May 20, 2008
Democratic Primary Matters
The closeness of the primary contest between Barack Obama and Hillary Clinton has made the mechanics of the primary process itself the central issue of the campaign. Politics is always about controlling narratives, and as the primary season comes to a close, the race has centered on two competing definitions of how the votes are to be counted and interpreted. The contest is increasingly focused on the question of how you do the math.
It takes a simple majority of the convention delegates to nominate a candidate for president. If delegates from Florida and Michigan are included it takes 2210 delegates and if they are excluded it takes 2026. Going into the primaries in Kentucky and Oregon, the total of pledged delegates (those that have been allocated by elections in state primaries or caucuses) stand at 1612 for Barack Obama to 1443 for Hillary Clinton. With about 200 delegates left to be determined from the remaining primaries in Oregon, Kentucky, South Dakota, Montana and Puerto Rico, neither candidate will be within 250 votes of reaching either 2026 (excluding Florida and Michigan) or 2210 (including Florida and Michigan). This means that the super delegates, those appointed by the party in each state, will decide the nominee.
Since early February, the mantra of the Obama campaign has been "Do the math". It has been their consistent claim since Super Tuesday that the only way Hillary could win the nomination was by getting the super delegates to overturn the will of the voters. Their point has been that because of proportional allocation of delegates, it would be impossible for Hillary to overcome Obama's lead in elected delegates in the remaining primaries and therefore Obama deserved to be the nominee.
Hillary has won most of the major primary contests since Super Tuesday, including the key states of Ohio and Pennsylvania, by substantial margins, but her margin of victory has not been great enough to substantially reduce Obama's delegate lead. Her victories suggest that Hillary would be a stronger candidate in the general election, a possibility that the super delegates cannot really ignore.
The Obama campaign's claim that the race was over was always an election strategy aimed at getting the support of the super delegates. The talk about doing the math was always more about "spinning the math". It was their side's way of framing the debate on how the Democratic Party was to look at several key issues that would decide the election: what to do with the votes from Michigan and Florida; how to count the overall popular vote; how to relate the primary system which reflects the popular vote to the general election that puts a premium on winning key states; which candidate had the best chance to win in the general election.
The national media has largely bought into "Obama math", his campaign's narrative on how the Democratic Party is to interpret the results of the Democratic primaries. Hillary hater-in-chief Dick Morris has been a frequent visitor on Fox News, promoting the view that Hillary has lost and it would be unimaginable for the Democratic Party turn its back on Obama's coalition of African Americans and enthusiastic young voters by having Hillary-supporting super delegates swing the election in her favor. The clear implication of this point was that there would be something sleazy and possibly racist about doing so. For the most part, CNN along with the major networks and their cable affiliates, have echoed this position.
Because the primaries have been reasonably close and delegates are allocated proportionally, Obama has maintained a small lead. While the Obama campaign may have been right about the outcome of the primaries, it is not hard to see how the Clinton campaign and its supporters have been short changed by the mathematics of the process. Even a cursory look at the voting totals from primary contests reveals that the contest has been essentially a dead heat in the states where delegates are apportioned from a primary and a complete rout in those states with delegates allocated from a caucus procedure.
Primaries and Caucuses
Obama's delegate lead in pledged delegates currently stands at 159 and is only twelve more than the 149 delegate lead from states that held caucuses rather than primaries. Hillary won only one caucus state, Nevada, and that by a single delegate. She lost the rest by substantial margins.
It is highly unlikely that in a race with such a close popular vote that delegate majorities of landslide proportions in the caucus states would have occurred had those states held primaries instead. Caucuses bring out the most motivated voters and their political views are often at odds with the far more numerous rank and file members of their parties. The results from the two states which held both primaries and caucuses, Texas and Washington, strongly support this view.
There were 2,800,000 ballots cast in the Texas primary and Hillary won by over 100,000 votes, a margin of victory of almost 4 percentage points. As a result she was allocated 65 delegates to Obama's 61. On the same day as the primary, 42,538 individuals cast a second vote in the Texas caucuses, which Obama won by a margin of 23,918-18,620 and as a result was apportioned 38 delegates to Hillary's 29. The combined delegate count from Texas was 99-94, a net gain of 5 delegates for Obama,
Do the math! Fully one third of the delegates were awarded to the 1.4% of the voters who participated in the caucuses, while two thirds were apportioned to the 2,800,000 voters in the regular primary. This distortion of the popular will meant that Obama, who lost the primary by 100,000 votes, came away with 5 more delegates.
