Censorship and Propaganda

To stave off mass rebellion, leaders focus their efforts on eliminating independent criticism that might spread dissent, while also currying favorable coverage that helps build popular support for the regime and its policies (Geddes & Zaller, 1989). Different types of dictatorship have different objectives in mind for controlling information. Totalitarian regimes seek monopoly control over mass communication to mobilize the citizenry behind the totalitarian ideology, be it fascist or communist (Linz, 2000). Personalist and monarchic authoritarian regimes may use the media and propaganda to construct a cult of personality around the leader or royal family, whereas military regimes likely focus on eliminating negative coverage that could threaten regime stability.

In Nazi Germany, Adolf Hitler and his propaganda minister Joseph Goebbels designed the massive propaganda machine with the objective to sell Nazi ideology to the German public, as if it were a brand (O’Shaughnessy, 2009). Other totalitarian regimes, such as the communist governments of Lenin’s Soviet Union and Mao’s China, relied on the careful regulation of information and government propaganda to develop new cultural identities and mobilize the masses behind that identity. In order to achieve this degree of control over messaging, Vladimir Lenin eliminated media competition within the first year of seizing power (Kenez, 1985).

In contrast, bureaucratic authoritarian regimes in South America wanted to depoliticize citizens and encourage mass complacency (O’Donnell, 1978). Bureaucratic authoritarian regimes dominated in the Southern Cone region of Latin America during the 1960s through 1980s. Led by professionalized militaries, they brought in technocrats, often trained abroad, to promote economic development directed at reducing these countries’ economic dependence on commodity exports and imported manufactured goods. To carry out these policies, which involved tactics that hurt the popular sector, the bureaucratic authoritarian governments’ primary objectives were to build public support for regime policies and thwart anti-regime collective actions (Remmer & Merkx, 1982). To do so, the regimes censored the media to ensure content that would prime citizens to evaluate the regimes based on their successes rather than their antidemocratic behaviors (e.g., triumphs over communist insurgencies vs. government responsibility for human rights violations) (Stein, 2013). Each country’s bureaucratic authoritarian regime employed different strategies for controlling publicly available information. Whereas Chile primarily relied on massive intimidation and the resulting self-censorship (Rosenberg, 1987), Brazil implemented systematic and formalized media censorship, issuing daily prohibitions to all media outlets and reserving the more punitive tactic of prior censorship for the opposition press (Marconi, 1980; Kushnir, 2004).

Intra-Elite Communication

Via censorship, repression, and a tight inner circle, dictators enshrine their regimes with secrecy in an effort to maintain an advantage over the opposition and potential rivals among the elite (Boix & Svolik, 2013). While government secrecy may create some opportunities for the leader, who is the most knowledgeable about his government, it also creates obstacles to cooperation, even between the leader and the group of elites who support him. According to several studies on dictatorships, the lack of reliable information and the asymmetrical distribution of it impede dictators’ ability to make credible commitments. These authors suggest that regimes adopt a variety of pseudo-democratic institutions, from semiautonomous legislatures to multiparty elections, to ensure that autocrats will not renege on their power- and resource-sharing commitments (e.g., Magaloni, 2008; Gehlbach & Keefer, 2011; Boix & Svolik, 2013). By formally ceding decision-making authority to other elite actors and offering paths for career advancement, the leader’s assertion gains credibility.

Institutionalization of authoritarian regimes also alters the strategies that authoritarian leaders employ against the opposition. Specifically, adopting institutions such as legislatures, political parties, and elections helps authoritarian leaders co-opt and negotiate with members of the opposition (Gandhi & Przeworski, 2006). However, incorporating these institutions simultaneously opens space for potentially public and destabilizing political contention (Frantz & Kendall-Taylor, 2014). While institutions may stave off most elite-driven transitions, regimes must employ other measures to curtail bottom-up efforts to pressure the regime for concessions or force regime change. Many authoritarian regimes utilize repression as a communicative tool to dissuade the masses from joining rebellions.

Repression as a Communicative Tool

Most research on collective action rests on the assumption that repression deters individual participation by signaling its high cost and thus impedes collective action. Repression, when coercive and observable carries with it a message that the regime will not tolerate dissent (Earl, 2003). Employing repression, while potentially persuasive in deterring participation in the short term, carries with it high costs for the regime. Apart from the direct costs, such as paying censors and soldiers, the outright use of physical repression may cost the regime its legitimacy internationally and domestically. A less obvious cost of repression stems from how it affects public expression and, in turn, the quality of information regimes can gather on their popular support, which they need to better anticipate and appease or suppress dissenters (Bratton & van de Walle, 1997). Repression likely leads citizens to self-censor their publicly shared opinions or to falsify their preferences. Ronald Wintrobe’s economic model of dictatorship describes the difficulty regimes have in gauging their true level of public support as the “Dictator’s Dilemma,” which Wintrobe suggests leads most dictators to employ both repression and co-optation (1998).

