LONDON/NEW YORK (Reuters) - Wall Street is sick of polls.
Leading up to the U.S. election on Tuesday, popular polls wavered about who would win and by how much. In the span of a couple months, they shifted from a likely victory for Republican President Donald Trump to a tight election contest, and then to a win for Democratic candidate Joe Biden.
On Wednesday, voting tallies offered another big bag of uncertainty about who will run the White House during the next term.
“I’m just absolutely perplexed as to why we ever believe these bloody polls,” said Stuart Oakley, global head of cash currency trading at Nomura, citing wrong polls during prior U.S. and UK elections, as well as Brexit.
“The polls are so wrong every single time. Why do we give them any credibility whatsoever? It’s just unbelievable, the way this keeps happening.”
Major global banks and hedge funds have hired polling firms to supplement what they were seeing in publicly available data, but without necessarily getting an edge.
“This is probably the Waterloo for the pollsters,” said Simon Maughan, head of Trading Alpha at Liquidnet, drawing a comparison between political polls today and the battle that ended the reign of French emperor Napoleon.
“Whatever they do to reach people is not working,” Maughan said.
Pollsters have struggled to accurately gauge people’s voting intentions during a rise of populist politics over the past decade. As candidates have become less moderate and more brash, voters may support someone who says controversial things but not want to admit that to their colleagues, friends and family, much less a pollster. [nL4N2HL13U]
Even Trump tweeted on Wednesday that his lead in polls had “magically disappeared.” His comment added to broader criticism of sites like FiveThirtyEight, which claim to have polling tactics that give them an advantage not always evident in the results.
Reuters conducts its own polls with a firm called Ipsos. The most recent predicted victories for Biden in Wisconsin, Michigan and Pennsylvania, and a narrow win in Florida. [nL1N2HK3G3]
The first three remained uncertain on Wednesday afternoon because not enough votes were counted, while Florida had gone to Trump.
After Trump beat his opponent in 2016 – the opposite of what preeminent polls had predicted – some financial firms hired their own people to call voters with unique questions and get an edge for themselves or clients. Though that helped some hedge funds reap profits around Brexit, successes have been rare.
For instance, an executive at one major bank told Reuters ahead of Tuesday’s election it had hired its own polling firm, which convinced management a Biden win was likely. The reasons? High turnout tends to be good for Democrats, the idea that people do not want to admit to supporting Trump in polls is false, and pollsters do not usually get things wrong in the same direction in consecutive presidential elections, the executive said. [nL1N2HL1IN]
A major U.S. hedge fund also had its own polling, which predicted Biden would be ahead, but in a much tighter race, an investor said. Their private polling also showed Republicans would hold onto the Senate, which was not a uniform assumption.
Although Biden is still in the lead, assumptions of a landslide or that Democrats would sweep the Senate have been incorrect. Though millions of votes are still being counted, Trump won far more support than pollsters expected in certain states, especially Florida.
“If I was in the polling business I would be seriously disturbed,” said Peter Kraus, a former Goldman Sachs executive who founded asset management firm Aperture Investments in 2018. “The samplings are just not precise enough and the margins of error are way bigger than they should be.”
Kraus does not believe that Trump supporters are reluctant to reveal their allegiance to pollsters.
“The ‘shy Trump’ thing doesn’t make any sense to me,” he said. “Trump voters put flags on their cars and drive down the street!”
Nate Silver, who runs FiveThirtyEight, defended the site’s predictions on Twitter as others posted criticism. He has long asserted that polling firms are fallible: they offer predictions with margins of error, rather than certainties.
Some investors went so far as to say the era of polling is over. Big firms will stop investing in customary polls and try to figure out election outcomes in a different way, they said.
Tom di Galoma, managing director at New York-based Seaport Global Holdings, predicted that strategies for gauging voter intention will move more toward internet searches, since people rarely pick up their phones for random numbers anymore.
“The polls can’t be believed,” he said.
Reporting by Simon Jessop, Lawrence White and Svea Herbst-Bayliss; Additional reporting by Anna Irrera, Matt Scuffham, Herb Lash, Ross Kerber, Alwyn Scott, Carolyn Cohn and Tom Westbrook; Editing by Lauren Tara LaCapra and Tom Brown
Our Standards: The Thomson Reuters Trust Principles.