Digital Authoritarianism in Sub-Saharan Africa. Internet Control Techniques and Censorship

Contents

1 Introduction
1.1 Research question and method
1.2 Limitations
1.3 Structure of the thesis

2 Background
2.1 Social Media in Sub-Saharan Africa
2.2 Internet Control Techniques and Censorship in SSA
2.3 Internet Governance Models

3 Literature Review

4 Theoretical Framework
4.1 Digital authoritarianism
4.1.1 International Practices Approach
4.1.2 Authoritarianism
4.1.3 Social media and the public sphere
4.1.4 Definition
4.2 Conditions
4.2.1 High Social Mobilisation
4.2.2 Strong Chinese influence
4.2.3 High Internet Penetration

5 Methodology
5.1 Qualitative Comparative Analysis
5.2 Selection of cases
5.3 Operationalisation and Calibration
5.4 Scope, method and data limitations

6 Empirical Analysis
6.1 Individual Necessary Conditions
6.2 Individual Sufficient Conditions
6.3 Analysis of Sufficiency (Truth table)
6.4 Solution Terms
6.4.1 Complex Solution
6.4.2 Parsimonious Solution
6.4.3 Findings

7 Conclusions
7.1 Policy recommendations

8 Appendices
8.1 Appendix 1: List of Sub-Saharan Africa (UNDP)
8.2 Appendix 2: Most common triggers for internet blockages in SSA (2016-2019)
8.3 Appendix 3: Chinese Investment in SSA (2016-2019)
8.4 Appendix 4: Freedom in the World Index 2016-2019
8.5 Appendix 5: Digital Authoritarian Practices
8.6 Appendix 6: Internet Penetration Index
8.7 Appendix 7: Raw Data
8.8 Appendix 8: Script Code for R
8.9 Appendix 9. Consistency and Coverage Formulae

Acknowledgements

I would like to express my gratitude to my supervisors PhD Patrick Mello for his support and the introduction to QCA, as well as PhD Hasnain Bokhari for introducing me to internet policy through the courses he imparted, to deliver the class in such a unique way that made me think of things beyond and build an opinion, and to encourage me during that time.

I am wholeheartedly thankful to my family for their support and trust. Most of all, to my dad, who provided me with the academic and personal foundations to go through all this process by myself. To my sister Natalia, who listened even when she is not familiar with the topic.

I would also like to express my gratitude to my friends, especially Sushobhan Parida, Alyssa Santiago, and Marcela Hernandez for their support, encouragement, and company through this process. Moreover, for listening and helping me order my ideas, as well as keeping me accountable to finish this work. To Usman Oyebamiji, John Kamoga and all my African friends, who were essential in my building insight into the African reality.

The last decades have witnessed an astounding empowerment of citizens in subSaharan Africa, who organise to demand political change and denounce mismanagements through social media; by the same token, the overwhelming power of states has also manifested itself in their exertion of surveillance controls and censorship. The African continent has been placed as the second region in the world, after Asia, with the highest number of blockages in the last 5 years. In this study, digital authoritarianism is defined as the state’s deliberate use of internet shutdowns, social media blocking and social media surveillance as a way to curtail freedom of expression and of assembly and the right to access information. Focusing on free countries, this study explores, with the use of a qualitative comparative analysis (QCA), three conditions that seem to be present in cases when digital authoritarian practices in SSA are undertaken: China’s soft power, social mobilisations and internet penetration. The wider use of authoritarian practices on the internet in SSA might be a a signpost of the lack of international consensus on internet governance and the recognition of digital rights.

Die letzten Jahrzehnte haben ein erstaunliches Empowerment der Bürger in SubsaharaAfrika erlebt, die sich organisieren, um über soziale Medien politische Veränderungen zu fordern und Misswirtschaft anzuprangern; umgekehrt hat sich die überwältigende Macht der Staaten auch in der Ausübung von Überwachungskontrollen und Zensur manifestiert. Der afrikanische Kontinent ist nach Asien die zweitgrößte Region der Welt mit der höchsten Anzahl von Blockaden in den letzten fünf Jahren. In dieser Studie wird digitaler Autoritarismus definiert als der bewusste Einsatz von Internetabschaltungen, Blockierung sozialer Medien und Überwachung sozialer Medien durch den Staat als Mittel zur Beschneidung der Meinungs- und Versammlungsfreiheit und des Rechts auf Zugang zu Informationen. Die Studie konzentriert sich auf freie Länder und untersucht mit Hilfe einer qualitativ vergleichenden Analyse (QCA) drei Bedingungen, die in Fällen, in denen digitale autoritäre Praktiken in SSA angewandt werden, gegeben zu sein scheinen: Chinas weiche Macht, soziale Mobilisierung und Internetdurchdringung. Die breitere Anwendung autoritärer Praktiken im Internet in SSA könnte ein Wegweiser für den Mangel an internationalem Konsens über die Internet-Verwaltung und die Anerkennung digitaler Rechte sein.

List of Figures

1 Social Movements (2009-2019) in selected countries

2 Most common triggers for internet blockages in SSA (2016-2019)

3 Imports from China (billion US$)

4 China vs US FDI flow to Africa (US$)

5 Social media penetration of eligible population - 2019 (%)

6 2x2 Relevant cells

7 Venn diagram - Complex Solution

8 Venn diagram - Parsimonious solution

9 Formulae of Consistency

10 Formulae of Coverage

List of Tables

1 Internet control techniques

2 Full and Service-based blockages

3 Beetham’s subversions of democracy

4 Internet penetration and usage by region

5 Selected cases

6 Operationalisation and Calibration Summary

7 Calibrated data

8 Necessary Conditions

9 Individual Sufficiency

10 Truth Table

11 Complex Solution

12 Parsimonious Solution

13 Directional Expectations

14 Appendix 1. Sub- Saharan Africa (UNDP)

15 Most common triggers for internet blockages in SSA (2016-2019)

16 Chinese Investment in SSA (2016-2019)

18 Summary - Digital authoritarian practices by country

21 Raw Data

1 Introduction

“If you insult the President, it is a matter of security” (Cross, 2019, p. 202), claimed a senior police officer in Tanzania. Insulting or criticising leaders on social media is there seen as a threat that potentially ignites protests. Thus, the Tanzanian state justifies any repressing practice on behalf of maintaining peace (Cross, 2019, p. 202). This feeling towards the internet and social media is spreading in the African continent. This Master thesis on digital authoritarianism examines some conditions that drive sub-Saharan African states to shut down the internet or censor social media. It focuses on democratic states that, in spite of this, have been found to curtail freedoms on the internet.

Networked and mobile communication have become an essential tool for today’s public participation and political mobilisation in Africa (Kalyango & Adu-Kumi, 2013). However, it has also posed an “information dilemma” (Göbel, 2013), in which governments must decide between protecting the freedom of expression to everyone, with all the risks and challenges that this entails, or applying more restrictive policies that limit freedoms on behalf of national security and political stability (Boas; Göbel, 2013), and, in the worst cases, to install dictatorships. In Africa, “shutdowns are not only growing in number but are also expanding in scope and affecting more and more people” (Access Now, 2020b, p. 6).

The motivation for this study is related to the ideals of freedom and democracy. The role of social media in the public sphere is the basis of such sphere’s freedom and neutrality. The digital arena “is an essential component of the sociopolitical organisation because it is the space where people come together as citizens and articulate their autonomous views to influence the political institutions of society” (Castells, 2008, p. 78), thus setting the cornerstone for democracy.

In this way, this research is situated within the internet governance and digital democracy research field. Internet governance is a complex, relatively new field involving different actors, levels, values and domains. The utopian and dystopian scenarios for what the endgame of societies can be, dependant on the ways in which the internet and technology unfold, have originated different normative approaches regarding the role of the state in the internet, as well as the boundaries of freedom on the net.

Two different approaches have emerged: liberal and sovereigntist. The former, led by Western Europe and the United States, encourages “the development of the internet as much as possible by giving individuals, firms and civil society organisations as much freedom as possible” (Flonk, Obendiek, & Jachtenfuchs, 2020, p. 365). In the latter, spearheaded by China and Russia, the state should “protect sovereignty and core domestic values and goals against domestic or international actors empowered by the internet” (Flonk et al., 2020, p. 365).

The sovereigntist approach has given room for authoritarian practices since it reinforces the political apparatus’ willingness to intimidate the opposition and quell citizens’ demands. Censorship is finding more spaces as the threats of misuse intensifies. Although many activists in Africa stand for the criminalisation of internet shutdowns and social media blocking, these practices are poorly studied outside the authoritarian regimes.

For this reason, this thesis aims to identify the underlying conditions that make subSaharan free countries more prone to undertake digital authoritarian practices. Likewise, the thesis attempts to contribute to the debate and academic research of this type of practices by expanding the lexicon and identifying relations that might set the ground for future research in the region or other regions. Lastly, some public policy recommendations that might hinder the expansion of digital authoritarianism in SSA will be drawn.

1.1 Research question and method

Although there exists a positive correlation between protests and shutdowns, not every government decides to disrupt internet services when there are protests and/or elections. Some countries, like South Africa, called “the capital of protests” (Runciman, 2017), have not been found to interrupt the internet in the analysed period. While others, like Nigeria, have acted in more restrictive ways, imposing blockages and prosecuting activists on more than one occasion.

This thesis presents an empirical study of digital authoritarian practices in SSA framed as a qualitative comparative analysis (QCA). QCA is a set-theoretical research method used to find the multiple pathways (equifinality) or combinations of factors (conjunctural causation) under which a social phenomenon can occur (Mello, 2020; Schneider & Wagemann, 2010)¹. This study claims that, beyond the specific events triggering digital authoritarian practices, it is necessary to look for international and socio-political configurations that can jointly impact a state’s decisions to curtail freedom through internet blockages. In doing so, three conditions are considered which have significantly increased in recent years in sub-Saharan Africa: social mobilisation, Chinese soft power, and internet penetration.

The choice of these conditions was the result of literature review and initial empirical findings. Similarly, the selection of the cases and the analysed period was determined by data availability, as well as following the logic of analysing free countries, going beyond regimes and specific events. Authoritarian countries are thus excluded, because digital authoritarian practices are an extension of the already established regimes; in other words, because the analysis of such countries would “suffer from a kind of tautological reasoning” (Glasius & Michaelsen, 2018).

Hence, the research question of this study is: Under what conditions do free countries in Sub-Saharan Africa undertake digital authoritarian practices?

This problem will be addressed through the following sub-questions:

- What are digital authoritarian practices?
- Is high social mobilisation a necessary or sufficient condition for states to cut off internet-based communication channels?
- Is a strong influence from China a necessary or sufficient condition for states to adopt the digital authoritarian approach to control the internet?
- Is a high internet penetration necessary for countries to restrict free access to this service?
- Is the configuration of the previous conditions sufficient to lead to digital authoritarian practices in SSA?

The hypothesis of this study is, then: Countries are adopting digital authoritarian practices in SSA where social mobilisation is high and Chinese influence is strong, regardless of the levels of internet penetration in such countries.

1.2 Limitations

It is not intended to claim that the conditions chosen for this study are the only configuration that affects the outcome. Legal frameworks, ISP’s ownerships, or regional authorities are some examples of other conditions that can influence the occurrence of digital authoritarian practices. However, given the temporal and material characteristics proper of a Master thesis, such other conditions will fall outside the scope of this study. Likewise, the availability of data is sparse, particularly with Chinese information and internet statistics. China’s government is very secretive with data related to aids, loans, and investments, and there is also no public (official) updated statistics for internet usage since 2017. Therefore, data from research centres and the UN were used to collate China’s investment figures and African internet figures. Data vary significantly from source to source. Thus, taking a different reference can change some sets’ memberships. Nevertheless, according to the individual case observations that were undertaken in this project, the proxies used are indicative of an acceptable qualitative direction for each case, and this renders fair memberships to the sets.