The state of Washington is equally revealing. There was a non-binding primary held 10 days after the caucus. No delegates were apportioned based on this contest that was won by Obama by a margin of 51-46. But in the caucuses, in which a much smaller number participated, Obama's margin of victory was much greater and he was awarded 53 delegates to Hillary's 25. Had the 97 delegates from the state of Washington been awarded proportionally according to the non-binding primary, Obama's delegate margin would be 51-46.
In Alaska, with a grand total of 405 reported caucus voters, Obama was allocated 9 delegates to Hillary's 4. This compares with West Virginia where a 67-26 percentage point win and a margin of over 147,000 votes produced a delegate allocation of 20-8, a net gain of 12 delegates for Hillary.
The Caucus States of America
Most of the caucus states have routinely voted overwhelmingly Republican in past presidential elections and the small percentage of Democrats who actually attend the caucuses tend to hold more liberal views than the average Democratic voter.
The election returns in the caucus states suggest the impact of "blue communities" in red states. While the states may be generally conservative, the college towns and communities of transplanted urbanites form their own subculture and their views tend towards the left wing of the Democratic Party. While their percentage in the general population is small, they make up a healthy share of the caucus participants.
There is no reason to believe that the views of the caucus goers in states like Idaho and Wyoming reflect the opinions of rank and file Democrats. But if you go on the election results website of the New York Times, you will see the results of caucus states reported in percentage terms rather than votes, leaving the impression that the voters of such states are overwhelmingly pro-Obama. While this is plausible, the small numbers of caucus voters do not provide anything resembling a random sample of typical Democratic Party voters. But this misleading representation of such votes is important to the Obama campaign because it supports the contention that the number of elected delegates accurately reflects the opinion of the Democratic voters in such states, and so the overall delegate total reflects the overall opinion of the Democratic Party.
Florida and Michigan
Finally, the decision to allocate no delegates to Florida and Michigan voters did significant damage to Hillary's campaign. This was particularly true of Florida, where the demographics were very favorable to her candidacy. Its large population offered a mother lode of potential delegates from the vote that occurred. If allocated according to the popular vote, it would have yielded a 43 delegate advantage for Hillary.
The initial decision to prevent all states but Florida and Nevada from holding primaries before February 5 was designed to give less well known candidates such as Obama the opportunity to compete against better financed and more well known candidates such as Clinton and Edwards. The states of South Carolina and Nevada were added to the pre-Super Tuesday mix because, unlike New Hampshire and Iowa that are primarily white, South Carolina and Nevada have substantial minority populations.
Indeed this helped Obama, as he didn't have to expend time and resources in the large states of Florida and Michigan, and could concentrate on South Carolina and a group of small Republican-leaning western states with caucuses rather than primaries. What's more, he was able to avoid facing Jewish voters in Florida and Arab-American voters in Michigan, who would have had the opportunity to force Obama to parse out what exactly are his "nuanced" views of the Arab/Israeli conflict.
Instead of attending campaign rallies in college towns, recruiting students with promises of "change" and shouts of "yes we can", he might have been forced to let the public know what he actually thinks the US policy should be about the future of Jerusalem, West Bank check points, the separation fence, Israeli military responses to rocket attacks from Gaza, what concessions he expects from the two sides in the dispute and what measures he would be prepared to take to pressure the two sides into making such concessions. Voters on both sides of this dispute deserve to have some idea what an Obama presidency would bring to this dispute. Right now all we really know is that Obama is enthusiastically supported by Martin Peretz and Louis Farrakhan.
It is telling that as a consequence of the Democratic Party decision to exclude the delegates from Michigan and Florida, the voter turnouts were low in both states. In Florida, the turnout was 48% of the number of votes for Kerry in the 2004 election. In other states the primary turnouts were mostly at 75% and above, and in Texas it was at 102%. This suggests that had Florida voters known that their votes would count in the national primary, Hillary's margin of victory might have been closer to 600,000 rather than the actual 288,617.
The Democratic Party mishandled the Florida and Michigan primaries. If they needed to punish the recalcitrant states, they might have been better advised to follow the lead of the Republican Party, which reduced the size of the delegations by 50%. This still allowed for a meaningful primary that compelled the leading candidates to run a meaningful campaign in two of the most important states in the general election.
When the Democratic Party made the decision to disallow the delegates, they anticipated one of the candidates securing the nomination long before the end of the primaries. But as neither candidate has been able to close the deal, the Democrats have a dilemma without a satisfactory solution. There is simply no way to resolve this fairly.
