Thursday, January 31, 2013

Insurance migration and counter-migration

I bet that if you were to look at the trends in medical insurance premium growth for companies in the US, you would find a bimodal distribution.  Some will have experienced a rate of premium growth lower than the average, and others will have experienced a rate above the average.  What might account for this?

It certainly is not the presence of wellness programs, sponsored by some companies but not by others. As Al Lewis and Vik Khanna documented in a recent Health Affairs Blog article, "the industry consistently mis-measures and overstates the direct healthcare cost savings" of wellness programs.  They continue, "The current wave of wellness programs are taking us wildly off course by promising substantial short-term reductions in health spending. This is clear whether one looks at the peer-reviewed literature, outcomes measurement generally, marketing claims, or the 'face validity' of the broader impact of wellness on population health status."

I'd like to suggest that a factor in the differential premium growth rates relates to whether companies have affirmatively counteracted the strategic plans of health insurers to migrate employees to plans that correspond to greater use of health care services.  Yes, you heard me.  Notwithstanding public pronouncements to the contrary, it is evident that insurers have persuaded plan fiduciaries (i.e, companies who offer health insurance to their employees) to adopt plan designs that are priced to diverge from the rates that would be based on actuarial calculations.  Plan designs for high-cost subscribers are subsidized by plan designs for low-cost subscribers.  I believe the insurers do this for strategic reasons, to migrate customers to those plans that create the most income for the insurers.  The plans that create the most income for insurers are the ones that generate growth in claims:  Insurers want larger groups to insure and they want to insure unhealthy populations.  After all, claim adjudication is the major source of income for the insurance companies.

Firms that are alert to this phenomenon are able to counteract it by modifying the relative contributions required by employees for the different plans offered.  A small amount of "counter-migration," especially among the high risk, high cost subscribers, goes a long way to mitigating the damage than can occur from pricing that is not reflective of actuarial risk.  Let me dive in further.

Let's start with the simple premise that health costs are overwhelmingly dominated by the average age of a risk pool.  Sure, there are other factors, but age trumps them all.  If we look at the various tiers of plans offered to employees, the average age of single people is roughly from 27 to 35, and their cost of health care is very low.  Moving to families, the average age of people rises to the mid-40s, and they are roughly 40% more expensive to take care of than single people.  For a family with two children, the rate should be roughly 3.8 times that of a single person.
But the underwriting benchmark prevalent around the country for the family plans is well below what would correspond to the actuarial predictions.  To make up the required amount of dollars, the premiums for singles are overpriced relative to their actuarial basis.  The rate for a single-plus-one policy, for example, should be about 2 times that of a single person's, but the industry benchmark is 2.5 or 3 times the single rate.    Strategically, the insurers want to shift business towards their preferred segments by subsidizing those segments and overpricing the other segments.

Firms often aggravate the problem by offering premium reductions for a deductible laden plan.  They hope by doing this to reduce their overall premium burden over time.  The problem with this is that consumers know what kind of claimant they are likely to be.  While consumers cannot predict accidents or the onset of an acute disease, they know if they have a chronic disease or cancer or some long-lasting malady.  Accordingly, those high claimant employees will never pick the high deductible plan, and so there are no savings to the employer.  (Why, then, are they offered?  Mainly, I'd suggest, so that the insurance company can then offer health savings accounts to the healthier families.  Administering those HSA's is remarkably profitable.)

In summary, risk is not a function of plan design.  Risk is a function of who is in the plan.  When plans are mispriced relative to actuarial risk, a less healthy population migrates into the pool, and that population is chronically underfunded.  This phenomenon is compounded when the low risk subscribers are encouraged by overpricing to move out of the risk pool, for it raises the cost of the residual population.  The insurer comes back at renewal time and says, "We're sorry, but your utilization has grown, and so we need to increase premiums by x percent," where x is dramatically above the average trend of health care costs in the region.