1.3 Structure of the thesis

This thesis is divided into three main parts. The first part, comprised of Chapters 2 and 3, provides the state of the art and a brief introduction into how social media has impacted sub-Saharan Africa, the current internet governance models and the most common internet control techniques used in the world, and particularly in Africa. The second part, Chapter 4, explains the theory behind the concept of digital authoritarianism and the conditions that may be associated with it. The third section, consisting of Chapter 5 and 6, establishes the structure of the empirical analysis and the interpretation of the QCA solution. Finally, Chapter 7 provides conclusions, general policy recommendations, and future research possibilities. Raw data, complementary data, code source, and output can be found in the Appendices, as well as in: https://bit.ly/3b9iGHO

2 Background

2.1 Social Media in Sub-Saharan Africa

In Sub-Saharan Africa, like in most parts of the world, social media have transformed societies in ways that have allowed people to share ideas, create content, organise, and participate in politics like never before. Networked and mobile communications have become an essential tool for today’s public participation and political mobilisation in Africa (Kalyango & Adu- Kumi, 2013). Facebook, Twitter, and Whatsapp, as well as mobile phones, are used to involve citizens in the political discourse in Sub-Saharan Africa. The engagement of citizens in debates through social media has been useful to denounce corruption and illiberal actions during legislative elections in Botswana, Sierra Leone, Nigeria and Kenya (Diepeveen, 2019; Dwyer & Molony, 2019; Masilo & Seabo, 2015; Nyabola, 2018; Orji, 2019), as well as to organise social movements to pressure states to make political changes in South Africa, Nigeria, Kenya, or Senegal.

In 2011, the Nigerian electoral process, which was already discredited, “led election stakeholders to seek solutions in digital technologies in general and social media in particular” (Orji, 2019, p. 152). Journalists and citizens began to use social media platforms such as Ushahidi (a Kenyan website to collect eyewitness reports), Facebook, Twitter and SMS to observe, document and report major events in elections. Additionally, the internet has become one of the main platforms for election campaigns in this country. Candidates use social media to make announcements, hold debates and publish poll results. Similarly, the Electoral Commission in Nigeria established a social media structure (YouTube, Facebook and Twitter) to “receive and respond to concerns raised by candidates, voters and election observers” (Orji, 2019, p. 167).

In the same way, social media has emerged “as one of the most potent political spaces” (Nyabola, 2018, p. 5) in Kenya. After a millionaire investment in the Kenya Integrated Elections Management System (KIEMS) and the Electronic Voter Identification Devices (EVIDs), Kenya’s 2017 general elections were invalidated because of the dissimilarities between what the election system showed and what the popular sentiment in social media—backed with images and testimonies—indicated. With a highly questioned electoral process and looming massive protests announced in social media, the incumbent government decided to nullify the election results and organise the second round in October of the same year (Nyabola, 2018). Moreover, the key role of social media transcends electoral processes—Twitter, Whatsapp, and Facebook have also been determinant in the general collective awareness and mobilisation in Kenya. For instance, on April 3th 2016, a rumour alleging fraud by Chase Bank was disseminated on Whatsapp and Twitter. Without proper evidence, in the lapse of four days, Chase Bank had all accounts frozen, was taken under control of the Central Bank, and the Heads were ceased and set under investigation (Nyabola, 2018). Although allegations ended up being partially true, the Chase case is an example of how fast a message in social media can scale and create massive overreactions.

In 2015, purported corruption, rise in the university fees, and political discomfort originated the #FeesMustFall and #ZumaMustFall movements in South Africa. Two years later, the reshuffle of the cabinet reignited the #ZumaMustFall movement demanding the resignation of the incumbent president Jacob Zuma, which—with the help of social media—turned out to be “one of the largest incidences of protest” (Bosch, 2019, p. 77) in this country.

However, political engagement in social media has also posed an “information dilemma” (Göbel, 2013): Governments may decide to protect the freedom of expression of every citizen, with all the risks and challenges this entails, or they can decide to apply more restrictive policies that limit freedoms on behalf of national security and political stability (Boas; Göbel, 2013).

Amid the Arab Spring upheavals in 2011, Egyptian authorities decided to apply unprecedented disruptions that served as measures to stem protests and cut internet connectivity for five days. The blackouts spread in the region as uprisings also happened in Syria, Libya, and Tunisia (Keller, 2011). The new era in which “millions of fingers plucking and poking at touch screen phones, logging in, posting and meeting up can bring down a government” (Courtemanche, 2011) began, and so did the eagerness of the government to control the internet.

2.2 Internet Control Techniques and Censorship in SSA

There is a plethora of technical paths to block or filter the flow of information in the network. The terms “filtering”, “blocking”, and “shutting down” the internet are seen as ways of “censorship” (Internet Society, 2015). Thus, these terms are used by some indistinctively. While “censorship” or “blocking” have a more negative connotation, usually “filtering” is used to indicate a harmless and more accepted way of censoring (Internet Society, 2015). Nevertheless, in this research, we will use all terms with the same overtone, since they all have the same effect on freedom.

In almost all the countries around the world that block content, censorship enjoys widespread support when it is related to child pornography or websites inciting violence or criminal acts (Internet Society, 2017). The most common internet control techniques used by countries—in different degrees and situations—are shown in Table 1 below. Although few exceptions are accepted by common consensus, the problem arises when political interests influence the interpretation of the words “illegal” or “dangerous”. In those cases, censorship controls are used to identify political opposition, quell protests, arrest inconvenient journalists, and silence activists. In this manner, this research will focus on those acts with an impact on freedom of expression (FoE), freedom of assembly (FoA) and access to information.

Internet shutdowns are defined as “intentional disruption of internet-based communications” (Access Now, 2019, p. 1). They occur when the Internet Service Provider (ISP)—by order of the government—fully blocks access to the internet in a delimited geographical area.

It is the most aggressive of these methods since it denies access to information and curtails freedom of expression and assembly through cutting off communication channels.

Content Blocking “is a practice in which end users are denied access to certain online

Table 1: Internet control techniques

Abbildung in dieser Leseprobe nicht enthalten

Source: Elaborated by the author from multiple sources (Internet Society, 2015; Access Now, 2019; Shahbaz & Funk, 2019)

Table 2: Full and Service-based blockages

Abbildung in dieser Leseprobe nicht enthalten

Source: Elaborated by the author with data from Access Now.

content” (Internet Society, 2015, p. 2). It occurs when the ISP blocks specific IP addresses or URLs such as blogs, news websites, or a service such as Twitter, YouTube, Facebook, Whatsapp, or Telegram. Thus, this method also limits FoE, FoA, and access to information. Despite the seemingly low internet penetration in Africa, governments have resorted to blocking the internet as a means of containing social movements when protests, elections, military actions, or political instability take place. Compared to 2018, in 2019 there were more shutdowns but also “there was a significant increase in the number of African countries that shut down the internet” (Access Now, 2019, p. 13), while ten countries decided to block the network in 2018, 14 countries did so in the subsequent year. Chad, Sudan, Democratic Republic of Congo, Gabon, Zimbabwe, Algeria, Benin, Malawi, Liberia, Ethiopia, Egypt, Eritrea, Cameroon, and Mauritania blocked the internet in 25 occasions, 39% more times than in 2018 (Access Now, 2019, 2020b).

Governments do not typically recognise internet blockages. However, when they do, they justify their actions appealing to national security, public safety, and hate speech arguments. In addition to this, and conversely to what happens in India, where Jammu and Kashmir local governments are responsible for the largest cut-offs in the region—and the world—shutdowns in Africa occur at a national level.

During the so-called “first social media election” (Dwyer, Hitchen, & Molony, 2019), Sierra Leone’s 2018 general elections experimented live conversations during candidates’ debates and intense monitoring of election results. Millions of messages spread online resulted in sharp engagements between political parties and citizens in social media. As a result, in the course of the second round and while information was flowing at its crux, President Koroma ordered a twelve-hour countrywide internet shutdown to slow down social hype (Dwyer et al., 2019).

Throttling refers to “practices that reduce the data throughput rates of delivered content to end-users” (Internet Society, 2015, p. 3). It occurs when an ISP narrows the bandwidth or sets a bottleneck in order to slow the connection speed so much that the targeted services are unusable. This technique has two drawbacks. First, it is not easy to detect or to prove it was on purpose. Targeted throttling might be the same thing as if one connects to a weak signal. Second, some may argue that since services are not entirely down, throttling does not limit freedoms. Therefore, this technique will not be considered in the analysis.

Social media surveillance “refers to the collection and processing of personal data pulled from digital communication platforms, often through automated technology that allows for real-time aggregation, organisation, and analysis of large amounts of metadata and content” (Shahbaz & Funk, 2019). Surveillance can be detrimental because “it may interfere with intellectual privacy—that is, freedom of thought, belief, and private speech [and] changes the power relation between the watcher and the watched” (Glasius & Michaelsen, 2018b, p. 3800) with the purpose of influencing and controlling individuals (Lyon, 2007).

According to Reporters without Borders, press freedom violations are widespread in the African continent. “They include arbitrary censorship, especially on the internet [...], arrests of journalists [...], and acts of violence against media personnel” (Reporters Without Borders [RSF], 2020, para. 1). In Tanzania, legislation has “been used to charge those who criticise the government on social media, including on public forums such as YouTube, in messages to Whatsapp groups, and in personal Facebook posts” (Cross, 2019, p. 202).

In this manner, in April 2016, Zimbabwean Pastor Evan Mawarire posted on Facebook an emotive video complaining about the economic, political and social crisis that afflicted his country. The video went viral, and after a few days, the cleric was leading the #ThisFlag national campaign, through which he “encouraged Zimbabwean citizens to speak out against the government’s mismanagement” (Karekwaivanane & Mare, 2019, p. 43) and coordinated “public acts of protests” (2019, p. 43). On July 6, “central business districts Harare and Bulawayo were virtually deserted as business owners, transport operators and members of the public” (Karekwaivanane & Mare, 2019, p. 53) followed Mawarire’s call for a national stay-away. Given the success of this event and the setting in motion of a second stay-away, Pastor Evan Mawarire was arrested on July 12. A few days later, due to public pressure, he was released and exiled.

Likewise, after approving Kenya’s Computer and Cybercrimes Law in 2018, which provides that anyone who makes an “online defamation” or spreads “fake news” can be prosecuted and jailed (RSF, 2018), activist Alenga Torosterdt was arrested because he “was attacking [Kenyatta’s] government and revealing secrets of his administration on social media through [his] blogs” (Nyambura, 2018). Furthermore, in Nigeria, Ibrahim Garba Wala, coordinator of a national advocacy platform for democracy, human rights, and anti-corruption (CATBAN), was convicted to 12 years in prison in April 2019 due to a Facebook post where he accused the Chairman of the National Hajj Commission of Nigeria, Abdullahi Mukhtar, of committing illicit enrichment. In request of Mukhtar, a Federal Court sentenced Wala “on the charges of ‘management of or membership in an unlawful society’, ‘public incitement’, and ‘criminal defamation’ ” (Frontline Defenders, 2019, p. 1). In March 2020, under the new administration, the activist was released under a presidential pardon (“Rights Activist, IG Wala Who Was Sent To Prison Over Facebook Post Receives Presidential Pardon,” 2020).

From these facts we can conclude that internet censorship is becoming a common practice in Sub-Saharan Africa since the role of social media in politics has increased in importance. As a reaction to this phenomenon, some states have decided to undertake repressive internet controls like those mentioned above, which will hereafter be called digital authoritarian practices.

2.3 Internet Governance Models

The internet is a social tool where hierarchies and boundaries are minimised. Public access to the internet and its impacts on the ways we relate to each other, to the economy and politics demanded the establishment of an order on the internet based on non-hierarchical norms, values, and processes that fit into the collaborative architecture of the web. This structure presented then, and now, a challenge to national and international institutions where the central actor is the state because, in the virtual sphere, they are “but one actor among others” (Hofmann, Katzenbach, & Gollatz, 2017, p. 1410).

In an attempt to establish some order on the internet, at the beginning of the 1990s, the Internet Corporation for Assigned Names and Numbers (ICANN) laid the institutional foundations and management principles of the internet’s infrastructure. However, these proved insufficient to face the various challenges that the dynamics between the different actors (public, private, and civil society) posed (Azmeh, Foster, & Echavarri, 2019; Hofmann et al., 2017) and raised new questions: What should be regulated? Who, how and where should these rules be formulated? What is the role of the states in internet governance?

