Discrimination has a transaction cost, and therefore in many cases non-discrimination between risk groups is efficient.Even when the cost of discrimination is less than the difference in the expected claims cost it may not be economical for some insurers to discriminate ( i.e. if a minority can benefit from discrimination, they can pay the costs of discrimination indirectly through rival insurers who do discriminate, while the majority have no discrimination transaction costs but pay for the higher risk).

Insurance coverage is also available for risks that are significantly affected by the will or choices of the insured, depending on the economic benefits of insurance versus the economic costs of higher risks generated by the altered choices. For example, I have insurance on my car for the risk that I will, through my fault, cause damage to someone else’s car. If I don’t pay attention to safe driving rules like driving at a speed that I can stop my vehicle within the clear distance I can see ahead of me, and following another vehicle only at a distance that I can stop my vehicle without impacting the other vehicle if it stops suddenly, then obviously my risk of having an at fault crash and causing damage to another vehicle is increased. Although I still follow these safe driving rules fairly diligently, and therefore know that my risk of an at fault crash should be lower than a lot of other drivers who don’t, I still choose to insure against the risk. Why? I am obviously transferring the risk of large losses away from me at a premium of less than its value to me, notwithstanding the mis-pricing of the risk, and the influence I have over the risk. Also there is indirect methods of discrimination by asking how many at fault crashes I have had in the last 5 years.

In New Zealand where I live, the insurance industry is delightfully unregulated. There is no form of licencing of insurers or insurance brokers and no regulation of the substance of insurance contracts of and kind. There are only five statutes specifically for insurance:
Insurance Companies (Ratings and Inspections) Act 1994
Insurance Companies’ Deposits Act 1953
Insurance Intermediaries Act 1994
Insurance Law Reform Act 1977, and
Insurance Law Reform Act 1984.

The substance of all these statutes is that insurance companies are required to deposit NZ$500 000 in approved securities into trust with the Public Trust (chicken feed for insurers), general and disaster insurers are required to have a current credit rating, insurers are required to disclose if they have a credit rating and what that rating is (if any), intermediaries are required to use a trust account for premiums received, and some minor changes to common law and equity under the insurance law reform acts.

The result is that the amount and types of risks included in insurance contracts, and the amount and type of discrimination, and the premiums are totally unregulated, and that entry into the insurance business is all but totally unrestricted.

Insurance available in this market includes credit insurance that makes repayments if you lose your job other than for your own misconduct, income protection insurance that primarily covers loss of income through illness or injury, health insurance that covers most medical costs related to injury or illness, and of course life, property, contents and vehicle insurance.

Regarding protecting bond purchases against default, I believe that there are derivative securities which transfer credit risk, but I can’t tell you much more than this.

For the record, most life insurence companies cover suicide — at long as it’s after 2 years of getting the policy. -Tracy SAboe

Tracy is entirely correct. After a two-year exclusionary period, suicide is indeed covered under all domestic life insurance policies that I know of. I believe the reasoning behind it, is along the following lines: suicide is not usually planned out more than two years in advance–it is usually in some manner more impulsive or immediate. With this lengthy exclusion, the insurance company can weed out almost everyone who wishes to “overuse” their life insurance policy (treating it as, in essence, a mortality-based lottery they are guaranteed to “win”), in much the same way as a health insurer limits their risk of overuse of a health policy by excluding claims arising from “existing conditions”.

If the insured commits suicide in the first two years, the legal situation in this case is that the insured is considered to have defaulted on the contract. The insurance company simply returns all premiums paid up to that point, and the contract is nullified–just as though it had never existed.

I don’t know exactly how the various parts of this situation are affected by regulation–whether or not they would be likely to exist on a free market–but I suspect that something very similar would indeed exist, if only because suicide is not terribly common. The risk of any individual person’s suicide–at some time in the future greater than two years from now–is statistically very small…and more importantly, unpredictable. Causing one’s own death, while technically under one’s own control, is such an extreme step that most people–even those who are severely depressed–are unlikely ever to take it.

At least with present actuarial methods, it is basically unpredictable as to which members of a given population will actually “take the final plunge” so far in the future. This makes the risk insurable. If there is a particular subset of the population that has a greater-than-average chance of suicide in that 2+ years timeframe, then the insurance companies will be the first ones to figure it out, and to mitigate their risk by charging the individuals in that high-risk group correspondingly higher premiums, or by creating broader exclusions, or by simply refusing to insure members of that group at any price.

Comments?

–Im sure Hoppe ment that in theory, you can’t insure for something that’s controlable. This does not mean that there are insurance to cover suicides. Its illogic for the insurance industry to cover suicide, however, the logic fails at predicting the life. Hoppe lectures about the true definition of insurance. Meaning that insurance should only insure the “uncontrollable.” That is why Hoppe pointed out that politics have interfere with the insurance industry, forcing them to insure what should be considered uninsurable.

I think you are right. However, i would expect such ratings affect only bond discounts and are used mostly by potential buyers of bonds, rather than insurance companies selling “business failure” insurance. Plus i don’t think that buyers of bonds buy insurance on the bonds they buy, based on these ratings.

