Kerry's chances

I have asked Roger to blog about the odds of a Kerry Presidency and he has graciously submitted the following post. If you reside in one of the states he identifies as crucial maybe you ought to volunteer your services with one of these organizations:

Move On

America Coming Together (ACT)

If you ain't got the time, then I'm sure your financial contributions would be appreciated here:

Contribute to Democrats

And now, on to Roger's analysis:

Moe,

There have been significant changes in the battlefield states in the past month. Two states, Iowa and New Mexico, have gradually changed their color from blue to red. The change seems to be statistically significant, beyond the margin of error. Wisconsin, which was never a "strong Kerry" state has become a see-saw between Bush and Kerry. The good news is small news: New Hampshire and Maine seem to be safely in the Kerry block.

A month ago there were many plausible combinations of states that would have led to a Kerry victory. That number now seems reduced only to a few. In fact there are only three states that are variables in the Kerry election calculus. These are Ohio, Florida, and Wisconsin.

If Kerry wins the Gore states, excluding Iowa, Wisconsin and New Mexico, and if he wins New Hampshire, then he will have 242 electoral votes. He needs another 28 votes to win the election. By itself, Florida (27 electoral votes) will not win the election since that would give him 269, putting him at the mercy of a Republican congress. So he must win Florida and some other state. If he wins Ohio (20 electoral votes) then he needs to win an additional state with at least 8 electoral votes.

A composite of recent polls gives Bush a 1% lead in Florida. Statistically, due to the large number of polls, this lead is significant and is beyond the margin of error. Based on this composite we would expect Bush to carry Florida by at least 50,000 votes. But the situation in Florida is volatile and frenzied. There is speculation that the turnout may reach 75%. Such a large turnout makes polling difficult. Pollsters use historical voting patterns as a template for their current polls. Their results can be skewed if their template does not fit the profile of the current voting population. With the infusion of many new voters and with the likelihood that many 2000 non-voters may become 2004 voters, I think there is an element of uncertainty with all Florida polls. For this reason I feel that Kerry has a reasonable chance in Florida, at least 1/3.

Ohio is a state that has good reasons to vote for Kerry. Cleveland has recently replaced Detroit as the poorest city in America. Over the past month there has been a gradual shift to Kerry. Most polling groups now expect Kerry to carry Ohio. I give him 3/4.

Wisconsin seemed safe for Kerry in early September. Since then there have been a number of polls that have shown Bush to be leading. Strategic Vision, a Republican outfit, has consistently shown Bush to be ahead. Zogby, except for one poll, has Kerry in the lead. Zogby probably has it right. I give Kerry a 2/3 chance of carrying Wisconsin.

There is a set of small Bush-leaning states that come into play if Kerry wins in Florida. The list includes Nevada, New Mexico, Iowa, Arkansas, and West Virginia. Colorado and Missouri might also be included. It is likely that Bush will win all of these states. I give Kerry a 1/5 chance of winning at least one of these states. The likelihood that Kerry will win two or more is virtually zero.

The probability that Kerry wins the election comes in two pieces. [1.] Florida and an additional state. This additional state could be Ohio, Wisconsin, or a state in the list of small red states. This has the probability of 0.31. [2.] Ohio and Wisconsin, or, Ohio and two of the small red states. This has the probability 0.50. Since these two pieces are mutually exclusive, we can add the probabilities. The likelihood of Kerry winning the election is 0.81.

This is a remarkable result. I am honestly quite surprised since I expected Bush to be the probable winner. This result is even more remarkable when one considers that Bush has a 1.5-2.0 point lead in the national polls.

A few last words. This calculation is optimistic since I have not considered the probability that a state that is predicted to vote for Kerry might vote for Bush. This might occur in Minnesota, Michigan, or New Hampshire. This calculation is pessimistic if the undecided voters break 2-1 toward Kerry. Many pollsters, including Zogby, have predicted this. Should the undecided voters break toward Kerry, Bush would have to take a deep breath and start looking for new job.

Roger

I would like to add that since the Red Sox have won the World Series hope no longer seems like a sucker punch.