In the early 2000s, a statement issued by the World Summit for Information Society (WSIS), promoted by the United Nations, defined the internet governance as “the development and application by governments, the private sector and civil society, in their respective roles, of shared principles, norms, rules, decision-making procedures, and programmes that shape the evolution and use of the internet” (WSIS, 2005). As broad as this definition can be in terms of scope and responsibility, it reflects the complexity of the internet and, more importantly, it embraces a multi-stakeholder approach. In this understanding, the WSIS mandated to convene an Internet Governance Forum (IGF) in 2006. This forum is meant to provide a space for stakeholders to dialogue, inform policymakers about risks and opportunities of the internet, and share good practices. However, it does not get formal outcomes or binding agreements (IGF, n.d.).

Internet governance is also multi-disciplinary because the internet, by its nature, apper- tains to different fields of study and jurisdiction in the offline world. The interrelationship between the actors, the norms and values, and, finally, the uses and misuses of the internet affect everything and everyone, even those who are not connected. It is this universality that makes internet regulation and controls extremely challenging.

Therefore, two major approaches or ‘spheres of authority’ (Flonk et al., 2020) have emerged. The perspectives and treatment of governments to internet-based platforms distinguish the twofold internet governance approaches. On the one side, the liberal approach encourages “the development of the internet as much as possible by giving individuals, firms, and civil society organisations as much freedom as possible” (Flonk et al., 2020, p. 365). The European Union and the United States share this perspective and support the core values of “free market and pluralist civil society thinking” (Flonk et al., 2020, p. 365). However, within this sphere, the nuances in the role of the state in providing security can also vary. By the time the General Data Protection Regulation (GDPR) was enacted in the European Union, this legal framework stirred up some tension between the EU and the US (Hoffman & Taylor, 2019), where most of the Western big tech firms dwell.

On the other side, the members of the sovereigntist approach “see the internet as a threat” (Flonk et al., 2020, p. 365). For them, the state should “protect the sovereignty and core domestic values and goals against domestic or international actors empowered by the internet” (Flonk et al., 2020, p. 365). This perspective is headed by China and Russia, who have established the most important technological and legal frameworks to control the flow of information on the internet on the basis of internet sovereignty, and comprises of internet filtering and data localisation. “China pioneered digital age censorship with its ‘Great Firewall’ of a state-controlled internet” (Polyakova & Meserole, 2019, p. 1) and is today “overseen by over sixty agencies with vast legal and technical ability to monitor and regulate online activity” (Polyakova & Meserole, 2019, p. 3). However, because the implementation of a comprehensive surveillance system like the Chinese is costly and infrastructure-dependent, developing countries keen to the giant Asian’s approach are creating versions on its resemblance, such as in the case of Russia. The Russian model is considered a low-cost surveillance framework that relies on data inspection and the criminalisation of online activities (Polyakova & Meserole, 2019).

This study aims to find some configurations or patterns that help understand why some states in SSA decide to undertake digital authoritarian practices, while others do not. In the next chapter, a review will be made of the most important literature studying the conditions under which internet shutdowns, content blocking, and social media surveillance might happen in SSA.

3 Literature Review

Digital authoritarian practices are widely associated with political instability in different formats, particularly in the Global South. Although the causal relations between digital authoritarian practices and free states or democracies are not well-established, the significance of the effects of repressive activities has opened spaces in many fields to study the conditions in which some states choose an authoritarian path. A comparative study and meta-analysis by Howard, Agarwal & Hussain revealed that states tend to disconnect the digital networks depending on “the degree to which different regimes feel threatened by social media, whether such tools are actively used to organise dissent, or passively used for producing and consuming culture” (2011, p. 230). Furthermore, it showed that internet shutdowns, as well as other digital authoritarian practices, are commonly executed for national security reasons, eliminating propaganda, mitigating dissidence, steering elections, or preserving cultural and religious morals (Howard et al., 2011). The way states react to these events, which are backed up by social media, may pose an “ICT dilemma” (Göbel, 2013) in which governments have to choose between freedom and repression.

This literature review is an overview of those studies that have tried to understand internet shutdowns, social media blocking and surveillance. In so doing, it will build upon the categorisation proposed by Glasius and Michealsen’s, albeit with an altered construal. The scaffolding identifies four main focus areas: economic-legal, freedom-advocacy, security- surveillance, and authoritarianism studies.

The economic and legal debates connect the appropriation of human experiences and massive data collection to incipient models of regulation, as well as individual and property rights. Under this approach, internet governance is viewed from the liberal and sovereigntist perspectives, but also, for some authors, the categorisation of digital experiences is seen as a trade issue, where private goods and services must comply with national regulations. Chinese and Russian models of internet sovereignty, internet filtering and data localisation as protectionist tools are explored in this legal-economic approach (Azmeh, Foster, & Echavarri, 2019; Gao, 2018). Moreover, Jaclyn Kerr proves that “even as overall internet repression levels have increased, the particular legal frameworks, technical systems, and other control practices used have been deeply influenced by complex regional interdependencies” (Kerr, 2018, p. 3814). Following this line, some researchers have theorised over an apparent diffusion of alternative governance models. McKune et al. point out an intentional diffusion in the case of China, a government that strives “to promot[e] an internet sovereignty agenda at home and abroad ... [envisaging] the regime’s absolute control over the digital experience of its population” (McKune & Ahmed, 2018, p. 3835). Other researchers, however, argue that authoritarian practices “depend more on local political conditions than Chinese grand strategy” (Chen Weiss, 2020).

The freedom-advocacy or democracy approach is headed by non-governmental organisations (NGOs), which create coalitions with local organisations to monitor and denounce internet shutdowns, blockages, and other kinds of abuses. Consequently, these organisations have been able to identify the most common causes related to internet shutdowns and content blocking practices. Empirical evidence shows that protests, military actions, communal violence, political instability, religious events, elections, and information control are circumstances in which some governments decide to repress internet-based communication, which typically these governments would justify referring to fake news, hate speech, precautionary measures, public safety, and national security (Access Now, 2019). Furthermore, researchers such as Tina Freyburg and Ben Wagner have attempted to understand the occurrence of internet shutdowns by analysing the context in which states cut off the internet in SSA and Pakistan, respectively. Considering elections as a relevant event, Freyburg explored the configurations of elections and the Internet Service Providers and proved that authoritarian states imposed internet shutdowns in Sub-Saharan Africa when they owned the ISPs and elections occurred (Freyburg & Garbe, 2018). Likewise, Ben Wagner (2018) claims that short-term internet shutdowns are used as a strategy to prevent mobilisation, whereas long-term events can have other explanations.

The security-surveillance literature focuses in the social and technical risks of mass surveillance that governments in the Global North exerted on behalf of national security such as the PRISM and NSA programs in the United States (Amoore, 2014; Andrejevic & Gates, 2014; Lyon, 2007; Lyon, 2014). This approach would argue that fighting terrorism is one paramount reason behind state surveillance.

Finally, authoritarianism studies build their inquiry upon the question whether ICTs make authoritarian regimes stronger (Hindman, 2008; Morozov, 2011; Murakami Wood, 2017; Sun- stein, 2008; Walker & Ludwig, 2017), or rather help undermine dictatorships by democratising societies (Cela, 2015; Diamond & Plattner, 2012; Fuchs, 2014; Iosifidis & Wheeler, 2016; Johannessen, 2013; Paparachissi, 2009; Shirky, 2011). This studies reveal that, concerning digital authoritarian causal relations, states make use of content blocking, surveillance, and shutdowns either to carry out repressive practices of the already authoritarian regimes or to stop and control the democratisation that threats them. Thus, this approach “suffers from a kind of tautological reasoning” (Glasius & Michaelsen, 2018) since the object of research only exists in the “purview of authoritarian regimes” (Glasius & Michaelsen, 2018), leaving aside other different types of socio-political organisation.

With the focus in sub-Saharan Africa, this thesis takes the economic-legal and freedom- advocacy approaches. It argues that governments in free states can also undertake digital authoritarian practices that menace freedoms whenever there are high social mobilisation and a strong Chinese influence, irrespective of their internet penetration rates.

4 Theoretical Framework

This chapter aims to define digital authoritarian practices and the possible set of conditions that make states in SSA prone to carry out digital authoritarian practices. Since a qualitative comparative analysis (QCA) will be made, this chapter will be divided into two sections. The first section, digital authoritarianism, sets out a working definition for the outcome and argues for the relevance of studying the relations between freedom and social media for strengthening democracies. The second section explains how the proposed conditions potentially affect the outcome. More specifically, we will discuss the probable impact of Chinese influence in subSaharan Africa by analysing policy diffusion and soft-power theories; the relation of social media and social movements; and lastly, internet penetration and African lifestyle.

4.1 Digital authoritarianism

In 2016, the Mercator Institute for China Studies used, for the first time, the term “IT-backed authoritarianism” to refer to China’s utilisation of “ubiquitous mass surveillance including the underlying IT systems linking and analysing multiple data sets and various channels to increase control over society and exert central authoritarianism more effectively” (Meissner & Wübbeke, 2016, p. 52). Later, in 2019, Polyakova and Meserole defined digital authoritarianism as “the use of digital information technology by authoritarian regimes to surveil, repress, and manipulate domestic and foreign populations” (Polyakova & Meserole, 2019, p. 1). Likewise, the Freedom House also used the term “digital authoritarianism” when referring to “a cohort of countries embracing the Chinese model of extensive censorship and automated surveillance systems” (Shahbaz, 2018).

Nonetheless, these definitions have a few drawbacks. On the one hand, they fail to acknowledge the existence of authoritarian behaviours outside the autocratic regimes since democracies are increasingly using control techniques in digital media that violate the human rights of all citizens (Glasius & Michaelsen, 2018a; Howard, Agarwal, & Hussain, 2011; Shirky, 2011; Walker & Ludwig, 2017). On the other hand, they fail to recognise that repression, control, and censorship not only happen through digital technologies but also by making them unavailable.

Therefore, the concept of digital authoritarianism in this analysis builds upon three ideas. First, the international practices approach, which will help to overcome the unbalance posed by the regime approach since authoritarian practices are not limited to authoritarian regimes, but are also exercised by democratic governments. Second, the essence of authoritarian actions derives from their constriction of freedoms, especially those related to communication and information access. The third idea is that failure to provide freedom in social media as the new space for the public sphere is against democracy.

4.1.1 International Practices Approach

Emmanuel Adler and Vincent Pouliot (2011) argue that the practices approach within the field of international relations allows for enquiring specific features or phenomena irrespective of the predetermined context in which it is embedded. Under this new ‘level of analysis’, the multiple practices that states decide to undertake are not necessarily attached to a regime but are the result of a particular local and global interaction, strategies, or the combination of elements. Furthermore, in the security field, Bigo and Tsoukala (2009) argue that the politics of terror—unbridled by the 9/11 attacks in the US—became a national security justification for using illiberal practices in liberal countries, and has eroded the democratic character of such countries. Although “they are not the result of exceptional decisions taken by the professionals of politics following a master plan, [t]hey are heterogeneous, globally incoherent, but highly predictable in their local effects” (Bigo & Tsoukala, 2009, p. 4).

Building on this approach, Glasius and Michealsen propose the concepts of digital illiberal and authoritarian practices and argue that the enactment of those poses three types of threats: “(1) arbitrary surveillance, (2) secrecy and disinformation, and (3) violation of freedom of expression” (2018b, p. 3796). The authors differentiate between illiberal and authoritarian practices for their case analysis, contending that illiberal practises “infringe on the autonomy and dignity of the person, and they are a human rights problem” (2018b, p. 3797), while the authoritarian practices “sabotage accountability and thereby threaten democratic processes” (2018b, p. 3797). Nevertheless, in line with the understanding of ‘authoritarian’ herein exposed, the authors also assert that “sustained illiberal practices may also come to constitute threats to the democratic process, and conversely, subversion of the democratic process typically also comes to threaten the autonomy and dignity of the individual” (Gla- sius & Michaelsen, 2018b, p. 3797). Contrary to Glasius and Michealsen’s study on the NSA surveillance program, in this analysis, there is no need to differentiate between illiberal and authoritarian; instead, it will claim that illiberal practices are ontologically authoritarian.

