … and on up to +100% (i.e. they collect a bunch of premiums, somehow have no costs, and never have to make payouts). We once again need to figure out what% of insurers need to be in each bucket to get USG profit to $8 billion on 2x winners/losers. That’s too many moving parts so I assume bucket-membership falls into a sort of bell-curve and play with the mean/variance of the bell curve till those numbers match CBO’s. Here’s the result:
If we live in this world, and if CBO is right, insurers will make 6.2% profit on average (again, see number in green). Hey! That’s the same number we ended up with in the simplistic 3-bucket model.
“But a bell curve isn’t just 21 buckets!” you protest. Fine, I’ll make the buckets tiny (0.1% each) running from -100% to 100% and repeat the exercise. That should be granular enough for this poor-man’s model. The resulting insurers distribution looks like this:
And the average insurer profits in such a world – the peak of that bell curve? 6.2%! It’s almost like you don’t need no fancy-schmancy bell curves or Black-Scholes integrals to get the big-picture answer.
1. The first takeaway is that I would like to confirm the substance of what ACA Death Spiral is saying: filtering CBO’s predicted $8 billion profits number through some common-sense feasible loss distributions implies insurers will be making at least a certain average profit. He says it’s 7+% and I say 6+%. In part the difference owes to technical reasons I think; I use normal distributions of [profits]/[premiums], he uses, I think, lognormal distributions of [costs]/[premiums], which gives his bell curve a certain skew. (When I do this in my spreadsheet version I get 6.8% for the insurer-profits answer.)
2. He says this is an unbelievably-large expected profit. I’m going to have to take his word for it, since I don’t have a feel for what’s high or low here. Health insurance seems like such a capital-intensive, pain-in-the-ass, heavily-regulated and litigation-risk-prone business to be in that I can’t imagine anyone wanting to do it for less than double-digits. But what do I know.
3. Since he wrote that post, we have some new info to incorporate into this model! Namely: the Obama Administration has, in their budget, asked for $5.5 billion (positive, not negative) to fund these ‘risk corridors’ in their first year of operation.
That’s right. This program that, we are told, is going to net us $8 billion in the first 3 years is also, we are told, going to cost us $5.5 billion in year 1.
How to make sense of that? I’d love to see ACA Death Spiral’s take, but for now let me take a stab. Note that per the assumptions of our models, in that first year there should have been 8 million signups and hence health insurers will have received $31.7 billion in premia. Obama is now telling us they’ll lose so much money in that year that they’ll get $5.5 billion net in risk-corridor payouts. Damn! That’s a bad year! What does it look like?
Let’s just use my 3-bucket model (since it seemed to be a decent poor-man’s approximation to the full bell-curve model). Here’s a way to get to that number:
So as you can see, apparently Obama is telling us that 80%+ of Obamacare insurers will lose money next year, and that insurers will lose an average of 25%. BUT WAIT THERE’S MORE. Since (obviously) we all believe the CBO number, this means insurers are due for a massive turnaround in the following 2 years: they’ll make so much that we will net $13.5 billion, more than making up for this year-1 $5.5 billion loss (and at the end of the day we will, again, have made 2x from gainers as paid out to losers). How will that happen you ask? Here’s the 3-bucket way:
So now you can see that by year 2 (once we have given insurers $5.5 billion), they will turn around and become a business that makes 12-13% in profits on every dollar of premiums they collect. Now that’s more like it, that’s a business model I can get interested in!
Disclaimer: all of the preceding depends on numerous assumptions. Big Assumption #1 is that the CBO number is not B.S. built on rosy and ridiculous assumptions. Big Assumption #2 is that Obama’s $5.5 billion budget-request is also not B.S. (like a made-up opening salvo based on nothing concrete), let alone some sort of outright fraud attempt. But of course, we all have no problem accepting those Big Assumptions. Right? Right.
UPDATE: Commenter Texan99 points out that standard lore says insurance companies make ~3% profit. In light of that, the fact that Obama claims to need $5.5 billion for corridor-losers is highly suspicious. How on earth can that number happen? I have laid out the simplest way (poor-man’s version). It involves insurers so severely mispricing their plans (or overestimating pool quality) that they anticipate taking ~25% in losses on average. Is that believable? And if it is, the idea that the Obamacare-insurance plans are anything but welfare is untenable. And in particular, the CBO estimate of $8 billion revenue within 3 years is a fairy tale that relies on year-1′s pools being dominated by adverse-selection but then curing themselves (and more) by years 2-3 as the healthy join and turn the pools into 12+% profit-generators for insurers. Again: 12+%, when the standard profits – in the pre-Obamacare environment, which presumably involved less competition – are 3%.
This number, if it stands, needs to be getting far more attention than I have seen.