The Competitive Advantage: Who Needs Insurance?

Who needs insurance? Not the perfect risk, other than for contractual and legal reasons. The perfect risk has absolutely no chance, not even remotely, of incurring a loss. If an account has almost no probability of incurring a claim, it does not need insurance. If predictive modeling identified the extremely best risks, then why should those consumers even purchase insurance?

Let’s use health insurance as an example. If DNA tests show person A is super healthy and is almost certainly likely to remain healthy for a long time, then why should person A buy health insurance? Maybe they should purchase accident insurance, but not health insurance. Or if multiple health insurers have this person’s information, then any policy written on that person will be profitable 99.9 percent of the time if the premium plus investment income exceeds the expense ratio. If insurance companies strive just for healthy people, and those likely to remain healthy, premiums will decrease by at least 50 percent, probably 75 percent, and the companies will cease to exist (unless it creates a monopoly, i.e., universal care or convincing healthy people they are not healthy).

Therefore, if the consumer knows insurance is not necessary, premiums decrease to $0. If the consumer does not know but multiple carriers know, rates decrease by 75 percent barring collusion. Either way, the company ceases to exist because it is no longer provides a service that has value.

Take person B. DNA shows person B will have a horrible illness requiring $200,000 of medical treatment in the next two years. Need for insurance exists. However, at normal premiums will the insurance company make money? No. At actuarially sound rates for person B, will person B likely be able to afford insurance? No.

Probability Near 100%

With DNA, predictive modeling, data-scraping and the multitude of data points combined with sophisticated algorithms, these scenarios are theoretically possible to realize today in life, health, auto, home, commercial auto, and workers’ compensation. For some lines, the probability is nearly 100 percent. In others, it may be as “low” as 50 percent, but the impact is high. What happens to insurance?

People do not need it, or they need insurance so much that it is not affordable. The middle is lost. One might argue most people will be of middling health (not true) and middling use of auto and losses will remain random, but random is just part of the equation. Going back to day two of underwriting school, the accident might be random, but to whom the accident happens is not necessarily random. The difference is monumental. Let’s assume though all is random. What happens if people likely not to incur a claim only need catastrophic insurance for that random event? The same result: very low premiums.

Round-and-Round

Insurance goes round and around because of uncertainty. Insurance is designed to protect against the uncertainty of mildly negative to horrendous events. Part of insurance is designed for frequency and part is designed to protect against full catastrophe. The more uncertainty is eliminated through knowledge or math (knowledge and math are not inherently the same), the less insurance is required.

I remember hearing a televised fundamentalist minister preaching many decades ago that knowledge was sin because it eliminates the need to pray. I am not here to make a religious point, but I am here to make an insurance point. A key provision in many insurance policies is coverage for acts of God. If acts of God are predictable, therefore, avoidable, manageable or too expensive to manage but known, then insurance’s role is extinct. Insurance is for acts of God because inherent to acts of God is unpredictability.

Too much knowledge eliminates uncertainty. Too much predictability creates a scenario like the Tom Cruise movie, “Minority Report.” In that movie, statisticians developed algorithms that predicted which people will commit which crime slightly prior to committing the crime. Future underwriting might be cancellations right before a claim occurs.

These are issues for the industry. At the individual company level, the situation is more complex because if one carrier develops much higher quality predictive modeling algorithms than its competition, it can take advantage of its slower peers. It can also take advantage of unsuspecting consumers. Such a company can write the risks that possess almost no chance of loss but charge more. The consumers thinks they are getting a great deal when they are just paying more for less. The out-of-date competitor thinks the company’s rate is too low and the company will soon go out of business like so many predecessors, but that will not happen if the algorithm is of high enough quality.

These advanced companies have an opportunity to make a lot of money not really out of brilliance, but by taking advantage of ignorance. Their window will not last, but neither will the traditional model. The industry is not sure the exact outcome, but the probability is high that rates will decrease substantially in many lines with partial offsets in other lines. For example, auto rates likely will decrease significantly, especially for the best drivers, but UM/UIM rates will likely increase because more marginal drivers will not find auto insurance affordable. Some aspects of life and disability will change to catastrophe or accident insurance for people identified as super safe.

Woe will befall the late adapters because adverse selection will find these companies. With less need for insurance, the market will shrink and fewer carriers will be necessary. (The market already has too many carriers given that approximately 10 carriers have 50 percent of the market share by net written premium, 90 carriers have 87 percent of the market, and the remaining 800 to 900 carriers are fighting over the scraps of 13 percent, based on A.M. Best data.) The slow companies are more likely to become insolvent, especially if they are too slow to sell themselves versus companies that are slow to adapt but quick to realize they fell behind and now need to sell.

Good News?

What is the good news? I am not typically a fan of regulators, but insurance commissioners have a vital role to play here. Finding the balance required for the public’s, carriers’, and agents’ best interests and technology will be hard work. We may need regulators to save the industry from too much knowledge and leave the essentials of insurance to acts of God. The more uncertainty, the greater the need for insurance. The less uncertainty, the less need, and I am not confident needing less insurance is good for society, at least not on a predictability scale. If this was due to safety innovations, then less insurance is better — but not on a predictability basis because inevitably, Mother Nature wins no matter how confident software developers and statisticians are of their models.