Brad DeLong sums up:

I would say that to first order cap-and-trade and carbon taxes are the same, that there are five first-order differences:

• Cap-and-trade involves less redistribution because the losses of
  the losers are partially offset by their initial awards of tradeable
  permits.
• Cap-and-trade runs the risk that the cap will be set at the wrong
  place and so the price will go damagingly above its social optimum
  value.
• Carbon taxes run the risk that the tax will be set too low and so
  the quantity emitted will go damagingly above its social optimum value.
• Carbon taxes have the advantage that the government gets money that
  it can use for good–either to cut existing taxes that have large
  deadweight losses or to expand underfunded programs that have large
  social benefits.
• Carbon taxes have the disadvantage that the government gets money
  that it can use for ill, and that the recipients and beneficiaries of
  that ill-used money will then dig in and defend their rent-seeking
  gains beyond death itself.

and that there are two third-order differences:

• It’s easier to get not-too-bright Republicans to vote against
  something that is actually in their long-run interest if you can
  demagogue it by calling it a tax.
• It’s easier to get not-too-bright Democrats to vote for something
  that actually is not in their long-run interest if you can demagogue it
  by claiming that it’s just a restriction on the behavior of corporations and not something that directly impacts people.

The fourth-order considerations would, of course, look at time consistency problems and irreversibility problems.

As far as I know, there has never been a rigorous
ex post evaluation of CGE [computable general equilibrium] models in practice, one that compares
predicted to actual outcomes.

Here is the broader argument, hat tip to Mark Thoma (there is more at the link), and comments are open so can you prove him wrong?

Robert Samuelson writes:

Unless we find cost-effective ways of reducing the role of fossil fuels, a
cap-and-trade system will ultimately break down. It wouldn’t permit satisfactory
economic growth. But if we’re going to try to stimulate new technologies through
price, let’s do it honestly. A straightforward tax on carbon would favor
alternative fuels and conservation just as much as cap-and-trade but without the
rigid emission limits. A tax is more visible and understandable. If
environmentalists still prefer an allowance system, let’s call it by its proper
name: cap-and-tax.

Mark Thoma gets upset at this passage, here is Ryan Avent, Brad DeLong and Matt Yglesias, all upset.  Avent was the fount of the opposition:

Yowza. As any economist worth his or her salt will tell you, a cap and
trade plan with auctioned permits is essentially identical to a carbon
tax. That also happens to be exactly what Barack Obama is proposing.
So, another way for Samuelson to have written this column would have
been to title it, “Barack Obama has a good plan to reduce carbon
emissions."

But Samuelson is correct here and Avent is misleading.  When there is uncertainty about the location of the social optimum, and uncertainty about elasticities, a carbon tax and cap-and-trade are by no means equivalent.  If you see very high costs from setting the binding cap too low and choking off growth — as Samuelson mentions — you should prefer the carbon tax.  The price of carbon is more certain and you bear less risk from uncertainty about how fast solar power and other technologies will develop.  Alternatively, you might say that risk is transformed into price risk rather than "you can’t exceed this cap no matter what" risk.

Of course the postulated uncertainties are realistic in this context and you don’t have to invoke uncertainty about the science of global warming. 

If there is very high environmental risk to having emissions above a certain level, and we are unsure about the relevant elasticities (again, uncertainty about the pace of technological development can drive this), that militates in favor of cap and trade.  It is then easier to ensure that emissions do not exceed a particular level.

You can see that we are comparing the "growth threshold problem" to the "environment threshold problem."  Samuelson is apparently more worried about the former than the latter.  Maybe he shouldn’t be so sure he is focusing on the right problem, but on the economics he is on the mark in the criticized passage.

Addendum: Here is Mark Thoma with more on the topic, here is Megan McArdle on same.

My poor aging eyes could not read the pdf (you, however, can print it out or perhaps you are young and vigorous), but I believe this article is interesting with a very high "p."

When I was in 5th grade I participated in a charity event called "Jump Rope For Heart". People donated a certain amount of money per rope jump. I found myself wondering why it was structured that way – after all, people didn’t really care whether I jumped or not.

Many charitable events are structured this way, though typically they involve public walking.

Why do they work this way? Why not ask for lump-sum donations rather than having a bunch of people dig fence post holes? Is it make-work bias? Is it the labor theory of value? Maybe instead my willingness to jump/walk or otherwise participate indicates my commitment to the cause and in some sense certifies the event? A band-wagon effect? Maybe the dollar amount per unit of effort (jump, miles walked, etc) is so low that it induces people to donate more money than they would otherwise?

Rather than paying somebody to do busy-work, why not instead pay people to do something productive, such as soliciting even more donations?

Consider publicity as the main scarcity a charity faces.  If you elicit volunteers to walk, run or skip rope for you, those persons will talk up the activity — and the charity — to their friends, both ex ante and ex post.  They’ll even wear your T-Shirt "Cystic Fibrosis Marathon."

Since most of the people are exercisers anyway, the charitable activity doesn’t cost them much on net.  In fact the exercise is one way of expressing a greater commitment to the charity and may encourage subsequent donations.  Commitment, of course, is not infinitely elastic in supply.  So some of the person’s commitment may be transferred to the charity and away from the ideal of personal exercise.  Counterintuitively, in the long run the person may end up less fit but more committed to the charity.  In other words you’re paying with some of your health and discipline rather than with your money.

Well…airplane drop.  Consider it another instance of zero overhead giving.