Projection: Clinton Wins Popular Vote, Obama Wins Delegate Count

Clinton could win the popular vote and Obama the delegate count, making it an even harder decision for superdelegates


The Clinton campaign has taken to boasting that its candidate has won states with more electoral votes than has Barack Obama. True. By my count, Clinton has won 14 states with 219 electoral votes (16 states with 263 electoral votes if you include Florida and Michigan) while Obama has won 27 states (I'm counting the District of Columbia as a state, but not the territories) with 202 electoral votes. Eight states with 73 electoral votes have still to vote. In percentage terms, Clinton has won states with 41 percent of the electoral votes (49 percent if you include Florida and Michigan), while Obama has won states with 38 percent of electoral votes. States with 14 percent of the electoral votes have yet to vote.

The Clinton campaign would do even better to use population rather than electoral votes, since smaller states are overrepresented in the Electoral College. By my count, based on the 2007 Census estimates, Clinton's states have 132,214,460 people (160,537,525 if you include Florida and Michigan), and Obama's states have 101,689,480 people. States with 39,394,152 people have yet to vote. In percentage terms this means Clinton's states have 44 percent of the nation's population (53 percent if you include Florida and Michigan) and Obama's states have 34 percent of the nation's population. The yet-to-vote states have 13 percent of the nation's population.

Thus the Clinton campaign could argue that Obama cannot win states with most of the nation's people even if he wins all the remaining eight primaries. Could argue—but I don't think that's going to persuade any superdelegates that Clinton is the real winner.

The Obama campaign has argued on occasion that its primary or caucus victories in Republican states means that Obama has a better chance to carry them in the general election than Clinton. As the Clinton people point out, that's ridiculous in some cases: No one thinks Obama's victories in lightly attended caucuses in Idaho or Wyoming mean that he can win them in November. Even in states like Minnesota and Colorado, Obama's caucus wins are less persuasive evidence than current polls that he can do better there than Clinton in November. Nor are Clinton's primary victories in states like Nevada, New Hampshire, New Jersey, New Mexico, and Ohio very strong evidence for the proposition that she'd be stronger than Obama. General election polls are better evidence; they buttress Clinton's case in New Jersey and Ohio, and refute it for Nevada, New Hampshire, and New Mexico. Interestingly, Clinton won primaries in only five states which went heavily for George W. Bush in 2000 and 2004—Arizona, Arkansas, Oklahoma, Tennessee, and Texas.

This has led me to ask what would have been the result of the Democratic primaries and caucuses if the party's rules tended to allocate delegates by winner-take-all rather than proportional representation. It would be an interesting exercise to apply the Republicans' delegate allocation formulas to the Democratic results. Interesting—but also time consuming, since those formulas tend to allocate many delegates by congressional district (or, in Texas, state Senate districts). So instead, using the summary, I simply assigned all of a state's Democratic delegates to the winner of the Democratic primary or caucus. The result: Hillary Clinton gets 1,430 delegates and Barack Obama 1,237. That's almost the exact opposite of's count of "pledged" (i.e., selected in primaries or caucuses): Obama 1,414, Clinton 1,247. It should be noted that the winner-take-all score would have been reversed if Clinton had lost Texas, which she carried by the narrow margin of 51 percent to 47 percent and which has 193 delegates.

That's an Obama margin of 167 delegates. And most of that margin came from caucus states and territories, where Obama's delegate lead was, by my calculation, 266 to 141—a margin of 125 delegates. (I'm leaving aside the minority of Texas delegates chosen by caucus.) In the primary states Obama's margin was just 1,148 to 1,106, a delegate margin of only 42.

It's at least theoretically possible for Clinton to overcome this lead in primary-chosen delegates in the eight remaining primaries. That would give the Clinton campaign another basis for arguing that their candidate is really the choice of the people. But the fact is that the Clinton campaign has only itself to blame for its weakness in caucus-chosen delegates. The caucuses were there on the schedule all along, and the Clinton campaign had as much time and about as much money to prepare for them as the Obama campaign did. The Clintonites simply did not prepare as well as I am sure they now wish they had. I suspect that some of the anger we see from Clinton backers comes from their own reflection that if they had planned and executed better they would be ahead in delegates now rather than behind. You get really angry when you have no one to blame but yourself.

While we're talking numbers, here are a couple of interesting charts. First, from the Democratic MyDD website, here is a projection of Pennsylvania voting based on the results in demographically similar counties in Ohio. It projects a 57 percent to 43 percent Clinton win. (Hat tip, Jim Geraghty.) And at, Jay Cost has prepared a spreadsheet on which you can put your own projections of the popular vote in the eight remaining primaries.

I couldn't resist using Jay Cost's spreadsheet to calculate the popular votes in the remaining primaries and my own old-fashioned legal pads to calculate delegate results. I used Cost's default turnout numbers and estimates of the two-candidate percentages which I consider optimistic from the Clinton point of view but not wildly unrealistic.


State Eligibility Kerry Votes Expected Margin Expected Margin Clinton Votes Net Clinton Margin
Pennsylvania Closed 2,938,095 63.0% 1,851,000 20.0% 370,200
Indiana Open 969,011 82.0% 794,589 20.0% 158,918
North Carolina Open 1,525,849 82.0% 1,251,196  -10.0%  -125,120
West Virginia Open 326,541 82.0% 267,764 40.0% 107,105
Kentucky Closed 712,733 63.0% 449,022 30.0% 134,707
Oregon Closed 943,163 63.0% 594,193 -10.0% -59,419
Puerto Rico Open N/A N/A 1,000,000 30.0% 300,000
Montana Open 173,710 82.0% 142,442 20.0% 28,488
South Dakota Closed 149,244 63.0% 94,024 20.0% 18,805
    Total Net Clinton Votes 933,684      

This would eliminate Obama's current popular vote margin, without including Florida and Michigan totals and even if you use imputed vote totals for the four caucus states (Iowa, Nevada, Maine, and Washington) where Democrats did not disclose vote totals. The current popular vote margin for Obama on is, under those favorable assumptions, 827,498. My spreadsheet numbers would give Clinton a 106,186 margin. The Obama margin if you don't give him his imputed margin in those four caucus states is 717,276. My results would convert that to a Clinton popular vote margin of 216,408.

