Today, on his New York Times blog, Nate Silver published the first of what will be a regular series of updates on the forecast for the November election. Silver makes no secret of his liberal leanings, but his statistical work (forged in his beginnings as an outstanding baseball analyst) is straightforward and generally reliable. In his first 2012 presidential forecast, he establishes President Obama as a “very slight favorite” as of the moment. That means he gives the president roughly a 60 percent chance to win re-election but, as he notes, his current estimate of 290 electoral votes for Obama is flexible with “outcomes ranging everywhere from about 160 to 390 electoral votes are plausible, given the long lead time until the election and the amount of news that could occur between now and then.” That sounds about right as does the criteria by which he seeks to judge trends via an economic index that will decline in importance, as we get closer to election day and an average of polls taken.
The problem, as Silver notes, is that national polls have trended toward Mitt Romney while state polls tend to favor the president, which has given the incumbent something of an intrinsic edge in Electoral College projections. That edge may be offset if current negative economic trends continue and Obama’s numbers decline everywhere. But it also points out that the wisest course for both candidates, but especially Romney, is to concentrate his resources in the key swing states of Virginia, Ohio and Colorado. Those three are the states where the probability of an Obama victory is the most narrow and may therefore decide the outcome. But while Silver’s detailed breakdown of the numbers is well worth readers’ efforts, the problem with this and any forecast is that trends have a logic all their own. If the public perceives that the economy is failing, then the president’s slight edge can be thrown out the window.
Distinguishing between short-term advantages (the “nowcast” as Silver puts it) and long-range trends, including how states have voted in the past, is vital to employing a coherent model. That means that although Democrats can certainly take some solace in their current small advantage, they need to be very worried about whether the arc of this election cycle is running against them. Romney has been gradually gaining strength, and if the economy doesn’t recover, he may be in a far stronger position at the end of the summer than he is today.
As he should, Silver sticks to the numbers, but there is more in play here than just polls and economic statistics. The president has the advantage of incumbency and the mainstream media’s liberal bias. Romney has more of the public’s confidence on the economy and that may play out as the factor that may be the difference in those tipping point states like Ohio and Virginia. So while it is certainly better to be the “very slight favorite” on June 7 than to be behind, as the last week’s terrible economic numbers and the Democrats’ Wisconsin disaster proved, the tide may be starting to turn against the president.
Nate Silver is conveniently forgetting that Obama and his top people are second rate. Romney's team is vastly superior. This will become more obvious within the next month.
Silver may indeed be rigorous in his analysis, but he also might do too much work. n nFour years ago he questioned some of Real Clear Politics' choices in which polls to include and which to exclude. I'm reasonably certain that RCP's averaging prediction was pretty close to Silver's in 2008. I know that in 2004 RCP's averaging predicted each state – except Wisconsin – correctly. n nIn other words you probably could get the same or similar results as Silver by election day even without all the work.
The Detroit Free Press has a poll today with Romney ahead in Michigan Obama 46%-45%. This was a poll of likely voters (people say this kind of poll is more accurate than the polls of registered voters). n nTo paraphrase Herman J. Mankiewicz & Orson Welles as the days goes by Romney has more than a chance, and the democrat has a little less than a hope.
The biggest trap in political forecasting is that it is far too reliant upon statistics and the movement of those statistics over time. Pundits often then unintentionally have to accept time an independent variable in the election. In other words, the simple passage of time drives statistics. While a trend is defined over a period of time, time itself is incidental and not a driver of the trend. This means that a trend will continue or not based on events and perceptions. In the end, historical data is just that, historical. It does not forecast the future, but tells the story of past events. Because those events are still in the consciousness of voters over a period time, the illusion of time as a driver of behavior becomes more apparent. Given recent events, and the trap that past is prologue in politics, I am not yet ready to accept Nate’s argument that Obama has a slight advantage.
The NYT will try to brainwash and mislead the readers. Their number is decreasing by the day.
To Jonathan Tobin: n> That sounds about right as does the criteria by which he seeks to judge trends via an economic index . . . n—– "as do the criteria" – that is, one criterion, two criteria. n