Why ‘just-in-time’ insurance would be a non-starter

A few times a year I get approached by someone asking the question: Why not allow consumers to purchase insurance just before a foreseeable loss is about to occur or immediately after it has occurred?

Such suggestions always come from folks who know virtually nothing about how insurance works and, thus, tend to be unrealistic and unworkable. Few – if any – truly understand the full implications of what it is they are proposing. (This partially highlights the problems that can – and do – arise when insurers aren’t at ‘the table’ when other groups are discussing matters that are of interest to them.)

Recently, I received an email from a friend who works in the area of resilience and business continuity:

I am in a strategic planning session and [the] topic has come up around feasibility of business interruption insurance that could be purchased on a just-in-time concept (i.e. just before the hurricane arrives). Are there policies out there being underwritten against this kind of peril? I’m thinking the premiums would be astronomical.

My response to him described the impossibility of operating under such a business model. Essentially, selling insurance on such a basis would not be doable, I explained, “not if the insurer wishes to be in business for any length of time.”

But before I discuss why this would be the case, it would be useful to take a very quick look at the conditions that must be in place for viable insurability of a risk. In the Swiss Re publication ‘Floods: An insurable risk’ (1998) Hausmann offered the following:

1. Mutuality: A large number of people must combine to form a risk community.
2. Need: There must be a need for insurance cover when the anticipated event occurs.
3. Assessability: The peril must be assessable in terms of possible losses.
4. Randomness: The event must be independent of the will of the insured, and the time at which the insured event occurs must not be predictable.
5. Economic viability: The risk community must be able to cover flood-loss financial needs.
6. Similarity of interest: The risk community must be exposed to the same threat and the occurrence of anticipated damages must result in the need for funds in the same way for each member of the community.

With these six considerations in mind, here are the key reasons why just-in-time insurance would be nothing less than a disaster in the making.

First, giving consumers the option of buying coverage just prior to or immediately after a loss could raise the spectre of moral hazard (i.e. where a party involves itself in a risky venture knowing that it is protected against the fallout because another party will incur the cost, monetary or other). Essentially, people might take undue risks if they knew that they could purchase insurance at the last minute – or shortly after the loss – in order to cushion the blow (kind of like how one might try a risky tee shot in a game of golf because they have a Mulligan as a back-up).

It is true that the existence of traditional insurance can, too, create moral hazard (eg. a person may take greater chances while driving a car knowing they have coverage for collision or may choose to build on a flood plain because they can easily purchase flood insurance). There has been a fair bit of academic research that has looked into this area (some of this work has found that the presence of reinsurance can have essentially the same effect on the risk-taking of primary companies). However, giving a person the option of purchasing insurance on a just-in-time basis would mean that the insurer wouldn’t have the time to properly analyze and understand the risk to be insured, and the potential losses associated with it. This speaks to Hausmann’s third and sixth requirements of insurability, Assessability and Similarity of threat.

Essentially, an insurer’s underwriting process allows a company to vet the risk, calculate the cost of coverage and the amount of the deductible, determine what should and shouldn’t be included in the coverage, and decide whether the company even wishes to take on the risk in the first place. Evaluating a new piece of business as a wildfire is coming over the top of the next ridge wouldn’t give the insurer the time to do its due diligence. This underwriting/vetting process, when done properly and ahead of time, is important because it can have the affect of tempering risk as the insurer may agree to take on a piece of new business only if the insured performs certain risk mitigations. Further, the insured may take it upon him/herself to temper risky behaviours knowing that they are under increased scrutiny by the insurer.

Second, most insurers suspend the issuance of new policies in a given area if a loss appears to be imminent. Therefore, you wouldn’t be able to buy a new homeowners policy if a wildfire was at the top of a nearby hill and threatening to burn down your house or if a hurricane was due to hit your area the next day.

A homeowner enters into the purchase of an insurance policy with no particular knowledge other than the determination that he or she is not willing (or able) to absorb the cost of a big loss to his or her assets (of course, they may also enter into an insurance agreement because their mortgage issuer has mandated it). Insurers, on the other hand, enter into the process with a greater degree of knowledge, about past claims experience of the insured, hazard and loss experience in the immediate geographic area in which the insured lives, and with some data of the expected probability of an insured peril taking place during the course of the policy period. Yet, there is still some degree of gambling that is taking place, with the insurer making more of an educated guess than the insured. Issuing a policy for what appears to be an imminent loss removes chance and replaces it with certainty. This speaks to Hausmann’s fourth requirement of insurability, Randomness.

What possible motivation would an insurer have to enter into such an agreement? Why would a company place its money on red when it knows for certain that the spin of the roulette wheel is going to come up black? Again, an insurer who opts for this model wouldn’t be in business for long.

Third, and further to the comments just made, insurance regulation in Canada requires that in order for a contract of insurance to be issued, clear risk transfer must take place. Again, if one is replacing chance with certainty, then no risk transfer would be taking place. This speaks to Hausmann’s first requirement of insurability, Mutuality.

Fourth, one of the cornerstone concepts of insurance is the idea of unforeseeability. Essentially, insurance generally only indemnifies against losses that are sudden and accidental in nature (though there can be exceptions for grey areas). Again, and further to comments made above, issuing a policy shortly before an insured peril is about to cause a loss (or has caused a loss) replaces that which is unforeseen with that which is foreseeable, even known. So the insurance policy stops being an instrument to indemnify against an event that is possible but whose time and place is unknown, and becomes an instrument to indemnify against the known. This also speaks to Hausmann’s fourth requirement of  insurability, Randomness.

Fifth, because consumers could purchase insurance only as a loss was bearing down on them, the risk pool would consist only of those about to make a claim. This ‘adverse selection’ – where only consumers that are expecting a loss would purchase coverage –would mean that the risk community would be too small to draw upon to cover all of the impending losses. This would likely mean that premiums would need to be exceptionally high, making them out-of-reach for many and possibly negating the very need for insurance. This speaks to Hausmann’s first and fifth requirements of insurability, Mutuality and Economic viability.

Finally, and perhaps most importantly, insurers have to build a bank over a medium to long term in order to pay for losses. Further, they must do so by collecting and investing the premiums of the many to pay for the losses of the few. Along with allowing the company to collect enough capital to pay for claims, this model also serves to keep premiums at reasonable levels and ensures continued availability of insurance capacity.

If we replaced this model with one where consumers were able to purchase insurance only immediately before or after a loss, premiums would have to be very high in order to make up for the absence of capital reserves that would normally have been built up over many years. The insurance would be too costly for most, and many insureds would be forced to absorb any losses they incurr. This also speaks to Hausmann’s first and fifth requirements of insurability, Mutuality and Economic viability.

So a business model centred around a ‘just-in-time’ method of insurance delivery would serve no one.

But insurance companies have lots of capital to pay claims, argue some. But they wouldn’t if such a model were adopted. No one would place money into the risk pool until they needed to. The pool would never be big enough to cover all losses because it would only be added to by those who were about to take out far more than they just put in. Like a well that was drawn from but never replenished, what was there would be quickly depleted and the insurance company would soon fail.

The current model under which the industry operates has been around for a few hundred years. And while it is not perfect, it works – quite well. Though new risk transfer mechanisms have risen over the years, most of these instruments are not designed – or intended – for use by the average insurance consumer.

And while there must be some kind of improvements that can be made to the current way that insurance is structured and delivered, a just-in-time model would most surely not be one of them. It couldn’t be.