Can these 6 worldwide Google search trends predict the 2020 US presidential election?

By Bhaso Ndzendze

Whilst international observers are commonly deployed in fragile or new democracies in the developing world, elections in the developed world are viewed from outside, partially out of a sense of stake-holding in the outcome by the rest of the world (due to the preponderance of the West in the affairs of these countries) and partially out of curiosity, and therefore are consumed as a piece of popular culture. Out of these, the US electoral cycle is perhaps the most closely watched the world over. This is no coincidence as in each successive year previous records are shattered in terms of expenditures on advertising by the various campaigns. The digital sphere has become the main arena in which the various campaigns reach out to potential voters. By one 2019 Forbes estimate, the current electoral cycle has seen “an increase of 59% from the 2016 election year when an estimated $6.3 billion was spent,” which represents nearly 16.5% of total local broadcast TV advertising revenue for this year, whilst digital media is forecast for 21% of political ads, whilst cable TV and radio both claim 14% and 5% respectively. This disproportional share for digital spending is indicative of what scholars have termed as the rise of “computational politics,” defined by one study as “the application of digital targeted-marketing technologies to election campaigns.” With this increase arises the question of which candidate will come out victorious, and whether expenditure is a predictor for which will win. In the last (2016) election, this proved not to be the case, as the Democrats, at $1.191 billion, raised nearly twice as much in dollar terms than the Republicans (at $646.8 million) but still lost. More recently, we’ve seen Michael Bloomberg being forced to drop out of the campaign, despite spending as much as $936.2 million, whilst Sanders and Biden, the last two Democratic contenders, had spent $162.3 million and $84.7 million, respectively. In fact, US elections are notoriously difficult to predict with a fair degree of certainty. Be that as it may, looking at some data in the previous sixteen years (i.e., four electoral cycles) makes for some insightful analysis, and potential projections.

Briefly, I trace over the 2004-2016 period Google queries for American presidential frontrunners and eventual winners and incumbents in the world. All data utilised in this article is obtained from Google Trends, a publicly available dataset of worldwide Google searches since 1 January 2004. All charts were generated by the author from sorted data. The Google Trends scores are values that are calculated on an index that places scores from 0 to 100, where, according to Google, “100 is the location with the most popularity as a fraction of total searches in that location, a value of 50 indicates a location which is half as popular.” The four following charts below show the growth for searches for each candidate between 2004 and 2012.

[Chart by author. Data sourced from Google Trends.]

[Chart by author. Data sourced from Google Trends.]

[Chart by author. Data sourced from Google Trends.]

[Chart by author. Data sourced from Google Trends.]

The undeniable trend from all four charts is that the candidate who gets the most searches goes on to win the election, despite both candidates getting an uptick the most amount of searches in November, the month of the election (with the eventual winner experiencing the most amount of searches). Given the results from the four election cycles, it is worth peaking assessing the amount of searches for the two Democratic frontrunners against Donald Trump.

Despite the field only being in a phase of narrowing down, past results show that the eventual winner starts getting the most amount of searches as early as January of the election year.

[Chart by author. Data sourced from Google Trends.]

The only exception was the 4th of March, the day after Super Tuesday, when Joe Biden got the most amount of searches, gaining 40% of the searches, whilst Sanders got 31% and Donald Trump got 29% (see pie chart above). Following the 4th of March, however, the ranking reverted to Trump being more searched than both Democrats, whilst among the Democrats themselves, Vice-President Biden began being more searched than Sanders, which continues to be the case at the time of writing.

[Chart by author. Data sourced from Google Trends.]

For the majority of the first quarter of 2020, therefore, has Trump having 77% of the searches, whilst the two Democrats only have 23% of the searches, split between Bernie Sanders (14%) and Biden (9%). If global Google search trends are any proxy for name recognition, therefore, we can reasonably expect Trump to be re-elected as President in November, regardless of which Democrat is eventually nominated. Indeed, any different outcome would go against the grain of the last four elections.