Josh Marshall thinks that's a "cop-out for the ages":
"The rationale for calling this a tie, according to the Des Moines Register, which has the story as an exclusive, is that 8 precincts’ numbers are lost permanently and will never be certified. So in practice it’s a tie, too close to call, etc. That of course probably applies to pretty much all recount type elections — Bush v Gore, maybe Franken v. Coleman, etc. The vagueries of the process itself is too imprecise in some sense to tell you who ‘won’ in some Platonic (the other sense of the word) sense. But in normal elections where the people holding it aren’t deeply invested in not letting one guy win we have a name for that kind [of] situation — Rick Santorum won."I guess I'll just have to add the views of Josh Marshall and the Iowa Republican Party to the immense list of things I don't get. Neither of them, as far as I can see, is making a lick of sense.
Let me see if I've gotten this straight. We have two vote counts, the original one conducted on the night the votes were cast and the recount undertaken over the last couple of weeks. You'd have thought that the original vote raised at least a rebuttable presumption that Romney won by 8 votes. Although I don't know it for a fact, I presume that the vote-counting methodology employed in the recount is at least marginally less fallible than the vote-counting methodology used on election night since, otherwise, there wouldn't be any point to recounting the votes in the first place. So let's stipulate that, to the extent the universe of votes counted and then recounted overlap, the recounted totals are more reliable than the totals yielded by the original count.
The question before us, then, is whether the recounted totals rebut the presumption that Romney won by 8 votes raised by the original count. Any vote-counting methodology, to be sure, comes with a margin of error. Yet we know one thing for sure: the universe of votes originally counted and recounted in this instance don't overlap, at least to the extent of the votes coming from the eight missing districts. So how can the recount rebut the presumption raised by the original count?
To join Marshall in calling Santorum the winner you have to subscribe to the preposterous notion that counting all the votes through unobjectionable methods on election night is a less reliable measure of an election outcome than recounting only some of the votes with marginally more accuracy. That's to say that you have to believe that vote totals compiled through normally acceptable methods that may be incorrect are less reliable than totals that are known to be incorrect with utter certainty. And to join the Iowa Republican Party in calling the election a tie you have to subscribe to the equally preposterous notion that the count and the recount are equally reliable measures of the election outcome that therefore cancel each other out.
Unless I'm missing something under these circumstances, there's still only one intellectually defensible answer to the question of who won the Iowa Caucuses: that would be Mitt Romney.