To which my friend Phil Goetz asked "What would someone less simplistic say?"
The general arguments for a higher gas tax ...
First the supposed problem "achieving energy independence" is without point, we do not want a world of atomistic countries that don't trade with each other, this will make us all poorer. If other countries can produce energy cheaper than the U.S., we should exploit that, and use our resources for something better.
Second the "budget deficit" requires taxes to reduce. Certainly more taxes without additional spending would reduce the deficit. One question is whether additional revenue would actually go to reducing deficit rather than additional spending. A second question is what is the problem with debt? Individuals and organizations borrow all of the time, hopefully for long-term capital investments rather than short-term operating expenses. Pay-as-you-go is not inherently better. If we expect our spending will produce positive long-run returns, deficits are ok. (That said, the federal govt. is wasting money hand over fist in adventures like Iraq and general mismanagement and pork, which should certainly be reigned in).
The rationales about air pollution and climate change are better. Air pollution (climate change) is a negative externality caused in part by burning of fossil fuels, and moving vehicles off of fossil fuels, or reducing their use, would all else equal, be a good thing. The economically right amount of pollution is not zero, however. It would internalize the external costs of pollution, so that polluters could make the appropriate calculation if the private benefits of polluting outweigh the private costs (which include the social costs). Identifying and monetizing those costs is not simple, and there is considerable debate as to their magnitude. Ideally we would charge per unit of pollution (output). Simplistic policy analysis views taxing per unit of gasoline (inputs) is an approximate surrogate. The problem is that cleaner burning cars pay just as much for inputs as polluting cars (assuming otherwise equivalent operating efficiency), thereby giving people no incentive to purchase cleaner cars (which will be more expensive because they have more complex technology like catalytic converters). I believe there are better mechanisms for charging for pollution more directly, particularly where vehicles are inspected annually already, like Maryland . In that case, we measure the output over a short period of time, can multiply by the annual miles traveled, and multiply that by a cost per gram or ton of pollutant. It is still a surrogate, and the system can still be gamed, and we are trading off collection costs against tax efficiency. But this has much better incentives than charging by input.
Finally, higher gas prices would reduce travel, which would reduce congestion. However, travel is in the short-term relatively inelastic, it doesn't change very much with increases in prices. Road pricing would enable us to distinguish prices by time of day, charging more during congested periods, and less in the off-peak, helping to spread.