4.1.2 Authoritarianism

According to Robert Dahl’s theory of democracy, a regime is democratic “when it allows the free formulation of political preferences, through the use of basic freedoms of association, information, and communication [...] without excluding [...] or prohibiting any members [...] by the use of force” (Linz, 2000, p. 58). Authoritarianism, on the other hand, implies a “submission [from the civil society] to authority, as opposed to individual freedom of thought and action [...] [or a lack of] freedom to create opposition political parties or other alternative political groupings” (“Authoritarianism,” n.d.a). Non-democratic rulers interpret the laws themselves “with a wide range of discretion” (Linz, 2000, pp. 58-59).

Furthermore, some authors understand authoritarianism as the absence of democratic values (Brooker, 2008). In his essay Freedom as the Foundation, David Beetham contends that “democracy without freedom is a contradiction in terms” (2004, p. 62), and with the purpose of guaranteeing them, modern democracies expect to preserve a set of freedoms by dint of individual rights. Thereby, authoritarian regimes hinder collective actions such as “taking part with others in expressing opinions, seeking to persuade, mobilising support, demonstrating, and all those other activities intrinsic to the democratic process as one of

Table 3: Beetham’s subversions of democracy

Abbildung in dieser Leseprobe nicht enthalten

Source: Beetham, 2004, P- 67-69

public rather than private decision making” (Beetham, 2004, p. 62). In his model to assess democracy, Beetham identifies the two types of subversions of democracy. The generic subversions refer to the general hindrances inherent to the abuse of power and impunity of the judicial system, whereas the specific subversions refer to the violations of individual freedoms (see Table 3 below).

While the specific subversion might be applied to individuals, when mass communication channels such as the internet are disrupted in order to restrict public debate, intimidate journalists, impede public assemblies, hide information or obstruct its access, it acquires more general dimensions that endanger those fundamental freedoms of democracy.

4.1.3 Social media and the public sphere

The relations between the internet, particularly social media, and the public sphere have been thoroughly explored (Castells, 2008; Cela, 2015; Fuchs, 2014; Iosifidis & Wheeler, 2016; Johannessen, 2013). The core concept of public sphere or Öffentlichkeit was described in 1962 by Jürgen Habermas as the space, within the domain of social life, where the public opinion is formed through a communicative process in which people express their ideas, engage in a discussion, and come to an inter-subjective agreement (Habermas, 1962). While exploring the origin of the public sphere, the German philosopher explained how the appearance of the printing press in the 18th century put an end to the aristocratic privilege to access information and included the incipient bourgeoisie to the conversation, which prompted deep social and political transformations.

Habermas’ ideas are distinctly more robust at present. Information and communication technologies (ICTs) have proved to be essential enablers of the public sphere since they are potential tools for discussion and propagation of ideas, as well as essential means for organisation and association in an extensive scale (Bennett, 2012; Castells, 2008). Social media have spurred the democratisation of knowledge, has provided collective awareness on public affairs to a higher number of people, and has given voice to those who were formerly not included in the public debate.

Social media—defined as all “interactive [internet-based] platforms via which individuals and communities share, co-create, discuss, and modify user-generated content” (Kietzmann, Hermkens, McCarthy, & Silvestre, 2011, p. 241)—comprehends not only the popular service- based platforms such as Facebook, Whatsapp, YouTube, or Twitter, but also websites, forums, and blogs that provide a space for online communication and collaboration. Evidently, both concepts conflate one another. Social media platforms are virtual spaces where people communicate, inform, debate, forge an opinion, associate, and organise, giving place to a digital sphere or virtual sphere 2.0 (Paparachissi, 2009).

The relation between the democratic values earned in the public sphere and social media goes beyond semantics. “The phrase ‘digital democracy’ was coined to provide evidence that technology enhances democracy [T]he fact that political discussion and resistance is happening online and on social media is provided as evidence that more access to technology is increasing citizen participation” (Nyabola, 2018, p. 32). Furthermore, Manuel Castells stresses that the digital sphere constitutes a cornerstone for democracy because it “is an essential component of the sociopolitical organisation because it is the space where people come together as citizens and articulate their autonomous views to influence the political institutions of society” (Castells, 2008, p. 78).

The debate on the impact of social media on the new digital sphere has been the topic of many pieces of research. Some authors contend that social media and the internet “can be tools of oppression rather than emancipation” (Joseph, 2012, p. 145), since social media sites have been used to entrench dictators and threaten dissidents, making it more challenging to enhance the public sphere and promote democracy (Hindman, 2008; Morozov, 2011; Sunstein, 2008). Others, however, argue that social media has empowered citizens to the point that they build the power to subvert autocratic regimes (Diamond & Plattner, 2012; Paparachissi, 2009; Shapiro, 2000). In this manner, Manuel Castells claims that “ICTs have been the critical tools of power throughout history because through ICTs people are connected” (Castells, 2017). The network strengthens communities while it creates new protocols of communication and connectivity, facilitating the construction of a new public sphere locally and globally. This rearrangement of socio-political organisations signifies a new structure of power that uplifts democracy; however, this can be endangered if the government captures the communication channels (Castells, 2008).

Thus, the digital sphere is a new space enabled by the new communication channels and social networks that enhanced the public sphere, previously co-opted by traditional media. This ecosystem has not yet found consensus on what the endgame would be and what structures are needed to guarantee neutrality, as well as to find ways to make all actors accountable without transgressing their freedoms in the digital sphere. The acknowledgement of the risks that a one-sided power on the internet can bring to the public sphere is the basis for understanding that any intrusion on this ecosystem is against democracy, and, is, hence, digital authoritarian.

4.1.4 Definition

In this study, “digital authoritarianism” or “digital authoritarian practices” will be understood as the set of actions that a state executes to (a) block access, fully or partially, to the network or any internet-based services, or to (b) surveil online social media activity with the intention to identify and prosecute users unaligned with the state. These actions are limited to the digital sphere and are authoritarian to the extent that they curtail freedom of expression, freedom of association, and the right to access to information.

Drawing on this definition, we shall delineate when and why blockages and surveillance practices are harmful and under what conditions they should be considered authoritarian. There is a myriad of methods that would not be possible to address here. However, we will concentrate on the internet control techniques that limit the fundamental freedoms of democracy; in other words, those practices that hinder the political communication in the public (digital) sphere systematically or in strategic events. As mentioned in Chapter 2, when governments recognise that they are controlling the internet, they do it under the justification of security. However, the state has a “privileged role in determining what constitutes a threat to security and deciding upon a subsequent course of action” (Cross, 2019, p. 197). Thus, internet shutdowns, content blocking or filtering, throttling, and social media surveillance have been used mostly for political repression and intimidation in social media. In order to complete the definition of digital authoritarian practices, we shall delimit those control techniques that constitute a digital authoritarian practice.

First, internet shutdowns render all internet-based communication inaccessible intentionally “for a specific population, location, or mode of access, often to exert control over the flow of information” (Access Now, 2019, p. 1). Shutdowns are characterised as the uttermost digital authoritarian practices since they violate freedom of expression and assembly and hinder the access to information simultaneously, which has the consequence of disabling an important channel of communication, disrupting the underlying network of the digital sphere, and silencing voices massively. Moreover, cutting off the means of accountability in specific events like election periods hampers the collective awareness and transparency of these processes.

Second, content blocking or filtering “is a practice in which end users are denied access to certain online content” (Internet Society, 2015, p. 2). Filtering information and blocking social media make unavailable alternative points of view and generate secrecy and misinformation, and this limits the public sphere. For instance, in the case of some countries in SSA where Facebook or Whatsapp are the leading social media, disabling these platforms restrains political communication for a significant part of civil society that does not have access to other sources of information or channels to speak their minds. To narrow the scope of this technique, we will refer to social media blocking as a digital authoritarian practice.

Third, social media surveillance “refers to the collection and processing of personal data pulled from digital communication platforms” (Shahbaz & Funk, 2019, para. 1) and can be perpetuated with artificial intelligence or, in a more straightforward way, through infiltrators and ‘online patrols’ on social media platforms. According to Shahbaz & Funk, “governments have long employed people to monitor speech on social media, including by creating fraudulent accounts to connect with real-life users and gain access to networks” (2019, para. 3). In this manner, activists and journalists are being arrested and prosecuted for posting or organising meetings through social media. This method discourages people from using the network to communicate if they know that authorities are monitoring their activities. Hence, it violates privacy, freedom of expression, of association, and to access information.

Figure 1: Social Movements (2009-2019) in selected countries

This image was removed by the editing department for copyright reasons.

Source: ACLED

4.2 Conditions

In the qualitative comparative analysis framework, the configurational variables or conditions are not—individually—deterministic causes of the outcome (digital authoritarianism). On the contrary, QCA will help us find patterns or combinations of factors that possibly lead to the outcome while acknowledging that other factors which are not contemplated might also play a part, and that the absence of the outcome could be explained differently. In the following sections, the relation between the conditions that will be tested and the outcome will be given.

4.2.1 High Social Mobilisation

In the span of 10 years, social movements in SSA have flourished. While in 2009 there were 402 protests in the selected countries, in 2019, the same region registered 2,602 protests. In the case of riots and battles, the increase was threefold.

Social movements are the materialisation of collective action and political dissent in a networked civil society. Protests aim to make public demonstrations and express objection, disapproval, dissent or complaint (“Protest,” n.d.b), typically in opposition of government actions or inactions, in the form of rallies, marches, civil disobedience, riots, strikes, and others. When these various ways of protesting exceed the size or impact of what governments can manage, some counter-measures are usually imposed to prevent or control political instability. Nowadays, as Tufecki argues, “social movements are empowered by their ability to use digital tools to quickly mobilise large numbers of protesters” (Bosch, 2019, p. 67). The capacity of the internet to reach thousands of people in real-time allows for immediate mobilisation within and across countries and challenges governments globally.

Authors like Malcolm Gladwell argue that social media is not able to create a massive community unity to make change happen in societies, since the connections in service-based platforms are based on weak ties (Gladwell, 2010). Instead, activism on the internet tends to become slacktivism², producing only small changes and not the radical movements expected from them. Despite the pessimistic perspectives that undervalue the social media effects on social mobilisation, evidence has proven that civic organising around the world by dint of the new ways of communication has thrived. Many authors have acknowledged that information and communication technologies are primary enablers of the public sphere and thereby play a key role in the citizens’ ability to communicate and organise (Bennett, 2012; Gerbaudo & Treré, 2015; van de Donk, 2004). Moreover, ICTs not only intensify communication within the community members but also forge temporary alliances, and have been found to encourage ‘global solidarity’ or to create coalitions standing for “human rights, social rights, poverty, and environmental issues” around the world (van de Donk, 2004, p. 16).

The Arab Spring was the result of the new communication dynamics, where people were able to organise protests and movements to make changes in their socio-political structures. It was one of the most representative social movements triggered by internet-based platforms in the last years since it evinced the power of social media as an enhancer of public participation, discussion, and communication. Furthermore, it became a beacon of inspiration for other countries in Africa. Twitter and Facebook first proved their relevance “as a news reporting mechanism during the Egyptian uprising [and] provided a continuous stream of events in real-time throughout the crisis” (Papacharissi & Fatima Oliveira, 2012, p. 1). Although these platforms are also robust in Africa, WhatsApp is particularly more useful and popular in this region because of its low data usage (Dwyer & Molony, 2019; “How WhatsApp is used and misused in Africa,” 2019).

“Digital tools have changed the ecology of the public sphere and [s]ocial movements were quick to adopt these tools and to use them to challenge power” (Tufekci, 2017, p. 225). As mentioned in Chapter 2, social media has been the facilitator for social movements across SSA, letting people come together for different issues such as elections, economy, education, and corruption in Kenya, Zimbabwe, South Africa, Nigeria, Senegal, and Tanzania, among others. However, “governments have learned how to respond to digitally equipped challengers and social movements” (Tufekci, 2017, p. 225), and have undertaken more restrictive measures like censorship, shutdowns, or leaders’ arrests.

