Friday, May 30, 2008

Delegate Math And Popular Vote Myth

Delegate Math And Popular Vote Myth

The Democratic National Committee's Rules and Nomination Committee is preparing to meet in Washington Saturday to decide the fate of calendar cheaters Florida and Michigan. Most observers are instead expecting delegates from the two states to be seated with a half vote each, though there's several variations on that theme.

One thing the committee won't settle is the debate on how to define the popular vote. How much of the Michigan and Florida fight about the delegates, at half or full strength, and how much is about popular vote claims? If the two states are seated in any way based on the too-early primaries, the votes "counted."

"Count all the votes" is Hillary Clinton's mantra, as she claims a popular vote lead in an effort to persuade superdelegates. But the formulas that show Clinton in the lead don't count all the votes -- and Iowa is among the states excluded.

"We are urging 100 percent of the delegations be seated and that each delegate have a full vote," said Hillary Clinton strategist Harold Ickes. That's an unlikely scenario.

MSNBC's Chuck Todd says one scenario gaining support involves seating Florida at half strength with results based on the Jan. 31 primary, as Clinton wants. But the Michigan delegation, also at half strength, would be evenly divided, as Barack Obama's campaign requests.

Todd also notes that giving all the delegates a half vote is a better scenario for Clinton than cutting the number of delegates in half. "If Florida delegates are seated in their entirety, but only have their vote counted as a .5, then Clinton will net approximately 19 delegates out of the state," he writes. "But if the delegation is cut in half, that's done in every congressional district as well as statewide. Then suddenly Clinton's advantage is only a net of six."

This solution makes it harder for Clinton to argue for counting Michigan in her popular vote math. But the popular vote math is impossible to calculate anyway, and is a matter of definition. We Iowans are one of the problems.

Iowa is among four caucus states (Nevada, Maine and Washington are the others) that release only delegate counts and not raw vote totals. The secrecy is part of Iowa's long-time truce with New Hampshire. In the 1980's Iowa and New Hampshire agreed that Iowa would have the first caucus and New Hampshire would have the first primary. If Iowa Democrats would release a raw vote total, the results would be too much like a primary for New Hampshire's taste, and the truce would crumble.

But with the national popular vote count now an issue in the campaign, and a linchpin of Clinton's last-ditch effort to persuade superdelegates, Iowa's quirky results are now more than just an annoyance to national reporters who want a straightforward vote total. You know, the kind Iowa Republicans provide. If the Clinton rhetoric is consistent, the lack of a raw vote total means... Iowa's votes aren't being counted!

Real Clear Politics adds the math several ways and tries to estimate the "vote totals" from Iowa. Those vote totals, of course, don't exist. Iowa's county convention delegates are based on realigned totals. Those don't reflect nonviable groups, like the five people out of 315 in my precinct whose first choice was Chris Dodd. Still, an estimate is something to go on.

Here's the scenarios, with various questionable totals included. This excludes the two states set to vote on Tuesday, Montana and South Dakota, and also excludes Puerto Rico which votes Sunday (though that begs the question of whether a commonwealth that does not vote in a general election should be considered as part of a "popular vote total.").

FLMIIA-NV-ME-WA estimatesObamaClintonSpread
nonono16,685,94149.1%16,227,51447.7%Obama +458,427+1.4%
nonoyes17,020,02549.1%16,451,37647.5%Obama +568,649+1.6%
yesnono17,262,15548.3%17,098,50047.8%Obama +163,655+0.5%
yesnoyes17,596,23948.3%17,322,36247.6%Obama +273,877+0.7%
yesyesno17,262,15547.5%17,426,80947.9%Clinton +164,654+0.45%
yesyesyes17,596,23947.6%17,650,67147.7%Clinton +54,432+0.15%

The two scenarios with Clinton ahead both include Michigan, and here best scenario excludes Iowa. Florida is straightforward enough: no one campaigned, and the vote totals may have been lowered by the widespread news that the state's delegates would not be seated. But at least there's a vote total for the Obama column.

Michigan is the problem, since Obama took his name off the ballot. Any scenario that argues a Clinton popular vote victory is based on a 328,309 margin out of Michigan: Clinton 328,309, Obama 0, Kevin Phillips-Bong naught.

That doesn't account for 238,168 uncommitted votes that were most definitely cast against Clinton. Now, it's probably not fair to credit those to Obama. John Edwards was also off the Michigan ballot and still in the race when Michigan voted on Jan. 15.

But here's a formula I've never seen applied elsewhere. Credit Clinton for her 328,309 Michigan votes, but then subtract the 238,168 uncommitted votes. That lowers her gain out of Michigan to 90,141 -- her margin of victory over uncommitted. Now apply that new math, and Clinton's best scenario (where Michigan counts but Iowa doesn't) slips from a lead of 164,654 to a deficit of 73,514.

In the end, of course, delegates choose the nominee, not a mythical national popular vote. The popular vote is only a tool for persuading superdelegates. But Clinton's strained efforts to selectively calculate the total make it a less persuasive tool.

No comments: