To understand the manner in which changing tort law has created a crisis in the insurance industry it is necessary to review the foundations of insurance underwriting.

Insurance pulls together a large number of risks in order to create a certain predictability. For example, a single homeowner may know that the odds are that only a few houses in his neighborhood will catch fire, but he cannot know whether or not one of those houses will be his. Thus he must either set aside enough money to rebuild his house in the case of a loss, or risk a total loss with no reserve. Saving individually, the homeowners would have to set aside a sum equal to the total value of all their homes. However, by combining with neighboring homeowners, each owner need only contribute an amount sufficient to assure that the total will cover the few expected losses, regardless of which owners suffer those losses.

While most people focus on the concept of large numbers, it is the predictability created by large numbers that is vital to the insurance function. That predictability comes not just from the law of large numbers but from the relative stability of the sources and causes of risks, in this case fires. A sudden but uncertain change in the statistical probability of fires would leave uncertainty as to how much money must be in the pool to cover losses. The prudent course for our group of homeowners, then, would be to charge each owner a bit extra to cover these possible extra losses – what might be called an "uncertainty tax."

But our owners face another problem. Some of the houses in the neighborhood are made of wood. They have old wiring systems and occupants who keep gas cans and old rags in the basement. Other homes are brick, with new wiring and tidy owners. If each homeowner is charged the same price, the brick owners, who are less likely to suffer a fire, end up subsidizing the wood home owners, who are at greater risk. Some of the brick owners will decide this cost is not worth bearing, and will choose to drop out of the risk pool (or they may join other good risks and form a new insurance pool). With the better risks leaving the pool, the remaining, poorer risks, must pay a bit more. This will encourage the best remaining risks to leave the pool, and the cycle repeats until the original pool falls apart. This process is known as adverse selection.

The insurance function controls this process by subdividing the large risk pool back into smaller pools. Lower risks (the brick owners) will pay lower premiums, while higher risks (the wood owners) pay higher premiums. If, for some reason, the insurance function is not able to break the larger pool back into smaller pools of similar risks, the better risks will either cease the activity (in this case, it is unlikely they will cease living in homes, but they may move to different neighborhoods, leaving only higher risks individuals in the original neighborhood) or they will find a method to self insure, i.e. cover their risks themselves. Either way, the original pool breaks down. If an insurer is required to charge each customer the same amount for insurance, adverse selection will be the inevitable result, with a corresponding collapse of the insurance market and rapidly rising premiums. Yet, as we shall see in Section IV, this is precisely what the new tort law requires insurers to do.

The insurance function also relies on risks being independent of one another. Suppose, for example, that a fire in one house automatically caused a fire in every other house in the neighborhood; any one fire would result in a total loss to all homeowners. Each homeowner would be back in the original position of having to insure the total value of his home, since the total probable loss would equal the sum of the value of all homes. There would be no way to spread the risk. Looking at it another way, it is important that the outcome in one case not affect the probability of accident for the other risks in the pool. Inter-related risks such as nuclear war cannot be insured against, as the purpose of insurance is to group together independent probabilities to predict losses statistically.

The insurer must also guard against what is termed "moral hazard" on the part of the insured. That is to say, the insured should not be able to benefit from his own "hazardous" behavior. Thus, fire insurance on a home is not written for more than the home’s value, lest the insured benefit from the destruction of the home. Likewise, health insurance usually includes deductibles and co-pays to give the insured an incentive to minimize expenses. A fundamental precept of insurance, then, is that the accident victim must have incentives to take precautions on his own.

Two other aspects of products liability insurance must also be kept in mind in considering the causes of, and solutions to, the liability insurance crisis. First, although insurers sell their products liability policies to manufacturers, the true insured risk pool is made up of consumers. In other words, liability insurance, unlike, says business interruption insurance, ultimately pays benefits to injured consumers, not the manufacturer holding the policy. Thus an insurance company underwriter, or a self-insured manufacturer, is attempting to guard against moral hazard, risk interdependence, and adverse selection not only in the products and manufacturers insured, but in the consumers using those products.

Second, in dealing with the effects of judicial decisions on products liability insurance, it is important to remember that rates for products liability are based on nationwide premium and loss data. This occurs partly because there are fewer state regulatory peculiarities to be met than in products such as workers’ compensation insurance, and partly to assure an adequate sample size from which to predict losses. However, the more obvious reason for this is that in a modern economy almost all marketing is done across state lines. This means that products liability cases may frequently be brought in multiple states. Thus rates in Michigan will be affected by legislative and judicial decisions from California and New Jersey, and vice-versa.

Yet liability changes at the state level remain important. Changes in Michigan may have a small but direct beneficial effect nationwide. Further, the pattern of law is such that changes in one state frequently wash over to set judicial precedent in other states. And, despite the growth of interstate commerce, Michigan corporations are more likely to be sued in Michigan – in particular, small businesses whose role in interstate commerce is limited can realize gains from action in their home state.

In Michigan, there is much that can be done to reform the tort system and restore liability insurance markets to good health.