Can data science help to mitigate rising populism?
If one headline were to describe the last months of politics in Europe and America, it would be “the rise of populism”. In the US, Donald Trump secured the GOP nomination without any substantial policy plan, and in Europe populist parties continue to shake up the political environment.
In Austria’s presidential elections, Norbert Hofer of the populist Freedom Party of Austria (FPÖ) just barely lost to Alexander Van der Bellen, a candidate backed by a consortium of established parties. In Great Britain, a majority of people voted to leave the European Union. At the same time, Germany’s “Alternative für Deutschland (AfD) is currently polling at 15% on national level, effectively challenging the standing of the Social Democrats (SPD) and Christian Democrats (CDU). Populism, a complex phenomenon, might have the potential to change political discourse for the years to come – how should established, moderate parties respond? Interestingly, data science offers some of the tools it might need to counter populism in the medium-term.
The media and its ambivalent role
A recurrent theme in most populist communication is a general scepticism against “the media”. Donald Trump, for example, points out with obscure regularity, how biased the media is against him. Recently, he even called CNN the “Clinton News Network” to underline his thoughts. Similarly, in Germany media outlets are oftentimes called “Lügenpresse” by rather right-wing movements, which voices the belief that newspapers and magazines alike are spreading lies and false statements.
While the media serves as a scapegoat, it is – on the other hand – also a great vehicle to bring the populist message across. As journalism and media have undergone a massive transformation in the last decade, online content, which goes beyond cat videos, needs to be geared to generate a maximum amount of clicks. Stories about controversial topics naturally resonate better in such an environment. Even if a majority of those stories might be critical of the populist message, the old rule applies: there is such thing as bad publicity. Donald Trump is credited with playing the news cycle in an optimal fashion, thus reaching a maximum amount of people without significant campaign spending. This “free publicity” helps to boost his prominence and spreads his message rather rapidly.
The silent majority and political messaging
The core of any populist party or movement is its sense of being deprived of right, identity or voice by the ruling elites. Donald Trump praises himself of finally giving a voice to the “silent majority”, whose interests, he claims, have been systematically ignored by politicians of both parties. Simple opposition to the status-quo is oftentimes the unifying force in the early stages of populist movements. Opposing certain policies and proposing simple solutions (“We will build a wall and Mexico is going to pay for it”, Donald Trump) is the best strategy from two perspectives: the message is clear and understandable and the continuous media coverage is guaranteed.
Interestingly, this also points to two of the major weaknesses in the communication of political policies of rather moderate, mainstream parties or contenders: (1) the difficulty to narrow their policies down to a crisp, distinguishable and comprehensible message and (2) the subsequent problem to get media coverage and reach a substantial part of the electorate.
Donald Trump has managed to incentivise millions of Americans to vote in the primaries. Also in Germany, voter turnout in recent state elections went up to 70%. Greater voter participation in elections is good news for democracy. However, the current media structure seems to favor simple messages (who could blame them?) and thus populist movements. Moderate parties fail to adopt their message to this new reality and risk to fall behind. How could the catch up again without changing their policy core?
Data Science as a mitigator of populism?
As attention remains difficult to grasp with moderate political messaging, other strategies need to be developed to counter the rise of populism – and data science could be its facilitator. Indeed, data science is already well established in US politics. In this year’s primaries, Ted Cruz managed to pull off a surprise victory against Donald Trump in Iowa and remained as his last challenger in the race until early May. His campaign was widely recognized as the most rigorous, data-driven campaign in the field.
In Europe, though, data-driven strategies are not that widely spread. On the contrary, there is a general reluctance in party headquarters to utilize data. However, in the absence of a populist message, complex and differentiated political messages call for communication strategies of the same kind. What does that mean in practice?
Identify swing voters: Populist movements are able to recruit voters from across the spectrum. For established parties, in order to maintain their regular voter base, it is critical to identify potential swing voters and consequently approach them with a tailored messaging strategy. In Europe, with the absence of large voter databases, this could be done through proxies such as income and other socio-economic factors. Pulling the data together, mainstream parties should be able to identify those societal groups, where they are most likely to lose votes. Segmentation is a key exercise and data science allows parties to do so in a fine-grained manner, not just trusting intuition and experience.
Identify new likely voters: While populist parties / candidates, speaking to frustration, outrage and opposition, motivate a lot of people to go to the ballot, this is a more difficult endeavor for established moderate parties. Nevertheless, identification of new likely voters is a crucial task for established parties. Modern technology and methods of data science help parties to test their messaging, identify target groups, iterate fast and tap into potential. In times where a large part of the electorate feels left out, parties should focus on targeting those people, clearly pointing out what has been done and will be done for them, making them feel included. This approach should be utilized across all channels, online and offline. Around their core message, those parties should be flexible enough to change focus, depending on the respective target group. As reaching the broad electorate should be in their general interest anyway, combining breadth with customization will allow them to mitigate the continuous erosion of their voter base.
Certainly, data science will not stop a party’s general decline in popularity if their messaging just does not resonate with the electorate anymore. However, even in those cases, data-driven methods will allow parties to detect this mismatch between their policies and societal preferences earlier than the next election results come in. Indeed, in the medium-term, data science could contribute to establish tighter feedback loops between societal sentiments and opinions and political policy. This will not prevent populist messages to resonate with the electorate, but helps established forces to effectively speak to those sentiments beforehand with policy solutions.