Wrong Again: Poll-Driven Wisdom Failed in Colombia. Where Next?

  • World likely to face more ‘disruptive elections’: pollster
  • Analysts say mood of revolt makes outcomes harder to call

First Brexit, now Colombia. For the second time in a few months, referendum results have made a mockery of conventional wisdom shaped by opinion polls, and investors who believed them.

Colombian stocks, bonds and the peso fell on Monday, a day after voters rejected a peace deal with rebels. At least three surveys published last week had predicted the government-backed accord would pass with 60 percent support or more. The slide in U.K. markets in June, after Britons voted to quit the European Union, was even steeper.

Each case has its own backstory. But professionals say there are lessons that apply across the board -- including for upcoming contests such as the U.S. presidential election, where most surveys show Donald Trump trailing Hillary Clinton. Cliff Young, president of U.S. public affairs at U.K. research firm Ipsos, says that “disruptive” votes are becoming more frequent.

“They’re driven by overall angst and disgust with the system, and they’re more difficult to predict,” he said in a phone interview in Washington. “We’re probably headed into a period where we’re going to have increasingly more disruptive elections.”

Judging Turnout

In both the Brexit and Colombian referendums, Young said, pollsters struggled to gauge the likely turnout -- overstating young people’s participation in the former case, and the general level of enthusiasm in the latter. “We’re only as good as our assumptions about who’s going to show up” at the ballot box, and that’s harder to calculate in a referendum, he said.

Colombia’s peace agreement, finalized last month after four years of negotiations, would have left some former guerrilla leaders in Congress rather than in jail. That was the price President Juan Manuel Santos accepted for ending a five-decade armed revolt that had left over 200,000 people dead.

There were early signs it wasn’t popular: One poll early in August showed half of Colombians planned to vote “no.” But the government appeared to be winning people over: several surveys in the second half of August showed the tide turning. By late September, a poll conducted by Centro Nacional de Consultoria found 65 percent support for the accord versus 29 percent against.

Joe Twyman, the head of political polling at YouGov Plc in the U.K., said that when he saw numbers like those, “it seemed like a prime opportunity for a disaster.”

For a QuickTake Q&A on Colombia’s peace process, click here

That’s because of the nature of the question, said Twyman. “It’s a classic social desirability bias: You’re asking people, do they want to support peace. People seemed obliged to say yes.” As in the U.K., there may also have been problems with obtaining a representative sample, for example by including enough rural voters, he said.

That said, polling is not completely broken. Defenders of the industry say most surveys predicted a close Brexit vote and three online polls in the final week of campaigning gave ‘leave’ the advantage. Still, investors chose to believe the majority of polls which showed a narrow victory for ‘Remain,’ resulting in painful losses for many when the pound plunged to its weakest level in decades.  

Obama’s ‘Confidence’

In the U.S., the Colombian results were immediately seized upon for potential insight into November’s election for the world’s most powerful office. Even President Barack Obama weighed in, cautioning that not all polling is likely to be equally wrong. Obama said Monday that he has “a lot more confidence in the kind of rigorous analysis that’s been done of the U.S. presidential election than obviously I would of the Colombian political system.”

But Nick Verdi of Standard Chartered Bank, the top Colombian peso forecaster, said he thinks there is a wider question that investors should ask after Colombia and Brexit: Are markets adequately pricing in the risks of non-consensus vote results? He thinks not.

“The only currency where Trump’s presidency is priced in is the Mexican peso,” Verdi said in an interview in New York. “The Canadian dollar and the Chinese renminbi maybe should be higher in terms of trade links. Markets have more to do in terms of pricing in potential for Trump victory.”

Young says the turnout question that’s confounded pollsters in the U.K. and Colombia is tough to call in the U.S., too. Analysts are weighing whether the “fear factor with Trump” will drive some voters to the polls, or if disgust with both candidates will keep them away, the Ipsos chief said.

“If we have above average turnout, that really helps Hillary,” he said. “If we go below 60 percent of registered voters, that is serious Trump territory.”

    Before it's here, it's on the Bloomberg Terminal.