So you balance out the any pigovian tax on gasoline with bigger income tax cuts.
There's no rule that says NO taxes should ever be added during a recession, just that you don't want to increase the total tax take.

In my original draft (that was too long, this was to be published elsewhere with more formal wordcount constraints), I explained that I thought that returning the money in the form of an income subsidy was a bad idea because of the lag time between the tax being instituted and people getting their money back. Folks like Krugman would argue that stimulus needs to happen now, not five months from now in order to have the greatest effect. Any waiting increases the amount of stimulus that will be needed.

I agree with the phase-in concept. Say 25 cents for six months, another 25 cents after that, etc. Lower the rate of increase and increase the amount of time before the next increase once you get closer to more uncomfortable prices. The way the NYTimes and WaPo bits were talking was that we should do it all at once, right now, and spend the money on infrastructure and healthcare.

One other important benefit of a tax, as long as it's a fixed amount, is that it helps even out fluctuations in the price of gasoline. Such fluctuations make it very difficult for both motorists and car makers to make sensible decisions about what sort of vehicles to buy/manufacture.

I agree, which I why I think any effective Pigovian gas tax should have a price floor built in so it doesn't send the wrong price signals to the marketplace which is happening now.

A pigovian tax on gasoline is all very well, but ideally what you want is a tax on pollution. If a car manufacturer can demonstration that a particular vehicle generates less pollution per gal. of gas used, then ideally the pollution tax should be lower. Given it's not really feasible to do it at the point of purchasing the gas, you realistically have to do it the point of registering the car, based on average usage, which has the added benefit of encourage car buyers to think a bit more long term about the real tost cost of owning and running the car.

Ideally you'd want both. Pollution is but one externality. Another is national security, and our dependence on foreign oil is a big problem. Any country that comes up with green energy technology first is going to have a leg up on the rest of the world in terms of being free from petro-tyranny. This is the main way I would sell a Pigovian tax to climate change deniers.

A pigovian tax on gasoline is all very well, but ideally what you want is a tax on pollution. If a car manufacturer can demonstration that a particular vehicle generates less pollution per gal. of gas used, then ideally the pollution tax should be lower. Given it's not really feasible to do it at the point of purchasing the gas, you realistically have to do it the point of registering the car, based on average usage, which has the added benefit of encourage car buyers to think a bit more long term about the real tost cost of owning and running the car.

FWIW, I think now is an excellent time to start phasing in a pigovian tax on gas, given its low price, and the fact consumers have already had to learn to deal with higher prices in the last year or so.

One other important benefit of a tax, as long as it's a fixed amount, is that it helps even out fluctuations in the price of gasoline. Such fluctuations make it very difficult for both motorists and car makers to make sensible decisions about what sort of vehicles to buy/manufacture.