Contagious protests

Abstract

This paper explores the spillover of protests across countries using data on nonviolent and spontaneous demonstrations for 200 countries from 2000 to 2020. Using an autoregressive spatial model, the analysis finds strong evidence of “contagious protests,” with a catalyzing role of social media. In particular, social media penetration in the source and destination of protests leads to protest spillovers between countries. There is evidence of parallel learning between streets of nations alongside the already documented learning between governments.

Appendices

Appendix A

See Table 

Table 12 Panel A: Countries included in the ACLED dataset with at least 1 protest

12

Panel B: Countries included in the news-based protest dataset

Afghanistan, Albania, Algeria, Andorra, Angola, Anguilla, Antigua and Barbuda, Argentina, Armenia, Aruba, Australia, Austria, Azerbaijan, Bahamas, Bahrain, Bangladesh, Barbados, Belarus, Belgium, Belize, Benin, Bermuda, Bhutan, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, British Virgin Islands, Brunei, Bulgaria, Burkina Faso, Burundi, Cambodia,Cameroon, Canada, Cape Verde, Cayman Islands, Central African Republic, Chad, Chile, China, Colombia, Comoros, Congo Republic, Costa Rica, Cote d'Ivoire, Croatia, Cuba, Cyprus, Czech Republic, Democratic Republic of the Congo, Denmark, Djibouti, Dominica, Dominican Republic, Ecuador, Egypt, El Salvador, Equatorial Guinea, Eritrea, Estonia, Eswatini, Ethiopia, Faroe Islands, Fiji, Finland, France, French Polynesia, Gabon, Gambia, Georgia, Germany, Ghana, Gibraltar, Greece, Greenland, Grenada, Guatemala, Guinea-Bissau, Guinea, Guyana, Haiti, Honduras, Hong Kong, Hungary, Iceland, India, Indonesia, Iran, Iraq, Ireland, Israel, Italy, Jamaica, Japan, Jordan, Kazakhstan, Kenya, Kiribati, Kuwait, Kyrgyzstan, Laos, Latvia, Lebanon, Lesotho, Liberia, Libya, Lithuania, Luxembourg, Macau, Madagascar, Malawi, Malaysia, Maldives, Mali, Malta, Marshall Islands, Mauritania, Mauritius, Mexico, Micronesia, Moldova, Mongolia, Montserrat, Morocco, Mozambique, Myanmar, Namibia, Nauru, Nepal, Netherlands, New Caledonia, New Zealand, Nicaragua, Niger, Nigeria, North Korea, North Macedonia, Northern Mariana Islands, Norway, Oman, Pakistan, Palau, Palestine, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Poland, Portugal, Puerto Rico, Qatar, Romania, Russia, Rwanda, Saint Lucia, Samoa, San Marino, Sao Tome and Principe, Saudi Arabia, Senegal, Seychelles, Sierra Leone, Singapore, Slovakia, Slovenia, Solomon Islands, Somalia, South Africa, South Korea, Spain, Sri Lanka, St. Helena, St. Kitts and Nevis, St. Vincent and the Grenadines, Sudan, Suriname, Sweden, Switzerland, Syria, Taiwan, Tajikistan, Tanzania, Thailand, Timor-Leste, Togo, Tonga, Trinidad and Tobago, Tunisia, Turkey, Turkmenistan, Turks and Caicos Islands, Tuvalu, Uganda, Ukraine, United Arab Emirates, United Kingdom, United States, Uruguay, Uzbekistan, Vanuatu, Venezuela, Vietnam, Yemen, Zambia, Zimbabwe.

Appendix B

See Table 

Table 13 Summary statistics

13

Appendix C Contagion and Distance

See Table 

Table 14 Panel A: Actual protests using reduced sample

14,

Table 15 Panel B: News coverage using reduced sample

15

Appendix D: Contagion, distance and social media—2010 facebook penetration

See Tables 

Table 16 Panel A: contagion, distance, and social media—Actual protests

16,

Table 17 Panel B: Contagion, distance and social media—News based protests

17

Appendix E: Contagion, distance, and social media

See Tables 

Table 18 Panel A: Actual protests–including countries with social media penetration below thresholds

18,

Table 19 Panel B: News based protests—including countries with social media penetration below thresholds

19

Appendix F

See Fig. 

Fig. 4

Distribution of Protests

4

Appendix G

See Fig. 

Fig. 5

Market Shares of Facebook. Panel A: Facebook versus other Social Media Platforms. Panel B: Facebook’s Market Share in Different Regions. Source: statcounter.com [https://gs.statcounter.com/social-media-stats#monthly-200903-202005]

5

Appendix H

See Fig. 