In the end, misinformed leaders frequently take actions that prove tremendously costly to them. For example, authors have attributed abrupt revolutions that bring down leaders and regimes to events that trigger some citizens to end preference falsification, which produces a bandwagon of protesters as more people express their true preferences (e.g., Iran in 1979 [Rasler, 1996], Eastern Europe in 1989 [Kuran, 1991; Lohmann, 1994], Tunisia in 2010 [Goodwin, 2011; Hussain & Howard, 2013]), and surprising defeats at the polls, as in the case of Chilean General Augusto Pinochet’s loss in the 1988 plebiscite and the forced departure of leaders of Eastern Europe and former Soviet states in the wake of post-electoral protests (e.g., Vladimír Mečiar’s government in Slovakia in 1998, Georgia’s then-President Eduard Shevardnadze in 2003, and Viktor Yanukovych in Ukraine in 2004‒2005) (Tucker, 2007; Bunce & Wolchik, 2010). In search of more enlightening data, dictators have turned to a number of counterintuitive mechanisms to elicit information, including elections, freedom of expression, and even permitting protests.

Elections as Information in Authoritarian Regimes

Elections in authoritarian regimes serve as more than just a credible commitment to share power with other elites.⁷ Many scholars attribute their adoption to the veneer of democracy that they offer a regime that might need to curry favor internationally (Calingaert, 2006; Hyde, 2011). Elections also allow the dictator to communicate publicly the strength of his support or at least his capacity to mobilize “supporters” to deter would-be challengers from plotting coups or discourage opposition groups from staging uprisings (Geddes, 2006; Howard & Roessler, 2006; Blaydes, 2011; Malesky & Shuler, 2011). Authoritarian elections represent one of many mechanisms that dictators adopt to gain information about popular discontent and identify opposition supporters. However, dictators face a trade-off between gathering good information, in which case they need to permit more fair elections that offer the opposition the possible chance of victory, and maintaining greater control over electoral outcomes, in which case dictators sacrifice some of the informational value of competitive elections (Malesky & Shuler, 2011). As Andreas Schedler notes, incumbents “want to bring opposition parties into the game but they also wish to keep them under control. After all, they are not interested in institutionalizing democracy, but in legitimizing their continuity in power” (2002, p. 114).

While some scholars argue that elections facilitate opposition coordination of mass actions via formalized political parties (Magaloni, 2010), others have suggested that such tactics help sustain the regime by dividing the opposition (Haugbølle & Cavatorta, 2011). Where the formal rules allow some contention, they encourage a portion of the opposition to play by the rules and push for change from within; alternatively, more radical opposition forces reject participation in authoritarian elections, preferring to express dissent through collective action or antigovernment violence.

Even fraudulent elections can prove informative. In noncompetitive elections, regimes invite election observers and subsequently engage in fraud to fulfill observers’ expectations and demonstrate their strength, while also lowering the cost of committing higher levels of fraud if necessary (Little, 2012). In multiparty elections, the regime can evaluate whether they needed to employ higher than expected efforts to manipulate the outcome. However, the use of fraud to ensure electoral victories has proven costly for many reigning incumbents because elections arguably help mobilize the opposition and activate latent participants (Tucker, 2007), which I discuss in greater detail in the subsequent section.

Other Information Tools

Beyond holding elections, some authoritarian regimes also have employed other counterintuitive information-gathering devises, such as permitting moderate levels of popular protest (Lorentzen, 2013) and allowing (adequate) freedom of the press (Egorov et al., 2009; Stockmann & Gallagher, 2011; Lorentzen, 2014). Protests can serve as “burglar alarms,”⁸ notifying the regime of growing dissent, allowing it to target particular groups with ameliorative actions to avert broadscale anti-regime campaigns. In addition to lending legitimacy to the regime, promoting media freedom and independence allows for investigative journalism that can help the regime monitor lower-level government officials who might engage in corruption (Egorov et al., 2009; Stockmann & Gallagher, 2011). Thus media freedom helps authoritarian rulers to improve governance and, in turn, reduce popular dissatisfaction. Just as staging elections generates information for the regime, permitting controlled protest, allowing freedom of the press, and easing restrictions on online expression also can inform authoritarian rulers about the nature of its opposition. Using data from content analyses of online discussion boards that captured user commentary prior to government censorship, Gary King and his coauthors show that the Chinese government permits selective criticism of the regime online, but quickly erases any comments that could help mobilize protests (King et al., 2013).

One should note that many of these studies on permissive freedoms of the press, expression, and assembly (e.g., Stockmann & Gallagher, 2011; King et al., 2013; Lorentzen, 2013, 2014) focus on China, a country with greater state capacity than most other regimes. Just as the literature on authoritarian elections highlights the risk of actual defeat (e.g., Howard & Roessler, 2006; Bunce & Wolchik, 2010), authoritarian governments that permit protests and free media also elevate the risk of generating common knowledge about the extent of popular dissatisfaction that could spiral quickly toward broader revolutionary action. Some authors have suggested that even in China, increasing space for online expression ultimately has empowered the country’s netizens and decreased the government’s ability to set the agenda (Esarey & Qiang, 2011). These informational devices might prove more costly than beneficial for authoritarian regimes that do not have the capacity to regulate them or shut them down should they grow increasingly contentious. The subject merits further empirical exploration across authoritarian regimes. State capacity also likely influences dictators’ approach to investing or not in ICT infrastructure, as well.

