In the debate over what to do about carbon emissions (and for today, let's start from the position that we are indeed going to do something) the general view has been that cap and trade is better than carbon taxation. Precisely because we can trade we can make the least cost reductions and thus solve the problem at that least cost.
I've always disagreed here and have always been in favour of carbon taxes: there's a number of reasons but George Monbiot (of all people!) highlights one here.
The EU promised that by 2020 all emissions permits would be sold at auction to the polluting industries. Now the heads of government have broken that promise: in 2020, big industrial polluters will have to pay for only 70% of the harm they do. Worse, companies will receive all their allowances for nothing if they can show that they're threatened by competition from firms outside the EU.
Cap and trade and carbon taxes should produce the same result at the same cost, but only if the politicians don't hand out pork to their favoured clients. And as George points out, and as even a passing knowledge of public choice theory would suggest, there's no way that politicians will not hand out pork to their favoured clients. Thus cap and trade is decidedly sub-optimal simply because politicians will do what politicians always do, buy support from certain groups and sectors by, well, being politicians. A carbon tax would be a decidedly better solution.
However, we do have one rather large problem here: how do we get it across to the politicians that they are the problem, not the solution?