Much has been made by the media and pundits about how close the 2012 election will be. They cite a multitude of reasons including the power of incumbency helping Obama and the weak economy helping Romney. Knowing how weak the economy is the Obama camp has tried to turn the election into a choice election between a “strong leader” like Obama and an “out of touch millionaire” like Romney. The problem with this strategy is that it never seems to work.
Nearly every election since 1952 (when modern polling got its start) has shown undecided voters break for the challenger on election day or within a three-day period. The theory goes that if you have not found anything good to vote for the incumbent on by election day odds are you won’t find it within 24 hours. These voters final choices tend to match their views on the job approval of the incumbent. For the most part, 70-80% of those who decide who to vote for on election day disapprove of the sitting president. The lone president to overcome this trend was G.W. Bush in 2004.
According to a Dick Morris analysis of how undecided voters act on election day when an incumbent president is in office his data shows from 64 to today that only one incumbent has benefitted. Johnson in 64 lost 3 pts to Goldwater, in 72 Nixon lost one point to an independent challenger, in 76 Ford lost four points to Carter, in 80 a three-point swing to Reagan or Andersen, in 84 Reagan broke even, in 92 Bush lost a point to Clinton. In 96 Clinton lost 5 points to Dole or Perot and in 2004 Bush actually gained two points.
So if this data is true then trying to turn an incumbent’s reelection from a referendum to a choice between the incumbent and the challenger is virtually impossible. But there is also another way to look at this. In recent elections pundits and analysts have paid increasing attention to a president’s approval rating on election day. The thinking goes if he is above 55% he is a shoe-in for reelection and if he is below 50% he is in trouble. Well by that standard Obama is in trouble. Even if we take the polls that show Obama above 50% approval his vote share in most surveys tracks his job approval rating.
Witness the Gallup tracking survey. In the latest addition to the survey Obama has a split 47/47 approval rating. Yet in the Gallup survey Obama trails Romney 48%-44%. It is possible that Obama’s horrid ratings on the debt and economy are the reason why his vote share is lagging his approval but what is more important is that 8% remain undecided in the tracking survey. If those 8% break for Romney, even if by election day they are only 5% the election is over. Rasmussen Report’s tracking poll also shows the same trend.
But let’s not just look at one survey for confirmation. Let’s look back to other data on this trend since 1972. Jay Cost at the Weekly Standard has helpfully posted data from the National Election Study showing few presidents running well ahead of their approval ratings. The obvious reason being few voters who disapprove of the president are willing to back him. And considering the data presented by Dick Morris the data is pretty damning. no offense to Ed Kilgore over at the New Republic. I love you but you are dead wrong on this election being a choice.
In this era of hyper-partisan and polarized politics there is also evidence to suggest that candidates and Congressional incumbents electoral prospects in swing districts are tied to the president. In 2006 President Bush’s job approval ratings were stuck at 42% and the GOP only won 44% of the popular House vote. Or also consider in 2008 that of the nine Senate seats Democrats picked up 7 were won by then Obama the candidate. And in all 7 states President Bush’s job approval was well below 50%. One must always be cautious of reading into such events but they do point to an interesting trend.
Now fast forward to 2010 and we can see this trend continue even more. Sean Trende at RCP has an interesting stat on all the Senate races in 2010 in each state with the president’s job approval in each and the Democrats actual vote total in the state. In the three states where the Democrat could not outrun President Obama’s approval below 50% the Republican won. In the one state where the Democratic candidate ran even with the president’s approval she was crushed and in only two races where Democrats outran the president’s approval did the GOP prevail. It is worth noting that in winnable races such as DE, NV, and CO the GOP electorate opted to go with the less credible, more conservative candidate and they lost. Thus this shows local factors still matter in voting.
All this shows that if the data and history are any guide this will be a referendum election. And if the president does not establish a substantial lead over Romney, undecideds continue to amount to 5-10% in the polls and the president’s approval is below 50% on election day he is very likely to lose and voters referendum on him will be for him to vacate the White House.