Hurray For High Gas Prices!

For a long time I have felt the price of gasoline in the United States was way too low. Pretty much all economists believe this. Greg Mankiw blogged back in October about the many reasons why we should raise gas taxes.

The reason we need high gas taxes is that there are all sorts of costs associated with my driving that I don’t pay — someone else pays them. This is what economists call a “negative externality.” Because I don’t pay the full costs of my driving, I drive too much. Ideally, the government could correct this problem through a gas tax that aligns my own private incentive to drive with the social costs of driving.

Three possible externalities associated with driving are the following:

a) My driving increases congestion for other drivers;

b) I might crash into other cars or pedestrians;

c) My driving contributes to global warming.

If you had to guess, which of those three considerations provides the strongest justification for a bigger tax on gasoline?

The answer, at least based on the evidence I could find, may surprise you.

The most obvious one is congestion. Traffic jams are a direct consequence of too many cars on the road. If you took some cars away, the remaining drivers could get places much faster. From Wikipedia’s page on traffic congestion:

The Texas Transportation Institute estimates that in 2000 the 75 largest metropolitan areas experienced 3.6 billion vehicle-hours of delay, resulting in 5.7 billion US gallons (21.6 billion liters) in wasted fuel and $67.5 billion in lost productivity, or about 0.7% of the nation’s GDP.

This particular study doesn’t tell us what we really need to know for estimating how big the gas tax should be (we want to know how much adding one driver to the mix affects lost productivity), but it does get to the point that, as a commuter, I’m better off if you decide to call in sick to work.

A more subtle benefit of fewer drivers is that there would be fewer crashes. Aaron Edlin and Pinar Mandic, in a paper I was proud to publish in the Journal of Political Economy, argue convincingly that each extra driver raises the insurance costs of other drivers by about $2,000. Their key point is that, if my car is not there to crash into, maybe a crash never happens. They conclude that the appropriate tax would generate $220 billion annually. So, if they are right, reducing the number of crashes is a more important justification for a gas tax than reducing congestion. I’m not sure I believe this; it certainly is a result I never would have guessed to be true.

How about global warming? Every gallon of gas I burn releases carbon into the atmosphere, presumably speeding global warming. If you can believe Wikipedia’s entry on the carbon tax, the social cost of a ton of carbon put into the atmosphere is about $43. (Obviously there is a huge standard of error on this number, but let’s just run with it.) If that number is right, then the gas tax needed to offset the global warming effect is about 12 cents per gallon. According to this National Academy of Sciences report, American motor vehicles burn about 160 billion gallons of gasoline and diesel each year. At 12 cents a gallon, that implies a $20 billion global warming externality. So relative to reducing congestion and lowering the number of accidents, fighting global warming is a distant third in terms of reasons to raise the gas tax. (Not that $20 billion is a small number…it just highlights how high the costs are from congestion and accidents.)

Combining all these numbers, along with the other reasons why we should tax gas (e.g. wear and tear on roads), it seems easy to justify raising the tax on gas by at least $1 per gallon. In 2002 (the year I could easily find data for), the average tax was 42 cents per gallon, or maybe only one-third of what it should be.

High gas prices act just like taxes, except that they are more transitory and the extra revenue goes to oil producers, refiners, and distributors instead of to the government.

My view is that, rather than bemoaning the high price of gas, we should be celebrating it. And, if any presidential candidate should come out in favor of a $1 per gallon tax on gas, vote for that candidate.


What would Jesus drive?

A hummer of course,

1.big enough for all the apostles
2.maximum load of loaves and fishes
3.indirect support for the army of god for Mary and the kids around town (See Da vinci Code) gallon gas lasts 8 days no matter how many miles are driven. (See Chanukka)


This would be just another nail in the coffin of so many families's fiscal survival. If they can't afford a car and gas, they may not be able to work. If they can't work, the results are obvious.


Report: 98 Percent Of U.S. Commuters Favor Public Transportation For Others


Based on the level of traffic I see blowing by my farm, I believe the demand for gas is inelastic and people will drive no matter what the price or how much you tax it. As anti-consumption tax severely penalizes people like myself that live in a rural area. For me it is 30 miles to the grocery store (one way) and 40 miles (one way) to check cattle. All an anti-consumption tax would do is suck money out of the economy. But I would not be wholly averse to an anti-consumption tax if it were only applied to the east and west coast states. Better to let markets determine the price of gas.



