Remember peak oil? Remember when oil prices were $140 a barrel and Goldman Sachs predicted they would soon reach $200? Now, the latest news is that oil prices have gone up all the way to $34 a barrel. Last fall, Goldman Sachs predicted prices would fall to $20 a barrel, which other analysts argued was "no better than its prior predictions," but in fact they came a lot closer to that than to $200.

Low oil prices generate huge economic benefits. Low prices mean increased mobility, which means increased economic productivity. The end result, says Bank of America analyst Francisco Blanch, is "one of the largest transfers of wealth in human history" as $3 trillion remain in consumers' pockets rather than going to the oil companies. I wouldn't call this a "wealth transfer" so much as a reduction in income inequality, but either way, it is a good thing.

Naturally, some people hate the idea of increased mobility from lower fuel prices. "Cheap gas raises fears of urban sprawl," warns NPR. Since "urban sprawl" is a made-up problem, I'd have to rewrite this as, "Cheap gas raises hopes of urban sprawl." The only real "fear" is on the part of city officials who want everyone to pay taxes to them so they can build stadiums, light-rail lines, and other useless urban monuments.

A more cogent argument is made by UC Berkeley sustainability professor Maximilian Auffhammer, who argues that "gas is too cheap" because current prices fail to cover all of the external costs of driving. He cites what he calls a "classic paper" that calculates the external costs of driving to be $2.28 per gallon. If that were true, then one approach would be to tax gasoline $2.28 a gallon and use the revenues to pay those external costs.

The only problem is that most of the so-called external costs aren't external at all but are paid by highway users. The largest share of calculated costs, estimated at $1.05 a gallon, is the cost of congestion. This is really a cost of bad planning, not gasoline. Either way, the cost is almost entirely paid by people in traffic consuming that gasoline.

The next largest cost, at 63 cents a gallon, is the cost of accidents. Again, this is partly a cost of bad planning: remember how fatality rates dropped nearly 20 percent between 2007 and 2009, largely due to the reduction in congestion caused by the recession? This decline could have taken place years before if cities had been serious about relieving congestion rather than ignoring it. In any case, most of the cost of accidents, like the other costs of congestion, are largely internalized by the auto drivers through insurance.

The next-largest cost, pegged at 42 cents per gallon, is "local pollution." While that is truly an external cost, it is also rapidly declining as shown in figure 1 of the paper. According to EPA data, total vehicle emissions of most pollutants have declined by more than 50 percent since the numbers used in this 2006 report. Thus, the 42 cents per gallon is more like 20 cents per gallon and falling fast.

At 12 cents a gallon, the next-largest cost is "oil dependency," which the paper defines as exposing "the economy to energy price volatility and price manipulation" that "may compromise national security and foreign policy interests." That problem, which was questionable in the first place, seems to have gone away thanks to the resurgence of oil production within the United States, which has made other oil producers, such as Saudi Arabia, more dependent on us than we are on them.

Finally, at a mere 6 cents per gallon, is the cost of greenhouse gas emissions. If you believe this is a cost, it will decline when measured as a cost per mile as cars get more fuel efficient under the current CAFE standards. But it should remain fixed as a cost per gallon as burning a gallon of gasoline will always produce a fixed amount of greenhouse gases.

In short, rather than $2.38 per gallon, the external cost of driving is closer to around 26 cents per gallon. Twenty cents of this cost is steadily declining as cars get cleaner and all of it is declining when measured per mile as cars get more fuel-efficient.

It's worth noting that, though we are seeing an increase in driving due to low fuel prices, the amount of driving we do isn't all that sensitive to fuel prices. Real gasoline prices doubled between 2000 and 2009, yet per capita driving continued to grow until the recession began. Prices have fallen by 50 percent in the last six months or so, yet the 3 or 4 percent increase in driving may be as much due to increased employment as to more affordable fuel.

This means that, though there may be some externalities from driving, raising gas taxes and creating government slush funds with the revenues is not the best way of dealing with those externalities. I'd feel differently if I felt any assurance that government would use those revenues to actually fix the externalities, but that seems unlikely. I actually like the idea of tradable permits best, but short of that the current system of ever-tightening pollution controls seems to be working well at little cost to consumers and without threatening the economic benefits of increased mobility.