Calculating all the externalities of driving

I and many others feel the best way to set urban and transportation policy is to properly price in the "externalities" into our travel, and to remove all other penalties and subsidies. If you can do this, then everybody is incentivized to improve the public good. In particular, entrepreneurs and companies are motivated this way, and it's their job to think of the new things nobody else thought of.

It is far from simple to accurately capture these externalities, but the result is valuable enough to give it a go. Even if you can't capture all of them, you can use traditional regulation for some things, and pricing of externalities for others.

So let's look at some of these and how to price them.


This is the most common area of discussion. Unfortunately there is debate, both scientific and political, on: * What harm pollutants cause and what it costs * Whether greenhouse gasses cause harm at all (ie. the global warming political debate) * How to properly account for "well to wheels" energy use in both fossil fuels and electricity * Contested externalities, like environmental effects of fracturing for natural gas extraction

Once you can calculate these, though, it's easy to price them into the fuel or electricity. One can argue some variation based on where you release the emissions. For example, small particulates kill many people, but far more if you release them upwind of population centers, and few if you release them in the country.

While we'll never get full agreement on the exact number, any serious number should be enough to get people to want to generate less pollution in their driving (and all other energy use.)


There is specific debate on the real costs of the battery packs in electric cars -- both the costs of the mining and processing of the materials, and the cost of recycling. One would want all the components of vehicles to have their externalities priced in, so that they are automatically factored in the depreciation cost of a vehicle mile.

Road usage

Roads cost money to build and maintain in ordinary use. A lot of that money was spent long ago, so one argument is not to bill for what taxpayers already paid for. Other arguments suggest that road usage is not in itself a negative externality when you factor in all the other things listed here.

Road damage

Road damage requires maintenance. Road damage is very different for different types of vehicles. It goes up roughly with the 4th power of weight. So yes, a 12,000lb truck is doing 80 times as much damage as a 4,000lb car. And bicycles and pedestrians are barely noticeable. Trucks do pay a tax based on their weight, but it is completely inadequate considering the difference in road damage they cause.

Congestion and impedance of traffic

Congestion is (like most of these) tricky. If you drive at a low-demand time, you really don't cause any congestion. You don't get in the way of others. All drivers near peak times contribute to congestion. Some would argue off-peak drivers are increasing the probability of congestion a little bit each, but if the road never congests, was there an externality?

All traffic does impede other traffic, because of the need for things like stop signs and traffic lights. The more flow on one street, the longer I have to wait to cross it. That's true for pedestrians as well as cross traffic. On a lightly flowing street, pedestrians can cross almost any time. With more traffic they should stick to crosswalks or even lights. Of course, the pedestrians using the crosswalk are also blocking traffic, but there is no way to bill them for their externalities (and politically it's close to impossible.)

You can get very detailed if you want. Everybody who stops or slows is impeding traffic. They might have a purpose, like dropping somebody off or parallel parking, but they do add to congestion. Big vehicles also tend to add more to congestion than small ones, but if they carry multiple people it can more than balance out. Vehicles with slow acceleration also are well disliked on the roads for slowing them down.

All of these factors could be measured in a robocar. Doing so in regular cars is difficult, though it could also be done with a small GPS device, if designed in a non-Orwellian way.

It is hoped that robocars and smart traffic management will actually reduce congestion for all, not increase it, so the cost of this will go down -- though it is still not paid for by the driver in most places.

Degradation of experience

The presence of cars on streets used for walking, shopping or anything but driving degrades the urban experience. They make noise, and they create a barrier to movement for people who must go to crosswalks or wait for lights. When traffic volume is high, they require space as roads become wider than would be desired for aesthetic or pedestrian uses.

It can be argued we would design our cities differently without cars on the surface, or without cars at all. Certainly cities of past centuries look very different, though a wide variety of factors went into that.

There is also visual blight to consider -- not just of roads, but of overhead wires, elevated tracks and bridges, traffic lights and more.

Street level transit would incur more costs -- the noise of trolley wheels taking a turn is much louder than most other vehicles, and the tracks create a hazard for bicycles, among other things. Buses are noisy and block streets and, if diesel, have high emissions per person.

Creation of risk

Driving creates risk for other users of the road. Both the risk that you will hit them, and for them an increased chance that their own mistakes could end in an impact. (ie. because the road is full of other cars, I must drive at a higher level of caution.)

It was noted that vehicles can be stolen and used as weapons. Fortunately this is extremely rare today but our reaction to it is very strong.

Consequences of accidents

Of course the risk leads to accidents. In theory this is paid for by insurance, but estimates suggest as little as 1/4 the cost of accidents gets paid that way.

In the comments it is also added that roadkill is an externality, not just from animal rights and death of wildlife, but because it is unpleasant and may need to be cleaned up.

The consequences

While I don't imagine we could come up with truly accurate and agreed upon values for all these costs, I don't think we need to be that accurate to get better policy results than just making hard and fast rules like banning certain activities or prioritizing others. In fact, in many cases we may simply start from a goal that might be a regulation and then put a price on it.

In the modern computer world, it is no longer untenable to have very complex formulae for calculating the price. There can literally be a different price per square-foot second of road occupancy for every different small piece of road, and computers can add it up, and optimize based on it. Calculations can have caps to make sure they remain roughly predictable.

