 | Obama a Favorite, but Not an Overwhelming One |
|
| Posted Under: Government |
Nate Silver's FiveThirtyEight blog recently posted his thoughts about the state of presidential race (see the original post here). In it he examines the history of polls, about six weeks (45 days) away from the election, versus election results. These, he says, support his claim that Obama's odds of winning are now roughly 80%. In particular, he says, "there has not been any tendency, at least at this stage of the race, for the contest to break toward the challenging candidate."
However, a careful look at the figures he provides suggests a different interpretation. Romney is not out of the race; he has a better chance than the conventional wisdom believes. We would place his odds of winning at 40% (well above Silver's 20%). The most likely finish is a 51 - 49 nail-biter.
The data above came from Silver's original post, but we lopped off a few years. We think going back too far in the past – he goes back to 1936 – isn't useful. Relationships between polls and election outcomes shift over time, and so going too far back contaminates the analysis with statistical relationships that are out of date. We think a more relevant starting point is 1960.
One simple way to look at the data since 1960 is that the candidate leading 45 days before the election has always won. We often see references like this in the major media. But that truth masks some major movement in the mood of the electorate between the polling period 45 days before the election and the actual voting.
Notice that when there was an incumbent vying for re-election, challengers typically gained, some substantially: 64, 72, 80, 96, 04. The exceptions were 76, 84, and 92 -- all GOP incumbents. Never has a Democratic incumbent seen a net gain from this point forward, and that includes LBJ and Clinton, who were both top-rate politicians.
Also notice that of the last 13 elections, the Republican gained nine times, Democrats only four, and that was only with non-incumbent Democrats as voters got to know them better (Humphrey, McGovern, Kerry, Obama). We doubt that voters are going to find out anything new and more positive about Obama in the next six weeks. This does not mean we expect something bad for Obama, just that voters already think as highly of him as they ever will.
In particular, look at 1996, with a popular incumbent Democratic president running on peace and prosperity. Six weeks before the election, President Clinton was up by 14.5 points. He won by 8.5, a loss of six points over the last six weeks. Much of that loss in support came in the last 10 days of the campaign, after the debates. If this were to happen again, Romney would win comfortably.
The bottom line is that Obama is the favorite, but not by as much as many seem to believe. If the Giants could beat the Patriots twice in the Super Bowl, Romney can still beat Obama.
|
|