It is becoming increasingly likely that there will be a disputed winner of the popular vote, because there are competing definitions of the popular vote. Some definitions include Florida, some include Michigan, some include estimated turnout in caucus states that do not record the popular vote, and some allocated varying amounts of Michigan uncommitteds to Obama. While Obama leads in virtually all of these metrics, there are a couple of counts where Clinton leads. The pro-Clinton totals are the ones that include Michigan and Florida, but give Obama zero votes in Michigan and do not estimate turnout in caucus states without a popular support count.
Right now, there is no solid ground for arguing that Clinton is ahead in the popular vote, since she only leads by 10,000 when Florida, estimated caucus turnout, and Michigan are included. However, since Obama supporters obviously made up ore than 10,000 of the 237,000 Michigan uncommitted, clearly right now more people who have participated in a nomination event have supported Obama than Clinton.
In keeping with the principle of one person, one vote, the only good metrics to use are the ones with the broadest popular participation. As such, when measuring the popular vote, it is best to throw the widest possible net. This means to include Florida. It also means to include the estimates from caucus states that did not release popular totals, which stand at Obama 334,084--223,862 Clinton. Finally, it means to include Michigan, but also to allocate Obama 72.91%, or 173,368 of the uncommitted vote. This number is derived by dividing Obama's exit poll support in Michigan by the combined exit poll support of Obama, Edwards and Richardson, and then multiplying that number with the total uncommitted vote. Adding in the totals from everywhere else, this results in a grand total of Obama 15,481,172--15,319,525 Clinton, or an Obama margin of 161,647 votes.
Problematically, these totals also results in a popular vote total with a margin of error, given the caucus and Michigan uncommitted allocation. However, if we are going to count everyone who participated in the process, we can't pretend that Iowa, Nevada, Maine and Washington had zero participants, or that no one in Michigan supports Obama. If we really intend to determine the popular vote winner, we have to count everyone who participated, and do so to the best of our ability. Otherwise, determining a popular vote winner is meaningless. As such, there is really no way of avoiding a margin of error. The difficulty arises from determining just how large that margin is, and what sort of popular victory would be outside that margin of error. I don't have an answer to that question right now.
A second problem arises now that it is clear neither Obama nor Clinton will win a majority of support from participants in nomination events. Without a majority vote winner, technically speaking, according to the principles of "one person, one vote," the will of the Democratic nomination campaign electorate is for no nominee to be determined on the first ballot at the convention. Should a narrow plurality winner be afforded the same respect as a majority winner? I don't think there is a clear answer to that question, either.
Overall, with the complications arising from competing popular vote totals, vote estimation, and that we are facing a plurality winner instead of a majority one, I have to conclude that the popular vote is actually meaning less and less as we go on. The actual nature of the process, as a series of several thousand small elections for delegate slots, is asserting itself. In order to avoid a post-voting convention fight, it looks as though Obama will need to make sure he is above the CF line, not that he is ahead in the popular vote count. |