AI and Pricing

A recent headline read, “How AI is Redefining Insurance Pricing Strategies” followed by, “AI in pricing represents a breakthrough, with some insurers already shifting to automated solutions that promise more accurate risk assessment and increased profitability.”

If someone is working in an Ivory Tower and not in the real world, AI probably is the solution. Otherwise, this statement and concept fail in light of two realities.

The first is simply summed up by the old truism about the insurance industry as a whole, “Carriers cannot stand to be profitable!” This means that the people running insurance companies, aided and abetted by agents, are too competitive for their own good. They want to win the sale even if they lose money in the process.

And while no rational person wants to lose money on every sale, this industry is not exactly populated by rational people and/or the executives running companies are seeking to enrich themselves even as they damage their company. This is maybe a bigger reality in mutuals and small privately held stock companies. How else is it possible to explain an average net income loss of $1 billion per year over the last ten years? And that is for one carrier, not a collection of carriers, and the leadership has never been fired.

No one needs AI to solve that problem. They need a backbone and a better board of directors. Possibly better accounting too.

Also, in my experience consulting and sometimes arguing or being threatened with being sued by carriers, I’ve learned that a material portion of carrier executives do not truly understand how they make money, or more important, how different carriers make a whole lot more money than they do.

The industry average combined ratio is approximately 100% over the last ten years. But some carriers possess cost and/or revenue advantages that enable them to run higher loss ratios. Assessing risk is implicitly measuring risk versus premiums and when the carrier loses their bet, the metric is the loss versus the earned premium. This is the loss ratio.

Let’s say though that I have a ten-percentage point expense advantage, which a few carriers possess. Then all else being equal, I can run a loss ratio ten percentage points higher and achieve the same profits.

The carrier that lacks this expense advantage has a choice. They can compete at the lower price and lose money or they can lose the accounts resulting cumulatively in the loss of market share, more adverse selection, and a series of other losses. Historically, carrier after carrier has chosen to compete on price resulting in long soft markets.

In other words, the right price can be the same for both carriers, but if one’s expense ratio is lower or one has materially more investment income with which to subsidize losses, the right price for a risk becomes secondary.

The second reason the AI concept fails is because it does not consider what the competition is going to do. Let’s use a simple example involving risk management that is happening today. The smartest carriers figured out the importance of identifying which houses were most likely to withstand wind/hail claims. They focused on writing those homes at a lower price than their competitors who pooled all their accounts assuming all had the same risk characteristics. AI can definitely help achieve this, but plain old-fashioned street underwriting can do much the same.

This strategy though only works up to a point and that point is where the competitors catch on and begin doing the same. Not all carriers are led by people with adequate insight so some will fall by the wayside, but others catch on. At that point, the pricing for specific risk characteristics becomes a commodity. To win business, someone will cut rates and become unprofitable.

Entirely missing from these strategies is consideration of consumers, but more importantly, competitors’ price elasticity and how this affects demand and profitability.

Then there is the regulatory issue. By some assessments, for example, drivers likely to have accidents of $X severity are quite predictable. But carriers cannot use these assessments without restrictions because the public health would be so severely harmed. If the bad drivers were truly charged the correct rate, they likely could not afford any insurance. This would cause the good drivers to carry much more UM and the UM rates would increase significantly. For public policy reasons, regulators have determined it is better to subsidize bad drivers to some extent so that more carry at least minimum limits.

The same happens or will happen with property insurance because states depend on property tax revenues. The states need those values to remain high regardless of the insurability of the properties.

In commercial, the factors are less public policy and more of a consideration of what competition will do. Which is another reason why the AI theory only works in Ivory Towers. Having worked with carriers for 35 years, I have rarely seen a carrier’s executive team’s strategic plan specifically address what the competition is likely to do in reaction to whatever their plans are. It’s like they go into a game with a game plan that does not consider what the other team will do. What’s a strategic plan that doesn’t consider the competition? A worthless strategic plan.

Whether it is pricing, a strategic plan, or adverse selection, most carrier management appears to work in their own little world regardless of the tools provided. AI is not going to fix that problem.

____________________

Chris Burand is known for creating highly accurate forecasts of carriers’ future actions and the industry’s overall future course.

---

NOTE: The information provided herein is intended for educational and informational purposes only and it represents only the views of the authors. It is not a recommendation that a particular course of action be followed. Burand & Associates, LLC and Chris Burand assume, and will have, no responsibility for liability or damage which may result from the use of any of this information.

None of the materials in this article should be construed as offering legal advice, and the specific advice of legal counsel is recommended before acting on any matter discussed in this article. Regulated individuals/entities should also ensure that they comply with all applicable laws, rules, and regulations.