So, the opinion polls don’t always get it wrong after all!
Thursday, June 8, 10 pm. I was sitting on my living room couch with my wife, switching on BBC One to watch the 10 o’clock news. I didn’t expect to learn much about the election results, as the polling stations had just closed, but I was curious anyway.
The first thing to appear on my TV screen was an image of the clock face on the Queen Elizabeth Tower (often mistakenly referred to as “Big Ben”, which is in fact the great bell within the tower, though that’s a different discussion), with the famous bell tolling 10 times, as the presenter announced the results of the first exit poll.
“The Conservatives are the largest party,” he began. So far so good, as that was as expected. But then the big shock: “No, they don’t have an overall majority at this point.” The presenter appeared to be as startled as me, along with a majority of financial market professionals.
In the first few minutes after the announcement, the pound lost more than 2 percent versus the dollar, and in Friday’s trading the 10-year Gilt yield fell by 4 basis points. The FTSE, on the other hand, gained more than 1 percent as the exchange rate depreciated, a pattern that has become quite familiar since the Brexit vote in June last year (another night I will never forget, watching BBC News in a hotel room in New York—but that, again, is another story for another time).
But was the election result really such a surprise after all? Didn’t we see the projected lead of the Tory party over Labour erode from over 20 percent on the day Theresa May announced the election to just 1 percentage point in one the of the last opinion polls right before the actual vote? Was it just wishful thinking that determined how financial markets positioned themselves? Then again, hadn’t Brexit and the US presidential election already taught us that opinion polls cannot be trusted?
Not necessarily. During the two rounds of the French election in April/May, much of the market movement actually happened before the electorate eventually went to the ballot boxes. For example, during the week ending April 21 (two days before the first round of voting), the 10-year German Bund yield had already risen by 6 basis points, a sign that markets had started to relax and risk appetites were increasing. Two weeks, 40 opinion polls and one televised debate later, the same benchmark rate stood another 17 basis points higher, driven purely by expectations. The outcome of the actual vote two days later had almost no effect on financial markets at all. The polls got it right!
Whether one trusts polls or not, one lesson we have learned from these recent referendums and elections is that voters cannot necessarily be relied on to do what politicians, and financial markets for that matter, expect them to do. So, it is important to be prepared for different outcomes. A good starting point is to compile a list of possible scenarios (e.g. “Tory majority”, “Tories biggest party, but no outright majority”, “Tories no longer biggest party, but no other majority either”, “Labour victory”) and estimate how the market might react to each. FX markets are often the quickest to react and a reasonable first step might be to implement exchange rate shocks within a stress-testing framework. When using a risk model in conjunction with the stress test, recent correlations or from periods that are similar to the current environment can give an indication where other parts of the market might go (e.g. pound down, FTSE up).
For more information and inspiration on stress testing, why not check out some of our recent research papers?