How pessimistic should we be about Iraq?

Barry Rubin, editor of the Middle East Review of International Affairs, argues in the Jerusalem Post that the U.S. lacks the ability to control events in Iraq and should turn over power in the next twelve months and leave.
Why the pessimism? Because, in Rubin’s view, “any elected or appointed regime in Iraq will face tremendous opposition, to the point of civil war. Even the best coalition representing the broadest possible number of interest groups with the most honest possible officials is not going to be able to govern the country.” The U.S. can attempt to prop up such a regime, but ultimately it must either sponsor a regime that will be overthrown, and thus end up on the losing side, or sponsor a regime that survives and wins by ruthless repression. Most likely, the result will be “an ongoing war in which radicals unite against the US presence, while American taxpayers finance an attempt to rebuild an economy for extremists to inherit.”
On the other hand, if the U.S. withdraws, Rubin believes that the outcome will be “a military strongman or a nationalist Shi’ite-dominated regime with some Islamic flavoring. Its attitude toward the US will depend on how much mileage the rulers think they can get by blaming America for all their problems compared to how much American help they will need for reconstruction and help against their own enemies.” Not the result we had hoped for, but probably not a disastrous one either.
Should we settle for this? Not, in my view, on the current evidence. It is just as possible to oversell pessimistic scenarios as optimistic ones. I have seen clever analysts demonstrate why the Soviet Union would never fall, and why Reagan’s policy in Central America would not succeed. How do we know that the level of resistance to any possible coalition government will be too great to deal with absent unacceptable levels of repression? The answer is, we don’t. Rubin’s doomsday scenario is as much an a priori one as Paul Wolfowitz’s rosy democracy-building one. I agree with George Will that we need to be empiricists about this. And to be good empiricists, we need more data, which means we need more time. Presumably, there will always be time enough to settle for a military strongman or a Shi’ite dominated Islamic regime, if it comes to that.