Jul 11, 2009

Mileage Tax

Last year around this time, I wrote a blog post suggesting we crank up the gas tax to $10/gallon while using the entire proceeds to give tax rebates to low income families. The reaction was generally plenty negative, although at least one person told me in person that he thought it might have some merit. No worries.

It turns out that there is a new idea for taxing vehicles on the horizon, one that is only recently possible due to the proliferation of inexpensive GPS technology. A tax based directly on mileage driven. I've found that very few people in the US have heard about this possibility, and most reject it at first glance without deeper consideration.

There was an interesting article last weekend in Reuters about Traffic jams in Manhattan.  In it, the author posited that the average car driven into Manhattan on a weekday causes 3.26 hours of delays to everybody else.  Those delays are felt by individual drivers in terms of a few seconds, but across all of the drivers combined, the delays are very high.  The article goes a bit further to claim that the average car has 1.97 people in it and the average per-car value of that lost time is $48.89/hour, so the societal cost of that average car being driven into Manhattan is as high as $160.

In major cities, many road networks have similar properties to a lesser scale than Manhattan.  Today I drove through San Francisco with no intent to stop in the city, but traffic slowed me down to a much lower speed than I was happy with.  The traffic alone added at least 30 minutes to my journey.  Ick.

A mileage tax at it's simplest adds a device to your car which tracks how many miles you've driven via GPS and charges you a tax based on this.  There are obvious reasons why this initial idea would have people roll their eyes.  The first is usually that a gas-tax has much the same affect and is cheaper to administer.  The second is that it seems to penalize folks who drive more fuel efficient vehicles (or at least not penalize gas-guzzlers).

The important part is that a mileage tax can be far more complex than this.  Instead of a flat tax per mile driven, the tax can vary based on where you are driving, what time, and the type of vehicle you drive (ie: a discount applied to lower-emissions vehicles).

Also, a small premium can be charged for driving in San Francisco during rush hour.  By charging say $1 to drive through San Francisco at lunch time, those willing to pay will have a shorter, faster trip because others are incented to use alternatives and get off the road.  There would be no need for your unit to communicate where it was and when (violating your privacy), it would simply need to have rules for different areas and calculate the tax internally.

A variable rate mileage tax could also nicely solve the problem of toll roads.  Toll roads in the US today are often built with the argument that state X shouldn't have to pay for the road that people in state Y are going to make use of for free.  Simply collecting the tolls has a cost though.  WSDOT did a study in 2007 (See Toll Roads Mean Billions in Extra Costs for Motorists) that indicated for every $100 in revenue raised by toll roads, there was an additional $22 overhead in collecting the tolls themselves.  This also doesn't include the drivers' lost time.

A mileage tax isn't a theoretical idea that can't be easily implemented in practice either, it is already in use in some places.  Last year Holland passed a law that will phase in a mileage tax over several years with variable rate pricing (See: Road Pricing).  In 2006, Oregon recruited 299 drivers for a year-long trial of a mileage tax (See: Oregon's Successful Mileage Tax Experiment).

Obama has however said he is not interested.  However, this may become a looming issue over time.  As fuel efficiency increases and drivers react to higher gas prices, transportation budgets have been feeling serious shortfalls.  It is politically difficult to pass or raise new taxes, which makes regularly updating the gas tax to match the transportation budgets sporadic at best.  The time hasn't come, but perhaps it someday will.


Ryan Moulton said...

A mileage tax at its simplest doesn't involve a gps at all. They would just check your odometer when you renew your registration and charge you for the difference. Variable pricing based on location seems impractical. You'd have to update a $100 device in everyone's car every time the law changed unless it phones home.

Gas taxes are simpler on every front. If you need policies to reduce urban traffic and a gas tax isn't enough, a use-tax for residents of the city and a toll on major highways for those entering should be sufficient.

Greg said...

I agree completely that gas taxes are simpler, which is the only reason they exist and a mileage tax doesn't.

However a gas tax has problems owing to its simplicity. I attempted to raise some of these points in the post.

Tolls have a high overhead and are difficult to enforce in areas with many entrances and exits. This is why they are used on long stretches of highway or bridges. Within most cities they are impractical. Not all large cities are islands.

Patrick Chapman said...

Posted this on FriendFeed, but not sure you caught it.

I wouldn't call Oregon's study successful. Recruiting volunteers is way different than implementing it in the real world. All it shows is that in an unmanipulated environment (i.e., in cars where the GPS device was not tampered with) that the calculation of gas prices based on reported miles traveled is possible. Also, from my perspective, rolling out something like this in Holland is much easier. Their population is only 16 million whereas the U.S. population is around 300 million. In terms of transition to the new system it would be way easier. And given that, its still going to take them about 7 years (projected 2016) in an ideal situation. I am not saying the idea should be abandoned though.

I'd like to make a suggestion/tweak to the idea and propose something different. First the suggestion: like the oregon study, use a sample of people to determine the tax rate for people in a particular area. Most people pay with a credit card and when you pay, you have to enter your corresponding zip code for the card holder. Just match that back up to the sample drivers in your zip. New idea: skip the GPS all together and use a RFID chip in drivers licenses. Each time you buy gas you have to swipe it. You can be charged based on the frequency of fillups and the amount you purchase. This way I can't spread my driving out over multiple cars to keep my costs down. The driver is taxed, not the car. Something like this can be rolled out relatively quick, I imagine.