With Hillary Clinton renewing her declaration of a popular-vote lead in the Democratic nomination contest, it’s worth checking in on the actual math. Deriving an exact popular-vote count is tricky, since there are several ways of computing it. Here’s a step-by-step look at how the various tallies are reached.
If you count all of the states and U.S. possessions that have held officially sanctioned primaries or caucuses in which exact popular-vote tallies were maintained, the current tally (with a scattering of votes in Oregon yet to be added) is:
Difference: Obama +415,976
Baseline + Florida
A case can be made that the results from Florida’s nonsanctioned January primary ought to be counted as well. Even though the candidates didn’t campaign in the state and overall turnout was much lower than in other states, the final result—a 17-point Clinton win over Obama (with John Edwards at 14 percent)—is consistent with polls taken before and after the primary. If you add the Florida numbers (and not Michigan, where the circumstances were different) to the baseline, then the current popular-vote total stands at:
Difference: Obama +121,204
Baseline + Florida and Michigan (without uncommitted)
This is the formula that Clinton uses to claim a popular-vote lead. The case to include Michigan’s primary numbers in a popular-vote tally is weaker and more complicated than the case for Florida. Not only was turnout even lower in Michigan than Florida, Obama’s name wasn’t even on the ballot. In the January primary, Clinton “won” by 15 points over an uncommitted slate that many Obama supporters voted for. However, polls before and after the primary showed Obama—had he been on the ballot—would have run roughly even with Clinton. The Clinton campaign, in its math, credits Clinton with all of the votes she won in Michigan, and Obama with zero, giving him no credit for his support in the state. If we accept this formula and add Michigan to the mix, Clinton passes Obama in the overall popular vote:
Difference: Clinton +207,105
Baseline + Florida and Michigan (with uncommitted)
If the Obama campaign does concede Michigan, it will no doubt be on the condition that their candidate at least receive credit for the votes that “uncommitted” won in the state. Almost all of them were meant for him anyway (since his name wasn’t on the ballot), and it’s almost certainly a conservative expression of his support, since polls before and after the primary showed him running even with Clinton—not losing to her by 15, as “uncommitted” did. Moreover, the Obama forces would say, it’s unfair to punish Obama for removing his name from the ballot, when most of the other candidates did so as well in an effort to honor the D.N.C.’s boycott request. If you count Michigan and give the 238,168 uncommitted votes to Obama, he regains the cumulative popular vote lead:
Difference: Obama +31,063
The Four Caucuses
And here’s the real wild card: Four states—Washington, Iowa, Maine and Nevada—held caucuses in which formal popular-vote tallies were not maintained. No one disputes, however, that hundreds of thousands of Democrats participated in these caucuses. Media outlets have estimated the popular vote in these states in different ways. Some have relied on exit poll figures and some have only counted the state convention delegates won by each candidate (tiny numbers compared to overall turnout). But the most widely accepted formula involves simply applying each candidate’s delegate share (i.e., the percentage that gets reported in the media) to each state party’s overall turnout figure. So in Maine, for instance, Obama would be credited with 59 percent of the 44,000 overall votes that the state party estimated were cast.
These numbers are inexact, but probably not by much. The volatility in caucus results, generally speaking, is due to large fields of candidates in which many voters switch their allegiance on a second ballot. But because three of these caucuses—Iowa is the exception—were basically two-way contests between Clinton and Obama, there really wasn’t much second ballot movement in them.
Using the above-described formula, here are the popular-vote estimates from Washington, Maine, Nevada and Iowa:
Difference: Obama +110,222
Now, the question is how to apply this data to the overall vote count. The Clinton campaign, obviously, would argue that inexact vote estimates do not belong in any popular-vote tally. Of course, they’d argue the exact opposite if these numbers worked in their favor. Conversely, the Obama campaign would contend that these caucus numbers are actually more worthy of inclusion than Florida and Michigan, since the four caucuses were all officially sanctioned contests where the candidates all had the opportunity to campaign and spend money.
Here is how Obama’s 110,222-vote plurality from the four caucuses would, if included, affect the overall margins under the four different standards described above:
Baseline + Florida: Obama +231,426
Baseline + Florida and Michigan (without uncommitted): Clinton +96,883
Baseline + Florida and Michigan (with uncommitted): Obama + 141,285
The Remaining Primaries
The final wrinkle is what will happen in Puerto Rico, Montana and South Dakota.
Obama should win South Dakota and Montana handily, resulting in a net gain of somewhere between 40,000 and 60,000 votes.
Clinton should win Puerto Rico decisively and supporters have been talking up the possibility of very high turnout in that commonwealth—perhaps 1,000,000 voters. However, a more reasonable turnout estimate is probably around 600,000, a figure provided by Puerto Rican election expert Manuel Alvarez-Rivera, who advises that U.S. presidential politics do not stir the passions of islanders nearly to the degree that their own politics do. Also, last Saturday’s El Nuevo Dia reported that the number of polling stations for the June 1 primary had been slashed due to low turnout estimates and a lack of volunteer poll workers. A 20-point Clinton win with a turnout of 600,000 would produce for her a plurality of 120,000 votes.
Roughly speaking, then, Clinton is likely to post a net gain of around 50,000-75,000 votes in the final three contests. This is just an estimate, obviously—high turnout in Puerto Rico could swell her number substantially, just as Obama could eat into it if he fares better in Puerto Rico than conventional assumptions say he will.
The bottom line is that if Clinton picks up a net gain of 75,000 votes in Puerto Rico, Montana and South Dakota, she will be able to claim a popular-vote lead over Obama, but only under two narrow standards: counting Michigan, and giving the uncommitted votes to Obama, but not counting the four caucus states; or counting the four caucus states, but without giving the uncommitted vote from Michigan.