The Democratic primary system awards delegates to the candidates according to the percentage of votes they receive in each state. When put together in aggregate, it amounts to a system in which the allocated delegate totals closely reflect the overall popular vote. This process contrasts with the Republicans' mostly winner take all primaries and more importantly, it differs from the electoral college system which determines the outcome of the general election.
The impact of proportional allocation of delegates to Democratic candidates limits the importance of winning states and increases the importance of demographic factors such as race, age, gender and income. There are no exact figures on how the various demographic groups distributed their votes. Exit polls are at best good approximations. But at least in theory, there is no reason why state-by-state analysis of voting is any more enlightening than demographic analysis. The impact of demographic factors on the total number of delegates produced in the primary states does not depend on their distribution within particular states. All that really matters is their percentages within the total voting population.
In the general election, however, this is not true. A candidate's vote total in the Electoral College, as measured by percentage, is not necessarily close to the percentages in the popular vote. While the Democratic Party primary system is probably more representative of their voters than the Electoral College system, it is sufficiently different that it is not a clear predictor of general election performance. For example, African American voters contributed greatly to Obama's margin of victory in many southern states, but because of the political conservatism of the same states, he is unlikely to carry any of them in November. Because of the policy of proportional allocation, the contribution of these voters to the delegate count in the Democratic primary would be unchanged if they were distributed to other states. However, disbursing them judiciously to say Ohio, Pennsylvania and Florida could easily change the outcome of the general election.
Compounding the Democrats' problem is that with almost 800 super delegates appointed by the party and bound only to vote their conscience, in a close election such as the current one, neither candidate is even close to obtaining an absolute majority from the elected delegates. This means that the super delegates are free to vote based on which candidate they believe is most likely to win in the general election. And in spite of what Howard Dean may hope, the super delegates don't have to make a final decision until they get to the convention in August.
Spinning the math
Both sides are looking at the math and coming to different conclusions.
1. The race isn't over until one candidate has secured an absolute majority of delegates. Delegates from Florida and Michigan should be seated at the convention and the number of delegates that constitute an absolute majority should be calculated with these delegates included.
2. Hillary defeated Obama in Massachusetts, New York, New Jersey, Pennsylvania, Florida, Texas, Ohio, and California, eight states totaling 215 electoral votes. Add in Michigan where she defeated uncommitted -- a proxy for Obama votes -- by a margin of 55-40, and you have a total of 232 potential electoral votes from states where she is considerably more popular than Obama. It likely that California, New Jersey and New York will vote Democratic in November, and it is equally likely that Texas will go Republican. But the remaining four, Ohio, Pennsylvania, Florida and Michigan, could go either way and Clinton seems like a much better choice to prevail in those pivotal contests.
3. Obama's margin of victory over Hillary is concentrated in states that almost never vote Democratic in presidential elections. These states lean politically conservative and Hillary is considerably closer to the political center than Obama, particularly on the crucial matters of national security and foreign affairs. Though he beat her handily in these states, neither candidate has much chance of winning in any of these and her chances of attracting independents and stray Republicans is better than his.
5. If the contest between Hillary and Obama were measured in electoral votes, the states she has won so far total more than 300 electoral votes and she would win easily.
1. From the beginning, the Obama campaign has argued that the will of the voters is measured by the number of delegates that are proportionally allocated from the primary elections and state caucuses.
2. Obama should be the nominee if he ends up with more of these delegates than Hillary.
3. Further validation of his rightful place as the nominee is that he has won more states in outright competition with Hillary.
4. Since the DNC decided to disallow the delegates from Florida and Michigan, it would a rewriting of the rules to seat those that were chosen in the January primaries, and as he hadn't campaigned there, it would be unfair to his candidacy.
5. If a compromise is not reached to seat the Michigan and Florida delegations, the number of delegates needed for nomination is a majority of the delegates from the other 48 states and various territories and the District of Columbia.
6. Once the number of Obama's elected delegates plus super delegates reaches the majority (2026 or 2210 depending how you factor in Michigan and Florida) the primary is over and he has won.
There is no way of resolving this dispute that will be satisfactory to both sides. There are genuine ambiguities here that must be resolved and the answer is not in the rules but in the political judgment of the officials of the Democratic Party. Any resolution on whether to seat delegates from Florida or Michigan, and how the super delegates decide to vote, is a political decision with a winner and a loser. Following a long and hard fought struggle, there will certainly be considerable hard feelings on the losing side. The Obama campaign, from the beginning, has attempted to frame the narrative around its own version of the math. For the most part the mainstream media has echoed their version.