“Social media have become coordinating tools for nearly all of the world’s political movements, just like most of the world’s authoritarian governments (and, alarmingly, an increasing number of democratic ones) are trying to limit access to it [sic]” (Shirky, 2011, p. 30). According to the #KeepItOn coalition, there is evidence that a stark relationship holds between protests and internet shutdowns. In 2019, this organisation “recorded at least 65 internet shutdowns during protests” (Access Now, 2020b). What is more, for the period 2016-2019, internet shutdowns and content blocking occurred 45% of the times amid protests, 28% in elections, 10% during political instability, 6% while communal violence, and 11% during other or unknown events in SSA, as shown in Figure 2.

Although blockages are sturdily linked with electoral periods (Access Now, 2019, 2020b; Freyburg & Garbe, 2018; Howard et al., 2011), this analysis is not limited to these particular democratic processes. A myriad of social issues can cause protests—all likely to be politicised—and, if suppressed, also likely to threaten democracy. Therefore, the present analysis will not be limited to those events only. Despite this, the selected period for this

Figure 2: Most common triggers for internet blockages in SSA (2016-2019)

This image was removed by the editing department for copyright reasons.

Source: Access Now, KeepItOn 2016-2019 (See Appendix 2 for more details by country)

analysis encompasses election years-that could have driven to protests—for all of the subSaharan countries except for Burkina Faso and Tanzania, which did not hold presidential elections between 2016 and 2019.

In conclusion, social movements are considered an underlying condition for free countries to undertake digital authoritarian practices. The purpose of governments is to break collaborative actions through the interruption of communication services at strategic moments, and their justification is the preservation of social order. In this logic, we expect to find a positive relationship that shows that governments are being digitally authoritarian when their civil societies are more active and politically mobilised.

4.2.2 Strong Chinese influence

This condition considers that a strong influence of China in the region can yield to digital authoritarian practices given (a) the absence of political commitments to democracy or international partnerships that could keep them accountable, and (b) the likelihood of African States to emulate and mimic authoritarian internet policies in virtue of China’s strong soft power in the region.

The debate on China’s political influence in Africa has taken multiple forms. Despite the development projects and significant investments that had boosted African economies, Beijing’s intentions in the region have been questioned harshly. From “new colonialism” to the “debt-trap diplomacy”, theories argue that China is a predator so that aids, debt, and investments are used to take control of African natural resources, influence local politics, and expand autocratic regimes (Zhao, 2014).

However, the Chinese non-interference policy has made it hard to prove that all its influence is a by-product of a master plan (Chen Weiss, 2019, 2020; King, 2013; Zhao, 2014). This does not mean that Beijing’s assistance is free of interest. It might not be “mercenary, but motivated by the combination of economic interests and the need of political influence” (Zhao, 2014, p. 1038) posits Zhao, “business in Africa has helped China to cement alliances” (2014, p. 1038) and expand its market. Furthermore, non-interventionist interests are mutual: it is convenient for African countries to have a partner that does not judge or interfere with how things are done locally; and for China to keep access to raw materials (Rotberg, 2008). For their part, Western countries have criticised China’s non-interference policy since it makes possible partnerships with any government, including UN-sanctioned states, irrespective of their attitude towards corruption, violence or repression (Rotberg, 2008; Zhao, 2014). Thus, China’s non-involvement strategy threatens international efforts to drive African states to democracy and “is regarded by human rights advocates as perverse and obstructionist” (Rotberg, 2008, p. 12).

In this manner, this study understands China’s influence in Africa regarding the exercise of digital authoritarian practices not as a Chinese foreign policy to install authoritarian regimes in Africa purposely, but as the willing adoption of an ever more attractive governance model of the internet as an effect of Chinese soft-power. Digital authoritarianism on social media— comprised by content filtering, internet sovereignty legislation, and advanced surveillance systems—is regarded attractive “over individual liberty [...] free expression, open debate, and independent thought” (Walker & Ludwig, 2017, p. 9) given its seeming success to achieve political stability and state power.

In the essay Understanding and rolling back digital authoritarianism, Jessica Chen Weiss argues that the fact that “China is subsidising and selling this technology” (2020) does not mean that China or Russia have an intention to overthrow democratic foreign governments and replace them with autocrats. “Ultimately, the demand for these technologies and how they are used depend more on local political conditions than Chinese grand strategy” (Chen Weiss, 2020). The author identifies three factors that help understand why digital authoritarianism is spreading: 1) emulation of what some consider a successful technique to keep social media safe; 2) diffusion of technology through selling, financing, and building surveillance infrastructure, being USA, UK, Japan, and China the main vendors, as well as providing training for censorship (China); and 3) proposing a more robust legal framework for the digital space—or internet sovereignty—than the Western laissez-faire approach (Chen Weiss, 2020).

In this study, the de facto digital authoritarian version of African internet policy is assessed within the framework of the interdependencies and international influences. States rarely act isolated, or without a purpose—their actions depend on partnerships, geographical location, and international agreements. Regionally and globally, governments deal with very similar problems that might be addressed in ways similar to those undertaken by other states. To answer why states might behave similarly within the same period, Elkis and Simmons (2005) give a threefold elucidation: (a) similar problems are solved independently in similar ways; (b) the policy is the result of coordinated actions among a group of countries, an international regulator, or a hegemonic power; (c) governments, independently, “make their own decisions without cooperation or coercion but interdependently [because they consider] the choices of other governments” (2005, p. 35). Thereby, a state reproduces policies taken by other nations which it admires through “learning, imitation, bandwagoning, emulation, and mimicry” (Elkins & Simmons, 2005, p. 35).

This attraction, or the capacity to influence others, is understood by Joseph Nye (2004) as a type of power. The author’s insightful distinction of hard power and soft power is instrumental in understanding the possible role of Chinese presence in digital authoritarianism. On one side, hard power is tangible when some countries influence others through their economic, political or military superiority in a coercive manner, i.e. by applying sanctions, threats or incentives. However, soft power emerges when “a country may obtain the outcomes it wants in world politics because other countries -admiring its values, emulating its example, aspiring to its level of prosperity and openness—want to follow it” (Nye Jr., 2004, p. 5). Thus, this subtler facade of power attracts followers through shared values to beget alliances, and its primary sources dwell in the transmission of culture, political values, and foreign policies (Nye Jr., 2004).

China’s soft power in SSA Chinese policies “resonate powerfully in Africa” (Huang, 2008, p. 298). Its particular authoritarian style and its extensive use of technologies have become a beacon for developing countries that need to alleviate poverty and achieve political stability. According to an Afrobarometer survey, the top five factors that most contribute to the positive image of China in Africa are the following: investments in infrastructure and development (32%), low cost of Chinese products (23%), business investments (16%), support for African countries’ international affairs (6%), and the non-interference in the internal affairs of African countries (5%) (Trines, 2019).

As Joseph Nye asserts (2004), culture—in the form of commerce, personal contacts, visits, and exchanges—is a source of soft power. Therefore, it stands out that China has increased its presence in culture, education, commerce, politics, and investment within the African continent, and has gained ground on Western economies (Rotberg, 2008). Indeed, as a result of the Belt and Road Initiative (BRI), “China has become the largest new investor, trader, buyer and aid donor in a select number of important African countries, and a major new economic force in sub-Saharan Africa as a whole” (Rotberg, 2008, p. 3). African imports from

This image was removed by the editing department for copyright reasons.

Source: CARI, UNcomtrade

China (Figure 3) increased from 25,043 mdd in 2009 to 61,508 mdd in 2019³ (China Africa Research initiative, CARI-SAIS, 2020) representing about 22% of the total imports of the selected countries and 20% of all sub-Saharan African imports. Likewise, Chinese investment flows to SSA grew 355% going from 1,077 mdd in 2009 to 4,905 mdd in 2019 (China Africa Research initiative, CARI-SAIS, 2020), and almost fourfold in the selected countries for the same period.

It is worthy to point out that American investment flows in Africa have fallen substantially, and as may be seen in Figure 4, it shows an outward direction in 2016 and 2018. Conversely, Chinese FDI flows grew faster since 2016. For the selected countries within SSA, 37 BRI agreements for investment and construction contracts reached a total value of 18.8 bn (US$) in 2019 (see Appendix 3).

Finally, soft power is also transmitted through education and training. In this regard, the Chinese Ministry of Education issued the Education Action Plan for the BRI, which is openly a factor of soft power. President Xi Jinping mentioned that the plan aims to “improve [China’s] capacity for engaging in international communication so as to tell China’s stories

This image was removed by the editing department for copyright reasons.

Source: CARI, Statistical Bulletin of China’s Outward Foreign Direct Investment

well, present a true, multi-dimensional and panoramic view of China, and enhance our country’s cultural soft power [sic.]” (Trines, 2019, para. 9). The plan consists in “expanding the network of Confucius Institutes (CIs) in Africa” (Trines, 2019, para. 9) as well as providing scholarships to study in China, skill training, and grants. “Every year some 10,000 African civil servants receive vocational training in China at universities, government agencies, and companies in areas like agriculture, health care, or poverty reduction” (Trines, 2019, para. 15).

In the digital context, some authors suggest that China is exporting “digital authoritarianism” through the sale of surveillance technology and training to African officials (Polyakova & Meserole, 2019). As part of the training programs, “Beijing is cultivating media elites and government ministers around the world to create a network of countries that will follow its lead on internet policy” (Shahbaz, 2018, para. 27). According to the Freedom House, in November 2018,

China hosted a two-week Seminar on Cyberspace Management for Officials of Countries along the Belt and Road Initiative. Visiting officials toured the headquarters of a company involved in ‘big data public-opinion management systems,’ including tools for real-time monitoring of negative public opinion and a ‘positive energy public-opinion guidance system’. (Shahbaz, 2018, para. 28)

In conclusion, training, culture, investment, and trade encompass the multiple tools of soft power by which China might influence African norms and values. More specifically, the diffusion of its ideas may create configurational relations that make other states more prone to take the sovereignty perspective on internet issues with the intention of protecting their interests, at the expense of individual freedoms and democracy.

4.2.3 High Internet Penetration

In 2017, 27% of the total population in Africa had access to the internet (ITU, 2018b) and only 30% of the eligible population in 2019 used social media platforms (We are Social, 2019). Given that almost 66% of the African territory has 3G coverage, the disparities in the use of social media are significant. According to the Global Digital Overview (2019), 51% of Southern African population (Botswana, Lesotho, Namibia, South Africa, and Swaziland) used social media actively, followed by Western Africa (20%), Eastern Africa (13%), and Central Africa (11%).

Table 4: Internet penetration and usage by region

Abbildung in dieser Leseprobe nicht enthalten

Source: (ITU, 2018a)

Despite the evident backwardness on the internet and social media usage, Africa has undergone 66 blackouts in the analysed period, which positions it as the second region with

Figure 5: Social media penetration of eligible population – 2019 (%)

Abbildung in dieser Leseprobe nicht enthalten

Source: We are social (2019)

the most digital censorship in the world after Asia. A closer look to African social dynamics and current hype with technological solutions for development may suggest that the impact of social media in SSA is significant even when “those who are making politics are a subset of a subset of yet another subset—those who have access to electricity, then those who have access to the internet, and finally those who have accounts on social media” (Nyabola, 2018, p. 35).

Therefore, we argue that internet penetration should not be a determining factor for digital authoritarianism when other aspects are considered, for instance alternative channels on mobile phones such as SMS, radial impact of individual devices, and the importance of elites.The use of SMS and oral communication are different ways of transmitting messages originated on the internet (Dwyer et al., 2019; Ekine, 2010). After her research in Sierra Leone, Maggie Dwyer (2019) reveals

[...] that information that originates in social media is often spread by word of mouth to those not using smartphones or the internet. This raises the importance of the role of ‘information brokers’ who spread (selective) information among their local interpersonal networks. (2019, p. 4)

In addition to verbal communication, people tend to “share smartphones” (Dwyer et al., 2019, p. 106) and the material devised on social platforms, such as Whatsapp, Twitter or Facebook, turns out to be a source of information for conventional media and is disseminated in the newspapers, radio or TV (Dwyer et al., 2019).

Furthermore, mobile communication is more accessible for people to connect to the network than fixed landlines connections, since it gives many advantages in terms of costs. When people have a mobile phone, they have the option of acquiring pre-paid services, do not need to make fixed payments monthly, may have access to the internet through free WiFi, can change residence without incurring in expenses, can share the device, etc. Hence, mobile connectivity is very high and it is relevant to understand internet usage, since it is through these devices that many access the internet. As reported by We Are Social, Southern Africa holds 162% of mobile subscriptions compared to its population, the highest density in the world, followed by Western Africa with 86%, 62% in Eastern Africa, and 53% in Central Africa (2019). If we consider only the population older than 15 years old (58%), in all cases the number of subscriptions exceed the eligible population, meaning that there are users with more than one subscription.

