At this year’s Manning Networking Conference I found myself having lunch at the same table as Sasha Issenberg, the author of the Victory Lab, a book that chronicles some of the awesome, new, cutting-edge, Big Data electioneering techniques the Obama campaign employed. He had just given an interesting presentation on this subject, which inspired me to buy his book (though I confess I haven’t read it yet, though I plan to).
I pointed out to him that while everything he said might be true, President Obama only won because of Hurricane Sandy. By hitting landfall a week before the election, this natural disaster banished Mitt Romney from the headlines in that critical last week of campaigning while it basked Obama in a presidential glow thanks to the fawning press coverage he received. In contrast to the treatment received by President Bush– who was immediately lambasted as soon as Katrina made landfall - the press praised President Obama’s handling of Sandy, in spite of the many disaster relief screw-ups (that were buried on the back pages and only highlighted in the conservative blogosphere).
As evidence, I pointed out that the same polls that correctly predicted Obama’s victory in the immediate days prior to election day, showed Romney leading, prior to Sandy’s landfall. This could be seen on Real Clear Politics’ (RCP) presidential election polling graph. Issenberg looked at me blankly and said he was not aware of such numbers.
I thought to myself: could I have imagined the whole thing? I mean, a data whiz and politics wonk like Issenberg surely couldn’t have missed numbers like that. Well, I finally got around to checking out the RCP archives and it turned out I was right. The politics geek got it wrong. Here is a screencap of the RCP polling graph for the final week and a half of the election (Sunday, Oct 21 to Tuesday, Nov 6):
It shows Romney clearly ahead of Obama from Oct 22 to Oct 30 in the public opinion polls making up RCP’s daily average. He wasn’t ahead by much (less than a percent), but this difference had persisted for over a week.
According to the Wikipedia page for Sandy:
“New York City began taking precautions on Oct 26. Governor Cuomo ordered the closure of MTA and its subway on October 28. … Train service and stations as well as the Port Authority Bus Terminal were shut down in the early morning hours of October 29… . On October 29, Mayor Michael Bloomberg ordered public schools closed. … Additionally, 200 National Guard troops were deployed in the city. Additionally, one of NYU Langone Medical Center's backup generators failed on October 29, prompting the evacuation of hundreds of patients, including those from the hospital's various intensive care units. U.S. stock trading was suspended for October 29–30.”
Meanwhile in New Jersey:
“Preparations began on October 26, when officials in Cape May County advised residents on barrier islands to evacuate… . Governor of New Jersey Chris Christie ordered all residents of barrier islands from Sandy Hook to Cape May to evacuate and closed Atlantic City casinos. Tolls were suspended on the northbound Garden State Parkway and the westbound Atlantic City Expressway starting at 6 a.m. on October 28. President Obama signed an emergency declaration for New Jersey. … Most schools, colleges and universities were closed October 29 while at least 509 out of 580 school districts were closed October 30.”
So it sounds like the biggest impact from Sandy in New York and New Jersey was felt between Oct 28 and Oct 30. Oct 30 was the last day Romney was ahead of Obama. Below is a screencap of RCP’s graph of Obama’s job approval polls for the same time period. Notice that the President’s job approval numbers increased from a low of 48.3% on Oct 27 to 49.8% on Oct 30, where they roughly stayed until the election. Prior to Oct 27, Obama’s job approval numbers had been falling.
From Oct 31 to Nov 3 there were hardly any election polls because of the storm, so the Romney vs. Obama graph hardly budges. Fresh polls started coming in on Nov 3 and showed Obama pulling away. The fact that both sides gained supporters was likely due to Low Information Voters making up their minds at the last minute. The difference between the final polls and the election results was likely because of the same trend continuing for another day.
Old political operators like Dick Morris like to point out that the undecideds almost always break against the incumbent. With high unemployment, a bad economy and Obamacare hanging around the President’s neck like an anvil, one would have thought this rule of thumb would have been especially applicable on Nov 6, 2012. But it wasn’t. The undecideds broke for Obama. Was the press’s hagiographic treatment of Obama’s performance during Hurricane Sandy a cause?
Did the virtual banishment of presidential candidate, Mitt Romney, from the headlines in the last full week of the campaign play a role in his defeat?
And might Mitt Romney be president today if Hurricane Sandy had decided to track up the mid-Atlantic instead?