But i am only speculating on this based on my own limited dealings with bonds. I’ve never bought bond insurance against my bonds tanking. Does anybody?

Isn’t this exactly what rating agencies like Moody’s and S&P’s do when they determine bond ratings? Companies which are perceived to have higher credit risk will pay more for borrowing, part of their cost being a premium to cover the possibility of default.

Paul and others above have pretty much explained the main issue I have with the article, but at the risk of generating even more confusion, I’d like to elaborate… (if only to make sure I understand it myself)

I particularly dislike his claim that “everything that is within either full or partial control of an individual actor cannot be insured”. Certainly, things that are within full control of an actor are uninsurable. Things that are partially within someone’s control, though, are insurable, but in effect only the uncontrollable aspects.

Any controllable part of an insured risk will be considered as “given” by an insurer. In a free market, this would lead to the further splitting of the “pools” or “classes”, i.e. those who intend to increase the risk in one pool, and those who don’t in another. This would alter the relative premiums paid accordingly.

Insurance against injury from a skydiving accident is a good example. Certainly being injured in a skydiving accident is within my partial control. I can choose to go skydiving or not, and this affects the risk of a skydiving injury. By Prof. Hoppe’s earlier statement, this is not insurable. In reality, the fact that I go skydiving is considered as given. The act that is actually covered is the un-controllable part, i.e. the risk of injury given that I jump out of a plane.

In a free market, this would lead to separate classes. Those who go skydiving in one class, and those who don’t in another. Those who go skydiving would pay a premium based on risk of injury per jump. Those who promise to not skydive would probably pay zero for this insurance. In practice, this would probably mean that those who want skydiving injury coverage (i.e. those who intend to go skydiving), would pay an additional premium on some other policy.

Only *risks* can be insured. There must be uncertainty for something to be insured. A *choice* cannot be insured against. Things that involve both risks and choice can be insured, but only the risk part will really be covered. Either the thing will be redefined to exclude the choice component, or the choice element will remain but will be assumed to be worst case scenario from the insurer’s point of view.

In a hypothetical, totally free, infinite size, zero transaction cost market, all of the choice elements that could possibly be influenced by an actor will be removed from insurance coverage, leaving only the actual accident risk being insured at the lowest possible premium. This is probably Prof. Hoppe’s point, but it doesn’t come across overly clearly in the article. In reality, discriminating choice vs. risk to this level is not practical, and thus we will always be able to influence what we are insured against to some small degree, and pay a slightly higher premium as a result.

For all the good of insurance and its proper role, it isn’t some magic or panacea. It is more like any other business that relys on fad or some other unquantifiable risk to be profitable.

I think Hoppe’s point is clear if we take one more look at what he said about “class probability”:

“…”class probability” implies the absence of any systematic redistribution of income: If I know nothing about any particular person’s individual risk except that he is the member of some group with a known group risk, then all redistribution must be random. It implies also that the individual cases that are grouped into one risk pool are homogeneous. Within the group, we cannot tell the difference between one individual and another. This implies also that the actual event comes in the form of an accident — and unpredictable event for the individual.

“…sickness is insurable only insofar as the health risk for a particular group is purely accidental. Such is the case with certain forms of accident insurance, or even for events such as cancer.”

From this it seems clear that one can’t insure one’s self against one’s own intentional actions because the only events that can be subjected to a class probability are accidental events. And anything that can happen that is intentional certainly does appear to be ruled out as being accidental. Suicide, as well as alcoholism, and obesity all seem to be events or affects which have risks that cannot be grouped into homogeneous risk pools for this reason.

Furthermore, it does sound rather dubious that one can ever fall into a class where the probability of one’s failure in business can be known. That sounds more like something that perhaps creditors would attempt to discern on a very individual case-by-case basis. It doesn’t sound like a class probability kind of thing.

Finally, I think the article was pretty clear that one’s lifestyle can move someone from one homogeneous risk level pool to another: “To put an individual client into the right group, the insurer has to discriminate according to various criteria….They might use certain behavioral criteria or lifestyles: smokers and non-smokers, people who are occupied in certain occupations that cause greater or smaller risks, and so forth.”

Regardless of the risk pool one ends up in, in the end, it is only an accidental event (from the insured’s perspective) that will be insured.

Well rated life insurance companies have very strict underwriting requirements. They discriminate against moral hazards, behavior with in one’s control and perils, incidents beyond control.

Smokers, Skydivers and diabetics will all get rated by the insurer for indemnification against increased risk.

Suicide is covered after 24 months and MOST people do not want to kill themselves. This is one reason financial records and a credit report are required before you can be underwritten. Also, someone who has a history of psychological problems may not be offered insurance, especially disability income insurance.

Let’s not forget that some people may develop a condition which may prompt suicide after becoming insured. That perosn’s beneficiaries still need the death benefits and the actuaries have worked that into the cost of insurance.

There are some minor flaws in the application of insurance; however, these can be overlooked becuase the piece stands by itself as a sound analysis irrespective of these minor errors.

In a state of true freedom, not only would insurance be far better it becomes far more necessary to manage risk in an uncertain world.

