To begin with, there's a big difference between the possibility of closed loops on rank-and-file conservatives compared to the effect on conservative elites. I think the former is no doubt true, but I'm not sure that it has the policy effects one might think. Consider the disastrous effect of the Iraq War on Republicans in the 2006 election. One might think that information loops are partially responsible: Republicans spend three years using Fox News and Rush to inform voters that everything is just peachy keen in Iraq, and then when politicians want to change course, they are constrained by a rank-and-file that strongly believes in staying the course. But I suspect that's wrong; had Republican elites wanted to change course earlier (either to a surge, or to retreat), it wouldn't have been difficult at all to rapidly educate Republican voters of the need for the new policy -- as was in fact the case for the surge after the 2006 elections, and the retreat agreed to prior to the 2008 elections. The very fact that the rank-and-file (of all parties) tend to follow opinion leaders means that making that process extraordinarily efficient may only give leaders more freedom, not less -- although to be sure those interested in public opinion may want to study whether that's actually true.
If there's a problem with policy, then, I think it's because conservative leaders are believing their own rhetoric -- something that may or may not be true. It's hard to tease out the effects of this stuff on voters, but it's even tougher figuring out what pols or talk show hosts really think, since they're apt to say what they think voters want to hear whether they believe it or not. However I do think that it's fairly obvious that if conservative policy-makers truly believe things that are not true, then their efforts to make good policy are likely to fail. So: for policy dangers, I'll argue that the real issue is information loops at the elite level.
On the other hand, if epistemic closure (fine, I'll say it -- it's just awful jargon, though) is helping the Republicans politically, then we're talking about the mass level, not the elite level. It can't possibly help Republicans to have John Boehner actually believe any fictional talking points he's using, although I can imagine it hurting them.
I disagree, however, with both Douthat and Drum on the question of whether Fox News et al. are good for the Republicans politically. The big data point everyone points to is the first two Clinton years (pre-Fox, but with many other pieces in place), but while I do think that Bob Dole's filibuster strategy hurt Clinton, I've always argued that it was Clinton's own weaknesses that hurt him. Health care reform in 1993-1994, in my opinion, wasn't defeated because of Rush and Betsy McCaughey; it was defeated because Ira Magaziner (and Hillary Clinton and Bill Clinton) did a terrible job, and the failure of health care reform hurt the Dems -- as did alienating labor over NAFTA, and various other policy and political mistakes. But after that...in which elections did Republicans do better than one would expect based on economic and other fundamentals? Surely not 2008 or 2006, and I'd strongly argue that 2004 doesn't fit, either. Douthat argues for very recent GOP political success:
In a sense, the last eighteen months have been enormously successful for conservatives: The polls have turned decisively against the Democrats, the Obama White House, and liberalism in general; the Republicans have won a series of elections they weren’t expected to win; and conservatives look primed for bigger gains in November.But that "series of elections" is really just the MA Senate race, and possibly the NJ contest for Governor; it was hardly a surprise that Republicans won in Virginia with a Democrat in the White House, and while Republicans might win the upcoming House specials, so far they've lost ground, not gained ground, in House special elections. As far as polls turning "decisively" against the Obama White House, in fact Obama is right around the same 50% he's been at for months now. Which, by the way, is better than Ronald Reagan was doing at this point in his presidency. And that's really the point, isn't it? There's no need to look beyond the economy to find reasons for Obama's tepid approval ratings, or the somewhat worse polling numbers for the Democrats. Really -- if all we knew was that unemployment would sit at 9.7% in April 2010, would we expect any president (or his party) to be doing well in the polls? If Republicans do manage to win 65 House seats and 10 Senate seats and a bunch of statehouses, especially if they do so with an improving economy, then I'll start looking for explanations. Right now? Those things haven't actually happened, and so crediting them to anyone seems awful premature to me.
So, no, to answer Dothat's question, I don't think conservatives "need Fox News." For all the hype, it's very possible that the political achievements of Fox News, Rush, and the rest of it are limited to winning a whole bunch of ephemeral news cycles -- and if you think that's a big deal, I suspect that the guy who lives in the big house on Pennsylvania Avenue would be glad to tell you some stories about Hillary Clinton winning the nomination, John McCain winning the election, and seven or eight deaths of health care reform.