How Did Insurance Get Priced Before Computers?

WV & OH ONLY

"Insurance used to seem simple. How did companies figure out what to charge for car insurance before computers?"

Plenty of folks miss when this was simpler. Honestly, so do I.

Before computers, car insurance was priced from a few facts on a paper rate card, starting with where the car was kept and what it cost, and later adding the driver's age and how the car was used. Actuaries worked those groups out by hand. A company could only charge for what it could collect and count on paper.

I wish it were still that simple and that cheap. Since it isn't, we work to get every client the best rate there is.

The old card was simple enough to read across a desk. Today every company has its own way of pricing you, and you can't see inside it. We can run your real information through 13 companies, which beats trying to figure out any one company's math.

My motive's no secret. I want to quote you. The history is fun. Paying more than you need to isn't. Let me check where you stand today.

WV & OH ONLY

Very little. Casualty actuary James Lynch wrote a history of rating variables for Contingencies, the American Academy of Actuaries' magazine. The earliest card in it, from 1919, covers private passenger liability and uses two things: territory, and four groups based on the car's list price.

Where it was kept and what it cost. That was the whole system.

They cared. They just couldn't measure it. Lynch quotes a 1922 actuary who called the driver "by far the greatest hazard" in auto underwriting. But nobody had a reliable way to tell a good driver from a bad one.

Companies had to record everything on paper and do the calculating by hand. If you can't gather a fact on a large scale, you can't price it.

There was an early try. According to Lynch, a 1929 plan gave a discount to drivers with two claim-free years. About 85% of drivers qualified, so it told companies very little about who was actually better. It was withdrawn in 1932.

I think about that one often. A discount nearly everyone gets isn't really a discount.

The numbers came first. Connecticut data from 1932 to 1936 showed drivers under 25 had 37% more accidents than the average driver, and 62% more fatal ones. By 1939, Connecticut had sorted cars into classes: low-mileage, non-business cars with no drivers under 25 in one, other non-business cars in another, and business vehicles in a third.

More than you'd guess. Gas rationing stickers limited how much people could drive, and Massachusetts data showed those stickers did a strong job of predicting losses. They were a stand-in for miles.

Right after the war, mileage looked like it mattered and age didn't. Lynch argues that was misleading, because millions of young men were away in the service. When they came home and started driving, age went back to predicting claims.

Massachusetts had companies collect age and expected mileage on signed questionnaires. It was slow, paper-heavy work. It also ran into a problem any agent could've predicted: people aren't very good at estimating how far they drive, and the self-reported mileage lost its value.

Records got easier to reach. The Federal Trade Commission's 2007 report says insurance companies started using credit-based insurance scores in the mid-1990s, and that use grew very rapidly. Now a device can report distance and speed directly.

Look at what stayed, though. Where the car is kept, what it's worth, who drives it and how far. People were arguing about those same questions a hundred years ago. There's just far more data behind them now, and companies don't all use it the same way.

For some people, yes. For others, no. A simple card puts a careful driver and a careless one in the same group, so the careful one helps pay for the other. Whether you come out ahead now depends on which company you ask.

A hundred years of arguing about price, and it still comes down to who you ask. Let me ask them.

WV & OH ONLY

Next in the series: Part 10, What Do Insurance Companies Check When I Ask for a Quote?