The way to fix this is to reverse the subsidization so that the effective pricing to consumers more closely reflects actuarial risks.  This can be done by adjusting the employees' contribution to the overall premium.  You don't have to do this by much to cause some counter-migration among the highest cost subscribers, and you don't need a lot of counter-migration to have a large dollar impact on the company's overall costs.  After all, the customers involved are the highest cost, so a movement of relatively few can make a big difference.

You may ask, where will they go? Well, here is an unpleasant aspect, but one that is inherent in the employer-based insurance system we have in the US.  As a colleague has said to me, "If I invite bad risk, it comes from somewhere.  If I encourage it to leave, it goes somewhere.  This is a zero sum game."  Specifically, a working couple will sit around the kitchen table comparing the health plans offered by their respective two employers.  They will transfer coverage to whichever spouse's firm offers the larger subsidy for their high-risk plan. That company will find itself shifted to a higher cost curve and will find itself moving up that cost curve at a rate greater than the average trend.  The company that has systematically acted to reduce untoward subsidies will find itself shifted to a lower cost curve and will find itself moving more gradually up that cost curve.  Simple mathematics drives this result. Firms have to understand that their strategic interest is at variance with that of the insurance companies.


Mitch said...

Two things:

-Thanks for sharing the Lewis and Khanna article. Because of a previous mention in your column I purchased and read Mr. Lewis' book. I have experience and close exposure to the kind of false or just plain bad claims made for behavior change, wellness et al programs.

-You are spot-on in your understanding of the mis-alignment of risk. As I pointed out to a German friend once, a PhD working at a behavior change company in Oregon, the insurers do NOT want people to be healthy. Nor do they want them to avoid the doctor. If they do so there is less activity, and less activity equals less for them to process and bill the payer (employer) for, assuming a self funded plan." As always, follow the money.

Keep up the great work.

Nancy Thomas said...

One anecdotal story first - I was at a board meeting of a visiting nurse group and a member of the board was a senior health insurance executive. When the executive director complained about utilization rates bringing up her health care costs for group insurance and looked to this executive board member for response he said, "here in this room, let me say this - stop hiring women under 40 who have babies. stop hiring women over 50 who get breast cancer. That's your answer." I was apalled, yet stunned into mental calculations. Of course, as a VNS this group employed 99.9% women. Between 30 and 60. Hmmm.

My comment on health wellness programs - "get up and move!" is that it pacifies only the public into thinking the health insurers care, and/or they are truly reducing their overall risks significantly. Certainly being healthier has its benefits but not on the actuarial calendar, I don't think it will show up in our lifetime.

Mark said...

Could you please explain: "claim adjudication is the major source of income for the insurance companies."


Paul Levy said...

Please see Mitch's comment.

Barry Carol said...

Self-funded employer health insurance plans pay insurers a fixed fee per member per month (PMPM) or per employee per month (PEPM). It doesn’t matter if average spending for medical claims is $4K per member per year or $6K or how many claims are filed. Age distribution doesn’t matter either in terms of how much profit the insurer makes on a given administrative services only (ASO) contract.

I worked for four different employers over my career all of which offered decent health insurance plans. As far as I know, none varied their required employee contribution toward the premium based on age or other actuarial risk factors other than single coverage vs. couple or family coverage. One of the firms did vary the contribution based on income. Low paid administrative employees paid a nominal sum while highly paid analysts, portfolio managers and executives paid $250 per month. That was in 1990.

Insurers will tell you that, at the population level, people between 55 and 64 years old use between five and seven times more healthcare than those in their 20’s in any given year. Yet Massachusetts only allows insurers to charge older people a maximum of two times what younger people are charged. The Affordable Care Act allows a three to one age rating band. In the individual health insurance market, insurers who use very strict underwriting criteria will have lower average claims and monthly premiums than insurers who are less strict about who will be offered insurance.

Employers issuing more than 250 W-2 forms are now required for the first time to report, for information purposes, the value of health insurance paid for by the employer as well as the employee contribution. Most employees are going to be hugely surprised at how large the employer contribution number is.