Finally, “internet access is often limited to wealthy social elites, but these elites have a key role” (Howard et al., 2011, p. 230). The authors of Why Do States Disconnect Their Digital Networks? determined that the internet-based architecture of civil society’s infrastructure to coordinate has empowered single actors to incubate political communication and social mobilisation, and that this is the utmost peril for political stability is the desertion of the elites. “When the cohort of wealthy families, educated and urban elites, and government employees decide they no longer wish to back a regime, it is most likely to fail” (2011, p. 230).

Among the elites, traditional leaders still play an essential role in political communication and democratic processes in Africa (Bertelsmann Stiftung, 2012; Logan, 2008; Ngwainmbi, 1995; Tonwe & Osemwota, 2014). These actors are closer to communities than politicians and have advisory power. They have sometimes even “compet[ed] with local government officials” who have then co-opted “traditional authorities as a resource for communicating with and mobilising populations” (Logan, 2008).

In this way, the third condition based on internet penetration aims to test if, in the proposed configuration, the level of the internet plays an essential role for a government to decide whether to cut off the internet. An intuitive inference would suggest that the fewer people connected to the internet, the less the state wants to control it. However, as we elucidated, there are local characteristics in African dynamics that might explain that this is not always the case.

What is more, in the wake of making large leaps towards universal connectivity, subSaharan internet usage grew 78%, reaching 46% of its population in 2019. In the last years, communities have hung on to new technologies (many times requiring network connection) to implement projects that promise economic development. “South Africa, Botswana, Kenya, Tanzania, Uganda, Ghana, and Nigeria have been noted to be making encouraging strides in establishing an e-government presence” (Amoah, 2014, p. 160). By 2019, Kenya implemented one of the most advanced e-government platforms (MPESA) and other platforms that require digital authentication in order to access public services and comply with their fiscal obligations. Furthermore, start-ups in Kenya (UjuziKilimo, iCow, and M-Farm), Nigeria (Zenvus), Ghana (Farmerline and AgroCenta) and Cameroon (AgroSpaces) have developed SMS and web tools to automate and improve productivity in local farms (Ekekwe, 2017).

Thereby, internet shutdowns have proved to have an economic impact because businesses are not able to complete transactions and are perceived as unstable or not trustworthy (Internet Society, 2019; West, 2016). Small and medium business that rely on e-commerce for national and international transactions are greatly damaged. For instance, in 2019, Zimbabwe’s internet shutdown “reportedly cost the country USD 5.7 million each of the six days it was unavailable” (Internet Society, 2019, p. 4).

In conclusion, despite the low connectivity, the impacts of the internet on African society and politics have begun to be troublesome to several states in the region. Hence, some questions arise: Does internet penetration levels matter? Does a relative higher penetration rate is enough for states to control it? Is it in combination with other factors when it matters? This seemingly paradoxical relation between internet penetration and digital authoritarianism will be tested in this study, where it is attempted to find whether it is in combination with other factors that the variable of internet penetration has a richer explanatory value.

5 Methodology

5.1 Qualitative Comparative Analysis

Social and political problems are the result of complex arrangements of conditions rather than isolated factors (Mello, 2020). Over the past few decades, researchers have been interested in other ways to analyse social problems. To that end, qualitative comparative analysis (QCA) has made headway to study a small or medium number of cases as an alternative to statistical approaches and comparative case analyses (Sehring, Korhonen-Kurki, & Brockhaus, 2013).QCA is a set-theoretical research method used to find the multiple pathways (equifi- nality) or combinations of factors (conjunctural causation) under which a social phenomenon can occur (Mello, 2020; Schneider & Wagemann, 2010). Moreover, QCA is a configurational approach grounded in the twin concepts of necessity and sufficiency (Grofman & Schneider, 2009, p. 662). A necessary condition for an outcome “is always present when the outcome occurs. In other words, the outcome cannot occur in the absence of the condition” (Ragin & Rihoux, 2008, p. xix). Similarly, “a condition is sufficient for an outcome if the outcome always occurs when the condition is present. However, the outcome could also result from other conditions” (Ragin & Rihoux, 2008, p. xix). To illustrate this, Box 1 shows graphically that the relevant cells to determine necessity are those in the right column, which represents the occurrence of the outcome because the necessary condition should be present every time the outcome is. On the other hand, Box 2 shows that the relevant cells for sufficiency are those where the condition is present (upper row) because a sufficient condition ideally would be enough for the outcome to occur. In this vein, this study is an attempt to test whether a strong Chinese influence, high social mobilisation and high internet penetration are necessary or sufficient conditions for free countries to undertake digital authoritarian practices through a crisp set qualitative comparative analysis (csQCA).

csQCA is one of the variants of QCA⁴, based in Boolean algebra, that helps identify

Figure 6: 2x2 Relevant cells

Abbildung in dieser Leseprobe nicht enthalten

Source: (Mello, 2020)

patterns for conjunctural causation through their qualitative differences using a binary code where 1 indicates the presence of the condition and 0 its absence. Subsequently, the minimisation of Boolean algorithms will let us reach simplified logical representations of the intersections between variables excluding those that do not make a difference. To exemplify this, combination 1 shows that the presence of condition A, condition B, and condition C 111 lead to the outcome 1. Additionally, due to equifinality, combination 2 indicates that the presence of condition A and condition B, and the absence of condition C 110 also lead to the outcome 1⁵.

Abbildung in dieser Leseprobe nicht enthalten

It can be observed that both combinations yield to the outcome, irrespective of the value of condition [C], meaning that [C] may be removed and a reduced expression is obtained, also called as ‘prime implicant’ (combination 3) (Ragin & Rihoux, 2008)

Abbildung in dieser Leseprobe nicht enthalten

.Hence, for the outcome [O] to be present, the combination of conditions [A] and [B] must be given. Both conditions are necessary but not sufficient for the outcome; however, the inter section or combination of those [A*B] could form a necessary and/or sufficient combination for the outcome.

To get to this point, QCA follows the next steps: (a) calibration of raw data into the code values; (b) identification of necessary conditions; (c) formulation of a truth table that will contain all possible combinations and memberships; and (d) solution or path that lead to the outcome. To strengthen the analysis, some ‘measures of fit’ will be considered. The main parameters used to measure necessity and sufficiency for individual conditions and combination of conditions are consistency and coverage6. The consistency parameter “indicates the extent to which cases that share a specific combination of conditions also show the outcome of interest” (Mello, 2020, p. 11). The coverage, for its part, “reflects ‘how much’ a condition or combination of conditions accounts for the occurrence of the outcome” (Mello, 2020, p. 11). Additionally, the parameters of relevance of necessity (RoN) and the proportional reduction of inconsistency (PRI) will be used to complement the analysis of necessity and sufficiency, respectively.

QCA is adequate for this study since we aim to unravel the possible combination of conditions and alternative paths that leads states to take digital authoritarian behaviours. It will, therefore, allow us to analyse, context-based, the different causes that explain why some countries are being authoritarian and others are not.

Furthermore, the reason why this analysis is suitable for QCA rather than a statistical method is that regression methods estimate the isolated magnitude of the effects of a dependent variable over an independent variable on average. Conversely, QCA focuses on “the identification of conjunctions of factors that may be regarded as either necessary or sufficient for a given outcome” (Grofman & Schneider, 2009, p. 666). A critical difference for the purposes of this thesis is that, in statistics, “it does not matter how the dichotomous dependent variable is coded. Flipping the values from zero to one or one to zero will yield the same coefficients (just with the direction of the signs inverted)” (Grofman & Schneider,⁶ 2009, p. 669). On the contrary, in QCA, the combinations of conditions for the outcome to occur are not necessarily the exact opposite from those when the outcome does not occur; this is also known as asymmetric causality.

Although the number of cases is not a reason to prefer QCA rather than logical regressions, statistics render better results when N is larger. For more specific, regional cases where N is medium to small, like those analysed in this project, a qualitative analysis is a better option, because even when statistically the results may not be significant, one cannot disregard the relationship that can “still be theoretically, empirically and substantively highly informative” (Grofman & Schneider, 2009, p. 666). Finally, as mentioned before, the QCA framework is appropriate to identify set relations of sufficiency and necessity as the core of the ontological argument of causal complexity that will be addressed with more insight in Chapter 6. In the following section, a description of how the universe of cases is constituted will be given as well as an explanation of the sources of the raw data and the criteria for their further dichotomisation or calibration. Additionally, some limitations concerning the method and data will be identified.

5.2 Selection of cases

According to the United Nation’s regional classification of the world, the African continent is composed by 54 countries from which 8 (Morocco, Algeria, Libya, Tunisia, Egypt, Djibouti, Sudan and Somalia) are part of the Middle East and North Africa (MENA) region. The remaining 46 countries embody Sub-Saharan Africa (SSA).⁷

As mentioned before, the interest of this study are not formally authoritarian regimes; thus, only free and partially free countries in the sub-Saharan region were selected. To do so, the classification made by the Freedom in the World (FIW) Report 2019 was taken into account. This report did not differ from the reports of 2016 and 2017, except in the case of The Gambia, which scored as not free in the first two years but reached the score of partially free in 2018 and 2019. Additionally, Uganda and Zimbabwe passed from partially free to not free and vice versa in 2018; however, they hold the same score the rest of the years (see Appendix 4). For this reason, 2019 is considered representative for the analysed period (2016-2019).

From the universe of 46 countries that comprise sub-Saharan Africa, nine scored as a free and 21 as partially free. Additionally, five countries (Cape Verde, Comoros, Mauritius, Sao Tome and Principe, Seychelles) are excluded due to lack of information to determine the outcome, meaning they were absent in the data sets for digital authoritarianism, or there was no information for either of the conditions. The final universe is comprised of 25 countries, as shown below.

Table 5: Selected cases

Abbildung in dieser Leseprobe nicht enthalten

5.3 Operationalisation and Calibration

The variables, or conditions, are essential to understanding the outcome. Shutdowns, censorship, and individual prosecutions take place when the occurrence of the conditions converge, yet they may not happen on a daily basis. Protests, for instance, depend on local events such as the passing of laws by Congress, presidential election results, corruption cases, or other issues that may outrage the general population. Likewise, to measure the Chinese influence in one country, the economic relations in a period of time—and not just one year’s activity—are taken into account. For this reason, the conditions will cover a lapse of 4 years (2016-2019); this period could not be extended to prior years due to lack of available information. In order to test the relation of the outcome and the conditions using csQCA, the operationalisation and calibration (dichotomise) will be done with the following criteria.

First, the outcome, digital authoritarianism (da_cs), consists of three elements that in-

Table 6: Operationalisation and Calibration Summary

Abbildung in dieser Leseprobe nicht enthalten

dicate a logical disjunction (yes/no), which considers that, if a state practised internet shutdowns, social media blocking, or social media surveillance, then it will be coded either as (1) digital authoritarian; otherwise, as (0) for not digital authoritarian. As mentioned before, the ways to block end-users’ access to certain sources of web-based information are manifold. It may be a complete shutdown, meaning there is no access to the internet at all, or content-based, when the ISPs block specific IP addresses, ASNs, or services, such as Twitter, YouTube, Facebook, Whatsapp, or Telegram.

Access Now, an advocacy organisation that “defends and extends the digital rights of users at risk around the world” (Access Now, 2020a), coordinates the #KeepItOn coalition, which is a network of about 220 civil society organisations from 99 countries around the world (Access Now, 2020a). In collaboration with the #KeepItOn coalition, Access Now follows media reports about network interruptions, verifies the drop in traffic, inquires with local partners to understand the context around this events (Taye, 2020). Finally, it collects the data to generate the KeepItOn Report every year.⁸ Using this report’s issues from the period of 2016 to 2019, the cases will be analysed within all the categories of shutdowns (shutdowns_type*) except for throttling. This selection includes information regarding broadband internet, mobile internet, and the most important service-based platforms. As a complement, the reports from the Open Observatory of Network Interference (OONI) will be used to gather information about services and news websites blockages. Finally, due to the complexity of the measurement of social media surveillance, this study will only consider the visible effects of these practices, that is, when human rights violations occur in the form of torture, prosecution or conviction of individuals that made declarations on any social media. The source of this third element is the Social media surveillance & digital election interference database from the Freedom House and individual reports from ReportersI without Borders (RSF)^{9 10} related to posts in social media. In Appendix 5, a detailed record of the events can be found. The presence of at least one of the elements leads to full membership 1, otherwise the code non-membership 0 will be given.

