Professor discusses betting markets vs. polls in predicting Scottish independence outcome
Tuesday, Sept. 23, 2014
After the historic Scottish independence vote on Sept. 18, the United Kingdom remains united, but many issues and questions remain. Before the vote, nationalists in favor of independence were certain they would win, while unionists against it were fearful they could lose. So what happened?
John Chilton, professor of economics in the Virginia Commonwealth University School of Business, has an interesting take. His research interests include voting, pre-election polls and election betting markets. He notes that the betting markets outperformed the Scottish polls perhaps because people do not always respond truthfully to pre-election polls. Also, he says, promises made by U.K. Prime Minister David Cameron may have shifted many votes from “Yes” to “No.”
Chilton recently discussed his take on the vote.
Why were the nationalists so confident that they would win? Did that notion have any validity at one point?
I don't know. I can speculate. It's common for insiders to a campaign to express more confidence than the evidence they have in hand would suggest. You want to convey the message that you will win if only your supporters turn out to vote. It's cheap talk, but to not overtly express confidence is to reveal you believe you're sure to lose.
Can you explain how betting markets work and how they differ from polls?
Pre-election polls provide a snapshot of voters’ intentions, asking, “If the election was today would you vote and how would you vote?” These polls may be an accurate reflection of current intentions, but intentions can change. That's part of their appeal, giving political races a feel of a horse race. In betting markets, players want to use all information available to forecast the outcome they are betting on. Whenever the current betting odds are out of line with forecasts the smart money moves in and as it does so the odds come into line with the forecast. Polls are information that goes into making a forecast, but not the only piece of information. Bettors may be incorporating future economic conditions that transpire before the election and may affect how people vote. In the Scottish case, the bettors would have had good reason to anticipate Prime Minister Cameron would make concessions if it appeared secession had a chance.
Another example relevant to Scotland are "shy Torys." The journalist and election forecaster Nate Silver suggested "shy Torys" could be why the polls were so different from the outcome. Scotland is heavily Labour and so Torys may be reluctant to reveal themselves to a pollster. Poll results aren't adjusted for this possibility, but that correction can and will be reflected in betting markets. The same effect is seen in the U.S. where it appears some white voters are shy to reveal they will not vote for an African-American candidate.
What did the betting markets predict in this specific case?
The economist Justin Wolfers took a look. He reports the market for betting on who would win gave “No” an 80 percent chance of winning. In the market for winning margin, a four-point win for “No” was most likely. Eighty percent isn't an expression of certainty, but the markets clearly outperformed the polls that were saying it was neck-and-neck and the margin would be very close.
Why do people not respond honestly to anonymous polls?
First, you never know if it is anonymous. The caller may be from a campaign and you may believe what you say will be revealed. Second, people may feel ashamed that their view is looked down upon and that shame is present even when stated anonymously. Third, voters may use polls to decide how to vote — for example, to determine whether the election is close and they should turn out on Election Day. In that case, you may be happy to mislead the pollster, and the public.
Finally, in the Scottish case, voters may have replied to pollsters with their heart and voted with their head. You could say they were of two minds, which is not an act of dishonesty, simply part of the human condition.
You’ve mentioned that Prime Minister Cameron's last-minute promises to give Scotland greater freedom may have swayed many people to vote “No.” Why do you think the “Yes” voters may have achieved their aims even though the vote for independence failed?
It is a mistake to think the desire for independence was due solely to “Braveheart” or North Sea oil. Scotland is a Labour stronghold. The majority of Scots prefer a more liberal government than the Conservatives in power. It's underappreciated that the status quo in the UK is that the government is more highly centralized than in other developed countries. As a consequence, if a majority in a region like Scotland holds preferences about the size and role of government at odds with the U.K. majority, that creates a desire for local autonomy. The Scots want to see greater income redistribution and a reversal of decisions made by Conservative governments going back at least as far as Thatcher to shrink the welfare state and privatize state industry. The devolution of powers promised by Cameron would allow Scots to take steps in that direction.
What would you like to add?
A major post-election survey of the Scottish vote for independence reveals the electorate wasn't convinced either side knew how independence would work out. There was great uncertainty. Many voters appeared to have concluded that taking a risk on “Yes” wasn't worth it, especially once Cameron made concessions. There's evidence that in the finance literature that women make better and more cautious investment decisions. A majority of men and women voted “No,” but the women's “No” vote was stronger. The post-election survey also shows the likely targets of greater taxation voted “No.” And older voters who voted “No” were particularly concerned with how independence would affect the National Health Service and other programs that benefit older voters.
Subscribe for free to the weekly VCU News email newsletter at http://newsletter.news.vcu.edu/ and receive a selection of stories, videos, photos, news clips and event listings in your inbox every Thursday. VCU students, faculty and staff automatically receive the newsletter.