Trading on the election polls

Have you ever considered basing your trading strategy on the results of election polls? In a time where it seems you would probably be better off predicting election results by flipping a coin, gazing at the stars or by looking at the dregs in your tea cup, you’re probably wondering “why on earth would I?” The utterly woeful ability for the polls to even ball park the results of recently passed elections and referendums, means that basing your trading plan on them, seems like a plan headed for financial disaster.

With the polls advising that England was bound to “remain” in the Brexit vote, that Clinton was sure to be the next President of the United States and that Theresa May was undoubtedly going to succeed in shoring up her electoral support, the only confident conclusion that one can glean from all this is that the polls are nothing if not completely unreliable. 

As to why the polls have been so completely off the mark has drawn a lot of speculation. Yet all factors point towards their random sample selections being not that random. Pollsters have acknowledged difficulties with obtaining accurate and random samples ever since the proliferation of mobile phones. Previously, the ubiquity of landline telephones made the process of getting a random and representative sample fairly straight forward as it was simply a matter of randomly selecting names from a phone book, call those potential voters and talk them through the interviews. In the process, a rich supply of context and human understanding was provided for which pollsters were more accurately able to analyze responses. This ‘old-fashioned’, if not ‘quaint’, method of a simple phone call also provided higher response rates and helped to moderate a non-response bias, where the polls themselves become skewed by the type of people who tend to answer.  

So why is blame lumped on the proliferation of mobile? Firstly, because landline telephone samples are no longer adequate. Secondly, the demographic differences in their adoption has entailed that mobile holders of a certain stipe, already represent a segmented group. Thirdly, mobile phone numbers are not as consistently publically listed as they were for phone lines. Yet, online survey methods don’t serve pollsters that much better, a seemingly more resourceful method to supplant the telephone, they have become particularly susceptible to bias and are regarded to have even lower quality results.

But coming back to our initial question, would you ever base a trading strategy on the results of election polls? The answer could be both no and yes, it really all depends on when. The ‘no’ or perhaps ‘no-way-in-hell’ camp would accurately point to the volatility in the markets that occurs just prior to national elections. This volatility reflects that there is no confidence being held in the reliability of the polls, but more than anything else mirrors the public’s indecision of how events will pan out. While forex traders do love their volatility, it still makes for a risky period as it is near on impossible to trade strategically.

Alternatively, those in the ‘yes’, or the ‘i-am-not-as-crazy-as-i-look’ camp, will point to the immediate aftermath of an election as a prime opportunity to execute their trades. In the last throws of the election, there is a lot more clarity on where the market is headed. If the polls were indeed accurate, a currency will normally increase in value, as this creates a sense of stability, especially if the election results are perceived to be a strength to the future of the economy. If the polls were inaccurate, a currency will normally decrease in value, as markets will respond badly to surprise shocks and the sentiment of instability that the polls infused. 

To test the theory, we measured it up against the spate of recent elections that include the US elections, the UK referendum and the subsequent snap general election, all of which experienced results that were in stark contrast to poll predictions. As election results were published the USD took a plunge against all other major currencies, dropping by as much as 8 percentage points against the Euro. If those other countries weren’t struggling with their own political turmoil there is reason to think the USD would have dropped even more.

Post Brexit the sterling took an absolute nosedive falling by 10 percentage points in the immediate aftermath. Though it’s arguable this happened in anticipation of the economic setback of England no longer being part of the EU block. The sterling did gradually start to pick up a little, especially when the general election was called in April, but the surprise results for the Conservative party appeared to be equally surprising for everyone as the pound plummeted yet again to a new low of 1.1278 GBP/EUR.

The rare occasion were the polls actually got things right was in the case of the French elections in May. Projections gave Macron almost two thirds of the vote, indicating that the path to the Élysée Palace was most certainly determined, even despite the fact that his party was barely a year old and that he was a relative unknown entity. But the success of the pro-EU campaigner came to enormous relief on the markets, which rose after the first round of elections and then again after the second round of elections, as the future of the EU appeared to be pulled out of its state of uncertainty.     

Keep in mind that trading the market is never this simple and traders would be well served to pay attention to interest and inflation rate changes along with other market indicators. Furthermore, elections don’t happen in isolation but especially in these cases, they occur in wider political contexts where there are multiple concerns at stake. Perhaps not all too coincidentally, the polls consistently appear to side with the electoral party that is considered to be of stronger economic benefit to the national economy, most likely a characteristic of its non-random selection.

And as to whether that’s a failing of Democracy, is another article for another time.

---