Would that regualtions and licensing disappear, we may end up with far better risk transfer and risk management tools.

First, let’s state that *all* risks can be controlled intentionally, to a some degree.

So what the insurance covers is not the risk of the adverse event happening, but rather the risk of the adverse event happening given the expectation of the insured party’s behaviour.

This expectation is based on the insurer’s knowledge of the “facts of life” as well as of the specific circumstances. It also takes into account the moral hazard (and, oppositely, the fact that at least some people have morals).

If follows that someone intending to suicide can insure his life just fine – if the insurer can set premiums (and their frequency) high enough to make it likely to have a profit. I.e. if someone holding a gun to his head asks me to insure him for one million, I’d agree – with premiums of $3000 payable every millisecond, with the expectation that he won’t be able to blow his head away faster than in half-second.

I used to sell life insurence.

Now, I don’t know if it’s a regulation or not that pushes this, but it could be argued that while a person doesn’t feel suicidal now, it’s possible the demener would change in the future and this insures in part against that risk.

TRacy

I think the way to look at it is this: you can buy insurance in situations in which you control the probability of the event, but this probability has to be known ahead of time so the premium can reflect the risk and thus your choice. On the other hand, you cannot insure yourself against actually choosing to make the event happen. So you can get fire insurance if you choose to live in a dry forest, but you can’t insure yourself against choosing to set your own house on fire.

Sometimes it is hard to distinguish between accidents and intentional behavior — this is where moral hazard comes in. In a community in which people closely monitor each other, more things can be insured, because it becomes possible to detect moral hazard. This is why the state gets much more intrusive when it tries to “insure” us against health problems.

Read the quote you lifted from Hoppe’s article. He explicitly writes that one cannot insure one’s self. The wife and the employer are third parties. However, even the wife will find it difficult to purchase a policy that will pay if the husband kills himself within two years.

For example, let us suppose company A sends out an ad stating that it will sell insurance policies even to wives whose husbands kill themselves. What will happen? Some husbands, who plan to kill themselves, for whatever reason, will, with the aid of their wives as proxies, buy this insurance, whose initial premiums will be less than the payout, thus guaranteeing a profit to the wife. Because of this unintended consequence, the premiums will rise to adjust, but in the rising, wives whose husbands do not plan suicide will find cheaper policies that explicitly exclude suicide. Eventually the market for wives whose husbands plan to commit suicide shrinks down to the unsubsidized pool, which is economically nonfeasible, which is precisely Hoppe’s point. What is irrelevant is the fact that the wife could send the husband the bill. Why do you think life insurance routinely includes suicide related exclusionary riders?

Your example of an employer and employee is even more irrelevant. No one has written that someone can’t purchase insurance against the actions, intentional or otherwise, of others. That the employer may recoup the cost by reducing the wages of the employee is a non-point since the insurance was purchased by the employer for the employer’s benefit.

As to your contentions with my point about liability insurance, I will restate what I wrote: it is not possible to insure one’s self against the consequence’s of one’s own intentional torts against third parties. You may file such claims once or twice, but once it becomes clear that you are a serial offender, you will be uninsurable without the forced subsidy by the noncriminal, which is Hoppe’s point.

Auto liability insurers are required to pay on intentionally inflicted injuries by vehicle only because of the court-enforced public policy behind mandatory insurance laws.

Outside of this narrow exception, intentional acts are routinely excluded because liability insurance against intentional injuries presents an unprofitable moral hazard.

This is not to say that you could never find an insurer willing to underwrite you for your own intentional conduct, but the premium would be equal to the entire loss insured against. Since insurers compete for premium dollars from good risks, not bad risks, you would be effectively priced out of the market.

I saw a disturbing segment on 60 Minutes last Sunday. It showed how hospitals routinely charge 2-4 times as much for the same treatment to the non-insured as they do to those who are privately insured or covered by Medicare or Medicaid. I imagine most viewers came to the conclusion that this is a just a case of the free market gone wrong, how the people who can least afford a big hospital bill are getting screwed over. As for me, I am inclined to believe that we are seeing the effects of hospitals trying to compensate for the inadequate reimbursement rates dictated by the feds.

Your counter-examples apply to policies purchased by third parties, or involve outright fraud and/or mistake.

The distinction is irrelevant. You agree that a wife could insure against her husband’s suicide, since she has no (direct!) control over that, and would be hurt by it. Why don’t you agree that she could send him the bill? You agree that an employer could insure against employee vandalism. Why don’t you agree he could impute the cost out of wages, like what happens all the time?

In addition, your point about liability insurance is also not a valid one. It is not possible to insure one’s self for the costs of intentional torts against third parties.

False. In many businesses, especially construction, they do this all the time.

Hoppe is pointing out that you cannot, in a free market, insure yourself against acts that are within your partial control.

And Hoppe is wrong in this respect. I have partial control over whether I crash into someone’s car. I have partial control over whether my factory burns down. All of these are insurable.

False. The liability insurer would be expected to pay. Please, learn about actual insurance markets. This happens all the time. You might as well claim liability insurers never have to pay the claims of drunk drivers. Hoppe’s claim is not supported. Or rather, he’s right, except when he’s wrong.