As for wellness programs, I’ve always felt that they don’t save money for the healthcare system over the long term. However, they do help employees to live longer and healthier lives. For that reason alone, they are worth pursuing.

Finally, if a married couple can choose between two employer sponsored health insurance plans, it’s not rocket science to expect them to choose the one that offers the coverage that best fits their perceived needs for the lowest employee contribution.

Mark said...

Of americans with health insurance, what % are on employer _self funded_ plans? Thanks.

Brad F said...

You refer to plans taking the risk, not those acting as TPA's, yes?

Incidentally, of the big insurers, approximately what % of their revenue stems from TPA activity vs acting as sole insurer? Is one more profitable or preferred?


Paul Levy said...

Mark and Brad,

I don't know those numbers. Perhaps someone else could provide or lead us to them.

To you and Barry,

If the pricing regime used by employers who are self-insured mimics the ones used by insurance companies who bear the risk--which they probably are--the same phenomenon described above will occur. Those companies will find themselves shifted to a higher cost curve and will find themselves moving up that cost curve at a rate greater than the average trend.

Barry Carol said...

Approximately 160 million people in the U.S. get their health insurance through an employer and another 18-20 million purchase it in the individual insurance market. 48 million people are insured through Medicare and about 55 million have Medicaid. 9.7 million people qualify for both Medicare and Medicaid (dual-eligible).

Of the 160 million people who get their insurance through an employer, I believe that between 55% and 60% are now in self-funded employer plans. For UnitedHealth Group, the largest health insurer by revenue and the 2nd largest by membership (after Wellpoint), 65.3% of its commercial membership were in self-funded employer plans as of 12/31/2012. The insurer calls this fee based business while 34.7% were in risk based plans where the insurer assumes all of the risk of paying medical claims.

The typical ASO contract fees for employer self-funded plans are in the $20-$25 per member per month (PMPM) range or $240-$300 per year. Profit margins can be in the high teens but it’s not a lot of dollars per member. By contrast, a mid-single digit profit margin on a full risk contract can produce a pretax profit per member that matches or exceeds the entire revenue from an ASO contract.

There is a strong trend toward employer self-funding which has persisted for years now. Administrative costs are lower for the employer, the plans are generally exempt from state benefit mandates and premium taxes and the employer can earn investment income on the reserves. They also own their claims data so they can more fully understand what is driving their claims and take steps like wellness and disease management strategies to mitigate their medical trend. On the flip side, under pure self-funding, the employer bears all of the financial risk of paying whatever covered medical claims its employees and their family members incur. Everything in life is a tradeoff.

me said...

@Barry (4:22 pm comment): I think that the rate ratio capping you mention among kinds of plans (i.e. single, couple, family) is a big part of the issue. Whether mandated by legislation or by plan design, if a rate is artificially lower than the actuarial calculation, an insurer - whose job it is to shift risk, not take it on - will simply artificially inflate the cost of a single person plan relative to a family plan. For example, if a single person plan should cost $100, then (taking the actuarial calculation that a family plan for 4 ought to have a rate 3.8 times that of a single person), the cost of the family plan should be $380. However, if this figure is capped at $250 (or some other number lower than $380), the insurer will simply raise the cost of the single payer plan. So, you might end up with a single payer being charged $150-200, in effect subsidizing the family plan. The numbers are just examples, but you get the idea.

Over time, this variation in a particular organization's plan - assuming that single insured members have a choice of alternate plan that is, from an actuarial perspective, correctly priced - will cause those single members to migrate to more cost effective plans, whether from an exchange/another employer/some other option. The result is a sicker risk/health pool of members who remain in the plan with artificial pricing. Since these members use services in greater proportions, the insurer, at the end of the year, will utilize that higher cost to calculate and support new higher premiums for the following year. That's what drives up a company's cost trend more steeply and to a higher level than the otherwise underlying cost trend.

