TCS Daily

When France Polls America...

By John Rosenthal - November 29, 2005 12:00 AM

The most recent AP-Ipsos poll, released on November 11, brought bad news for President Bush. The headline told the story: "Poll: Most Americans Say Bush Not Honest". Coming just after the indictment of vice presidential aide "Scooter" Libby for perjury in the so-called CIA leak affair, the implication was clear: the majority of Americans were beginning to get what Democrats and Frenchmen had understood all along (or almost): "Bush lied!"

But this was not the first time that an AP-Ipsos poll had been the bringer of bad tidings for the President. Back in September, it was an AP-Ipsos poll that in the aftermath of Katrina first showed the President's approval rating falling below the landmark 40 percent barrier, thus confirming the widespread perception -- widespread, at any rate, in France -- that Americans were unhappy about the President's handling of and/or responsibility for extreme weather phenomena. Curiously, other polls conducted at the time (Rasmussen, CNN/USA Today) showed the President's ratings holding in the high 40 percent range. Before that, in June, an AP-Ipsos poll announced that a majority of Americans now considered the Iraq War to have been a "mistake" -- just as so many a prominent French commentators, starting with President Jacques Chirac himself, predicted it would be. In between such blows for Bush, AP-Ipsos polls managed to reveal, furthermore, that Americans in large numbers found other Americans ill-mannered (October 14) and their own kids obese (October 24).

Why, if one were to judge by AP-Ipsos polling, one would have to conclude that American attitudes toward their President -- and indeed themselves! -- were beginning to seem positively... well, French. Americans were finally acknowledging that they were mistaken for re-electing the malevolent boob -- and that they were themselves uncivilized and fat to boot.

But, then again, if one were to judge by AP-Ipsos polling, George Bush would not have been re-elected in the first place. On October 22, 2004, just ten days before the presidential election and at a time when other polls almost all showed Bush in the lead with just a smattering of ties, the AP released an Ipsos poll showing John Kerry with a three-point lead. As Pennsylvania's Lebanon Daily News put it in an October 28 survey of the polling data: "Best news for Kerry: Up three in AP-Ipsos". Moreover, inspection of all the presidential job approval numbers since the run-up to the 2004 elections until today shows the AP-Ipsos approval number persistently tracking at the low end of the data range, sometimes flanked by the likes of the openly politicized Pew surveys and the NYTimes/CBS News polls. And if one takes rather the "spread" between the approval and disapproval numbers, the AP-Ipsos finding persistently tracks below the rest of the data: frequently so far below as to represent an obvious statistical outlier.

So, maybe Americans are not turning French, after all. Maybe the anomalous AP-Ipsos results have to do rather with the firm that is doing the polling.

What exactly is Ipsos?

AP press releases identify Ipsos coyly as an "international polling firm". Ipsos's own releases on its AP work describe the company as "a leading global survey-based market research company" -- as well as "non-partisan" and "objective". One would hardly expect them to say otherwise. But here is what neither AP nor Ipsos want Americans to know and assiduously avoid saying: Ipsos is a French polling firm. Not that this should matter per se. But AP and Ipsos undoubtedly fear that to many Americans it might or that, in light of the current climate of Franco-American relations, it might at least raise some doubts about Ipsos's impartiality and objectivity.

And what is worse: about this particular French polling firm, these doubts would be highly justified. On its home market, Ipsos is well known precisely for the unreliability of its polls and for being especially tight with the French political establishment.

Here's how a November 2001 profile in the French economics weekly l'Expansion described the cozy relationship of Ipsos co-President Jean-Marc Lech to the occupant of the Elysée Palace:

During the two seven-year-terms of François Mitterrand, he was one of the advisors to the prince and he held open house at Copenhagen, the famous restaurant on the Champs Elysées not far from the "castle". Since he began working for Jacques Chirac, he has left the Champs and stays put in the XV arrondissement at lunchtime. Now, he merely delivers his confidential polls personally to the antechamber of the President.

According to the latest Ipsos financial report, a holding company controlled by Lech and his partner Didier Truchot controls 35 percent of Ipsos capital and nearly half of the voting rights in the firm. Ipsos's international expansion in the late 1990s was, incidentally, largely financed by the Artémis investment group of French businessman François Pinault. This is the same Artémis and the same Pinault that were heavily implicated in the Executive Life fraud and that only avoided being indicted in US courts presumably through the intercession of Pinault's close personal friend Jacques Chirac and by coughing up some $185 million. Artémis sold its stake in Ipsos when the firm went public in 1999.

French public opinion firms have a long tradition of seemingly minimizing or exaggerating trends -- if not outright contriving them -- to suit the interests of their "prince" or, more generally, of the French political elites. This regularly leads to great "surprises" when votes are actually held: such as, for instance, the April 2002 "shock" that saw Jean-Marie Le Pen of the National Front taking second place in the first round of the French presidential elections in front of Socialist candidate Lionel Jospin. Ipsos arguably has the worst track record of them all and, unlike the more discreet Truchot, Lech has made something of a name for himself in France by regularly putting his foot in his mouth before important votes. In the run-up to the April 2002 ballot, Ipsos polling and Lech had been building up Trotskyist candidate Arlette Laguiller as the potential "third man" in the presidential contest. On the eve of the vote, Lech predicted that Le Pen would receive 13 percent of the vote and confidently asserted that "if there's one political party whose score we can predict days before the election without worrying about making a mistake, it's the National Front". Le Pen finished with 17 percent and Ipsos's "third man", the politically non-threatening Laguiller, finished well back in the pack with less than 6 percent.

Even though the last Ipsos survey before the May 29 French referendum on the EU "constitution" hit the mark on the 55 to 45 percent official result, Ipsos polling in the run up to the referendum provided still further confirmation of its seeming penchant for accommodating the designs of the French elites: which in this case clearly involved pulling for a "yes" vote. Thus, on May 2, less than a month before the referendum and after opinion polls had already been showing the "no" vote steadily ahead for several weeks, Ipsos pulled a veritable rabbit out of the hat and announced that the "yes" had suddenly recaptured the lead -- and by a comfortable six-point margin no less! As this startling news evidently did not manage to convince the majority of the French that they did, after all, support the EU "constitution" that they were against, Ipsos, rather than risk embarrassment, quickly went back to reporting a majority for the "no".

It should be noted that Ipsos risks no embarrassment in publishing its recent figures on George Bush, since the latter will not be running in any further elections. A closer look at Ipsos polling data, however, gives considerable cause to pause regarding the company's methods. Thus, for example, for it latest poll -- the one allegedly showing that "most" Americans consider George Bush dishonest -- Ipsos somehow managed massively to oversample Democratic voters. As the raw data reveal, fully 51 percent of the Ipsos respondents identified themselves as Democrats as compared to only 40 percent who identified themselves as Republicans.

On June 27, 1939, Prime Minister Eduard Daladier went before the French Chamber of Deputies and alluded gravely to the evidence of German preparations for war. "While all these events are occurring on our borders," Daladier continued,

at the interior of our country we are witness to an active campaign of propaganda, the links of which to foreign influences have now been demonstrated. The point of the campaign is to break the unity of France and to open a breach in the combined energies of the French nation, through which all sorts of intrigues and maneuvers will be able to pass.

Hardly anyone today would question the perspicacity of the French prime minister's words. Americans might be well advised to recall them the next time they read the results of an AP-Ipsos poll.

John Rosenthal's writings on international politics have appeared in Policy Review, the Opinion Journal, Les Temps Modernes and Merkur. He is the editor of the Transatlantic Intelligencer (  


TCS Daily Archives