The climate policy debate is rapidly moving away from the question of whether action should be taken and toward the question of what sort of policy should be implemented. In this context, many advocate market-based emission control strategies as a means of reducing the inevitable costs of controlling carbon emissions, but what sorts of market-based policies are truly desirable?
In theory, there is not much difference between a pollution tax and a tradeable emission credit regime. In theory this is correct because a supply limitation can operate as a tax, and vice-versa. In practice, of course, the two programs are not equivalent. Among other things, policy makers lack the necessary information to know what tax level would be equivalent to what supply restraint, but this is hardly the only difference.
As is so often the case, transactions costs are important in choosing between available environmental policy instruments. There are significant transaction costs to creating either a pollution tax or an emission trading scheme, but the costs are unlikely to be the same. As the economist Ronald Coase pointed out in his work on firms, hierarchical command structures are sometimes less costly than competitive market transactions due to transaction costs. This is why we see non-market firm structures in the marketplace. The prevalence of these firm structures will diminish (or shift) as entrepreneurs find ways of reducing transaction costs. In the meantime, however, we should expect to see many successful firms with hierarchical command structures.
In the environmental context, the persistence of transaction costs means that market-mechanisms are not always more efficient or cost-effective than their command-and-control alternatives. This is particularly true when one is dealing with pollution problems that are highly dispersed and involve a wide range of different types of pollution sources. It may be relatively easy to design and implement a trading regime for sulfur emissions from utilities or the lead content of gasoline. It is much more difficult to have a reliable and enforceable trading regime for carbon dioxide emissions from non-industrial activities. The upshot is that a trading regime will not always be more cost-effective than a more traditional regulatory regime if it is more costly to monitor and enforce (see, e.g., carbon offsets). Pollution taxes have their own problems, to be sure, but in my experience the drawbacks of a new tax are more readily apparent, and more quickly seized upon, than the potential drawbacks of trading schemes.
Another consideration in choosing between various emission control strategies is that there is reason to believe that cap-and-trade programs are more vulnerable to rent seeking than are emission taxes designed to achieve equivalent reduction levels. Implementation of a cap-and-trade regime requires many more decisions about regulatory design than a tax regime, and that each decision presents the opportunity for rent-seeking behavior. While a tax can be designed to be relatively uniform, implementing a trading scheme necessarily requires many decisions about how to allocate and value allowances -- e.g. are the allowances to be allocated by auction, lottery, or past behavior? If by lottery, how is participation determined? If by past behavior, what behavior counts? What is the relevant time period? Is it purely retrospective, or partially prospective? What metric is to be used to evaluate comparable, but not identical, activities? Must some allowances be discounted in certain sectors to account for monitoring or enforcement problems? And so on. Users of allowances are not the only with something to gain through rent-seeking, those who seek to trade or broker allowances can also capture rents by influencing program design. (This was one of the reasons Enron fought so hard to get the Bush Administration to endorse carbon trading.) Insofar as rent-seeking involves a socially wasteful (and at times, even destructive) use of resources, the vulnerability of a system to rent-seeking should be a relevant consideration when choosing between various policy instruments.
The bottom line for me is that there is no one-size-fits-all answer to how to control harmful emissions. I like markets more than most, and readily advocate property rights solutions to environmental problems where they are viable (see here and here). Among other things, property rights approaches appear to create more powerful incentives to reduce transaction costs over time than do regulatory interventions. Lacking enforceable property rights in the relevant resources, or a ready means to move in that direction, we are inevitably forced to choose between a variety of imperfect policies. What sort of policy makes the most sense in a given context will depend upon a wide range of context-specific factors. In some cases cap-and-trade will work best. In others, traditional command-and-control or a pollution tax is the least bad option. In still others, we are best off adopting some sort of liability scheme or even doing little or nothing until other alternatives become more cost-effective.
In the climate context, at present, I would be inclined to support a carbon tax over a cap-and-trade scheme. If I had my way, a carbon tax would replace some other tax(es), such as the corporate income tax, and would be designed to be revenue neutral. But before implementing such a tax, I would do other things, such as replacing all energy subsidies, including those for renewable energy, with "http://volokh.com/posts/1170343187.shtml">prizes" for meaningful technological breakthroughs. For more of my recent thoughts on climate policy, see here and here.