Dictatorship in the Information Age

In light of events from post-election protests in Ukraine, Moldova, and Iran to the massive mobilizations in Tunisia, Egypt, and other Middle Eastern and North African countries, the potential outcomes of expanded Internet access, either via computers or mobile phones, have attracted attention from governments, media, and scholars alike. Early research on the Internet in authoritarian regimes (e.g., Kalathil 2003; Kalathil & Boas 2003) illustrated that authoritarian regimes embraced a broad range of approaches to ICT adoption. Some countries, including Cuba, North Korea, and Burma, kept a strong grip on controlling access, investing little, and permitting only government insiders and some academics to have access to the World Wide Web (Kalathil 2003). Other countries, such as China and Egypt, invested considerably in providing Internet access to citizens, though China also invested in developing controls on Internet content, developing what has become known as China’s Great Firewall (MacKinnon, 2011).

More recent research on ICT access and use suggests that dictators face a double-edged sword: they must decide if they should accept the risk of ICTs (e.g., increased protest and information sharing), for the expected economic rewards (e.g., economic growth). Espen Geelmuyden Rød and Nils Weidmann note that “the Internet is not imposed on a particular country from the outside; rather its introduction relies critically on the permission and support of the domestic government” (2015, p. 4).

Since the commercialization of the Internet in the late 1990s, the Internet has become increasingly integrated into the world economy. To thrive economically, countries that lack natural resource wealth, require a wired society. Studies have shown that countries with better connected societies—those with higher Internet penetration rates (the number of users per 100 people)—enjoy higher economic growth, on average (Milner, 2006; Vu, 2011).⁹ Authoritarian leaders who govern during economic decline or stagnation—like their democratic counterparts—suffer from lowered rates of public approval. Thus, if they fear investing in ICT infrastructure due to the threats these technologies may pose, they also put their economic performance in jeopardy. For dictators, constraining the expansion of ICT access could potentially lead to a loss of performance legitimacy.

Data suggest that dictators and their regimes prefer to err on the side of fewer protests, instead taking risks with regard to economic performance. This falls in line with literature that suggests that some dictators have short time horizons and highly discount future gains (Wright 2008b). Table 2 compares data on average Internet penetration rates between autocracies and democracies for the period between 1995 and 2014; on average, authoritarian regimes had lower Internet penetration rates than did similarly developed democratic regimes, across all income categories.¹⁰ Figures 1‒7 illustrate the difference in how Internet access has expanded over time in democracies versus autocracies given a countries average per capita income. On average authoritarian regimes delayed investing in the Internet; although at higher levels of development—where democracies have reached near-universal access—autocratic regimes have narrowed this gap in recent years. In my estimation, many of these regimes have shifted their strategies from trying to limit access to the Internet toward focusing on restricting available content. While these summary data do not establish a direct causal path from dictators’ calculus of the risk versus reward of investing in ICT infrastructure, the data align with prior studies that suggest that Internet penetration rates are endogenous to regimes, with authoritarian regimes resisting expansion more so than democracies (Milner, 2006).

Open in new tab

Figures 1–7. The growth of Internet access in democracies versus autocracies, 1995–2014, controlling for income categories.

Note. As in Table 2, Internet Access (Internet users per 100) and income categories (based on GDP per capita_ppp) rely on data from the World Bank World Development Indicators (accessed in April 2016). Regime type comes from Autocracies of the World, “demo_r” variable (Magaloni et al., 2013), which I have updated for 2013 and 2014. The lines in these graphs are derived via lowess smoothing and fill in gaps for missing data.

Wealthier authoritarian regimes and those with high state capacities not only seem better able to tolerate the risk of expanding Internet access, but they also can capitalize on informative benefits for the regime. Some authoritarian states use the Internet as a new apparatus for monitoring citizens and identifying dissenters. In the Middle East and North Africa, for example, countries like Saudi Arabia and the United Arab Emirates that have high-tech, well-financed tools for monitoring information technologies and sanctioning online dissent better withstood contentious episodes associated with the Arab Spring relative to countries such as Egypt and Jordan that tolerated (or were unable to limit) online dissent (Hussain & Howard, 2013).

Following on Georgy Egorov and his coauthors’ work (2009) that suggests that authoritarian states without oil permit media freedom to help monitor officials and improve good governance, I broke down average Internet penetration rates for authoritarian regimes, comparing countries that derive more than one-fifth of their GDP from oil rents. Table 3 supports the idea that petro states are less permissive with Internet access than those that may rely on them to improve principal‒agent problems. Again, controlling for income groups, for all but one income category, which has very few observations, petro states have lower average Internet penetration rates than do countries that derive a smaller percentage of their GDP from oil rents. An alternative explanation that might account for lower ICT penetration rates in petro states aligns with Joseph Wright’s (2008a) explanation for why non-oil-dependent authoritarian regimes more likely adopt legislatures; because they are “more dependent on the production of the domestic economy to generate the revenue necessary to sustain their rule, they have an incentive to establish governing institutions that constrain their own power” (Wright, 2008a, p. 322). This logic could carry over to explain why non-oil-dependent states look more favorably upon Internet access than petro states, which I define as any states that derive 20% or more of their GDP from oil rents; the Internet has the potential to serve as a constraining agent.