Today, a large fraction of people use a tool like Waze to plan their trip. It puts a focus on travel time based on a very complex analysis of the factors on every road segment and intersection. It also considers distance and the costs associated with distance. Deeply complex rules need not impede route planning, and because they will be simpler in aggregate, they can still affect vehicle choice purposes. We can know the emissions of every kwh of electricity in every region. We can know the weight and size of every vehicle. Software will know our travel patterns and can tell us, "with your pattern, this vehicle will cost $X and that vehicle $Y, but if you are willing to change your travel pattern as such you could lower those prices as follows.

What might change?

  • More energy efficient transport (per passenger) would become the norm, but the solution would be chosen by people, not policy.
  • The use of tunnels might become much more competitive
  • VTOL aircraft (flying cars) might also be much more competitive, except on noise
  • Some transit modes would suffer from the loss of their subsidies and their noise, size and overhead wires.
  • Strong incentives against congestion would distribute traffic for more efficient and predictable flows
  • Most of all, things net yet imagined would get invented under the better economic rules.

So, what other externalities do you think should be priced into travel?


If damage really goes up as 4th power of weight, then a12000 pound vehicle does 81 tines the damage of a 4000 pounder. Or is it only the second power?

My brain wrote the wrong number, I've fixed it.

Roadkill as an externality:
  • Aesthetic is bad.
  • People are sad.
  • People are paid to clean it up.
  • Ecology:
    • Lots of dead animals.
    • Isolates populations.
    • Some predators require large, easy to traverse territory, but high density crisscrossing roads means less of these territories. Lack of predators can lead to huge swings between animal composition equilibria in an area.

I think the "creation of risk" is huge, and so prevalent that we are blind to it. A huge portion of the education of a child is learning how to protect himself from traffic. Imagine transporting a person from hundreds of years in the past into a modern city. If he accidentally walks five feet in the wrong direction, he could be violently killed. The constant vigilance that we all must practice is a huge cost, hard to measure in dollars.

Well, in the future, if you had a street with all robocars, you actually would be safe walking out into traffic so long as you didn't do it from a hidden location (ie. jump out from behind a parked van.) Of course, all the people in the cars who had to brake hard for you would be rather upset at you.

This is also true for buses and street cars and at-grade trains. Of course as kids we played along the train tracks all the time. It's also true in the subway station, until they started installing walls with gates. If we did put a very high price on this risk though, it might well cause a push to use tunnels and flying vehicles for transport by making the more competitive.

An upcoming external cost may be the electrical grid upgrades needed for to cater for widespread adoption of electric vehicles. Instead of fuel tankers refilling gas stations we may end up with the energy being delivered via the electrical wires and recharging points. To make it more complex, some of the generation could come from local solar although generation will not always be reliable.

I don't think there needs to be much in the way of updates to the grid except when new generation capacity is needed. Robocars will drive to the charging, you don't need to build infrastructure to bring the charging to them. Put most charging next to existing transformer substations if parking can be found nearby.

However, in time, the total power load to the substations might need upgrade. But the price of that will be incorporated into the price of electricity. That makes it not an externality because you are paying for the cost in your expenses. Unless it is suggested that the government or other users would subsidize this cost. While that's not impossible, it would not be the natural course of things. In fact, when shared infrastructure is expanded, the economies of scale usually make each person's share of the infrastructure get cheaper.

There have been a number of papers on how electrical grid networks would cope with the widespread introduction of EVs. There are a few issues, especially if multiple vehicles are on fast charge in the same area on low voltage lines.

If autonomous vehicles need to make their way towards existing substations to recharge then there is a trade off. It somewhat works against widely distributed vehicles which could otherwise be waiting to be summoned and creates more miles driven by empty vehicles adding to congestion. Having plenty of oncall vehicles is critical if response times are to be kept short. Vehicles that have been recharged near substations would make their way back to optimised parking areas but you would still need a few extra vehicles.

It may be possible in cities with plenty of sunshine to make use of empty garages, shopping centre car parks etc.. and offer locally generated solar electricity. It would be a complex trading system and not always available, however it could move much surplus generation directly into the transport system.

Agree it is not an external cost if the full cost of changes and recharging points are recovered from the end user.

A vehicle that is charging is probably not in the on-call fleet. I mean it could be, if it's charged enough, and the need is desperate. But since need is not that unpredictable I don't see this as a big issues. If the car has to travel to charging, that has a cost, and the fleet manager can calculate that cost, and decide if it's more cost effective to put in some charging in new locations. There will be cases like this. But nothing like what human driven electric cars need with charging in every parking lot, every home.

A lot of charging is done at night, when power is (today) cheap but not at all solar. Mid-day charging could come from solar, of course, but not every day.

Some modes of transport can be used as weapons more effectively than others. Some are more effective in spreading disease than others. Both of these seem to go beyond the description under the "creation of risk" heading.

Also, different health impacts for users. Could be considered an externality source, if one considers things like obesity to be "epidemic" or that the costs of same are socialized.

This has effect directly on the user/owner of the car. We're looking for effects on other people who are strangers.

Negative effect on user/owner and on strangers aren't mutually exclusive, e.g.,

But you specifically named the health effects on the user. Other than pollution, and accidents, what are the health effects on 3rd parties?

How about the positive externalities? Happier neighbors and colleagues, workers who can work more hours in a day, more cops on the streets, more organ donors...

Add new comment