Freakonomics in the Times Magazine: Not-So-Free Ride

Dubner and Levitt have written a column for the Times Magazine‘s inaugural “green” issue of Sun., April 20. The subject: how auto travel produces a variety of negative externalities — pollution, carbon emissions, congestion, accidents — and how one new strategy may be able to help: pay-as-you-drive (P.A.Y.D.) auto insurance.

The strange truth is that most auto insurance in the U.S. is bought on an all-you-can-eat basis, which means it’s costless to drive additional miles, at least from an insurance perspective. But P.A.Y.D. insurance would change that.

Here is some of the research that went into the column:

1. Various writings by the economist Aaron Edlin on P.A.Y.D. insurance, including:

2. A draft paper on P.A.Y.D. insurance by Jason E. Bordoff and Pascal J. Noel of the Hamilton Project at the Brookings Institution, and an op-ed-style summary of the paper, written by Bordoff for the journal Democracy.

Also worth reading: a discussion of the Democracy piece from the Marginal Revolution blog.

3. U.S. auto statistics on accidents, miles driven, etc., from the Bureau of Transportation Statistics and the Fatality Analysis Reporting System, and a report on congestion costs from the Texas Transportation Institute.

4. An earlier blog post by Levitt, “Hurray for High Gas Prices!” that touched on the issue of mis-priced auto externalities.

5. Here’s the company that is about to roll out P.A.Y.D. insurance on a large scale.


Back in 1996, California briefly flirted with the idea of pay-at-the-pump no fault insurance. In one swoop, this would have gotten rid of uninsured motorists (if you gas up, you have insurance), accident-related lawsuits, and the externality associated with the standard all-you-can-drive plan. Overall, insurance would have been cheaper for everyone, and would become suddenly accessible to the working poor (no redlining!).

I thought it was a brilliant idea, but I was in the minority. Proposition 200, the initiative that proposed this, got pummeled at the polls.


Ironic how the inaugural "green" issue is coming out on 4/20


Didn't Andrew Tobias run a one-man crusade for this many years ago? I used to have his book. IIRC, he tried to get it on a ballot in California ...



I suggest looking at MileMeter ( and watch their video from the contest. They will begin offering cents-per-mile insurance this summer. And they have done it all using private (no taxpayer) funds.


Nuclear Mom

Well, I am one of the elusive drivers who believe they are worse than average. Although I have not had any accidents, I am not particularly comfortable behind the wheel and don't think I have good judgment, therefore I drive as little as possible. In fact, I telecommute, hate to shop, order online, etc. I am lucky to have a husband who LOVES to drive, is a good driver (with several not-his-fault accidents, BTW) and drives all over for our vacations.

I love the PAYD idea at the gas pump. Like universal health insurance, it removes the headache of dealing with the economic ramifications of the uninsured. I think people who cause accidents, or perhaps who are driving under the influence, etc., can be dealt with with license suspensions, as is done now for really bad drivers.

Good point about the older drivers. I expect to be an even worse driver as I enter my golden years. But you are going to have to pry my car keys out of my cold, dead hands before I give them up, because the public transportation sucks or is nonexistent in Orange County, California. I loved taking the bus & subway when I lived in Boston, but can't imagine taking forever to take a bus to the grocery store here and schlepping all my groceries back.

There are those of us who would happily program in a destination, put our car on autopilot and give up driving control when that glorious technology arrives. Hurry up!



Mark, I do think that if Zeke reduces his mileage that he'll become a worse driver. How are your driving skills after even an all-too-brief vacation during which you don't drive? It takes a little while to get back in the rhythm. People who only drive in their neighborhood aren't as skilled at driving in more challenging surroundings (unless their neighborhood is rough). And the only times I have ever feared for my life while riding with friends is when I ride with people who drive as little as they can get away with. (I grant that they may drive little because they know they aren't a good driver.) I assert that the overall risk rate per mile isn't uniform. That said, someone who only drives in their neighborhood is going to me most safe there because all they have to watch for are differences. The weird turns or yields are routine for them, so they really only pay conscious attention to the cat running in front of them that a driver new to the area might not see.

I would assert that most people who drive a lot get better and their risk rate is lower per mile-hour than someone who drives very little. The people in the middle of the bell curve are probably all about the same, though. There is the "random" risk factor that should be equal for everyone, but I see a lot of "random" accidents or near accidents in other cars that I handily avoid due to paying attention to situational awareness and having more response options than hitting the brakes.