Even authoritarian regimes with high state capacity or an abundance of natural resources remain imperfectly informed, particularly with regard to underlying mass support (Kuran, 1991; Ginkel & Smith, 1999), and these regimes cannot completely control information. Nevertheless, dictators generally have more and elaborates on the struggles dissident leaders face with regard to information acquisition and diffusion when mounting antigovernment campaigns.

Organizing Protest amid Repression

Nearly all theories of mobilization based on rational actors presume that repression suppresses individuals’ willingness to participate in mass actions absent additional incentives (e.g., DeNardo, 1985; Lichbach, 1987, 1994a, 1994b; Tucker, 2007). Political opportunity theory, for example, suggests that changes in repression play a role in leaders’ decision of whether or not to hold a protest, and if one is occurring whether or not more individuals will join the protest (Opp & Roehl, 1990; Brockett, 1993, 2005; Kurzman, 1996). Activists not only consider the probability that the planned action will achieve gains but also the likelihood that anyone who chooses to participate could be arrested or beaten as a result. The logical implication is that leaders of opposition movements in authoritarian regimes must seek moments of reduced risk to attract participants. In order to do so, leaders implicitly must be able to gauge the degree of risk, presumably by monitoring government repression in other arenas, and then communicate this information to a broader audience without attracting the attention of the regime (Stein, 2016).

An extensive literature exists about the relationship of repression and protest,¹² yet little consensus has emerged about the effectiveness of repression as a communicative tool for the regime in deterring protests. The breadth of hypotheses regarding the effect of repression on protest presented in this body of research runs from predicting monotonically positive or monotonically negative effects to an inverted-U pattern, or countervailing effects of higher levels of repression on the incidence of collective action. At best, one could conclude that repression affects mobilization both positively and negatively depending on the (a) intensity of repression (DeNardo, 1985; Lichbach, 1987, 1994a; Muller & Weede, 1990), (b) target of repression (Mason & Krane, 1989; Siegel, 2011), (c) timing of repression (Brockett, 2005), and (d) consistency in the application of repression versus concessions (Lichbach, 1987; Rasler, 1996). Other studies point out that findings of a positive correlation between protest and repression might be an artifact of observation (Shadmehr & Bernhardt, 2011; Ritter & Conrad, 2016). Where repression is high enough to deter protests, no protests occur (and thus no incidental repression follows). By contrast, protests might take place in countries that employ more moderate levels of protest; hence, we observe more protests and more repression.

One factor accounting for these differences stems from variations in the conceptualization and operationalization of repression, and thus what it means for a country or for protesters to experience more or higher levels of repression. Some authors focus on types of repression, distinguishing between repression targeted at individuals versus repression targeted at society (e.g., Rasler, 1996; Frantz & Kendall Taylor, 2014), while others highlight the government’s strategy of selective versus random repression (Mason & Krane, 1989). Studies distinguish between the degree of violence involved, for example, distinguishing between arrests and deaths (Khawaja, 1993), or the incident rate of repression (Francisco, 1993).

Despite the large number of studies that suggest that repressive environments invariably reduce individuals’ willingness to show up for mass actions, empirical evidence repeatedly shows that under many circumstances more repression leads to more protest or has no statistically significant relationship. These results suggest that scholars of collective action in authoritarian regimes, particularly those who employ rational choice, may need to revisit their underlying assumptions. One alternative perspective employs prospect theory, arguing that citizens are more inclined to participate in risky actions when they perceive the status quo or negotiated agreements as a loss in comparison with their objective outcome; members of the opposition who perceive government offers of accommodation as an improvement to the status quo more likely will refrain from costly actions with uncertain outcomes (Masters, 2004). This indicates that the framing of dissident causes as well as their planned mass actions could have a large effect on mass participation as well.¹³

Signals Sent, Signals Received

Many studies of collective action employ some form of signaling game to suggest how dissident leaders receive information about expected repression (e.g., Lohmann, 1993, 1994; Ginkel & Smith, 1999; Bueno de Mesquita, 2010). Several authors indicate that dissident leaders observe signals from government (either privately or publicly), though authors do not fully explicate (a) what type of signals they receive, (b) from where or from whom, and (c) how they receive these signals.¹⁴

Building on prior evidence that the relationship between repression and protest is less than straightforward, a few authors suggest that the role of repression in provoking protest exists because of the costly signal these actions send to latent participants. Ginkel and Smith (1999) who incorporate three players in their game—the government, dissidents (i.e., opposition leaders), and the mob—suggest that participation should grow in repressive environments. They claim that opposition leaders who initiate and participate in early antigovernment actions under the threat of repression lend credibility to their cause; the mob will more likely join movements where leaders demonstrate their degree of commitment through the high risks they accept. Ethan Bueno de Mesquita (2010) suggests that when first-actors, who he refers to as “revolutionary entrepreneurs,” engage in violent actions they likely increase mobilization, though he contends that the growth in participation results not from the power of numbers per se, but from information the entrepreneurs intentionally communicate about antigovernment sentiment to manipulate other’s beliefs about the benefits of action.

