British citizens will vote Thursday on whether to leave or remain in the European Union.
1. In the past, polling for “separation” votes has been wrong, while prediction markets have been right
These results follow the pattern we’ve seen in predictions about previous separatist votes in established democracies: Polls tend to believe that the separation will happen, but markets predict it will not — and, so far, the markets have been correct.
For instance, in 2014, Scotland voted on whether to separate from the United Kingdom, while in 1995, Quebec voted on whether to separate from Canada. In both cases, polls predicted more votes for separation than were actually cast.
Here’s why: Within a given geographic area, separation usually has considerable social desirability bias. People want to sound like part of the crowd; it is socially desirable to claim to love one’s country and people. On the other side, the reasons people have for staying are generally more mundane and less sexy than the reasons for leaving.
So if someone calls with a deep Scottish or Quebecois accent and asks whether a respondent wants to leave the union, he or she will feel a lot of internal pressure to say yes.
But in the privacy of the voting booth, many of the same people choose the less sexy but more stable option of staying.
2. Undecideds break for the status quo
Second, even looking at the polls, roughly 9 percent of people said they were still undecided with only two days before the vote. In such cases, undecided voters tend to break toward the status quo. A voter who is truly not certain will find it easier to vote for continuity than major change.
3. Turnout will make a difference.
Third, the base voters for staying are traditionally “looser” voters, or citizens who do not always find the time to go to the polls. That means that if there’s high turnout, voters will most probably be opting to stay; if not many people go to the polls, a “leave” vote is more likely.
Taking a closer look at the polling, younger, liberal, Scottish/London voters want to stay, while older, conservative, Welsh/Northern England voters want to leave the E.U.
Polling is much better at understanding current sentiment than it is at assessing likely voter turnout. It tends to assume that roughly the same number of voters will come to the polls as they have in past elections. With extremely high coverage and stakes, the markets are assuming that voter turnout will be strong.
4. Keep a close watch on a second market: The exchange rate between the British pound and the U.S. dollar.
For the past couple of months, the currency exchange rate has been tracking the prediction markets’ assessment of the probability of Brexit. The more likely the U.K. is to stay within the E.U., the more dollars it takes to buy a pound. The market believes that if Britain votes to leave, people will be desperate to exchange their pounds for dollars. And right now, the currency market has made pounds very expensive in relation to dollars — which again suggests that Britain is about to vote to stay.
David Rothschild is an economist at Microsoft Research. Find him on Twitter @DavMicRot.