The bigger problem with PAYD is that it doesn't address the fact that suburbs are set up to require driving and insurance agencies don't pay for environmental damage. I, myself, chose to live in the one walkable neighborhood in my whole town where I can conveniently walk to food and services. Everyone else must drive. The pollution remediation is done by taxes, not inusrance agencies. The health negatives of sitting in congested traffic are paid by health insurance, not auto insurance. I also chose to live somewhere with a short commute to reduce fume breathing, but the tradeoff is a much smaller living space for the same money. How about charging people who commute more than 30 minutes during rush hour more money? They're the ones in stop and go traffic, idling their engines for no good cause.

Reducing the number of miles people MUST drive to live an average life, and ticketing people for obstructing traffic rather than for speeding would do a lot more to alleviate congestion, pollution and accidents and would ultimately reduce hours on the road. Charging people more for miles they have to drive anyway is already done through gasoline taxes.


Michael Replogle

I'm a civil engineer with three decades experience in transportation planning and policy. To my mind, getting widespread implementation of pay-as-you-drive insurance is the single most effective thing that America can quickly do to reduce traffic and pollution from driving. It would save the majority of households money, cut dependence on oil imports, reduce accidents and health costs, and provide early action to address global warming.

It's a big step forward that insurance companies, like Progressive, are starting to introduce products that offer deeper discounts based on how much and how people drive. Driving fewer miles and driving more calmly, with fewer quick stops and starts, cuts fuel use, pollution, and accidents. If all car insurance premiums were put on a pay-as-you-drive basis, it could cut greenhouse gas emissions by 8-15%. That's huge. But getting this result will require government action and incentives to overcome the inertia of the insurance marketplace and state insurance regulations.



Mileage-based insurance is an excellent idea; however, giving taxpayer funds to insurance companies to provide mileage-based insurance is asinine.

The insurance companies clearly have a large economic incentive to create a mileage-based product on their own. Giving them taxpayer money only dampens the effects of an open and competitive market.

Case in point:
The referenced Brookings draft paper states "In 2006 Progressive was awarded $1.3 million in Federal, state, and local funding for a mileage-based (but not true PAYD) pilot in North Central Texas." Yet only 93 data points were collected that could provide mileage-based analysis. So our government gave $1.3 million to a private company to give us a tiny data set that borders on being statistically useless. And we have no mileage-based insurance product to show for it.

Rather than spending taxpayer money on handouts to big insurance companies that give them an unfair advantage relative to their competition (as well as potential new companies), perhaps Freakonomics/Brookings could argue for more open-market reforms in insurance? Say, by reducing the regulatory burdens that stifle innovation and prevent new ventures? If the economics justify "PAYD", then the financial incentive will cause insurance companies to offer the product.



What this fails to take into account is that drivers who drive very little tend to be terrible, timid drivers, which puts them at higher risk being in or causing an accident. Driving is something that gets better with practice. When you're no longer panicking about the mechanics of merging and just do it, you can spend the extra brain power on situational awareness.

Also, it's not clear to me that the monitoring plan would be able to tell good drivers from bad ones. My approach is to maintain a consistent speed as much as possible. I routinely get stuck behind people go go 65-75-65-75, so I add a burst of speed to get around them, then go back to my comfortable steady pace. Would I be penalized for that traffic? And if my insurance company knows that I drive over the speed limit, will they be required to ticket me, even if the behavior is justifiable?

Also, it's not just mileage - it's hours on the road. I used to drive 2.5 miles in 15 minutes to get to work in Boston. Now in SoCal, I drive 7 miles in 15 minutes. My car is going 3X as far, but using my skills for the same amount of time, and burning through fuel more efficiently. I gained about 4-5 mpg in the commute moving from stop and go surface roads to freeway.

I guess you can tell that I think the PAYD plan is too simplistic and is a bad idea. I especially think bulk monitoring of individuals' driving from a distance is horrific because most of our laws don't encourage safe driving but congestion.


