The Volatility of Politics in a Digital Age: Opinion Attention Trends Over Time in Britain and Germany

It is now conventional wisdom that politics has become ever more fast-moving and unstable, leading to electoral shocks and even regime changes. But there is not much systematic evidence that modern politics has become more volatile. This paper seeks to address this gap by analysing public opinion and media data over the last twenty-five years in the United Kingdom (which has experienced considerable electoral change and large fluctuations in support across parties) and Germany (which appears to exhibit greater stability). The paper seeks to answer whether increased volatility can be explained by greater use of social media. The argument is that social media platforms inject instability and volatility into social and political life. They exert social influence on users by showing in real time what other people are doing (social information), creating feedback influences that have been shown to introduce instability into cultural markets [1], and the authors have started to identify this pattern in US and UK politics labelled as "chaotic pluralism"[2]. This disorganized environment may exhibit known features of natural complex systems, such as 'self-organized criticality', characterized by fat-tailed distributions of system parameters rather than distributions approximating ‘normal’, and system properties [5] of volatility. Instability manifests, for example, in highly unequal distribution of participants across mobilizations, exhibiting ‘leptokurtism’ as identified by Baumgartner and Jones [3] for various types of policy change in the US Policy Agendas project. But in contrast to their 'punctuated equilibria' model the 'chaotic pluralism' model predicts that punctuations will be continual, the new normal, while stasis is the exception. We test this theoretical model with empirical data and data science methods to establish whether leptokurtic distributions on social media are translating into political volatility. We systematically track attitudinal surveys (such as responses to 'what is the most important issue facing the country today?' asked by IPSOS-Mori) and traditional (news)media articles in both countries, before and after social media became widely used, from the mid-1990s to the present. We propose to measure volatility using information theoretic approaches including entropy and K-L divergence [4], which we connect with discontinuities in volatility patterns in attitudinal data and indicators of 'pull' factors such as petition signing, focusing on time periods before and after use of social media became widespread. We correlate measures of turbulence with measures of social media use and other predictors in a time series model. [1] Salganik, M., Dodds, P., Watts, D. (2006). ‘Experimental study of inequality and unpredictability in an artificial cultural market.’ Science 311.5762: 854-856. [2] Margetts, H., Hale, S., John, P. & Yasseri, T. (2016). Political Turbulence: How Social Media Shape Collective Action. Princeton University Press. [3] Baumgartner, F., and Jones, B. (2010) Agendas and instability in American politics. University of Chicago Press. [4] DeDeo, S., Hawkins, R., Klingenstein, S., and Hitchcock, T. (2013). Bootstrap Methods for the Empirical Study of Decision-Making and Information Flows in Social Systems. Entropy, 15(6), 2246-2276; doi:10.3390/e15062246 [5] Bak, P. and Chen, K., 1991. Self-organized criticality. Scientific American, 264(1), pp.46-53.