Thursday, April 3, 2008

The US Elections: The Popular Vote Myth

Things seem relatively quiet today on the Democratic front. As a result, a little news snippet that otherwise might have only have led to a comment or two in passing has been given some attention, if still not enough.

It's the assertion by two stalwart Clinton supporters, John Murtha and Jon Corzine, that Clinton needs to win the popular vote to have any real chance of winning the nomination.

Now the relevance of this seems obvious enough. Clinton can't realistically hope to overtake Obama in terms of the pledged delegate count. To stay alive and have a shot at convincing the superdelegates to back her, she needs to be able to argue that the voters, overall, prefer her to Obama.

And if that argument could be made truthfully, it might be persuavive (although not in itself conclusive). After all, quite a few Democrats are still smarting from the fact that Bush bested Gore in 2000, in spite of the fact that Gore led the popular vote by about 500,000. What could be worse than that the Democrats, after having to accept that grievous injustice, were then to turn around and inflict it upon themselves?

You can see how this could be a seen as a strong argument. It is, however, fundamentally flawed for a number of reasons.

The first flaw is as blatantly obvious as it is convincing. The Democrats set up their own rules when it comes to choosing their nominee. What they've chosen is a delegate system, not a system based on the popular vote. There's no reason on earth why they couldn't have switched to a popular vote system if they'd wanted to. They didn't. They chose for a system whereby the delegate count was the determining factor.

The second flaw, though, is the one that I find more interesting at the moment. It's this: how on earth does one actually determine the popular vote?

In the presidential election, that's easy enough. The count can certainly be hampered by technical issues - which is what happened in Florida in 2000 - but the idea is that everyone gets to vote in basically the same way, all the votes get counted and that it.

Not so in the primary season. For starters, when it comes to the popular vote there's a huge difference between primaries and caucuses. Turnout in caucuses is obviously very much lower than in primaries, so how does one compare the two? It would, perhaps, be possible to set up intricate number crunching models that would allow for some sort of extrapolation of a caucus turnout to a virtual primary. Alternatively, one could just accept the actual numbers. In both cases, however, it's essentially comparing apples and oranges.

Take, for example, State X. It's in the mid-West, and it looks a bit like Colorado. Or perhaps Oklahoma. It's got 5 million inhabitants; it's predominantly a red state, but it might just turn blue in November.

Now say it's a caucus state (like Colorado). Comes the caucus date, 100,000 people show up to vote for their Democratic candidate of choice. 60,000 vote for Obama, 40,000 for Clinton. In the total popular vote tally, that's a net gain of 20,000 for Obama.

But now let's assume X is a primary state (like Oklahoma). Suddenly, the turnout isn't 100,000. It's 400,000. Assuming the same divisions apply, Obama wins the popular vote by 80,000.

That makes for a significant difference in the popular vote tally. But that difference has nothing to do with the candidates or how well they're doing.

Now, the above example assumes that voters vote the same way whether it's a caucus or primary (the only variable being the number of voters). We all know that's not true. Caucus states favour the grass-roots candidate; primary states, by comparison, favour the more establishment candidate (whilst Texas, in an maudlin fit of madness, favours both).

In state X, if Obama would have won by a margin of 20,000 in a caucus, he may well have lost in a primary by, oh, 50,000 or so. And the vote tally would be totally different again.

So it's apples and oranges; oranges and apples. And that's just for starters.

Another thing to consider is that some states don't even release voter numbers. This is true for Iowa, Nevada, Washington and Maine. We simply do not no how many people voted for their candidate of choice in these states. If you think this is amazing, consider that these states are - quite rightly - working on the basis of the assumption that the popular vote isn't the deciding factor. So how to determine the popular vote in those states? You can't; the only option is to guesstimate.

A third issue centers on the debacle of Florida and Michigan. In the popular vote debate, does one count the voters in these states or not? And if so, how? In particular, it would certainly seem weird to include Michigan, where Clinton was the only major Democratic candidate to leave her name on the ballot after the DNC sanctions were levied. By what means could anyone assume the outcome in that state was a fair and balanced representation of the voters' views? Besides, the major non-Clinton vote went to a candidate called "Uncommitted". What is one to do with those votes in a popular vote tally? Any choice made is inevitably a wrong choice.

So, what does all this mean? Well, in the end, it's very simple. Given the Democratic nomination process in general and the specifics of this campaign, there simply is no way to accurately determine a relevant outcome of the popular vote.

As a result, the entire concept of Clinton potentially winning such a vote and thereby having a legitimate shot at the nomination is essentially a myth. With apologies to all those doing their very utmost to prove the opposite, it just can't be done. And that's as it should be: it's not what the Democrats wanted in the first place.

No comments: