President Obama's odds of winning the Electoral College fell for the second consecutive day on Friday according to the FiveThirtyEight forecast. His chances of winning the Electoral College are now listed at 76.2 percent, down from 78.6 percent in Thursday's forecast.
The most significant reason for the decline was a poor report in one of the economic data series that forecast model tracks, industrial production. A government report on Friday found a sharp decline in industrial production in August, which has otherwise been among the more positive economic data series.
I have been emphasizing the economic component of the forecast model less in recent weeks - in part because there has been a lot of polling news and in part because of the way that the model is designed. The assumption built into the forecast is that, while the polls have a tendency to revert to the economic mean early in the election year, this is less likely to occur as Election Day draws nearer.
This is not to say that the economy will not have any influence on voters' decisions - it almost certainly will have quite a bit. But by the very late stages of the race, whatever effect it does have will presumably be reflected in the polling, meaning that there is less marginal value in considering the economic numbers in addition to the polls.
However, we are not quite at the end stage of the race yet, so substantial shifts in the economic data will continue to have an influence on the forecast for the time being.
The economic component of the model has tracked the polls quite closely over the course of the year and would currently project a narrow win for Mr. Obama by three or four percentage points - about where the polls now do.
There is very strong empirical evidence that voters weigh recent economic trends rather than the direction of the economy over the whole course of a president's term - which is why it is not necessarily as much of a negative for Mr. Obama as the conventional wisdom might hold. And there have been some positives in the numbers lately, like the substantial gains in the stock market in recent months and a rebound in personal income.
But these industrial production numbers were very poor - much more so, really, than last week's jobs report. In contrast to the jobs report, the industrial production numbers showed an outright decline in economic activity - by 1.2 percent, the largest such decline since March 2009. (The government attributed some of the decline to the effects of Hurricane Isaac.)
Apart from the economic data, Mr. Obama's polling numbers were fairly average on Friday. Pol ls put him ahead in Colorado and Virginia, but by just a percentage point in each state - numbers more like what we were seeing before the party conventions than during Mr. Obama's stronger phase of polling earlier this week.
There is little question that Mr. Obama received a bounce in the polls from his convention and that his overall position is stronger than it was a few weeks ago. But the higher expectations for Mr. Obama mean that he will have to do more to maintain his standing in the FiveThirtyEight forecast.
The question we are now trying to determine is whether Mr. Obama is ahead in the race by four or five points - as polling in the immediate aftermath of his convention seemed to suggest - or by more like two percentage points, as he was before the conventions. The forecast model is splitting the difference a bit, projecting a 3.5-point win for Mr. Obama on Election Day.
No comments:
Post a Comment