Fig. 6

Facebook Penetrations and Protests Panel A: Average Number of Protests per Million Persons Per Month. Source: ACLED. Note: Include countries with at least 1 protest. Data coverage starts from January 2010. Panel B: Facebook Penetration (% of Population, December 2018). Source: www.napoleoncat.com

6

Appendix I: A simple model of contagious protests

The model is inspired by Edmond (2013) who studies informational drivers of protests including social media. The model presented here is a simplified two-country perspective. We assume two symmetric neighboring countries, namely country 1 and country 2. The focus of our model is on the spillover effect of protests from country 1 to country 2.

In each country, each citizen decides whether to protest against the domestic regime (${s}_{i}=1$ if she participates in protest; ${s}_{i}=0$ if not. The aggregate protest is as follows: $S={\int }_{0}^{1}{s}_{i}{d}_{i}$.

The regimes in the two neighboring countries have the same strength: ${\theta }_{1}={\theta }_{2}$. The respective strength is private information for the two regimes. The true strength is not known to citizens. A regime is overthrown if enough people in the respective country protest: $\theta <S$.

A citizen’s payoff is as follows:

• In the case where a given individual does not protest, the individual payoff is: $u\left({s}_{i},S,\theta \right)=0$

• In the case where a given individual protests, the individual payoff is $u\left({s}_{i},S,\theta \right)=Prob\left(\theta <S\right)-p$, where $p$ is the individual cost of protesting.

A closed-economy equilibrium

In the case of the closed economy equilibrium described in Edmond (2013), the information structure is as follows. Each citizen observes a signal of inflated regime strength ($\theta +a$) with noise: ${x}_{i}=\theta +a+{\varepsilon }_{i}$, where $\theta +a$ is the inflated strength advertised by the regime. Advertising an inflated strength is costly to the regime. $\theta$ is the true strength that citizens do not observe.

Solving for the global game, Edmond (2013) finds that there exists a unique equilibrium strength ${\theta }^{*}$ such that the regime is overthrown if $\theta <{\theta }^{*}$, in which there is a unique threshold ${x}^{*}$ such that a citizen participates in protest when${x}_{i}<{x}^{*}$. Using Edmond’s framework, one can interpret the trigger for protests and regime change could include either a decline in $\theta$ or a decline in the cost of protesting,$p$.

A two-country equilibrium

In our simplified two-country framework, suppose protests break out in country 1 (a decline in the cost of protesting, $p$, for example). The new feature introduced here is that there are learning spillovers between countries stemming from individuals’ participation in protests.

Protesting individuals in country 1 update their beliefs about the true strength of the regime. Each participant receives a new signal of regime strength that is closer to the true strength after she or he protests:${x}_{1i}={\theta }_{1}+{\varepsilon }_{1i}$. Note that $a$ becomes zero after the learning from the protests took place. In other words, protesters realize the regime strength is weaker than it was advertised by the regime.

If social media is available in both countries, individuals in country 1 can diffuse their updated signal about the regime strength. In equilibrium, a fraction $\gamma$ of country 1’s population protest. Suppose a fraction ${\delta }_{1}$ of these protesters participate in social media and diffuse via social media the new signal they observe with some noise ${x}_{1i}={\theta }_{1}+{\varepsilon }_{i}$.

Assume that a fraction ${\delta }_{2}$ of country 2 participates in social media. Hence, a fraction $min({\gamma \delta }_{1},{\delta }_{2}$) of country 2’s population can update their signal about the regime true strength ${x}_{i2}={\theta }_{2}+{\varepsilon }_{i}$ (note that ${\theta }_{1}={\theta }_{2}$). The remainder of population of country 2 without social media access does not update their signal. For them, the signal remains the same ${x}_{2i}={\theta }_{2}+a+{\varepsilon }_{i}$.

We have a new global game where the population of country 2 receives different information structure about the regime strengths. One portion receives ${x}_{i2}={\theta }_{2}+{\varepsilon }_{i}$ (updated via social media), and the remaining portion receives ${x}_{2i}={\theta }_{2}+a+{\varepsilon }_{i}$.

From Edmond (2013)’s main result, we can see that the portion receiving updates via social media will be more likely to protest after receiving their updates about the regime’s true strength. Therefore, the new equilibrium${\theta }_{2}^{*}$, the threshold below which the regime is overthrown (that is $\theta <{\theta }_{2}^{*}$), will be smaller than that in the closed-economy equilibrium. This explains why waves of regime changes follow contagious protests amongst regional neighbors.

A prediction from the simple framework we will test in our empirical analysis is that social media penetration (in both origin and destination countries of protests) can encourage more participation in protests in a neighboring country (country 2) to the country where protests initially take place (country 1).