Some studies suggest that participation grows from information about actual protests that cascades through society. As the information that others protested and to what effect becomes common knowledge, it influences subsequent participation sending new information to still-latent protesters who join later protests, and so on (Kuran, 1991; Lohmann 1994, 2000). These arguments neither specify how the information reaches new members of the public nor how they become common knowledge in limited information environments.

In other research (Stein, 2016), I suggest that the missing factor from many of these arguments may be the mass media even when censored, both as the indirect provider of signals about likely repression, as well as a generator of common knowledge.¹⁵ I suggest that dissident leaders monitor media content and the government’s reaction to changes in content¹⁶ to gauge the government’s tolerance for public dissent. Empirical evidence drawn from the period of political opening (1974‒1982) during the Brazilian military dictatorship indicates that if the media challenged the regime by publishing stories on taboo subjects without suffering consequences, dissidents took this as sign that the regime was more tolerant of dissent or unable to repress it and they would subsequently initiate more mass actions. If, however, the regime responded with the suppression of content or repression of journalists in the face of media challenges, dissidents held off on initiating new mass actions, perceiving the regime’s disposition to be unfavorable toward public displays of dissent. Furthermore when this information appeared in mainstream media, it generated common knowledge about participation and consequences that influenced participation in subsequent protests.¹⁷

Dissident leaders’ ability to assess the probability of repression and its likely effects on participation and movement success represents just one school of thought about dissident leaders’ role in mobilizing the masses. Just as literature on authoritarian leaders’ behavior has adapted to evolving types of hybrid authoritarian regimes, so too has the literature on antigovernment protest in such regimes. Specifically several studies have looked at how elections, electoral fraud, and voter mobilization help opposition leaders in competitive authoritarian regimes to organize antigovernment actions.

Mobilizing Participation in Electoral Authoritarian Regimes

The distinguishing feature of competitive authoritarian regimes from other types of authoritarian regimes is that they permit some degree of real contestation with the possibility, though low, of an incumbent defeat (Howard & Roessler, 2006). While the literature on institutionalized authoritarianism highlights the benefits of pseudo-democratic institutions for authoritarian survival, the inclusion of political parties and elections inherently facilitates opposition coordination by eliminating the need to overcome obstacles to coordination present during other times, such as deciding when to protest.

Debates among the opposition about whether or not to participate in elections that the opposition and public anticipate to be less than free or fair hinge on the message the opposition’s participation would send to the regime and the outside world. Some opposition leaders see participation as a way to express their dissent (and the breadth of it among the population); while detractors see it as sending the message that the people view the incumbent regime and its institutions as legitimate. Where the opposition can overcome this hurdle and coalesce around a single candidate, Marc Howard and Philip Roessler (2006) contend that this cooperation helps mobilize people to vote against the incumbent by providing “signals to the electorate that the incumbent is vulnerable to defeat,” increasing the possibility of a “liberalizing electoral outcome” (Howard & Roessler, 2006, p. 372). Valerie Bunce and Sharon Wolchik (2010) believe that the opposition’s ability to collaborate among themselves and with civil society groups informs the regime of the degree of the threat it faces, and thus pressures the regime to make concessions about electoral procedures, voter registration, and transparent vote counts, which can improve the opposition’s chances of challenging the status quo. As the normative value placed on democracy increases, competitive authoritarian regimes have adopted an increasingly standard set of strategies to lend legitimacy to elections that includes many of the above tactics along with inviting international election monitors (Hyde 2011).

Opposition leaders are not alone in trying to mobilize participation. Regimes that incorporate competitive institutions to improve their durability also must mobilize enough of the opposition to participate in the election to appease international election monitoring organizations that procedures were fair enough for the observers to sanction the results and, thus, force the domestic opposition to accept the results. Knowing this to be the case, the presence of international election monitors actually encourages the opposition to boycott the election, sending the message to domestic and international audiences that the institutions are rigged. Because not all opposition parties willingly participate in elections that they perceive to be rigged from the start (Beaulieu & Hyde, 2009), partial boycotts, which weaken the opposition by publicly demonstrating the opposition’s disunity, occur with some frequency (Lindberg, 2006).

Many opposition groups in electoral authoritarian regimes support the presence of international election monitors believing the threat that observers might publicly allege fraud should limit the regime’s ability to renege on results in the case of an opposition victory (Kelley, 2011). Anticipating fraud, many opposition groups preemptively organize post-election protests, as was the case of the Green Movement in Iran following Ahmadinejad’s pronounced electoral victory in 2009 (Morozov, 2009b).

Opposition leaders’ ability to capitalize on events such as fraudulent elections requires them to act strategically. Part of the effectiveness of the successful Color Revolutions, a series of post-electoral protests that forced the departures of several incumbent leaders in former Soviet states, stemmed from opposition leaders’ ability to frame events in ways to attract participation of key sectors, such as youth voters (Polese & Ó’Beacháin, 2011). Once messaging is set, dissident leaders face the next challenge of reaching their audience, who increasingly may be found online.

Do Dissident Leaders Actually Lead in the New Information Age?