Second, the condition of strong Chinese influence (china_cs) will be measured in terms of imports from China, investments and construction contracts that could be indicative of a strong Chinese presence in the host country, and consequently a stronger soft power. A classification of ‘strong influence’ or 1 will be given either when the share of China’s investment in the total FDI flows is greater than 40%, or when the share of the host country’s imports from China are greater than 70%; otherwise, the condition will be computed as 0. Both indicators (FDI and imports) aim to work as a proxy of the degree of Chinese presence in comparison with other nations’; therefore, even when there are some disparities on what each country and China report, and within sources, these numbers are still useful to indicate the magnitude of Chinese influence. The China Global Investment Tracker (CGIT)10 database from the American Enterprise Institute and the Heritage Foundation was used to obtain the value of China’s investments and construction contracts on technology, energy, agriculture, transportation, real estate, and other industries including the BRI related projects. Additionally, the total foreign direct investments (FDI) inward flow to each African country was retrieved from UNCTADstats¹¹ for the period 2016-2019. It is important to point out that (a) FDI flows are taken instead of stocks¹² because the investments and construction contracts payments are also ‘flow’ variables, and (b) the relation can be greater than 1 because “Chinese firms typically do not report spending increments on an annual basis [and] when broken down by country, by year, our investment figures are ‘lumpy’ as compared to actual flows” (D. Scissors, personal communication, April 19, 2020). Besides, contracts are not totally paid the year they are accounted but will be distributed in x number of years. To estimate the share of imports from China in the total imports, the value of exports from China to each African country was retrieved from the China-Africa Research Initiative^{13 14} data set, elaborated from the UNcomtrade reports. Lastly, the value of total imports was retrieved from UNCTADstats.

Third, data on social movements (smo_cs) was obtained from The Armed Conflict Location & Event Data Project (ACLED)14, which is a non-profit organisation that collects data of all political violence events and protests reported across the world. Due to the diverse nature of the mobilisations, it is not possible to establish the number of movements per year that are organic; in other words, there is not a determined number of protests that a country should have. Therefore, the median of the number of protests and riots¹⁵ in SSA (205) will be considered as a comparative threshold to determine whether social mobilisation is high 1 or low 0 in the region. This study excludes the rest of the categories (battles, explosions, violence against civilians, and strategic developments) because they are related to terrorist or paramilitary groups and, in most cases, cutting off the internet to avoid a terrorist attack can be considered a justified reason.

Fourth, internet penetration (intpen_cs) rates are obtained from the International Telecommunication Union (ITU)¹⁶, the UN agency for ICTs which allocates the worldwide spectrum for radio and satellites, and develops technical standards for international connectivity. Statistics provided by the ITU have some drawbacks: for instance, the agency collects official data from its members, which can have different methodologies to measure the individual internet usage, plus many low-income countries there might not be a complete record of these statistics. According to the ITU, South Africa had 56% of individual internet usage while Kenya, a highly digitalised country, had 18%. To overcome this, and acknowledging that the main form of connecting to the internet in Africa is through mobile phones, we will also consider mobile internet subscriptions. Hence, the compound index for internet penetration will capture individual usage (70%) and mobile ownership (30%), see Appendix 6 for more detail. The use of these weights is an attempt to keep the focus in the individual internet usage while also accounting for the connectivity through smartphones. According to the ITU stats, few countries hardly reach 90% of individual connectivity (the United States 87%, Germany 84%) and the average in the world was 55% in 2017. Internet penetration in Africa is low; therefore, the threshold to determine ‘high’ penetration is indicative and comparative within the region. In this thesis, this threshold will be the median of the compound index (40%).

The operationalisation of the variables is the preparation for the calibration. Following the thresholds established above to determine the membership of each case, the next table shows the calibrated values for all the cases¹⁷.

5.4 Scope, method and data limitations

This MA thesis attempts to better understand why free states in SSA decide to control by blocking, fully or partially, the internet; it does not attempt to make bold generalisations. While employing the QCA method and retrieving data for the analysis, some limitations were encountered which are noteworthy to underline.

Table 7: Calibrated data

Abbildung in dieser Leseprobe nicht enthalten

Perhaps the biggest critique of csQCA is the dichotomisation of values (Schneider & Wagemann, 2010). As mentioned before, crisp sets only give two values, 0 or 1, to the variables to categorise them as ‘fully in’ or ‘fully out’. This method of calibration ignores some nuances of membership and, thus, could render the analysis deterministic. The fuzzy sets method was developed to overcome this limitation; however, in this study, (a) the qualitative traits of the outcome [digital authoritarianism] could not objectively be classified as more or less authoritarian, (b) the challenges to document the outcome’s elements could bias the assertion of an ordinal value, and (c) an attempt to run fsQCA was made, but the results did not differ significantly, and the thresholds were unsubstantial.

Along these lines, the three elements for measuring digital authoritarianism cannot entirely capture what governments do to censor information. This limitation is especially significant in surveillance practices where we are not able to identify surveillance if it is not openly related to legal repercussions, such as an arrest or prosecution. Furthermore, governments can exert intimidation before the information is published or can accuse claiming other charges different than the information posted on social media; in those cases, the identification of censorship on social media falls out of the scope of this project.

Another sort of limitation for this research was the issue of the availability of data. As detailed in the preceding sub-section, the inconsistencies among the existing data for China—due to the differences in the reported period, the definitions of the concepts, and their interests to hide information—make cumbersome to measure the variable accurately. China is characteristically very secretive with its information. It does not publish complete information related to investments, loans, or official development assistance provided to other countries. Therefore, data sets for China were retrieved from specialised institutes that track Chinese activities, yet the investment amounts are estimated, and some may be misleading if taken individually rather than by periods. Another peculiarity is that China triangles some exports and imports with Hong Kong, making it more difficult to track. Likewise, the availability of internet statistics is seldom public, detailed, or transparent in its methodology.

The ITU, stopped publishing statistics in 2018 by country for public access, hence, the last internet figures were taken (2017). The rest of the sources vary significantly from one another. Therefore, to minimise the error, the data from the same source was used when available.

This analysis has a sample size of 25 countries of sub-Saharan Africa, which share particular realities and social structures among them but are very different from the rest of the world. For this reason, the conclusions of this study cannot be generalised to all free or partially free countries. In all cases, there might be more conditions around digital authoritarianism that, due to the scope and space of this research, it is not possible to cover here. For instance, ISP Ownership, specific social structures, or digital legal frameworks were contemplated but required of more extensive research, which was not possible in the timeframe and given other limitations imposed by the pandemic. Nonetheless, it is feasible to make reasonable generalisations about patterns of causality identified in the selected countries in the sub-Saharan region. More importantly, the results can be used as a starting point for further studies related to the incipient topic of digital authoritarianism, and amount to the current investigations on internet shutdowns, social media censorship and surveillance.

6 Empirical Analysis

This Chapter presents the qualitative comparative analysis between digital authoritarian practices and the conditions previously described in Chapters 4 and 5. The objective is to test the configurational relationships of social mobilisation, internet penetration, Chinese influence and digital authoritarianism in sub-Saharan Africa. From the theory and some empirical evidence, we expect that, regardless of the internet penetration, the higher the rest of the conditions are present in a country, the more likely the state is to take authoritarian control of the internet. In this study, the software R was used to conduct all the estimations. The complete code script, as well as the original R output, can be found in Appendices 8 and 9.

Taking as a starting point the calibrated data of Table 7, the QCA follows the standardised steps described by Schneider and Wagemann (2009; 2010) and the research design of Mello (2020). First, individual necessary conditions and individual sufficient conditions will be identified. Second, through the construction of the truth table, configurational sufficiency will be determined. And finally, the solution terms will be explained.

6.1 Individual Necessary Conditions

The necessity and sufficiency of conditions, or a combination of conditions, are key in the qualitative comparative analysis. Following the standards for QCA stated by Schneider and Wagemman, the first step is identifying the individual necessity, in other words, looking for conditions that are present whenever the outcome occurs, which does not mean that this condition causes the outcome alone, but its presence is required for it to happen (Mello, 2020). To do so, the parameter of consistency will be used. In crisps sets, this parameter measures the percentage of cases with outcome 1 that also have the condition 1. As it is usual, typically this value should be greater than 0.90 to claim that a condition is necessary. Table 8 shows the values of consistency for each condition, as well as the value of consistency

Table 8: Necessary Conditions

Abbildung in dieser Leseprobe nicht enthalten

for the absence of the conditions (J. Note that the consistency of china_cs has a value of 0.875, while the rest of the conditions have a consistency lower than 0.56. The proximity of china_cs consistency to the minimum threshold of necessity (0.90) suggests that it is “almost” a necessary condition.

To strengthen the necessity test, it is worth observing the other two measures that appear in the Table: coverage and relevance of necessity (RoN). This is only needed for those conditions that are close enough to 0.90 of consistency. On the one hand, coverage of necessity will indicate “the degree to which a condition ‘accounts for’ instances of an outcome” (Mello, 2020, p. 130). Because crisp sets work with whole numbers, this parameter can be measured as the share of condition X1 that also present the outcome in the total of cases with condition X1 [1].

On the other hand, RoN or relevance of necessity helps detect when a necessary condition is relevant or trivial, in the sense that the consistency may indicate as necessary a variable that is always present, irrespective of the outcome, so that these variables are ‘ubiquitous’ or ‘static’ (Mello, 2020) and, therefore, misleading. To do so, the RoN estimates the share of cases with condition X1 absent 0 in the total of cases that either does not show the outcome 0 or does not present the condition 0. For china_cs, the coverage and RoN parameters are high, 0.875 and 0.818 respectively, meaning that there is ground to claim that strong Chinese influence, in this study, is an ‘almost’ necessary condition. Furthermore, as shown in Table 7 of calibrated data, from the 16 cases coded as digital authoritarian 1, 14 cases

Table 9: Individual Sufficiency

Abbildung in dieser Leseprobe nicht enthalten

(Benin, Gambia, Ghana, Guinea, Ivory Coast, Kenya, Liberia, Niger, Nigeria, Sierra Leone, Tanzania, Togo, Zambia, and Zimbabwe) also have a strong Chinese influence 1, meaning that “almost every time” that a country turns to be digital authoritarian, the China factor is present.

6.2 Individual Sufficient Conditions

As for the analysis of individual sufficiency, the pof() function was run for each condition, which also resulted in that the condition china_cs had a high consistency (0.875). This indicates that “almost every time” that a country has a strong Chinese influence, that country is digital authoritarian. Although this result is significant and certainly gives a hint of what to expect, the sufficiency of the combination of conditions will be the focus of the analysis rather than the individual sufficiency.

6.3 Analysis of Sufficiency (Truth table)

This is the core of “Ragin’s original version of QCA is the truth table” (Sehring et al., 2013). The truth table is constructed with all the combinations of conditions that are logically possible. To estimate the possible combinations, the formula 2k is used, where k is the number of conditions. This study comprises three working conditions; thus, eight possible combinations. By the same token, the truth table also provides a consistency assessment of each configuration and “information about the empirical distribution of cases, and their connection to the outcome. Therefore, each row of the truth table is also a statement of sufficiency” (Mello,

Table 10: Truth Table

Abbildung in dieser Leseprobe nicht enthalten

2020, p. 145).When programming the truth table in R, we set the consistency threshold at 0.80, whereas the typical minimum standard is 0.75 (Mello, 2020). This parameter will “indicate which rows of the truth table should be treated as positive instances of the outcome and included in the analysis” (Mello, 2020, p. 152). In the cases where the incl or consistency parameter is greater than 0.8, the output (OUT) will be 1, and when it is below that value, it will be 0. As a result, we get the following table.