Well, actually congestion and accidents are externalities. The problems of congestion and the risk of accidents you cause yourself are factored into your own internal decisionmaking, but you do not consider the external costs imposed on others by the reduced capacity of the roads when you drive on them, nor do you consider the risk imposed on others of an accident when you drive.


I agree with gevil. A higher tax on petrol will not likely reduce congestion much - people will just move to more efficient cars, which is a much more acceptable substitute for them than quit driving. Consider Europe - the reason why people are driving less than in the U.S. is because they don't need to drive more (shorter distances and better public transport), and because they don't want to drive more (congestion, the same problem as in the U.S.). It is not because gas would be expensive. The effect of the gas tax on the amount of driving is extremely limited (and the gas tax for instance in the Czech Republic is in the range of 3.7-4 USD per gallon). However, as I said, people were very good in switching to more efficient cars, which is a good thing after all.

If you want to reduce congestion, you most likely have to introduce per-mile congestion fees, possibly depending on the time when you drive, so that you balance traffic. And the experience of London says that in order to have a significant impact, these fees have to be heavy (entry to London = 8 GBP ~ 16 USD).

The project of the British government to introduce time-dependent fees on highways sounds interesting, but it is still a very long path to be implemented.



You mentioned only global warming but in my opinion it is a quite abstract externality, whereas the negative effects of the toxic fumes cars produce are much more tangible: asthma, noise etc. So apart from reducing congestion, the extra money can be spent on helping the effects above. However, the government rarely spends money properly...


"I think I can safely assume I feel the pain at the pump a little more than an a certain someone who has a book on the New York Times bestseller list."

But do you drive as much? I'm not sure what it's like in the US, with so much rural land, but in the UK the richer people are, the more they drive. So increasing gas prices don't necessarily impact on poorer people worse.

"Taxes on gas (petrol) are a blunt instrument for dealing with the negative externality of congestion, although ideal for dealing with the externality of global warming. A driver in an isolated area will cause no congestion but use petrol."

They're blunter than per-mile road pricing, but they're not entirely blunt; a car will burn more gas per mile in a congested street than on a wide open highway. The ideal is still per-mile tax, but that does raise privacy concerns.


"Well, actually congestion and accidents are externalities."

Hmm. Let's see.

"you do not consider the external costs imposed on others by the reduced capacity of the roads when you drive on them,"

But this is true of everything. I don't consider the impact on others of the reduced supply of peanut butter when I remove one from the store shelf, and I don't consider the impact of the reduced supply of money when I decide not to buy peanut butter today (and keep my money in my pocket). Again, not an externality. If there's congestion, I pay the price the same as anyone else.

Oh, there's an externality in road maintenance -- a friend of mine once claimed that trucking was effectively subsidized versus trains by the trucking firms not having to pay as much as it truly cost to maintain the roads. I don't know about such things, and he was from a railroading family so he may have been biased; but anyhow, that may be a true externality, and a stronger one than global warming (I agree that pollution is also stronger, good point).

"nor do you consider the risk imposed on others of an accident when you drive."

But my driving increases my risk of accident FAR more than it increases anyone else's risk. My risk goes from almost zero to x; theirs goes from x-delta to x (where delta is very small). Again, I pay the price in terms of risk (although if I'm "lucky" in an accident I may not pay the final price as high as the other car's occupants). Even if I'm a REALLY bad driver (so delta is big) my own danger of accident is _still_ high.



Unfortunately, I can't remember who wrote it, but I read an editorial that challenged the idea that higher gas prices will necessarily reduce gas consumption. His main argument being that gas prices *have* increased significantly in the past couple years, and it's had virtually no effect on how much gas we use. I'm not an economist so I can't verify his statistics, but as a layman I do seem to be reading a lot of articles about how, despite high gas prices, Americans will be traveling as much as ever this summer for their vacations.

If gas is a dollar more per gallon than it was a few years ago, and we drive just as much, I wonder how much difference an additional dollar per gallon in tax will really make. It seems to me the problem is that we are a driving culture, and it will take significant changes to change the culture. Significant meaning gas being double, maybe triple what it costs now.

Just a thought.



Good point about railroads. Trucks do not have to pay roads maintenance (the government looks after everything), whereas trains must carry the entire burden alone. So, the trucking industry really is subsidized.


How about a progressive tax on vehicle size? Larger vehicles do more damage to other cars in an accident, generally get lower gas mileage, and put proportionally more wear on infrastructure.


I agree with you, if people are paying more for their gas, they will drive less making it more convenient for me. It would be nice to see more segways and other modes of transportation, especially in cities.


s-side, I have heard there are places that tax cars via proportional registration fees based on vehicle mass. can't recall exactly where though.

