The first of many Presidential debates took place on September 26th and what we witnessed was a less than prepared Republican learning a valuable lesson in politics from a seasoned Democratic candidate. Our focus here isn’t to make the case of which candidate is better suited than the other but to understand better the market sentiment towards one candidate over the next.
Leading up to the presidential debates, markets sold off approximately 1% which was interpreted as concerns that a Donald Trump Presidency could in fact materialize. The next trading day, market participants seemed to have been assured that not only was Donald Trump incoherent and unprepared but lacked the temperament and poise presented by his democratic counterpart. Market participants, for a brief moment, had a sigh of relief that a Trump president was less likely than first thought. Markets responded trading up and closing ahead by 0.7%.
A single debate is not entirely indicative of an election outcome and from now till the presidential elections on November 8th, we expect more volatility to come. As we get closer to election time, we may see Trump take more measured steps in preparedness and if the markets get any sort of whiff that Trump has what it takes to take on Hillary, it will manifest itself through negative market movement. We at TrendVesting were interested in seeing what predictive value (if any) that Google Trends may have in determining presidential election outcomes and extrapolate from this what we’re to expect in the market trend post elections.
Our thought process begins with the assumption that voters will make an attempt to educate themselves on each of the candidate’s campaign platform, that is, what is the candidate all about? We suspect that as searches for one candidate surpass those for the opposition that it would suggest internet users (isolated in the U.S.A.) are more interested in the platform of one candidate over the next and thus act as an opinion poll for the presidential race. I’m sure there are many flaws that can be pointed out in the rationale. One in particular that the current Republican candidate is hardly your run of the mill career politician and as a result of his antics and celebrity status, search interest could be skewed in his favor for more than his campaign platform.
We paired up each of the Democratic and Republican front runners from respective elections as far back as November 2004 and noticed that the candidate that had the greater search count ended up winning the election. We also note that when using the candidates first and last name the relationship broke down slightly, in that searches for ‘Kerry’ were greater than those for ‘Bush’ during the November 2004 period (We all know about the controversies that transpired then!). We could chalk this up to Google being in its infancy during that time period and that today’s Google data is more representative of the public’s interests but it’s just as likely that this is simply an exercise of data mining which is why readers should note that Google Trends Data should not be interpreted as being predictive in nature. Nonetheless, we note that search count has correctly predicted the last three election outcomes.
We focus in on the current electoral candidates Clinton V Trump, and taking what you will from the above, the results should raise some concerns among investors banking on a Clinton presidency:
That data shows that search interest in the Republican candidate Donald Trump is materially greater than searches for his Democratic counterpart with nominal searches for the month of September in Trump’s favor 70 – 52. While this may not sit well with market participants, we point out that searches for the term ‘Clinton’ continue to grow and could soon surpass ‘Trump’ searches.
With a little more than a month left till elections, it will be interesting to see where search count is at and whether or not Google Trends maintains its perfect track record of predicting election outcomes.