Prior to Facebook and Twitter and even the popularization of mobile phones, in 1994 the not-yet-known Zapatista National Liberation Army (EZLN), better known as the Zapatistas, launched an attack against the Mexican government from the southern state of Chiapas. The EZLN leader’s then-savvy use of the Internet distinguished the Zapatistas from similar insurgent movements. At the time, the majority of Internet and email use occurred through government and university networks. Subcomandante Marcos, the EZLN’s main leader—who the Mexican government believes to have been a university professor and thus would have had access to email and early online content—used the Internet to draw international attention to the indigenous community’s plight in Southern Mexico and to skirt the framing by traditional Mexican media (Schulz, 1998; Garrido & Halavais, 2003; Darling, 2008). The Zapatista leaders also used online resources to raise financial support for their campaign from sympathetic foreigners (Bob, 2005), perhaps representing one of the first incidents of crowdfunding online.

As access to the Internet in developing and non-democratic countries has grown exponentially, dissident group leaders’ perception of the utility of ICTs has evolved. Scholars claim that ICTs have two major functions: (1) they generate global awareness and (2) they mobilize domestic actors (Ritter & Trechsel, 2014; Little, 2015).

Critics of the so-called Twitter Revolutions point to the relatively low rates of diffusion of these technologies—particularly for foreign-based sites such as Twitter—among the populations in Iran, Tunisia, and Egypt (see, e.g., Morozov, 2009b; Gladwell, 2010). Other scholars challenge the quick dismissal of the importance of these technologies, highlighting the fact that those who could and did access these sites were likely “opinion leaders” (Kavanaugh et al., 2013). Andrea Kavanaugh and her coauthors contend that “opinion leaders (also called influentials) play a central role in finding and evaluating information from mass media and other sources, and sharing it actively with people in their social circles” (Kavanaugh et al., 2013, p. 4).

While people generally believe that with regard to collective action, a single person’s participation is unlikely to make the difference, under certain circumstances that might not be the case. As Timur Kuran’s behavioral cascade model suggests, depending on the underlying distribution of individuals’ threshold for participation, in some cases even small changes among a few citizens’ actions could trigger a domino-like effect. Keeping that in mind, for ICTs to affect rebellion they do not necessarily need to reach the majority of citizens. Rather, a few citizens who access them could trigger a mass rebellion if they spread the information via face-to-face communication or by becoming early participants whose actions could trigger a participatory chain reaction. The now well-known Egyptian Google executive Wael Ghonim may have been one of these influential actors when he set up a Facebook page to memorialize a murdered blogger, “We are All Khaled Said,” which served as an online arena for organizing initial protests early in the Egyptian uprising (Ritter & Trechsel, 2014). Citizen journalists and bloggers thus became the initial leaders of the early revolutionary movements (Ritter & Trechsel, 2014).

Not all Egyptians who participated personally would have accessed the Facebook page; many would have heard about it indirectly from friends or family or passed from cell phone to cell phone. The first participants in the Egyptian as well as Tunisian protests that triggered revolutions were part of the “youth bulge”—well-educated, yet unemployed and disillusioned youth—and were likely to have had higher rates of access to online social networks than subsequent participants (Howard & Hussain, 2011). The online activists may have been crucial to establishing the critical mass, after which traditional modes of mobilization took over (Ritter & Trechsel, 2014).

Furthermore, traditional media interact with and report on content in new media, amplifying the reach of independent news producers and commentators. To keep pace with technological development and commercial demands, news organizations have gone through a process of convergence in both the production and the distribution of news (García-Avilés, Kaltenbrunner, & Meier, 2014). Major media organizations often publish blogs by their columnists on the outlet’s website. Journalists who work for traditional news outlets may pick up stories from independent bloggers and social media outlets. Erin Snider and David Faris (2011) note the symbiotic relationship between a nascent independent press and online social media in promoting the events that ultimately led to the protests known as the Arab Spring.

However, do these individuals count as dissident leaders in the traditional sense? They may have initiated the spread of information, but they did not lead established interest groups, political parties, or social movements. Do the Internet and social networking tools eliminate the need for organizations and leaders? Scholars have noted that the new communications environment changes the obstacles to collective action, reducing the need for traditional organizations (Bimber et al., 2005). Furthermore, David Siegel’s (2011) work on the structure of networks and repression suggests that particular network structures that occur in online social networks (e.g., the “opinion leader” network) may help movements survive when the regime cannot quickly identify and target leaders for removal. Regardless, no mass action can succeed unless it attracts the support of a sufficient percent of the masses. In the next section I discuss the literature on information, public opinion, and political mobilization among the masses.

How Public Preferences Cause the Dictator’s and Dissidents’ Dilemmas, and Create Difficulties for the Masses

Not only do authoritarian regimes often—though not always—engage in censorship, but so too do members of the public living under dictatorial rule. In repressive environments, people often hesitate to share publicly their true preference if they believe they hold a minority opinion or if they fear they may be punished for expressing opposition to the regime (Kuran, 1991, 1995; Ginkel & Smith, 1999). Since many people only feel comfortable publicly expressing what they perceive to be the majority opinion—even if they do not share this opinion—it reinforces other people’s perception that the position they expressed is, in fact, the dominant opinion in society, which in turn leads others to misstate their publicly expressed opinion (Noelle-Neumann, 1974). In this environment neither the public nor the regime can gauge accurately the degree of support that the regime enjoys.

