In Which I Am Confused

Obamacare, much in the news lately, has noted that it depends on millions of young healthy individuals who are not currently insured to sign up for insurance, in a large part to pay for the coverage that they will extend other uninsured people who are not-so-healthy. This strikes me as an actuarial accounting error.  Or a dishonest tax using crooked actuary tables.

Actuarial methods assign costs and risk based on statistical advantage of your place in a pool of subscribers. An honest actuary prices your insurance premium at the same level as your cost. Given a large enough subscriber pool, your premium averaged over the large number of people is exactly matched by the insurance payouts for your pool.

Thus adding a new group, young uninsured healthy 20 y/olds should have zero impact on the larger picture. Their premiums should be the same as the payouts for those in the same pool. But .. this is apparently not the case.

Why? Because the designers of Obamacare are crooks. Is there another explanation? ’cause it seems the only explanation I can see from here.

Leave a Reply

Your email address will not be published. Required fields are marked *

11 comments

  1. Boonton says:

    Well yea, but then by this logic all of insurance is a theft.

    Ultimately if you assign premiums based on each persons costs then ultimately you simply assign premiums based on what they spend.

    For example, your house didn’t burn down this year? Your homeowners insurance premium is $0. Ohhh, your’s did? Well you’re $300,000! The only thing you seem to add is maybe trying to predict this before the year is over, but then assuming the insurance company got really, really, good at predicting then it wouldn’t matter.

    Of course that doesn’t happen because if it did what would be the point of selling or buying insurance to begin with?

  2. Mark says:

    Boonton,
    No. The insurance company taking a bit off the top is not theft. Stealing from one actuarial category to pay for another is theft.

  3. Mark says:

    Boonton,
    As a homeowner you are in a category of living in a certain region with given weather patterns and climate and risk for fires. You pay a rate for that. If you’re payment is designed to lessen the payment for those living 100 in oil soaked canvas houses (a higher fire risk) so that their payment is lower than their risk should suggest then you are being robbed.

  4. Mark says:

    Boonton,
    Group. Risk group.

    Ultimately if you assign premiums based on each persons costs then ultimately you simply assign premiums based on what they spend.

    You assign a persons premium based on their actuarial risk assesement. Their statistical risk. You want a lot in a group to flatten out the stats, so that in average their cost = their aggregate premium.

    Faulty actuarial assessment to shift cost burdens from one risk group to another preferred group is what I am calling theft. It is up to you to defend this theft.

  5. Mark says:

    Boonton,
    Look at it another way. Say you are a non-smoker. Smoking has a higher actuarial risk for a bunch of cancers and heart diseases over the general population. But some politician decides, gosh health costs for smokers are really high. So we’re going to insist that non-smokers pay 30% higher premiums and that money will be used to lower the premiums for smokers. Is that right to do? How is that justified?

  6. Boonton says:

    No. The insurance company taking a bit off the top is not theft. Stealing from one actuarial category to pay for another is theft.

    What the hell is an ‘actuarial category’?

    Ahhh

    As a homeowner you are in a category of living in a certain region with given weather patterns and climate and risk for fires.

    Group. Risk group.

    So an insurance company can get 1000 people and note that every year their claims total $565,000. Hence it charges them $565 premiums (assume for the moment nothing is ‘taken from the top’ to cover profits and administrative expenses).

    But suppose a kid with some paper interning for the summer notes that the 1000 people consist of 500 men and 500 women. The men’s claims totalled $300,000 and the women $265,000. Then it makes the premium for men $600 and women $530.

    Now note what happened before. Men were getting a ‘bargain’ of $35 and women ‘screwed’ for $35. Was that theft?

    But now imagine the kid interns next summer. Now he has learned to master Excel and churning thru the stats discovers bald men have different claim rates than non-bald men. Left handed women are different from right handed. He calls his friend in the University Computer department and they fire up the mainframe, they discover instead of breaking the group of 1000 into two groups (men/women) they can break them into many groups (men/with college degrees/30’s/late30’s/who drive a BMW/who drive a red BMW/ who are geminis and so on). They end point of this process is to end up with so many acturail groups that every person ends up in their own unique group. In other words they know exactly who will have claims and what they will be.

    At this point insurance is destroyed. Why? Suppose you applied to this company and they tell you they will insure your house for $1! Turn it down. Since they know who will file a claim they know you won’t need too so it makes no sense for you to even sign up for $1. Likewise suppose they tell you $300,000. Again turn it down, they know the future, they know your house will burn and you’ll need $300K. You might as well, if you have it, just set aside $300K yourself instead of going through the policy.

    Now of course regression analysis and Big Data is not going to be powerful enough to become an Oracle of Delphi. But you can see the logically suicidal path insurance is set for. The insurance company wants to predict and control claims. But the better they get at it, the more the business is destroyed. Kinda paradoxical isn’t it? Not really, insurance is really a business of buying and selling risk. The more you narrow down probabilities the less risk there is. If you eliminate risk, you eliminate your business of risk management just like McDonalds would eliminate itself if it invented a hamburger that was so good you wouldn’t ever need to eat another one again.

