A few letters in this week's NYT Book Review call Judge Posner for task respecting his latest NYT Book Review essay, this one on modern media. He showed his intellectual asshole side -- some wanted the guy on the Supreme Court; the place has enough people a bit too full of themselves already, thank you, and I again feel sympathy for the people who have to come into contact with him in a professional way.
[Update: Krugman furthers my point by defending himself in a follow-up editorial. I was a bit more specific -- adding bits in brackets. I'd add Krugman is too cautious in saying that Bush "would still have been declared the winner" (unless the overseeing judge revised his order) if the Florida Supreme Court ruling stood. A few too many factors could have kicked in to be that sure of oneself. His point about prettifying history, however is right on the money. Hasen might have a follow-up too, I guess, but I think the below comments are useful since his viewpoint is representative of others.]
Rick Hasen, who has a well received blog on Election Law, unfairly (imho) criticized a recent "misleading" Paul Krugman column. The column referenced a new book on the elections of 2000 and 2004, which also apparently discusses the problems of federal elections these days as a whole. Krugman noted that the author argued that a full manual recount would have clearly resulted in a Gore victory. This was misleading to Hasen because Gore did not ask for such a recount and (here, he wasn't sure) the undervote recount that was authorized would have brought a Bush victory, according to press accounts etc.
Krugman did not say otherwise. Hasen fails to note that the actual recount ordered by the Florida Supreme Court was not asked for by Gore either: Gore focused on undercounts in a few counties with faulty machines, but the court authorized a full [manual statewide] recount [of all ballots in which a machine failed to register a vote for President]. Also, Hasen fails to note reports that the judge appointed to be in charge might very well have included overvotes as well. Hasen also suggests Krugman implies that Ohio was stolen too:
Krugman also makes claims about the vote being stolen in Ohio in 2004. From what I have seen so far (including the Conyers report), I am not convinced that intentional action by state officials cost John Kerry the vote in Ohio.
Krugman cites the "less cautious" (Krugman's words) Conyers report as saying that there were many problems, but not enough for a Kerry win. He does not imply that he thinks that report was wrong to think this. Krugman cites the stolen Ohio argument to in effect show that those who think the system is broken are not all the same. In particular, the book he cites separates 2000 and 2004. It is interesting that Hasen speaks of "intentional action" (what does this mean? it implies unintentional action might have, which is a compelling statement). Still, Krugman receives enough unjustified criticism from the usual suspects for Hasen not to add himself to the mix.
Hasen also sets forth a few general reactions. First, noting that even Judge Posner (who he fails to note supported the Supreme Court action in effect out of distaste with trusting the people with the messy recount) acknowledges the point, it is "pretty clear" that a majority of Florida voters intended to vote for Gore. The problem is that not "all" of them translated their intent to votes that could be counted by existing voting technology.
Fair enough. Next, he says that for "all practical purposes, the Florida count was a statistical tie." From this, he suggests various things (legal or otherwise, including the Nader candidacy, partisan state election officials, and the oversea military vote not being counted via actual Florida regulations at the time) could be point to in various counterfactuals. And, finally, "From all of this, I don't think the question of who 'really' won Florida (opposed from the legal conclusion that Bush won) is one that can be answered in any sensical way."
This doesn't follow. The "misleading" tie argument actually helps Bush and Posner sorts a great deal. Though a structural argument can be made that the federal courts should not have stopped the count (the process was primarily left to the states and Congress), a "tie" would mean the end result didn't really matter. Sure, the Supreme Court helped stop the count etc., but since it was a tie, there is at best a 50/50 chance that the wrong person benefited. What's the big deal? And, stopping the count had other values, so heck, we come out ahead.
I'm not sure what "practical" purposes he is talking about. Statistically, it is clear that Gore won. One can be "pretty sure" about this, perhaps even via a clear and convincing standard (some would go up a notch). To point to but one issue, the overvotes went Gore's way by a ratio of about 2:1. Now, using current voting law, an election official might not be able to take this into consideration. But, he uses the word "statistically." If the vote went to Congress, could they not determine that an equitable result would be a Gore victory? Arguably, their "count" could take statistical probability into account. Confusing.
Also, why cannot we discuss who "really" won? The question would turn on how we define the term. Gore might have statistically been ahead beyond the margin of error, and in some real sense it wouldn't matter. He's not President. But, since various factors can be pointed to that suggest it is much more likely that he won than Bush, Hasen's statement is dubious. In fact, he himself points out to the intent of the voter -- wasn't that actually the legal standard?
Annoying really. Actually, maybe next time you can actually read the book being cited as well as not disparaging an editorial for doing something it actually didn't do, especially one that adds force to the "2004 was not stolen" brigade by someone the stolen faction respects.