Wednesday, October 20, 2004

Guest Post: Ian’s Electoral Analysis

This election is coming down to only a few key states., a leading online betting site which allows wagers on the Presidential election on a state-by-state basis, only nine states are “close,” defined as the market being over 1/3 sure that Bush will win the state, but less than 2/3 sure than Bush will win. Those nine states (with their electoral votes in parentheses) are:

Florida (27)
Iowa (7)
Minnesota (10)
Nevada (5)
New Hampshire (4)
New Mexico (5)
Ohio (20)
Pennsylvania (21)
Wisconsin (10)

All other states are either solid Kerry states, ones where the market feels there is less than a 1/3 chance that Bush will win; or Bush states, ones where the market gives Bush a 2/3 or greater chance of winning. People will quibble about the resulting list of “swing” states – many feel that Colorado, West Virginia, or Arkansas could go Kerry, and Karl Rove probably secretly believes that Bush can take New Jersey, Oregon or even Washington.

But I would submit a couple of arguments here. One, the markets on this kind of thing have proven remarkably efficient, and the markets don’t give these “quasi-swing” states much chance to unexpectedly swing. Second, and more importantly, if any of the “quasi-swing” states actually swing to the non-expected candidate, then enough of the “true” swing states will switch that way as well, making the swing of the “quasi-swing” state irrelevant. I simply don’t see how Colorado could swing the election to Kerry – if Kerry wins Colorado, he will have won enough of the above 9 states to win the election anyway; similarly, if Bush takes New Jersey, the election will have been long over before New Jersey’s electoral votes are officially assigned to Bush. I therefore firmly believe that analyzing only these nine states allows one to do a very accurate set of scenarios around the election.

Under this analysis, Bush starts with 26 states and 222 electoral votes. Kerry starts with 16 states and 207 electoral votes. The above 9 states have 109 electoral votes. 270 electoral votes are required to win the election. So my question is: what are plausible scenarios for Kerry to get the 63 required votes? (In looking at the question this way I betray my loyalties – I admit I am a Kerry supporter. But I hope to make this analysis as non-partisan as possible.)

I start with one fundamental observation: there are 3 big states on this list – Florida, Pennsylvania and Ohio. I believe Kerry will have to win 2 of these 3 to have a realistic chance. Mathematically he could win just Florida and lose Ohio and Pennsylvania – for example, he could take Florida, Minnesota, Wisconsin, Iowa, and 2 of 3 of Nevada, New Hampshire and New Mexico, and this would give him 63 or 64 votes. But I find it highly unlikely that Kerry will be able to stitch together so many small states. And mathematically, there’s no way for him to win by JUST winning Ohio or Pennsylvania of the big states – the most he could get would be 61 (if he won only Ohio) or 62 (if he won only Pennsylvania), even if he won all the small states on the list. So realistically, Kerry MUST win 2 of these 3 big swing states.

Of course, this “2 of 3” analysis is not new – many pundits say whoever wins 2 of 3 out of Ohio, Florida and Pennsylvania wins the election. I strongly believe this is NOT the case, at least for Kerry. Should Bush win 2 of these 3, he likely wins the election. But should Kerry win 2 of 3, he still faces many realistic scenarios where he loses the election.

If Kerry wins 2 of 3, to get his magic 63 electoral votes he’ll still need somewhere between 15 and 22 more votes from the small states. (He’ll need 15 if he wins both Florida and Ohio, 16 if he wins Florida and Pennsylvania, and 22 if he wins Ohio and Pennsylvania). The 6 small states total 41 votes, so we are looking at Kerry needing about a 50/50 split. This may seem easy, but is harder than it looks.

According to recent polls, Kerry appears to be slightly ahead in Minnesota and perhaps New Mexico, and Bush appears ahead in Iowa, Nevada, New Hampshire and Wisconsin. So Kerry has to pull at least one of these states his way. Iowa or Wisconsin by themselves are likely enough, but if he pulls Nevada or New Hampshire, he’ll likely need to pull both of them (or take Iowa or Wisconsin, but as I said that would likely be enough on its own anyway).

To recap, my recipe for Kerry is to win 2 of 3 of Ohio, Florida and Pennsylvania, to hold onto Minnesota and New Mexico at all costs, and to “steal” Iowa, Wisconsin or the New Hampshire/Nevada combination.

To be honest, and to borrow one of President Bush’s favorite phrases, I see this as “very hard work.” I believe it’s fairly likely that Kerry will be able to take Pennsylvania and Ohio – the Democratic machine in both states has done an incredible job in registering voters and motivating the base – but I think Kerry faces a huge uphill battle in Florida. And the vote in Ohio is especially close. Furthermore, Kerry would still need to steal a small state or two – I see Iowa as the most likely, and currently believe that both Wisconsin and Nevada will go to Bush.

Under this scenario, Kerry wins the Presidency with 270 electoral votes to Bush’s 268, making it even closer than the 2000 electoral vote. There are a multitude of other combinations, of course, but the overall message is clear – Kerry needs to be very targeted, and really focus on winning 2 of the 3 big states. Importantly, he can’t neglect the small states, but I also believe he can’t hedge his bets and try to win all of them. They are too geographically dispersed and care about different issues for him to campaign effectively in all of them. I would really focus on the Midwest, especially in Iowa and Wisconsin. I think Kerry must take one of these states.

Which makes me think of an issue I didn’t mean to write about – don’t you think Kerry wishes he had taken Dick Gephardt as his VP nominee? I certainly do, and this analysis shows exactly why. Wins in Ohio and either Iowa or Wisconsin will likely give Kerry the election. Edwards adds little in either state; Gephardt would have.

And one further prediction: in 2000 Al Gore won the popular vote but lost the electoral vote. I see this as highly plausible again in 2004, but in reverse. Kerry will need to win the majority of “close” states to take this election, and it’s very plausible that Kerry’s small margins in 5-6 key states, combined with Bush’s roll-up of so many Republican-leaning states, would mean that Bush could actually get more individual votes than Kerry but lose the election.

* Note: In 2000, the author predicted the electoral college vote 2 weeks before the election, and had the election going 271-266 for Al Gore. The end electoral vote was 271-266, but for George Bush. The author correctly predicted 49 states and the District of Columbia, and only missed the state of West Virginia.


Return to Top