A win for the pollsters: French election predicted accurately

The first round of the 2017 French presidential election produced one clear winner: the polls.

After getting it wrong in high-prized contests such as the Brexit vote and the Trump election, polls got it right — and uncannily so — in predicting the outcome of the French vote Sunday.

Here are the results, as officially reported, followed by what the polls of the eight major organizations (Odoxa, BVA, Ifop, Ipsos, Elaba, OpinionWay, Harris, Kantar Sofres) averaged in the week before election day:

      23.9%                   22.2%                         19.4%                    18.8%                   7.6%  

What an astounding performance! The polls correctly picked the top-two finishers, who will face each other in the run-off on May 7. They also got the order of finish right for the five major candidates, and they hit a near bull’s-eye with their 23.9 percent prediction for Macron.

This is all the more impressive given the presence of a right-wing populist like Le Pen in the race. Such candidates or issues, as in the case of Brexit, have befuddled pollsters, leading to major upsets. Not so in France this year. Unlike Trump, Le Pen’s support was not underestimated by the opinion polls. If anything, she slightly underperformed in the end. Why is that?

For one thing, Le Pen is a familiar brand in France. Marine Le Pen already ran in the last presidential election — in 2012 — placing third in the first round with 17.9 percent. Her father, Jean-Marie Le Pen, has been running in presidential elections as far back as 1974.

He scored a stunning upset in 2002 as one of the top two finishers, edging the Socialist candidate. Whatever spell might have kept French voters from admitting support for an extreme-right candidate was probably broken long ago.

The polling method used in France may also have helped get it right for seemingly “politically incorrect” choices like Le Pen. French pollsters do not rely on telephone interviews, with their dismal response rate and resulting bias. Instead they contact a sample of respondents online and use familiar demographic categories as weights.

Derided by some as “quota sampling,” it nonetheless overcomes the low response rate of telephone polling. Voters are more inclined to express their political preferences by clicking online than telling an interviewer over the phone.

Sunday’s vote, of course, was not the final decision. It’s more of a top-two primary than a general election. Polling has shown for some time that Macron would easily defeat Le Pen in the run-off on May 7. But these were hypothetical scenarios until Sunday. Now, voters who preferred other candidates face a real choice.

This is the way it has been in every French presidential election since 1965, with no candidate winning a majority in the first round. What does history tell us about which one of the two-top finishers in the first round will become the next president of France? The winner of the first round has gone on to win the second round in six of the nine elections since 1965. That makes Macron the 2-1 favorite over Le Pen on May 7.

History, however, is not a perfect guide for the 2017 contest. The typical top-two competition in French elections is between a Republican on the political right and a Socialist on the left. Not this year. For the first time, neither of the candidates in the run-off is the nominee of a mainstream party.

The closest parallel is the 2002 election, in which Jean-Marie Le Pen faced Jacques Chirac of the mainstream right. It ended in an 82-18 rout in favor of Chriac, with Le Pen barely exceeding his first-round total. 

Such a landslide does not seem in the cards for Macron. He is a candidate without an affiliation to a party on the right. Hence, he does not command the support of mainstream conservatives the way Chirac did in 2002. Even though Francois Fillon, the conservative who missed the run-off, has endorsed Macron, a good portion of Fillon’s vote will likely wind up for Le Pen. 

Macron can probably count on supporters of Hamon, the Socialist candidate, who also endorsed him, but Hamon’s percentage of the vote in the first round was in single digits. A big question is what supporters of extreme-left candidate Melenchon will do in the run-off. Though the National Front is anathema to the French left, hostility to globalization and the EU may appeal to voters on either extreme.

In the end, Marine Le Pen will make it a lot closer in the run-off this year than her father did in 2002, but she will be lucky to pick up enough votes to avoid losing by less than 20 points on May 7 — an ominous anniversary of defeat for right-wing extremists.  

Helmut Norpoth is a professor of political science at Stony Brook University. Norpoth developed the Primary Model, a statistical model he uses to predict the results of United States presidential elections based on data going back to 1912. He accurately predicted the election of Donald Trump to the presidency in March 2016. 

The views expressed by contributors are their own and not the views of The Hill.