Autonomous Vehicle Safety

I had a conversation with someone yesterday in which I asserted that a shift toward more autonomous vehicles will improve safety. He immediately objected and listed a bunch of computer related disasters.

What I don't think this person realized is that computers have steadily taken a bigger and bigger role in the transportation they use every day. The biggest automotive safety advance in years, electronic stability control (ESC), is a computerized system. The National Highway and Traffic Safety Administration (NHTSA), estimates that ESC reduces accidents by 35 percent. Every generation of commercial airliners includes more and more sophisticated avionics software. The airline industry is going through it safest period in years.

In contrast, the NHTSA reported 5,973,000 U.S. car accidents in 2006, which caused 42,642 deaths and 2,575,000 injuries. In addition, motor vehicle crashes were the leading cause of death for every age from 2 through 34. I don't have any statistics available to prove it, but I'm confident the vast majority of these accidents share one primary cause -- human error.

Links: NHTSA 2006 Fact Sheet | NHTSA on ESC

Tolls Everywhere - Marginal Revolution Edition

Links: Marginal Revolution | New Transporation

Safe Mining

Thank you to my father for alerting me to this story.

The Washington Post ran a fascinating article about the airline industry this past Sunday. The airlines have had an amazing safety record over the past few years. They attribute much of this success to a surprising source -- data mining. In one example cited in the story, analysts discovered an unknown bulge in a Vermont runway by mining takeoff angle records.

This success story is only possible because planes, air traffic control, and airports track an amazing amount of information. Commercial airplanes have been the safest form of transportation for a long time. Yet for some reason, we have not made similar investments into data collection for cars, the least safe major form of transportation. I suspect if we did we could make a major dent in car deaths.

Link: Washington Post Story

Congestion Pricing

New York City is considering various schemes to cut down on traffic in Manhattan. A New York state commission just released an interim report evaluating four plans, each of which combine one or more of the following: congestion pricing, bridge tolling, pricing of parking and taxis, and license plate rationing. I strongly support any effort to reduce traffic. However, I think congestion pricing is the most direct and logical choice. Ideally, all drivers should be charged for the delay they are causing by being on the road. The best way to do that is to build a comprehensive system of tolls whose fees change dynamically with the traffic. One of the four plans considered by the commission, the one proposed by Mayor Bloomberg, seems to be far closer to an ideal system than the others.

Link: Commission Report

Advertisements

I put a small text advertisement on the left navigation. I will probably insert text ads into the feed and a graphic ad somewhere on the main page. The money I make from the ads will be used exclusively to advance the cause.

I have a few ideas already that require money: attracting people to this blog by advertising on other blogs, registering a domain, printing posters, registering as a non-profit, and flying to Chicago to be on Oprah.

Hopefully, you won't find the ads to be too obtrusive. If you do, leave a comment.

Tolls, Tolls, Everywhere

There are two ways to end the direct subsidy of roads. The first way is to privatize the roads, but this doesn't make much sense to me. The second is to dramatically expand tolls. Ideally the fee will correspond exactly with the cost the driver is imposing on the government. Heavier cars will be charged more because they cause more damage to the roads. Driving on roads that were expensive to build or are expensive to maintain (e.g. bridges) should cost more than driving on inexpesive roads.

I think there is interesting research and creative entrepreneurship needed to make this work well. We need a toll collecting mechanism (probably similar to EZPASS) that is inexpensively deployable and does not slow traffic. We also need a way of telling drivers how much they are being charged. I don't think a sign at every street corner is the answer. Better would be to store toll information in GPS navigation devices so drivers can choose from different routes based on toll cost. We also need an accurate and comprehensive model of the damage different cars do to roads and the cost of construction and maintenance of different roads.

Do you have any ideas for making ubiquitous tolls work? If you do, please post a comment.

Economic Costs of Cars

I found a paper, by economists Ian Parry and Kenneth Small, that discusses optimal gasoline tax rates. The paper is very technical, but I think I was able to glean something useful from Table 5. The paper estimates the external cost of four problems with cars: Global warming, pollution, accidents, and congestion. They respectively contribute: 5¢, 16¢, 24¢, and 29¢ to the optimal gasoline tax. The United States consumed 126 billion liters of gasoline in 2003. Thus, if I understand the paper correctly the external cost of global warming is $6 billion, pollution $20 billion, accidents $30 billion, and congestion $37 billion, for a total of $93 billion.

I will have to dive deeper into the paper to dive deeper into the paper to be confident that my understanding is correct. Hopefully, I can find an economics graduate student at Carnegie Mellon willing to help me out.

This paper seems to have a very right leaning slant. I think its primary political purpose is to argue that the gasoline tax in Britain is too high. However, Greg Mankiw, an American economist, uses it as a source to advocate raising the tax on gasoline in the United States.

Do you know of any other papers that estimate the economic cost of cars? If you do, please post a comment with links.

Links: Ian Parry and Kenneth Small | Greg Mankiw | Gasoline Consumption by Country

Moderating Comments

Comments will no longer be moderated. I think it confuses users when their comments don't appear immediately. Commenters will still be required to have Google accounts and solve CAPTCHAs. I will delete inappropriate comments.