What a Teen Driver Actually Costs — and Why Doing It Right Doesn't Lower the Bill

I went into this expecting to be rewarded.

Not lavishly. I wasn’t looking for a windfall. But we had done the work — all four kids well past fifty hours of supervised practice, and we paid extra to put them through separate defensive driving programs, the kind that let a teenager feel a skid or an emergency stop somewhere it can’t hurt them.

I assumed that would show up somewhere on the bill.

It did not. There was a small driver’s ed discount, which every teenager gets for enrolling in a course and which has nothing to do with whether they learned anything. Past that, my kids were priced exactly like the kid down the street who had done the legal minimum and not one minute more.

So eventually I asked

Some time later I found myself at an event with a room full of actuaries from several of the largest carriers in the country. Not the people who explain the prices to you. The people who set them.

This was eight or nine years ago and nothing about it was on the record. I’m recalling the substance of a conversation, not reciting a transcript. The substance has stayed with me.

I asked what seemed a reasonable question. Given what’s known about experience and risk, how do you think about a teenager with more and better preparation? Could that earn a lower rate?

The answer was direct, and I have thought about it ever since.

Pushing the cost of the risk onto the families who aren’t willing to put in the work isn’t feasible. The good have to carry the bad.

So I pushed a little. I brought up the numbers — how many teenagers die in a year, what those crashes cost, what a real difference in preparation appears to be worth.

The chief risk officer of one of those companies answered without hesitating. They had already factored in the casualties.

I want to be precise about why that astonished me. It wasn’t that the man seemed cold. He didn’t.

It’s that I had walked in holding what seemed to me an obvious proposition: if better-prepared teenagers crash less, then the industry with the most money riding on fewer crashes would be interested in producing more of them. It felt like the most logical thing anyone could say out loud in that building.

It simply was not part of the program. It had never been part of the program.

And here is what took me longer to accept: he wasn’t being callous. He was being accurate. That is what a loss model is. Somewhere inside the pricing of every teen policy in this country sits an expected number of deaths, and its presence there is not a flaw in the model. It is the function of the model.

The system isn’t broken. It is working exactly as designed. The design simply doesn’t have your kid’s life as the thing it’s trying to optimize.

They aren’t wrong, exactly

Insurance prices a pool, not a person. With an adult driver you have years of history, so the price converges on the individual. With a sixteen-year-old you have nothing — no record, no claims, no miles. So you price the category, and the category is “sixteen-year-old.”

There is also a verification problem I would have to concede if I were in their chair. Think about what that fifty-hour number actually is. The agency that asks for it — your state’s licensing office — does not verify it. No audit, no check, no mechanism. A parent signs a line attesting to a number, and that signature is the entire evidentiary basis for the requirement.

A form that says fifty hours might represent fifty. It might represent twenty-five. It might represent five. Nobody knows, including the office holding the form.

So when I ask a carrier to price my kid’s preparation, I’m asking them to price something nobody has measured. Any discount built on that number wouldn’t reward preparation — it would reward filling in a form, and the families most willing to round up would collect it first.

All of that is defensible. I would probably say the same thing in their seat.

What they measure, and what they can’t

Here is where I’d push back, and it isn’t about motives. It’s about measurement.

Everything a carrier rewards on a teen policy is a proxy. Grades are a proxy — a B average correlates with conscientiousness, which correlates with not doing something stupid at eleven at night. Telematics is the closest thing to real measurement in the business, and it’s a proxy too: braking force standing in for judgment.

Not one of them is a picture of how well your teenager actually drives. They are what’s available. I suspect everyone involved knows it.

So here is what a family can actually get a discount for today:

• Enrolling in driver’s ed. Not doing well in it. Enrolling.
• Grades. A B average typically earns something in the range of 8 to 25%.
• Going away to college without the car, once the school is more than 100 miles from home.

What appears nowhere on the list, at any carrier: how many hours your teenager practiced, in what conditions, or how well they actually drive.

And it isn’t that the industry doesn’t know. The most rigorous study we have — the one showing that structured, varied practice cuts a teenager’s on-road failure rate from 15% to 6% — was funded by one of the largest carriers in the country. That research is more than a decade old. It has not produced a discount anywhere, including at the company that paid for it.

Telematics is the industry’s real answer

To be fair, there is one place behavior does get priced: telematics. An app watches braking, speed, phone handling and time of day, and your rate moves accordingly. The savings are real.

Two things about it. It starts after the license, so it grades the driver you have already produced rather than helping you produce a better one. And it is monitoring rather than credit — it can raise a rate as easily as lower one, and it commonly penalizes night driving, which is precisely the practice your teenager most needs. A kid who works a closing shift can do everything right and score badly.

Read the terms before you enroll a seventeen-year-old.

What it actually costs

Adding a teenager to a policy runs about $3,600 a year on average — roughly $4,000 at sixteen, sliding down toward $3,100 by nineteen.

Where you live matters more than almost anything you do. In Hawaii, adding a teen raises a premium about 4%. In Maine, it can be 137%. Same kid, same preparation.

Four kids across two states is enough repetitions to say this with confidence: the number comes down with time and a clean record, and very little else you do moves it much.

What is in your control is smaller than you’d like and worth doing anyway: take every discount you qualify for and re-shop annually, check your liability limits before the license is issued rather than after, and understand what one at-fault claim does to a teen-rated policy over the following three to five years. I’ll cover each of those properly on its own.

And the thing that actually pays

It took me a while to see that I was looking for the wrong reward.

The insurance company was never going to pay me for preparing my kid. The premium is a bet on a population, and my household is one data point inside it. That number on the bill is not a grade on my parenting and was never designed to be one.

It would be easy to read all of this and conclude the deck is stacked, so why bother. But the insurance company was never going to help you, and neither is the licensing office — it is going to hand you a number and a form. Both of them sit downstream of a competent driver that somebody else has to produce.

That somebody is you. It was always going to be you.

And the return on those fifty hours and those defensive driving courses doesn’t arrive as a discount. It arrives as a crash that doesn’t happen — the one I’ll never know about, that never generated a claim, never showed up on a renewal, and never cost anybody anything.

That is an unsatisfying sort of payoff. It’s invisible by definition. You can’t point at it, and nobody sends you a letter about it.

It was also the only one ever on the table.