    Anyway, there is a logical way out of the trap, a ‘veil of ignorance’. If the insurance company agreed to charge all 1000 people the same price, then it would continue to exist *even if* it somehow achieved perfect knowledge. Yet you would call that theft….which is strange considering that you would choose to buy such a policy on your own free will.

    Your assertion doesn’t make sense because ‘actuarial category’ is not a stable concept. One day the category might be gender, the next it might be gender and age, the next gender/age/schooling and so on. For any insurance system you give me, I can subdivide their categories into yet more categories and discover that my particular category might be the victim of ‘theft’ or might be ‘stealing’ from other categories….unless I find an insurance company with so many categories that each person sits alone in one. In which case such a company would be as useless to me as I it.

  7. Mark says:

    Boonton,
    Well,l we can dispense with discussions of “risk analysis” that entails perfect knowledge of the future. Your “big data” digression ignores the fact that actuarial tables (probably now algorithms, but let’s stick with the old terminology), is also suspect because law prohibits many inputs, like health insurance adjusters cannot ask a persons if they are gay frequenting multiple partners a year even if that impacts their risk. The point is risk related criteria are valid (non-theft). Non-risk related criteria are what I’m terming theft. If you add people to a group not in your risk pool your premium should be unaffected. If it is, then you should suspect theft.

    The point you are ignoring is (oddly) the one you are ignoring and not defending. Your premium should depend on your calculated risk, what column/category you fall into on the actuarial tables. It is not legal for Aetna to base your insurance premium on non risk related data. It would be illegal for Aetna to have a policy of adding $500 a year to your premium because you blog, or are Jewish, or vote Democrat (that last one is probably difficult for a guy you cannot imagine why the IRS audits based on political affiliation might be a bad idea).

    My simple contention is as follows
    (a) non-risk related premium adjustments are theft.
    (b) adding a new group which is not in your risk group should not affect the premiums assessed to your risk group.
    (c) If it does, then that is evidence of theft or insurance malpractice.
    and finally
    (d) adding a unrelated risk group (young single) to the premium pool is alleged to reduce premiums for other unrelated risk group in Obamacare, ergo, Obamacare is actuarial theft.

    I suppose an alternative to theft is it is a tax. But there are lots of Constitutional criteria which have to be applied to judge whether this is a legal tax. I suspect not. Race is a valid health/life insurance criteria (sickle cell &c) … it is not a valid tax criteria … except now it supposedly is.

    (NOTE: edited)

  8. Boonton says:

    The problem is that a perfect risk analysis will produce perfect knowledge down to the individual level of future claims which destroys the value of insurance.

    Now you argue that ‘non-risk’ criteria make ‘theft’ but this isn’t how you began the discussion.

    Again, consider an unsophisticated insurance company who sets premiums simply by dividing the claims over the pool evenly. That’s not theft. If you sub-divide the pool, though, you will find each smaller pool has different levels of claims so you can create different premium levels. Again you’re ok with that.

    So for any given insurance company, it’s probably a given that they could add another variable to their analysis, produce smaller pools with more precise averages for each new pool. This would mean some premiums would go up and others go down. Does this mean those whose premiums would fall were previously the victims of ‘theft’?

    It is not legal for Aetna to base your insurance premium on non risk related data.

    This is a different issue not related to the post. The purpose of creating tiny pools out of bigger pools is to fine tune the tables to get a better view of the risk (the ultimate table, of course, would be able to precisely predict risk down to the individual level). Clearly in this endeavor you would only want to add a variable to the analysis if it allowed you to produce a more fine estimate of what each pool would cost in claims. See the example of males and females above having different claim rates.

    (b) adding a new group which is not in your risk group should not affect the premiums assessed to your risk group. (c) If it does, then that is evidence of theft or insurance malpractice. and finally

    Again

    1. What exactly is ‘your risk group’? One company may use two variables….male or female or 8 age brackets (0-10, 10-20, etc.). In that case they would have 16 risk groups. Suppose they decide to add state? Then there are 800 risk groups. Suppose they add # of years of schooling (say 1-16). Now there are 12,800 groups. The risk groups seem to be a function of much the company wants to pay database programmers and mathmeticians.

    2. ‘Theft’ requires a property right in something that is taken from you against your will. With insurance the legal concept is you own your own ‘insurable interest’. For example, you are sending a boat full of gold off to the far east intending it to return with spices. You own the boat and the gold and will own the spice. You also own the risk that the boat will sink, that the gold or spice will be taken by pirates, that the crew will mutiny etc. You can sell this risk just as much as you can sell the boat. Because the risk is a bad thing, you sell it for a negative price to a willing buyer.

    So here’s the problem, where’s the theft? If you decide to ‘sell’ your risk of getting sick for -$200 a month (in other words you pay an insurance company to take on the risk of you incurring medical bills) how exactly is it ‘stealing’? You signed up with the company, you opted to buy the policy. You’re saying the insurance company should offer this risk transfer for $50 a month because you’re very young and very healthy. OK that may be but you have no ownership interest in what a buyer is willing to buy at.

