Ali, a friend who spends a significant amount of time researching climate science, asked me recently about environmental tax shifting.
A condensed version of my response is reprinted here:
So tax shifting (as I understand it) is "Pigouvian tax one thing you'd like to discourage and then take the money and return it by cutting broad-based corporate or income taxes" (or, in reverse, subsidizing one thing you'd like to encourage and raise the money by raising broad-based taxes).
Though I don't see this discussed anywhere, I think a more interesting variation is the idea that you can even return the money in a lump sum to the industries you're taxing - basically, by punishing companies based on per-unit behavior and returning the money to the industry in a lump sum, you give each company an incentive to do better than its peers without crippling the whole industry. For example, if Nike and Reebok are competing, and Nike's a lot more efficient about carbon than Reebok, basically, Reebok will have to pay Nike proportionate to the amount by which Reebok is worse. Reebok has an incentive to improve so they don't have to pay money to their biggest competitor - which hurts way more than paying society generally - Nike has an incentive to improve so they can cripple their biggest competitor, and the US still has a very viable shoe industry.
Environmentally, I see the more general idea proposed all the time with carbon taxes and gas taxes accompanied by corresponding cuts to income taxes. The idea seems, to me at least, to make a lot more sense with gas taxes than with carbon taxes. That's partially because I see gas taxes as incentive compatible with behavior we want, whereas I think carbon taxes are a worse alternative to a combination of gas taxes and smart grid/battery charging station buildout (largely because they ignore international concerns about industry and assume widespread Coasian equivalence). If carbon taxes were industry-targeted, I'd be much happier about carbon taxes (though still accompanying, not to the exclusion of, a more general gas-tax-for-income-tax-and-smart-grid shift).
Note that all of this is relative to other carbon-mitigation options, and none of it tries to compare to the status quo, except maybe a gas tax (which is similar enough to what we've seen that we can reliably estimate effects). Whether or not a standard carbon tax or cap and trade would be better than the status quo (mostly relying on supply-side R&D and marketing) is beyond my ability to assess.
As with everything, tax shifting only works if the Pigouvian Taxes you're placing are more intelligently designed (from an incentive perspective - both in behavior you're trying to prevent, and in general economic deadweight loss) than the corporate/income taxes they're replacing.
The problem is that measuring effects of carbon taxes is really hard. We know so little about what climate change is actually going to do, and effects of a carbon tax are not linear, at all. Going from 0 to 1 on carbon taxes may produce different results than going from 9 to 10, and it's hard to calibrate the right level. At a certain point, carbon taxes are no longer useful because the economic impact of taxing that much impedes economic activity, which in turn reduces a lot of necessary investment and R&D to actually get out of the patterns of producing carbon emissions. It's the same fallacy as what you can observe in the people who want to jack up income taxes a lot - not everything is scalable the way activists would like them to be.
If it's not obvious, I generally support intra-industry tax shifting, gas taxes and smart grid/battery charging stations, and am very skeptical that the cap-and-trade or carbon tax options brought up publicly can be effective without major side effects.