I work for an oil company, and have no idea where the notion that they are subsidized come from. We get nothin', pay a lot of royalties and taxes with no preferential rates. maybe this is just in the U.S. (I'm in Canada), or just a myth. I'm not arguing for subsidies by the way, in any industry.

If global warming is to be confronted, it may be more efficient to use a CO2 emissions cap-and-trade system than just more and more taxation. Taxes just end up being wasted by governments and distort markets, with no real solving of the problem.


gevil is right. in the short run, gas demand is pretty inelastic. after a few years when people start buying new cars, maybe they'll make fuel efficient choices, so in the longer run, demand will be more elastic. but still not *very* elastic, to the extent of affecting congestion. I think the entire infrastructure of most cities, towns, and suburbs is set up for a car culture, and that car culture developed in large part because cheap gas made driving really easy. Once that gas is no longer cheap, maybe people will be less happy with the car culture, but they've still got the same infrastructure where affordable housing is a 45 minute drive to work and a 15 minute drive to the supermarket. That will need to change too, and denser development will take a long time. In the mean time, maybe the gas tax will lead to more telecommuting and a greater preference for renting (if you rent, it's easier to change your housing when your job changes).



Anyone can do a cost analysis, but what about the benefits? Many people including those arguing for driving curfews of teenagers talk about the costs that are reduced when less people drive. The fact is we could ban everyone, but emergency vehicles from driving and then there would be very very few road deaths, no contribution to global warming and we wouldn't have any congestion.

But what would be the reduction in benefit from all those goods and services not running around the country. You have to assume that all those people who are currently driving are applying some sort of cost benefit analysis to their driving and that raising the price of gasoline would reduce some benefit and that the benefit in question would likely be greater than the congestion, fatality and gasoline usage that is involved.

The reduction in benefit might be much higher than you have anticipated and this could be a really bad idea, but by ignoring the loss of benefit you are missing it.



People will stop driving only after they don't need to drive anymore. People drive a lot because of the way most metropolitan areas in North America are structured today: distant suburbs gravitating around a central business area. And some of those suburbs don't even have a regular bus connection to downtown.

In Germany for example lots of people live in villages and small towns (less than 10000 people) but they don't need to commute too much because many of the businesses are also located in similar places. Combine that with an excellent rail road infrastructure and get the complete opposite of North America.

I agree that rising fuel tax will have positive impact on the environment as people will move towards buying more fuel-efficient cars. This however will not stop commuters drive to work if it is located 20, 30 or 50 kms from home (if no decent public transport is available).

Another example from Europe: in some countries they use what they call "vignette" - a sticker that you put on your windshield and that allows you to drive outside of the premises of your place of residence. In fact that's a permit to use the intercity roads. This sticker can be bought almost everywhere and its price is different for different types of vehicles - heavier trucks and buses pay more than a standard car.
So if you drive only inside your town, you don't pay for a vignette. If you need to go to a different city or drive for your vacation, you buy the vignette for the period of time you need: one day, one week, month or a year. The money collected through vignettes is not considered tax money and goes directly for the road maintenance budget.



to dchai
Yes it is true that prices are higher and driving habits haven't changed much but the main reason for that is that prices have not been consistently higher. Because prices may come down, there is litte incentive to change habits. You bite the bullet and hope. Besides like others have said, it is very difficult to change driving habits overnight. It you live in a suburb with little or no mass transit, you have no other short term options. What a gas tax would do is raise the floor so that when people are making economic decisions, they would do it with the knowledge that prices are going back down so I have to adjust.


You obviously have no understanding of an economic externality -- the demand and supply of peanut butter is not at all like travel congestion. Just pick up an econ 101 text and read the chapter on externalities.

Second, the paper cited by Levitt on accident externalities supports the argument that the risk one imposes on OTHERS (in aggregate) by driving is in fact quite large, despite your apparent intuition that it must be a "very small" delta.


A blanket tax doesn't seem to be a good fit for solving the congestion problem. Marginal congestion is minimal for people who work and travel in off peak hours; it seems inefficient to tax them at the same rate as their counterparts.
As far as crashing is concerned, isn't that why we pay insurance. The risk is an inherent cost with traveling by any means. If a bus crashes then the costs will be passed on to the patrons of said bus in some way.
Aren't we already paying taxes on the gas for the environmental concerns, as well as suffering from higher prices due to environmental regulation on the gas production.
It frustrates me to read that anybody thinks higher taxes are an answer to any problem.