Political scientists and all around smart guys Carl Scott (in this thread) and James Poulos make the argument that Michael Steele should be fired regardless of procedural and public relations obstacles. I agree, but let's add up the costs and benefits of firing Steele.
If Steele is fired he might well become a critic of the GOP and go on liberal-leaning media to peddle whatever anti-Republican narrative puts him in the best light and plays to the prejudices of those running those media outlets. So from that perspective, it might make sense to have Steele making a fool of himself inside the tent than causing trouble outside. But as Rich Lowry points out, even if Steele isn't fired, he probably won't be reappointed when his term expires in 2011 and he "will be sorely tempted to run to MSNBC to tell the world how awful his party is." So not firing Steele now doesn't so much avoid the risk of criticism from Steele as much as shift it from now (or next month) to next year. Maybe there is good reason to want to risk that criticism next year rather than later this year. But what are the costs of keeping Steele? Those costs include a) more Michael Steele gaffes and b) losing out on whatever good an effective RNC chairman could accomplish. It would seem to me that the costs of keeping Michael Steele far exceed the costs of dumping him. I know that this analysis partly depends on the assumption that Steele's replacement would be competent. A guy can hope.