Regulating Medical Loss Ratios
I was just thinking about a new paper in the American Journal of Managed Care:
The Robert Wood Johnson Foundation thought enough of it to post it on their website and I did a post over there
but sometimes they don’t like them, so I’ll edit it and redo it here because I think the conclusion, that a lot of health insurers and health benefit plans cannot possibly comply, is really important.
While really, really appealing, you can’t impose a “medical loss ratio” (I will bite my tongue since I really think it ought to be a “nursing loss ratio” or even better, a “nursing benefit ratio” but that is a fight for another day) because the whole point of insurance is that neither individuals, nor insurers, know what their losses will be next year – that’s why people buy insurance, and it is why insurers succeed and fail on a regular basis.
Insurance is all about the Central Limit Theorem (CLT) and even more so, about standard errors, which is why my book is going to be called “Standard Errors: Life, Health & Death When Hospitals, Long Term Care Facilities, Home Health Agencies, Physicians & Nurses Are Insurers.”
The medical loss ratio for any insurer, or health benefit plan, is a random variable that estimates the population medical loss ratio (I will call this PLR) for the population from which the insurer selected its policyholders. The operating results (Profits, losses, insolvencies, policyholder benefits) of insurers or health benefit plans, large and small, are simple linear functions of their Portfolio Loss Ratio Estimates (PLREs).
Large insurers’ estimates (PLREs) of the PLR are much more accurate than small insurers’ PLREs, so their PLREs are packed more tightly around the PLR than the PLREs of small insurers.
Large insurers are less likely to earn huge profits than small insurers, but way, way more likely to earn the profit margins assumed in their premium calculations and also way, way more likely to be able to convert a significant portion of their risk premiums to profits, at year end. They don’t have the roller coaster thrill of low loss ratios, but they have a really nice, safe ride through the amusement park. They watch the people screaming on the roller coaster from a really safe spot.
Small insurers are far more likely to incur huge losses and become insolvent than large insurers for the same reason that some of them will earn huge profits: their estimates (PLREs) of the PLR are far less accurate than the estimates of large insurers.
Small insurers, and this is the key to understanding why we can’t legislate or regulate the 80% medical loss ratio, are incapable of offering benefits anywhere near that high.
When you do the math, which is really simple (which makes you wonder why we are spending hundreds of millions of dollars a year trying to figure out why very small portions of insurer or managed care benefit plan premiums seem to get converted to nursing benefit ratios [Oooops!] patient services), you can see that if an insurer with 1,000,000 policyholders can offer benefits of $0.75 per premium dollar, while maintaining a probability of 0.9987 of avoiding insolvency, an insurer with 10,000 policyholders cannot provide any benefits at all “throughout a policy year,” if it wants to avoid insolvency with the same probability.
Yes, of course, the small insurer can start providing benefits late in the year, but it can’t provide any benefits at all “throughout the year” because it might just have a really bad year, but they don’t know, in January whether they are going to have a good year, or a bad year.
Most policyholders/patients prefer access to care year round, not just when their insurer or health benefit plan feels it can start providing benefits. Why would anyone want a health benefit plan or health insurance policy that you start paying premiums on in January and sometime around September you can start getting your benefits?
The smaller an insurer, the more inefficient it is in terms of converting premiums to benefits. There are things small insurers are very good at: Turning premiums into profits or losses, for example.
What makes an efficient insurer? If “Efficiency” means higher probabilities of profits, lower probabilities of losses, lower probabilities of insolvency, and higher policyholder benefits. The answer is: PORTFOLIO SIZE. Big portfolio size good – Small portfolio size bad. Sorry about that.
A very, very large insurer is more efficient, economically, statistically, mathematically, and financially, at all these desirable characteristics, than any smaller insurer. This greater efficiency is not by a few fractions of a percent, the sort of difference you might have to look for with an electron microscope, but by astronomical differences, the sort involved in deciding whether you are looking at moon, or the sun.
So, why do we have tons of research studies that cannot manage to get this right?
You can’t legislate 80% MLRs because, extending our example above, the only insurer that can actually meet, and exceed, an 80% MLR (NBR – ooops there I go again) and can even do much better, is………….
An insurer with 308,000,000 policyholders.
A national health insurer can actually offer benefits that high and would be able to earn profits of close to 5% while doing it, if everyone paid the same average premiums that go to far less efficient insurers currently.
Lots of small, competing insurers and health benefit plans are a waste of precious health care resources. It is really hard, actually downright impossible, to improve risk management through insurance beyond what we have known about the Central Limit Theorem for the last 200 years.
Implement an 80% medical loss ratio and you will have a lot of small insurers and small health benefit plans failing long before they honor their contractual obligations to cover the health care expenses of their policyholders.