According to the dominant logic on collective action in high-risk environments, if an individual opposes the regime, before she would join an antigovernment mass action, she likely would first assess the relative strength of the opposition and the regime, and estimate the breadth of the underlying opposition among her compatriots in order to gauge how others might react if protests were to turn into a revolutionary movement (Shadmehr & Bernhardt, 2011). For example, she may know other people who, like her, also oppose the regime; yet she knows none of them has admitted this publicly. Therefore, she might conclude that the discord between her circle’s underlying and publicly expressed opinions probably occurs among other groups in society. In assessing underlying public support for the opposition, she therefore might discount the degree of publicly expressed support for the regime. The problem returns, however, to the lack of reliable information available to her that could help her determine the discount rate she should apply to arrive at the truth. Her tendency should still be to avoid participating because she likely would pay a higher cost if she were to mistakenly overestimate the discount rate—and thus underestimate the regime’s true support—and join a mass action that lacked support as opposed to underestimating the discount rate and remaining safely at home but discontent.

For a variety of practical and psychological reasons, people’s willingness to participate in costly actions often depends on the number of other people they anticipate will also participate. As James DeNardo’s (1985) book title suggests, there is “power in numbers.” As first actors protest, this information cascades to other citizens whose own participation was contingent on some threshold of prior participation (Granovetter, 1978; Kuran, 1991; Lohmann, 1994). Each person’s threshold can depend on a variety of factors including the degree of opposition she has for the regime, her tolerance of risk, and the degree she suffers psychologically from preference falsification. Participation in protest campaigns reaches equilibrium when the remaining latent participants’ thresholds remain higher than the actual level of participation. According to these arguments the unpredictability of the outcomes of protest cascades—whether they produce revolutions or peter out—does not stem from the uncertainty of the government repression, per se, but rather from the unknowable nature of the distribution of people’s true preferences throughout society (Kuran, 1991).

In competitive authoritarian regimes, however, information and coordination problems pose less of an obstacle to collective action, changing individual’s calculus to participate. Elections offer citizens an institutionalized route to mass mobilization, at times facilitated by the existence of political parties.

Why Vote in Authoritarian Elections? When Do Post-Electoral Protests Evolve into Nonviolent Revolution?

In earlier sections, I have discussed why dictators and opposition leaders would support and participate in elections in authoritarian regimes. However, for the elections to serve their purpose for each of these groups, enough citizens must turn out to vote at the polls. If, however, elections are noncompetitive or the electorate anticipates massive fraud, why do people turn out to vote short of being threatened by brute force? Few studies of competitive authoritarian regimes directly address the motivations of individual voters.

The studies that discuss individual motivations for voting behavior propose that two primary motivations drive voter turnout: voters expect some form of payment for doing so (Blaydes, 2011), or because voters see elections as a venue to express their dissatisfaction with the status quo (Miller, 2015). Evidence from Mubarak’s Egypt supports the former case, showing that the government manipulated budgets during election periods to appease different constituents, and even the average caloric intake of citizens increased during the run-up to the elections (Blaydes, 2011). In the latter case, although voters may not get to vote out the leader, with their vote they signal their dissatisfaction with the regime effectively demanding policy concessions in exchange for forgoing material benefits associated with voting. Regimes that experience negative electoral shocks to the vote count subsequently spend more on education and social welfare after the election (Miller, 2015).

Some people who have experienced years of authoritarian rule may have minimal faith that an election itself could bring about major policy change or a change in leadership. However elections can facilitate mobilization at a time when international observers have an eye on the country. Hence, some citizens may go to vote hoping the government does as expected and manipulates the results. The anticipation of fraudulent elections provides a rallying cry for the opposition and effectively lowers the cost for individuals of participating in mass protests (Tucker, 2007). Doug McAdam and Sidney Tarrow note that “disputed elections have become one of the most common catalysts of protest movements in non-democratic states” (2010, p. 534). Several authors concur, noting a coincidence between post-election protests and allegations of fraudulent outcomes (Calingaert, 2006; Kuzio, 2006; Tucker, 2007; Bunce & Wolchik, 2010). When fraud reaches such extreme levels that it leads election monitors to call into question the declared winner, the cost-benefit ratio of participation becomes even more favorable for individuals (Tucker, 2007).

In recent years as the lines have blurred between institutional characteristics of democratic and non-democratic regimes, so too have the distinctions between online and offline opposition activity. When protests appear to erupt, it may be difficult to decipher whether they were organized by a traditional social movement or if they evolved from communication among participant’s online social network.