    Remember your ‘theft’ analysis has to apply both in and outside the exchange. That means no insurance company can fail to use every ‘risk group criteria’ available to set premiums. No insurance company may simply set premiums based on total expected claims divided by total policy holders. Why not?

  9. Mark says:

    Boonton,

    Now you argue that ‘non-risk’ criteria make ‘theft’ but this isn’t how you began the discussion.

    Yes it is. Adding people to the pool who are unrelated to your risk should not affect your premium. Your premium should be calculated by your expected statistical cost. If your pool is big enough, then adding other people shouldn’t touch your premium. But this is not the claim, the claim is these people are required to pay for the cost of the healthcare of others. This means the cost of their risk is being shifted to your premium. This is evidence of actuarial malpractice.

    This would mean some premiums would go up and others go down Does this mean those whose premiums would fall were previously the victims of ‘theft’?

    No. It means your algorithm has changed. This is different than adding new people to the pool with a very different risk calculus affecting your premium.

    This is a different issue not related to the post.

    It is not different. It is the logical conclusion. The explanation for why adding people to a already very large pool changes your premium is that there are factors not related to risk changing your premium.

    The risk groups seem to be a function of much the company wants to pay database programmers and mathmeticians.

    Those experts are called actuaries. It is a profession. It is a closely watched profession, by federal and state regulators, who have some notion of actuarial malpractice. How can you tell if there is such malpractice. My claim is that your insurance costs changing when other people join is a sign of that. Look at an example.

    You are in health insurance pool. Your insurance premiums suddenly rise. You find out that the reason they went up was because a bunch of old NFL vets joined in a merger. Why would your rates rise in that case? Well, it is because your rates are not tied to your risk but you (and the bunch like you) are paying for a group that was added whose premiums are not in proportion to their actual risk. Factors other than risk came into play.

  10. Boonton says:

    Yes it is. Adding people to the pool who are unrelated to your risk should not affect your premium.

    Why?

    Your premium should be calculated by your expected statistical cost. If your pool is big enough, then adding other people shouldn’t touch your premium.

    I’ve given the real examples. Consider my pool of 1000 people versus my pool of 1000 people split into 500 males and 500 females. Are you telling me the first method of calculating premiums is theft?

    But this is not the claim, the claim is these people are required to pay for the cost of the healthcare of others.

    Actually most people are. That is how insurance works. Almost everyone who sends a premium check in this month for homeowners insurance will be getting nothing and instead pay the few people whose homes burned.

    No. It means your algorithm has changed. This is different than adding new people to the pool with a very different risk calculus affecting your premium.

    So you’re ok with changing algorithms. If an insurance company decided to simply have one mega risk pool and divide premiums evenly instead of multiple tiny pools you’d be ok with that. Great.

    Problem, life insurance is often written on terms ranging from 5 to 30 years. Health insurance (and others like auto, home, etc.) are written on a year to year basis. Every year you ‘re-up’ your insurance or opt to change it. No one is ‘being added’ to any pool. Everyone who buys a policy via the exchanges will be, by definition, new to the pool. So where’s you claim of theft?

    BTW, how exactly does someone get to ‘own the pool’ here simply because they are a policy holder? I don’t own my insurance company’s pool of people! I simply own an individual contract with that company. It’s not my responsibility to price that policy, for all I know they messed up and gave me too low a premium. Not my problem, they signed the contract.

    It is not different. It is the logical conclusion. The explanation for why adding people to a already very large pool changes your premium is that there are factors not related to risk changing your premium.

    Since the concept of insurance is pooling risk, it seems kind of strange to assert that it doesn’t matter who or what is in the pool.

    Those experts are called actuaries. It is a profession. It is a closely watched profession, by federal and state regulators, who have some notion of actuarial malpractice. How can you tell if there is such malpractice. My claim is that your insurance costs changing when other people join is a sign of that. Look at an example.

    You mean “look for an example”. A google search results seems to indicate ‘actuarial malpractice’ seems to mean nothing like you claim it to mean. It seems instead to consist of cases where actuaries said $X would be needed to handle claims for some program or project and the true amount turned out to be much larger (for example: http://www.nytimes.com/2009/12/20/business/20gret.html?pagewanted=all&_r=0)

    You are in health insurance pool. Your insurance premiums suddenly rise. You find out that the reason they went up was because a bunch of old NFL vets joined in a merger. Why would your rates rise in that case?

    You are in an apartment building. A bunch of tenants are horrible paying the rent. The landlord discovers that of the 50 apartments he is only collecting 35 rent checks per month since deadbeats score a few free months of rent before they can be kicked out. As a result he raises rents by 40%. Now he is ok since even though he still suffers deadbeats, the higher rents collected offsets for the rent checks he is cheated out of.

    You object, you have been there 20 years, never been late, have a great job and are a very responsible person. Since *your* risk of being a deadbeat is no higher than before, why should you suffer a rent increase?

    By the time you get to court, the judge will calmly tell you you’re out of luck. If you don’t think the apartment is worth a 40% increase in rent, then move.

  11. It’s great that you are getting ideas from this article as well as from our discussion made here.