But note a couple of other things. One is that this popular vote margin is exceedingly small when measured in percentage terms. With my estimate of 6,444,230 turnout in the remaining primaries, that yields a total Clinton-Obama turnout (with the four imputed caucus states included) of 32,995,378. The Clinton popular vote margin with the imputed caucus result was, as noted, 106,186, which is 0.32% of the total.

The other thing to note is that all of Clinton's popular vote margin and more comes from Puerto Rico. The turnout in other extraterritorial jurisdictions was very small: 1,921 in the Virgin Islands, 22,715 among Democrats Abroad and 284 [sic] in American Samoa. I'm projecting a turnout of 1 million in Puerto Rico, which has a population of 4 million. 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 Puerto Rico has not had a presidential primary before, so no one knows what turnout will be like. Puerto Rico will also be a challenge for the candidates. How do you campaign for the June 1 primary there and also campaign for the June 3 primaries in South Dakota and Montana?

Are my projections for Clinton's share of the vote too optimistic? Quite possibly. But I think they're at least defensible. I have her carrying Pennsylvania by 20 percent--a 60 percent to 40 percent margin of the two-candidate (Clinton and Obama) vote. That's better than she did in Ohio, where she won 55 percent of the two-candidate vote. But her showings there in the 6th congressional district (70 percent to 27 percent), the 17th congressional district (63 percent to 35percent) and the 18th congressional district (66 percent to 31percent) have influenced me; those areas are a lot like most of western and central Pennsylvania, where you also find very few blacks and upscale whites. Those results have also influenced my projections of even bigger percentage margins for Clinton in Indiana, West Virginia, and Kentucky. I projected a 10 percent margin for Obama in North Carolina; the average of recent polls has him ahead 57 percent to 43 percent in the two-candidate vote. I have Clinton losing also by 10 percent in Oregon. That’s roughly comparable to her showing in the nonbinding February 19 primary in next-door Washington, where she got 47 percent of the two-candidate vote. I have Clinton winning Montana and South Dakota by 20 percent margins, when the conventional wisdom seems to be that these states lean to Obama. It’s true that Obama did very well in caucuses in Minnesota, North Dakota, Nebraska, Idaho, and Wyoming. But my hunch is that the wider primary electorate will go the other way. The closest comparable I can come up with is the nonbinding primary in Washington, where the vote in eastern Washington, the heavily Republican area east of the Cascades, went 50.3 percent to 49.7 percent for Obama. I don’t think he’ll do as well in Montana or South Dakota as he did in his halcyon days in February in this nonbinding contest. In any case, the popular vote margins in Montana and South Dakota are so small that they’re unlikely to make much difference in the bottom line. My projection for Puerto Rico is a guess, nothing more. Clinton has done well with Latinos in other states, but they’re a diverse group and voters in Puerto Rico may be different. Governor Anibal Acevedo, who has endorsed Obama, has just been indicted; other leaders of the two major Puerto Rico parties, the Popular Democrats (PPD) and New Progressives (PNP), are, according to this post, for Clinton.

My projections on Jay Cost's spreadsheet put Clinton ahead in popular votes, however they're measured. But my projections on my legal pads leave her behind in delegates. Each of these contests allocates most of a state's delegates by congressional districts, except for South Dakota which has only one congressional district; Montana also has only one congressional district, but it allocates most of its delegates in the two congressional districts it had in 1980, before the apportionment following the 1980 Census reduced its number of House seats to one. I give Obama small delegate edges in North Carolina (5) and Oregon (6), and Clinton relatively small edges in Pennsylvania (22), Indiana (12), West Virginia (10), Kentucky (17), Montana (3) and South Dakota (3) and a relatively big edge in Puerto Rico (20). Even so, that reduces Obama's current lead among "pledged" delegates (those selected in primaries and caucuses) from 1,414-1,247 to 1,655-1,565.

These two projections, if they come to pass, seem likely to cause maximum pain among the superdelegates. Clinton will be able to claim a lead in popular vote. But only because of Puerto Rico—and because Puerto Rico this month replaced its caucus with a primary. Obama will be able to claim a lead in pledged delegates. But only because he gamed the caucuses better. His lead in caucus-selected delegates is currently 125, as best I can calculate it; that would mean Clinton would have a 35-delegate lead among delegates chosen in primaries. Both sides will be able to make plausible claims to be the people's choice.

Let me add that my projections don't leave much room for a cascade of superdelegates to Obama. On each day's contests I have Clinton leading Obama both in delegates and popular votes (because North Carolina would be outvoted by Indiana on May 6 and Oregon outvoted by Kentucky on May 20). She would be getting closer to the nomination, not farther away.

Of course my projections could just be plain wrong. Clinton could win Pennsylvania by an unimpressive margin on April 22 and get clocked in Indiana as well as North Carolina on May 6. Then you might see a cascade of superdelegates toward Obama, and the race might effectively be over. But if all those three things don't happen, then I am sure the contest will go on through June 3. And in that case I think my projections are within the realm of possibility.