U.S. Politics

How Much Do ‘October Surprises’ Move The Polls?

Former New York Rep. Anthony Weiner in New York in 2013 |RICHARD DREW / AP

FiveThirtyEight

People are already calling it an “October surprise” — an unexpected moment late in the campaign that could change the trajectory of the election. The news that the FBI is looking into a new batch of Hillary Clinton’s emails (found on the computer of Anthony Weiner) is certainly surprising. And it came in October. But it will take several days to measure its effect on the race, and the real surprise would be a wild swing in the polls. That’s because even the most memorable October surprises of recent history weren’t the game-changers they’re sometimes portrayed to be.

There’s no official list of October surprises, a term that is loosely defined, but I chose six events from past campaigns that would seem to meet the definition, using the benefit of hindsight.

President Lyndon Johnson announces a halt to the bombing of North Vietnam on Oct. 31, 1968, leading to peace talks.

Polling average the week before: Richard Nixon +3.

Polling average in the final week: Nixon +1.

Result: Nixon +1.

Vice President Humphrey couldn’t unify the Democratic base, largely because Johnson, who had escalated the increasingly polarizing Vietnam War, was so unpopular. Johnson’s bombing halt is sometimes referred to as the “first” October surprise, and so you might expect it to have had a clear effect on the polls. The truth, though, is more muddled. Although Nixon had held a large lead through most of October, it was down to just 3 percentage points even before Johnson’s announcement. The polls moved towards Humphrey by another 2 points after the announcement. That’s not nothing, but it wasn’t enough.

News breaks on Nov. 2, 2000, that George W. Bush was arrested for drunk driving in 1976.

Polling average before: George W. Bush +4.

Final polling average: Bush +3.

Result: Al Gore +1. (But Bush wins in the Electoral College.)

Bush, a born-again Christian, was running as a compassionate conservativewith the heavy backing of evangelical Christians. There was talk that the DWI news, coming five days before the election, would erode Bush’s support among members of this key voting bloc. But it’s not entirely clear what impact Bush’s driving record had on the race. Bush underperformed national polls conducted the week before the news broke. And Karl Rove, Bush’s chief campaign strategist, believes it may have cost Bush 2 percentage points nationally. That’s certainly possible; from the week before the DWI news to the final polls, Bush lost about 1 percentage point off his margin over Gore. Gore, of course, won the popular vote but lost in the Electoral College. If the DWI was the cause of the late Gore polling bump — and of Gore outperforming his polls on Election Day — it very nearly cost Bush the election.

Osama Bin Laden releases a video tape on Oct. 29, 2004, talking about the U.S. election.

Polling average before: Bush +3.

Final polling average: Bush +2.

Result: Bush +2.

Bin Laden ridiculed Bush in the tape, but also said it didn’t really matter whether Bush or John Kerry were elected. At the time, there were argumentsabout which candidate would benefit from the tape — Bush, because he was running as the national-security candidate, or Kerry, who could remind people that Bin Laden had not been caught. Months later, Kerry said the tape had clearly benefited Bush, and said the tape had a noticeable effect on Kerry’s polling. But the public polls showed little impact from the tape. The national polls conducted before the tape was released were very predictive of the final result. The final polls in 2004 (after the tape’s release) were about as accurate as they’ve ever been.

The stock market suffers large losses the week starting Oct. 6, 2008.

Polling average before: Barack Obama +6.

Final polling average: Obama +8.

Result: Obama +7.

There can be little doubt that the economy hurt Republican John McCain’s chances of winning the 2008 election. Indeed, there were so many bad economic developments — from the Lehman Brothers collapse in mid-September to the $700 billion bailout of the financial system on Oct. 3 — that you could choose any of these moments as a September or October surprise. The key thing to remember, though, is that the economy was struggling through most of 2008, as indicated by a rising unemployment rate; the recession, it later turned out, actually began in December 2007. That’s why it shouldn’t be too surprising Obama held a consistent lead after wrapping up the Democratic nomination in June (with the exception of a short McCain surge immediately following the Republican convention). Whatever effect the crashing economy had on the election seems to have mostly been baked in by early October.

Hurricane Sandy hits the East Coast on Oct. 29, 2012.

Polling average before: Obama +1

Final polling average: Obama +1

Result: Obama +4

Some pundits, mostly on the right, had confidently predicted Mitt Romney would win the 2012 election, and when he did not, a few blamed Sandy, pointing to a rally-around-the-flag effect for Obama after the superstorm made landfall. The theory makes some sense, and Obama did do better than the polls before Sandy indicated he would. But the final result was also better than his post-Sandy polls. That suggests the polls were simply off, rather than any sizable effect from Sandy. Obama held a small lead before and after Sandy hit. In the swing states, the polls were consistent in showing an Obama edge. They were also pretty consistent in the states that were directly affected by Sandy.

The hostage crisis in Iran remains unresolved in the final weeks of the 1980 presidential election.

This was the surprise that never really happened. For months, Ronald Reagan had warned that President Jimmy Carter would try to create an October surprise by freeing the hostages held by Iran’s revolutionary government since November 1979. But talks with Iran stalled, and the hostages were not freed until after Reagan’s inauguration. (Allegations later emerged that Reagan’s team had secretly negotiated with Iran to delay the release, but repeated investigations never turned up proof of the charge. Even if the allegations were true, there’s no sign Carter lost ground because of them.)

The hostage crisis initially caused a rally-around-the-flag effect in late 1979 — Carter’s popularity surged. By late 1980, however, as the crisis dragged on with apparently no end in sight, it only added to Carter’s campaign woes. Carter clearly did worse in the election than he was doing in the polling before negotiations broke down, but there was a lot going on late in the 1980 campaign, making it hard to attribute Reagan’s performance to any single event. Carter and Reagan, for example, held only one debate, and it was only after the debate that Reagan’s standing jumped, and even then the polls underestimated his final margin of victory. The final polling average for Reagan was +3, but he won the election by 10 points.


Some of the October surprises listed above (the halt in bombing in North Vietnam and Bush’s DWl) appeared to have a modest effect on the polls. Others, less so. All told, these surprises moved the polls — from the week before to the final week — about 1 or 2 percentage points, on average. None of the surprises on this list moved the polls by more than 2 points.

Again, this isn’t a full list, but it makes sense that late campaign news would have a limited impact. The later in a campaign an external shock occurs, the more voters have already made up their mind and the more impressions of the candidates are fixed. October surprises, in other words, may have less of an effect because they come in October.

That doesn’t mean, however, that the 2016 presidential race is decided. Clinton currently has a national polling lead of about 5 points, according to FiveThirtyEight’s forecast. In three of the campaigns listed here (1980, 2000 and 2012), the final polling average was off by 3 percentage points or more. Even if the polls don’t move, they could just be wrong.

Harry Enten is a senior political writer and analyst for FiveThirtyEight.

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s