UPDATED: When the Last shall be wrong

Pro-Hillary op-ed uses fuzzy math

by Will Bunch | Columnist

Published May 23, 2008, 3:51 p.m. ET

Conservative Inquirer pundit Jonathan Last is in "Operation Chaos" mode, trying waaaaay too hard here to make Hillary Clinton the Democratic nominee, or at least prolong the nomination fight to the advantage of his presumably favored GOP candidate John McCain. He has an op-ed piece today that's currently the most widely read story on Philly.com, entitled "In most inclusive count, Clinton has the numbers." It attempts to make the case that by June 3, Clinton will have more popular votes than Barack Obama, assuming you use a formula that's not the most pro-Clinton but is still fairly pro-Clinton (not using Michigan, where Obama wasn't on the ballot, but also disregarding the caucus results from several states that Obama handily won.)

For one thing, the whole popular vote argument is a little silly, since -- and this is coming from a hardcore, lifelong political junkie -- this is a measurement that has never, ever been used before. Winning the nomination is all about getting delegates, just as winning in November is all about getting electoral votes. If I have time later, I'll be sure to dig up all the columns that Last wrote in 2000 declaring the illigitimacy of George W. Bush and annointing Al Gore our 40th president.

But even so, I would have let the matter drop until I saw the fuzzy math that Last used to give Hillary those "numbers" from the headline:

The third and fourth counts - the ones which include Florida - seem more fair. Here, Obama is clinging to a slight lead of 146,786 votes (257,008, with the caucus estimates). However, with Puerto Rico, Montana, and South Dakota remaining, he will almost certainly finish behind her in these counts, likely by a few hundred thousand votes.

But could Clinton take over the lead in all of the popular-vote tabulations? Quite possibly. In Puerto Rico's last major election, two million people voted. Let's assume that turnout for this historic vote - Puerto Rico has never had a presidential primary before - will be equal to or greater than that turnout.

If Clinton were to win Puerto Rico by 20 points she would pick up at least a 400,000-vote margin. This would allow her to swamp Obama in the popular-vote counts, which include Florida, making her the leader in four of the six permutations of the popular vote. At that point, Obama would be left clinging to the least-inclusive count, which he now leads by 441,558 votes (551,780, including caucuses).

OK, it's been a while since my poly-sci major days, but a couple of things leaped out at me here. One, he notes that the last three primaries are Puerto Rico, Montana and South Dakota, but then never mentions Montana or South Dakota ever again. Could that be because Obama is heavily favored to win those two states, an inconvenient truth for Last's premise?

But what really got me going was his vote estimate for Puerto Rico, which assumes a turnout equal to a recent election there. Except a) not everyone in Puerto Rico is a Democrat and b) primary turnout is usually much lower than general election turnout and c) the biggest predictor of voting is habit, and if it's a first-time primary that would actually depress turnout, not increase it. So I immediately searched to see if anyone else thinks 2 million people will vote in Puerto Rico, the cornerstone of his whole premise.

It took a very short time to learn the truth: No.

But scant attention has been paid to Puerto Rico and most analysts seem to have simply accepted the nice round estimate that 1,000,000 residents will participate in the June 1 Democratic primary.

The estimable election expert Michael Barone, in his much-discussed analysis of the popular-vote possibilities, attributed this forecast to the fact that "turnout in Puerto Rican elections is, as a percentage of those eligible, higher than anywhere on the Mainland, something on the order of 80 percent as compared with 61 percent in the 2004 presidential general election."

But that fact might not have much bearing on this year's primary vote. Manuel Alvarez-Rivera, an expert on Puerto Rican politics, told me that for several reasons "the estimate of one million voters for the upcoming June 1st primary is way too high." He suggested a more reasonable turnout figure of 600,000, "give or take 100,000."

Wow, so Last is probably off here by a factor of 3, or even 4 -- even the most optimistic forecast is just half of the number he pulled out of thin air, apparently.

I do agree with him about one thing, and that is that Clinton, who's popular with Latino voters, will probably win in Puerto Rico -- but not by nearly enough to erase Obama's lead in any calculation except the one that counts a large state, Michigan, which did not even have Obama on the ballot. So the entire premise for this "widely read" article is wrong, with easily available facts.

So I'm not sure is Last is being just lazy here, or intellectually dishonest.

But in this case, the Last shall be wrong.

UPDATE: This will not help Hillary with the math:

Hillary Clinton today brought up the assassination of Sen. Robert Kennedy while defending her decision to stay in the race against Barack Obama.

"My husband did not wrap up the nomination in 1992 until he won the California primary somewhere in the middle of June, right? We all remember Bobby Kennedy was assassinated in June in California. I don't understand it," she said, dismissing calls to drop out.

That's an interesting strategy, huh?