But there are serious problems with their narrative. One sixth of the delegates are super delegates who are bound to no candidate until they vote at the convention. In other words, unless one of the candidates has an absolute majority of pledged delegates, the primary season is not really over until the convention.
The super delegates are chosen to play an independent role in the process. They can interpret the primary results however they wish and vote accordingly. There are some issues that they are going to have to confront.
The popular vote
There is no one best way to tabulate the popular vote. At the Horserace blog on Real Clear Politics, Jay Cost has 15 different ways of tabulating the popular vote. The issue is which votes to include and how to count them.
The count hinges on several issues: whether to include the Florida or Michigan votes, how to count the Michigan vote if they do, and how to count the votes from the four states with caucuses where no overall vote was tabulated. Then there is the question of how to count the votes from Washington. Should they use the numbers from the caucuses or the numbers from the non-binding primary?
Cost looked at previous results and using demographic data from earlier primaries and making estimates of expected turnouts, he came up with likely election totals in the remaining six contests including West Virginia. Using this hypothesized date, he offered 15 ways to count the popular votes, 9 of which favored Hillary and 6 of which favored Obama. In the one primary that has now been held, West Virginia, Cost's numbers underestimated Hillary's margin of victory by 25,000 votes.
The Democrats' dilemma
There is no compelling reason for the super delegates to ignore the popular vote in making their decision. And there is also no compelling reason to pick one way of counting the popular vote over another. The reality of the situation is that they are under no compulsion to make a decision before the convention, and even if they declare for a candidate before August, they are not bound by such a decision until the first ballot at the convention.
Though ahead in the delegate count, Obama's total of elected delegates is nowhere close to an overall majority. Looking at the pledged delegate count alone, Obama is the winner. But viewed from the perspective of optimizing the number of potential electoral votes, Hillary appears to be stronger. And from the point of view of the overall popular vote, there looks like there is no clear cut winner.
There are serious problems with deciding against either candidate. With Obama there is the long term potential damage of taking the nomination away from the first African-American candidate to reach the end of the primary season with the greatest number of elected delegates. This would have unfortunate consequences for both the Democratic Party and for the country.
There has been a strong "I never thought I would live to see the day" vote from African Americans, and even many who strongly oppose his candidacy on political grounds, share the view that on racial grounds alone, an Obama victory would be good for the country. Obama's supporters have been stressing this on talk shows and news programs since they started declaring the inevitability of his nomination back in February. So there are compelling reasons not to deny Obama the nomination, particularly if he has a slight lead in elected delegates.
On the other hand, the Obama lead in elected delegates is entirely due to the allocation from the caucus states and the decision to exclude the delegates from Florida and Michigan. The great irony of this campaign is that Obama, if he wins, will have fashioned his victory primarily from a complete dominance of the caucus process in states that are predominantly white.
Underlying the conflict between the candidates is the ongoing rift between the Party's most liberal members, who are heavily in favor of Obama, and its centrists, who are supporting Hillary much as they supported her husband in 1992. Though the candidates' official positions on issues are similar, their supporters represent two very different poles within the party.
Obama is the candidate of those who think the war in Iraq is immoral, Israel should be pressured to make more concessions for peace, guns should be outlawed, abortion should be available on demand, free and without restrictions, US foreign policy was responsible for 9/11, the United States should not have a military until it abandons the policy of "don't ask, don't tell", religion can be ridiculed unless it is done in a mosque or Trinity United, the gravest threat to individual liberty is the government eavesdropping on phone calls from Al Qaeda, captured terrorists should be Mirandized, lawyered up and released to blow themselves up in Baghdad or Tel Aviv, and you can fill in the blanks from reading a visit to any number of left wing blog sites.
In short, Obama is the candidate of the academic left. Though I don't believe that Obama is an elitist or even shares all of these views, he is the preferred candidate of the faculty lounges and the activist students, an insular group whose attitude toward most Americans is somewhere between condescension and contempt.
What is at stake in the Democratic primaries is political primacy in the Democratic Party. Hillary Clinton may be a big government liberal but she is a pragmatic one. She is in favor of things that conservatives don't like, but she will work with them and she is not going to do anything drastic. What she is not, is the Michael Moore/DailyKos wing of the party. And right now she is putting up a determined fight to keep them from taking control.
This fight is important. Even if she loses, as it now looks like she will, the strength of her fight will pressure an Obama presidential campaign and possible presidency toward the political center. For that I am grateful.
Jonathan Cohen is Professor of Mathematics, DePaul University