Public Opinion and Protest in the Internet Age

Images of protests in Tahrir Square in January 2011 that eventually brought down the 30-year reign of Egypt’s then-dictator, Hosni Mubarak, remain indelibly etched in our minds. News coverage in the West highlighted the role that social media played in overcoming coordination problems. While many referred to the Egyptian revolution, and those of neighboring countries, as Twitter or Facebook revolutions (Zuckerman, 2011; Hudson, 2011),¹⁸ they were hardly the first such events in which citizens subjected to authoritarian rule turned to “new” ICTs to draw others to the streets. Scholars have argued that similar use of cellphones—first for texting and more recently to access social networking and video-sharing websites—helped citizens mobilize to force regime change in the Philippines in 2001 (Santner, 2010; Shirky, 2011) and Ukraine in 2004 (Goldstein, 2007), and publicly challenge results of elections in Iran (Kavanaugh et al., 2013) and Moldova (Morozov, 2009b) both in 2009, just to name a few instances.

Those who support the potential of ICTs contend that ICT access can facilitate antigovernment actions in authoritarian regimes through an informative mechanism, and via a tactical mechanism (Breuer, 2012; Diamond, 2012; Little, 2015). First, they increase access to information and a diverse array of perspectives in closed societies. Exposure to new, independent content opens the eyes of citizens, influencing their opinion of the regime and its performance—altering the cost-benefit ratio of action—potentially leading to a growing pool of prospective participants in antigovernment actions. Second, ICTs—text messaging, microblogging, and content-sharing, in particular—ease coordination problems inherent in mobilizing mass protest events. They facilitate the transmission of coordinating information among a network of opposition supporters and latent participants connected to these supporters.

Some attempts to demonstrate that new ICTs promote offline mobilization rely on country-level data to test broad implications of the relationship (e.g., higher Internet penetration rates or cellular subscription rates correlate with higher frequencies of demonstrations and riots) (e.g., Meier, 2011). To date, relatively few data exist at the organizational level (in part because many recent rebellions seem to lack “traditional” organizations behind them) or on individual participants, particularly in countries where mobilizing carries high costs. Given the lack of hard empirical data at the individual or group level, several scholars interested in the subject employ formal models. Alternatively, studies published in the wake of major international events relied on single case studies or comparing two to three, perhaps atypical, cases.

A few studies stand out for their effort to look at individual-level data through surveys and interviews with participants and, importantly, nonparticipants. Evidence drawn from surveys of people who participated in protests in Egypt in Tahrir Square in 2011 supports the importance of social media in shaping citizens’ decision to participate (Tufekci & Wilson, 2012). Other studies also have relied on content analyses of microblogs to show upticks in online communication prior to major protest actions (Howard & Hussain, 2011).

Skeptics of the power of ICTs point out that dissent long preceded the rise of the Internet, and that other nonviolent revolutions, such as those in Eastern Europe in 1989, occurred amid dictatorships prior to the Internet revolution. Students of Middle Eastern and North African countries contend that for the underlying dissent to take shape in the form of mass rebellions like those of the Arab Spring it required a mechanism to unite individual sentiments: “Digital media helped to turn individualized, localized, and community-specific dissent into a structured movement with a collective consciousness about both shared plights and opportunities for action” (Howard & Hussain, 2011, p. 41).

Not only does the information these mediums provide alter the dynamics of information control and information gathering in authoritarian regimes, but these technologies also change the methods through which people can access them, lowering the cost of reaching larger numbers of citizens. Increasingly, people access the Internet via mobile phones which overcomes the lack of personal computer ownership in developing countries. Mobile access to interactive communication technologies can aid in the rapid transmission of last-minute changes about logistics to avoid government forces. These factors may make ICTs, particularly cell phones, more influential for residents of dictatorships who cannot organize freely than for social movements in democracies.

To test whether the influence of ICT access on protest differs by regime type, I employ data on a smaller subset of countries that have better available information on protests. Many large-N studies on protests, which frequently rely on content analyses of newspapers to generate counts, ultimately produce relatively poor quality data. The Social Conflict Analysis Database (Salehyan et al., 2012) hosted by the Robert S. Strauss Center for International Security and Law at the University of Texas, Austin, released higher quality data on countries in Africa and parts of Latin America from 1990 to 2014. Using these data I compare the correlation between Internet penetration rates and the incidence of organized and spontaneous protests (i.e., demonstrations and riots) in democracies versus dictatorships (still relying on regime data from Magaloni et al., 2013 data). These data provide preliminary evidence that higher rates of Internet diffusion in authoritarian societies are associated with an increasing average numbers of protests in a given year, whereas little difference occurs in the average number of protests in democracies as countries’ Internet penetration rates increase. This distinction results in part because among the countries included more protests occur in dictatorships, but even taking that into consideration, the difference between regime types holds. The data also indicate that Internet access matters far more for spontaneous protests (where leaders are not clearly identifiable) than they do for organized ones, findings which suggest interesting implications for future research.¹⁹

Early empirical studies on the relationship between ICT use and mobilization improve on impressions people reported in the immediate aftermath of particular events. Studies suggest that ICTs offer some promise for certain forms of mobilization in authoritarian regimes. However, to date, many authoritarian regimes have managed to catch up to the basic tactics, using measures like slowing down Internet speeds to inhibit uploading of content to video-sharing sites, or jamming cellular signals. Since these tactics affect not only active supporters of the opposition, but all residents in the area, a regime hoping to sustain whatever popularity it maintains and hold on to power must use these tactics sparingly.