The insurer doesn't take on any of this risk; it simply transfers it. And, by this kind of plan design, it ends up with exactly the kind of pool that it wants: one that uses a higher amount of care services.

Barry Carol said...

I don’t think very many employers offer much choice in health insurance plans beyond an HMO or tiered network product vs. a PPO and/or a high deductible plan vs. a lower deductible plan.

If all health insurance were priced purely on an actuarial basis at the individual or even age group level, it would simply be unaffordable for older and sicker people. At the other extreme, in Switzerland, all people over the age of 25, including the elderly, pay the same rate for comparable coverage in a given canton (similar to a state in the U.S.). The net effect of that is to force young people to drastically overpay for their health insurance in order to subsidize the old and the sick. There is a mandate to purchase health insurance in Switzerland but 45% of the population qualifies for a subsidy.

Similarly, most towns rely heavily on local property taxes to finance K-12 education. If you have no children or no longer have children in school, you still pay the same property tax on a house of similar value in a given town as a family with three or four kids in school. That’s the way it is and, I think, should be.

Health insurers need to cover their costs and earn a profit at least sufficient to cover their cost of capital over time or, in the case of non-profit insurers, to maintain adequate reserves and capital levels. I frankly don’t see any viable alternatives to allocating the cost of health insurance across the population without making it prohibitively expensive for older and sicker people.

Brad F said...

Getting back to Paul's statement, and was the reason for my question above:

Insurers want larger groups to insure and they want to insure unhealthy populations. After all, claim adjudication is the major source of income for the insurance companies.

Seems, based on Barry's data, approximately a third of ESI comprises risk based contracts. Lucrative as they might be, the phenomena Paul describes not the majority case. Fee-based TPA activity wont generate any more profit in a high vs low adjudication universe. I clarify, a) because the distinction has import, and b) migration concern only applies to a slice of the total market.


Paul Levy said...


To repeat:

If the pricing regime used by employers who are self-insured mimics the ones used by insurance companies who bear the risk--which they probably are--the same phenomenon described above will occur. Those companies will find themselves shifted to a higher cost curve and will find themselves moving up that cost curve at a rate greater than the average trend.

Brad F said...

I am sorry Paul, I am not understanding.

Insurance companies have difft plans and they will migrate insured individuals--one from the other--to accomplish the end result you envision (risk based contracts).

HOwever, in a self-insured plan, the TPA collects PMPM fee. Perhaps I am misunderstanding the profit motive or how they arrange the contract, but how does the migration effect them, the insurer. The company has one pool of folks-- their workers, and they choose amongst a menu of plans. The company accrues all the risk and only pays a fixed administrative fee. What motivates the TPA to do what you propose?

Again, sorry for beating a dead horse, but I want to understand the arrangement.

Paul Levy said...

Sorry I am not being clear.

Two cases, as you suggest. The insurance company has a general set of prices that they use for their risk clients, where they clearly have a financial stake in promoting a greater number of claims. But they suggest the same relationship among the various premiums when they advise the self-insured clients. You are right that they have a different financial interest in that case (putting aside the extra profitabilility if they administer the health savings account). But the effect of an improperly subsidized rate structure on the self-insured company's future health care trends will nonetheless exist.

Nate O said...

It's not just insurers who have pushed this. We have a whole list of preventative services that must now be covered at 100% and the medical community pushes "seeing" your doctor for just about everything as well. All the drugs advertised on TV suggest you see your doctor.

In general terms what insurers have done is design plans that replace out of pocket spending with up front premium. Studies have shown in 1965 50% of spending was OOP and now it is under 13%.

Until recently when employers tried to save money they made foolish decisions like raising co-pays or deductibles. If you take the number of co-pays a groups has in a year and times that by $5 or $10 it is pocket change. Even increasing the deductible from $1000 to $1500 is pennies. The real real bang comes in co-insurance % and max OOP which are less seldom an option to change.