Internet Access & Political Movements: Measuring Protests Globally

• In Event: Digital Authoritarianism

Abstract

Does an increase in internet access (penetration) lead to more anti-government protests around the world, in both democracies and non-democracies? In a meta-study covering more than 300 studies published worldwide in the past twenty years, Boulianne (2015; 2020) reports that the general finding is that the internet and social media platforms increase participation in civic and political life. But much of the previous research has focused on democracies. We know much less about the effects of digital information on democratic protests in non-democracies.

The past three decades witnessed a dramatic expansion in internet use for news and communication around the world. In 2019 more than 56 percent of the global population used the internet, according to the World Bank. The number of protests has also been rising since the early 2000s in all world regions in countries with all levels of income (Ortiz et al. 2022, 14). During Arab Spring, the internet and social media played an important role in enabling and organizing protests (Howard and Hussain 2013). More recently, social media helped to upscale the Black Lives Matter movement (Mundt, Ross, and Burnett 2018), and messaging app Telegram helped to facilitate mass mobilization during protests in Belarus (Mateo 2022). The literature emphasizes digital media’s potential for enabling social movements and increasing the number of democratic protests (Manacorda & Tesei, 2020; Zhuravskaya et al., 2020; Anderson, 20) by lowering the barriers for collective action, facilitating the spread of information, and decreasing coordination costs (Boulianne, 2020; Enikolopov et al., 2020; Shirky, 2009).

The literature shows that the impact of internet use on protests varies based on the political regime, and the results are conflicting. In autocracies, for example, Weidmann & Rød (2019) find that places with higher internet use are less likely to have protests. However, if protests do erupt, higher rates of internet penetration make it easier to sustain the ongoing protests. Using cross-national time-series data from 1990-2013 and the Freedom House to measure regime types, Ruijgrok (2017) finds that an increase in internet use increases the expected number of protests in autocracies but not democracies. We build on Ruijgrok’s work and investigate the relationship between the changes in internet penetration and the number of anti-government protests globally using an expanded time frame and causal inference designs.

We make several improvements to the previous work. This study extends the time frame for analysis from 1990 to 2020, an important period of political unrest. Mass protests are measured using the Mass Mobilization dataset that covers the anti-government protests in more than 150 countries. As a measure of internet penetration, we use the percent of the population with internet access from the World Telecommunication/ICT Indicators Database. We alternatively measure access to mobile and fixed broadband on the occurrence of anti-government protests. To analyze the relationship and uncover causal links between internet use and the number of protests, we use time-series cross-national data, fixed effects, and negative binomial regressions to model protest data in more than 150 countries from 1990 to 2020. Alternatively, counts of the number of protests are divided by the population and logged, using regression analysis. Rather than the Freedom House, which has been shown to be more qualitative, we use the electoral democracies index from the Varieties of Democracies (V-dem) project.

Estimating the causal effects of internet access is difficult because it cannot be randomly assigned across countries. Countries with and without widespread internet access are likely to have political, social, and economic pre-existing differences, making it difficult to draw causal inferences by a simple comparison. This study overcomes these challenges by leveraging sharp year-to-year changes in internet penetration at the country level. We compare country-years when internet access expanded rapidly over a short time to country-years when internet access expanded more gradually. Country years that experienced a jump in internet penetration of plus two standard deviations from year to year are considered treated cases (86 countries) compared to countries that had a more gradual explanation of internet access (67 countries). We replicate our models of internet penetration interacted by regime type (V-dem scores) for subsamples of treated and non-treated countries.

Using this causal inference design and expanded time series analysis, the results corroborate the finding from Ruijgrok (2017) and show that an increase in the internet penetration rates increases the number of protests in non-democracies. The number of protests in democracies remains unaffected by the changes in the internet penetration rates.

Authors

• Daria Kuznetsova

• Caroline J. Tolbert, University of Iowa