Mark Brucker

Erika gets one thing wrong-low mileage drivers who aren't good ARE penalized for that, but high mileage drivers now, even if they're bad, do NOT pay for the added risk they create. A high mileage bad driver pays only as much as a low mileage bad driver, but creates far more risk.
What she and Stephanie are arguing for is something similar to saying everyone should pay the same for home insurance, regardless of the risk they create. If Amy is a really good driver and covers 5k miles a year, she shouldn't pay the same as Zeke, who is a good driver but drives 5 times as much. If their risk per mile is the same, Zeke will average a wreck 5 times as often as Amy. But another thing I think Erika is wrong about is that she apparently assumes that if Zeke reduces his mileage that he'll become a worse driver. That's unlikely. Mileage-based insurance would encourage everyone to drive less, but that biggest incentive would be for the bad drivers, so roads would get much safer.


devils advocate

Random note for the sake of argument with reference to the above comment:

"let's first look at those who truly do create a greater burden to society"

Speaking in the long term, as baby-boomers reach to retirement and beyond it is logical to expect them to keep driving and to have a reduced ability to do so safely (eye sight, motor skills, reflexes, etc.).

How is there no mandatory driving test for people over X years? Is there data for the likelihood of elderly drivers to be in accidents?

Personally, I would prefer to have a 30 year old with 3 beers in them on the road over a 90 year old with arthritis and awful eyesight any day.

Also, old people love voting and I expect them to defend their ability to drive without testing and passage of such a law is unlikely, but that doesn't mean it's not a good idea.

How about differential insurance rates for drivers who voluntarily take a driving test above a certain age? I could see Geico pulling this off before any policy led initiative getting anywhere.

Just throwing out ideas, but let the outrage ensue :)

Looking forward to any comments


Mark Brucker

Great column. They note that they address only 3 externalities-congestion, collisions and carbon emissions and report estimates of $300 Billion a year. Numerous studies have attempted to estimate other externalities. E.g., some have estimated that there are 6 parking spaces for every vehicle, with an estimated cost of another $300+ Billion a year.

Charles D

Excellent idea. The 'all you can eat' policies lead to so many problems with overconsumption. I'd love to be able to save money if I drove less.

Marc Brodeur

One thing is that this would likely be implemented with GPS or something similar. The privacy advocates would be up in arms about that one.

Otherwise, great idea.

To Stephanie #5, There are many things the insurance companies would take into account. Your driving record would count in your favor, but the high mileage would count against you (more chances for random bad stuff to happen). And do you not already tell the insurer how many miles you drive in a year?


I really like this idea. If it didn't save me $ on auto insurance, I think it would at least save me $ on gasoline - b/c I'd be constantly monitoring my mileage & would inevitably cut back.


As someone who puts many miles on my small car (mainly because I live 4 hours from my hometown and return frequently to visit friends and family there) and has /never/ caused an accident, I'm appalled at the insinuation that my insurance should be greatly increased because somehow I am more likely to cause accidents. I have always carried car insurance on my vehicle and I certainly pay more for gasoline than many people since I travel. I also rarely drive to places within walking distance (going out to eat, etc). I see no reason why I should be considered a "high-risk" driver and be charged higher insurance premiums when I am obeying all laws associated with driving a vehicle and have never caused an accident.

If you insist on creating greater burdens, let's first look at those who truly do create a greater burden to society - those who drive without insurance, those who continually disobey traffic laws, those who drink and drive, those who cause accidents, etc.



I didn't read all the comments so I hope this hasn't been brought up before, but aren't most accidents within a short distance from the home and normally at lower speeds? Wouldn't that imply that it's people on short trips, since they happen far more frequently, who drive less that are in accidents? PAYD would simply just penalize people who drive on the highway to commute, which statistically is safer than city streets if I am remembering correctly. It just seems to not make any sense to me.


The P.A.Y.D Insurance sounds interesting, I am going to check out all these links. I have never heard of this before so thanks for writing about it.



Last year, Progressive offered me a 5% discount if I would report my odometer readings to them. I took them up on it, gave them my initial reading, and since then they've only asked one more time, about six weeks before renewing our policy.

As I usually bicycle to work three to four days a week year-round, mileage on my car is pretty low. And they rewarded us with slightly lower rates for the next year.

I'd love to see insurance paid for at the gas pump myself, as a per gallon charge. Then everyone would have insurance, and perhaps America will realize they don't need a Hummer to drive to work by themselves.


"how one new strategy may be able to help"

What a nice idea! Forget providing fast, safe and convenient public transportation. Just set the insurance rates so high that poor people can't afford to drive to work. Not like we need their kind moving around anyways.