The distribution of cases with empirical evidence can be observed in column n, where the number of cases that fall in each configuration and that reached the inclusion cut of consistency is indicated. In this manner, the Truth Table provides perspective to observe whether the cases are clustered or very disperse, or whether there are configurations without empirical evidence (Mello, 2020). Note that rows 4, 3, and 8 show the output 1; this means that the solution will encompass 3 out of 8 configurations, and will comprise 16 cases. On another note, it is worth observing that row 7, with the combination 110, shows a question mark (?) in the output. This issue is known as limited diversity, and it occurs when there is empirical evidence associated with a particular combination of conditions (Mello, 2020; Schneider & Wagemann, 2010). Since only one row does not have empirical evidence, this should not represent any problem for the solution; however, it may affect the solution in the next section, as will be seen.

According to our truth table, three configurations lead to digital authoritarianism:

(1) ~smo_cs * china_cs * intpen_cs

The absence of social mobilisation AND a strong Chinese influence AND high internet penetration is a perfectly sufficient configuration, since both of its members—The Gambia and Zambia—have undertaken digital authoritarian practices. This is indicated by the incl or consistency parameter in the truth table (1.000).

(2) ~smo_cs * china_cs * ~intpen_cs

Similarly, the absence of social mobilisation AND a strong Chinese influence AND low internet penetration is also a sufficient configuration in seven cases (Benin, Guinea-Bissau, Liberia, Niger, Sierra Leone, Tanzania, Togo). However, its consistency parameter is slightly lower (0.875) due to a contradictory case (Guinea-Bissau) that does not show the outcome.

(3) smo_cs* china_cs * intpen_cs

Finally, members with the configuration high social mobilisation AND a strong China influence AND high internet penetration (Ghana, Guinea, Ivory Coast, Kenya, Nigeria, and Zimbabwe) are also strongly associated with the outcome (excluding Senegal). This combination presents 0.875 consistency on account of the contradictory case.

In addition to the consistency parameter, the truth table also displays the PRI or proportional reduction in inconsistency. This measure helps to “identify simultaneous subset relations in the analysis of sufficient conditions” (Mello, 2020, p. 133). Given that these kinds of relations are more likely to “happen with fuzzy-set data when a condition or combination of conditions is both a subset of the outcome and a subset of the non-outcome” (Mello, 2020, p. 133), this parameter is not relevant for crisp sets. This can also be inferred by looking at the formula of the proportional reduction in inconsistency. In crisp sets, the values are only 1 or 0; therefore, the minimum value of condition X (X;), the outcome (Y;) and the non-outcome (~Y;), represented in the second factor of the formula below, will always be 0, leaving only the factors in the left (X)min(X;,Y;) / EX;), which is the formula of consistency of sufficiency (see Chapter 5). For this reason, values in the incl and PRI columns of the truth table (Table 10) are the same.

Abbildung in dieser Leseprobe nicht enthalten

Thereof, we can conclude that the expressions ~smo_cs * china_cs * intpen_cs + ~smo_cs * china_cs * ~intpen_cs + smo_cs* china_cs * intpen_cs are sufficient for the outcome and, consequently, can be reduced according to Boolean logic to complete the analysis.

6.4 Solution Terms

The Boolean minimisation of the expressions drawn in the latter sub-section will lead to a solution term. There are three different types of solutions, depending on the treatment of the logical remainders. The complex (or conservative) solution does not consider the logical remainders because it only works with sufficient configurations. The parsimonious solution includes the logical remainders assuming they are sufficient. Lastly, the intermediate solution gives the option to select which logical remainder to include and under which assumption (Mello, 2020).

6.4.1 Complex Solution

The conservative or complex solution will only take into account the rows with empirical evidence with a consistency greater than the specified threshold (in this study, 0.80). The complexity coined to this solution originates from the commonly long solution terms it pro-

Table 11: Complex Solution

Abbildung in dieser Leseprobe nicht enthalten

duces (Rihoux & Meur, 2008).

Table 11 shows the reduced expression of the sufficient configurations and throws one solution path: the absence of social mobilisation AND a strong Chinese influence (Benin, Guinea-Bissau, Liberia, Niger, Sierra Leone, Tanzania, Togo; Gambia, and Zambia) OR a strong Chinese influence AND high internet penetration (Gambia, Zambia; Ghana, Guinea, Ivory Coast, Kenya, Nigeria, Senegal, and Zimbabwe) lead to digital authoritarianism.

This solution can be better grasped through the Venn diagram below (Figure 7). The Venn diagram generated in R shows the solution shaded in green. First, observe that the cases outside the set high social mobilisation (~smo_cs) and inside the set strong Chinese influence (china_cs) are covered in areas 2 and 23. Second, note that the cases inside the set strong Chinese influence (china_cs) and inside the set high internet penetration (intpen_cs) are areas 123 and 23. The darkest green, area 23, indicates that the combination china_cs AND intpen_cs is covered twice by both prime implicants. This is also visible in the solution table (Table 11) where Gambia and Zambia can be seen in both rows.

Albeit areas 123 and 2 are part of the solution, Senegal and Guinea-Bissau appear to be not digital authoritarian. “Senegal has a reputation of being one of the few stable democracies in West Africa” (USAID, 2018, para. 1). On the contrary, even though the government of Guinea Bissau has not repressed freedoms on the internet yet, this country has proved to be very authoritarian in traditional media. “In 2020, the military occupied the headquarters of

Figure 7: Venn diagram – Complex Solution

Abbildung in dieser Leseprobe nicht enthalten

the state radio and TV, accusing its journalists of ‘bias’ in favour of the opposition” (RSF, 2020). Therefore, this is a compelling case to follow in the next years in questions like: if any of the other conditions change, would Guinea Bissau’s government decide to control the internet? Would a large social movement trigger such control? Does this government even need anything else to change to do so?

According to the solution terms, it is possible to claim that—in the selected countries and with the current data—countries with a strong Chinese influence and high internet penetration are undertaking digital authoritarian practices. Moreover, under these circumstances, when there is a significant number of social mobilisations, many countries have also opted to impede access to social media or prosecute individuals.

Area 12 stands out for not having any empirical evidence to analyse (logical remainder), making it a decisive combination that gives room to other possibilities and could weaken the Chinese factor. Additionally, in the white area, note that Botswana and Mozambique, who are not digital authoritarian, are also not members of any set. These countries do not have a significant Chinese influence, social mobilisation, nor high internet penetration. To further support the empirical evidence provided by the indicators used in this analysis, underline Botswana’s close relationship with the UK and South Africa must be underlined due to the former’s membership to the Commonwealth of Nations. Mozambique, on the other hand, is partially close to China since it receives tourism and educational support from that country, yet it still has more robust relations with India, Netherlands, and South Africa (World Bank, 2020).

Lastly, area 1 and 13 both contain contradictory cases. Malawi and Mali were found to be digital authoritarian but do not fulfil any of the solution terms. An interesting fact is that even when these countries do not show Chinese influence, Mali recently joined the Belt and Road Initiative and Malawi’s media highlight the country’s closeness to China. What we can infer from these areas is that, according to the data collected and in the selected countries, having only high internet penetration and high social mobilisation has not been enough for states to restrict internet-based communications.

6.4.2 Parsimonious Solution

As mentioned before, the parsimonious solution includes the logical remainders in the solution minimisation assuming they show the outcome. When more expressions are included in the reduction, the formula provides a broader zone for generalisations (Rihoux & Meur, 2008). This solution leads to what is called “counterfactual reasoning” because it implies that we assume that if the logical remainder had a case related to its configuration, then

Table 12: Parsimonious Solution

Abbildung in dieser Leseprobe nicht enthalten

this would be associated with the outcome (Mello, 2020), which in fact is not proved. Thus, the hypothetical treatment of conditions may bias the solution. In this case, we only have one logical remainder the configuration 110, which represents high social mobilisation AND strong Chinese influence AND low internet penetration. Therefore, when it is assumed that this configuration presents the outcome, the following solution is obtained:

Unlike the complex solution, the parsimonious minimal expression draws only one—more general—condition. China_cs was already present in both paths of the complex solution, which means it was a superset that contained them. However, when we include the logical remainder (area 12), and assume its positive association with the outcome, the condition china_cs completes its presence and becomes the only superset of all possible paths. The following Venn diagram shows visually the superset, where the other configurations that lead to the outcome are encompassed by china_cs.

Figure 8: Venn diagram – Parsimonious solution

Abbildung in dieser Leseprobe nicht enthalten

6.4.3 Findings

Before delving into the findings, the following table summarises the theoretical relations exposed in Chapter 4. It is important to remember that the conditions are not being assessed individually, so the expectations here expressed only attempt to associate the qualitative attribution of each condition with the outcome.

Table 13: Directional Expectations

Abbildung in dieser Leseprobe nicht enthalten

The application of the QCA in this study aimed to find a configuration, rather than a solely variable, under which SSA states choose the digital authoritarian path. Acknowledging that social phenomena are the result of international and national configurations of factors, the complex solution provides us with a more insightful answer to the relation between social mobilisation, internet penetration, and Chinese influence in the region, and avoids making any additional assumptions. Considering this, the empirical analysis conducted in this study allowed to test the expectations shown in Table 13

The conservative solution supports the theoretical assumptions of this study partially. Based on theory, it was expected that high social mobilisation and strong Chinese influence would be present in digital authoritarian cases, which turned out to be true; however, more cases in other configurations not related to the outcome could have strengthened this conclusion. In the solution, we found that social mobilisation is only a decisive factor to restrict the internet when accompanied by high internet penetration and China’s influence. In this way, social mobilisation was not necessary for the Gambian and Zambian States to come out as digital authoritarian.

Furthermore, while testing if internet penetration played a role in shutdowns—and other ways of censorship—empirical evidence showed that the number of people able to access the internet was not an influential condition for some countries to undertake digital authoritarian practices. Even though internet usage indicators are not a straight forward, the results indicate that indeed particular social dynamics in African societies facilitate the communication within the communities, so that one smartphone with internet access impacts way more than one person. People without internet access is actually affected by social media they consume indirectly. Moreover, that political campaigns are directed to the elite.

Although these results are indicative of an interesting effect of the increasing influence of China over the African continent, it demands further research. Supranational institutions, such as the Economic Community of West African States (ECOWAS), are emerging as observers of online censorship with binding solutions to discourage digital authoritarian practices in West Africa. In 2017, Togo registered a social media blocking during an antigovernment protest (Krapiva, 2020), and later, in 2020, the same government shut down social media amid the elections. The Court of Justice for the ECOWAS declared that 2017 shutdown was illegal, fined the Togolese State, and mandated the protection of internet access under the law (Krapiva, 2020). Sino-Togolese relations are still firm. Recently, the African country evinced its political commitment with the Asian giant, when Togo backed up China on the Hong Kong national security Law that many have condemned. If the Togolese government complies, it may pinpoint the possibility that regional institutions could influence national attitudes towards the internet.

Additionally, Western influence derived mostly from post-colonial links, especially with the United Kingdom, and this turned out to be an interesting characteristic of those countries that have not taken the digital authoritarian perspective, such as South Africa, Namibia, Lesotho or Botswana.

7 Conclusions

The perception of the internet and social media as a threat to peace and security is spreading in the African continent. Sixteen out of the 25 selected countries have opted for restricting social media or prosecuting and arresting members of the opposition based on online publications. In Benin, The Gambia, Ghana, Guinea, Ivory Coast, Kenya, Liberia, Malawi, Mali, Niger, Nigeria, Sierra Leone, Tanzania, Togo, Zambia, and Zimbabwe internet control has had a political backdrop for restraining freedom of expression, preventing protests or silencing adversaries. Some of them, like Tanzania and Kenya, are making these practices legal by including punishments to those who disseminate “fake news” or incur in “defamation” on social media.

The misuses of social networks are no trifle, and neither are the authoritarian control exerted over them. Curtailing freedoms by means of service blockages or intimidation in social media threatens the public sphere and democracy. Habermas, Fuchs, and Castells coincide when stressing that the digital public sphere should be “free from state censorship and from private ownership” (Fuchs, 2014) to guarantee its neutrality. However, nowadays, sub-Saharan social media is being menaced.

Internet governance is one of the new challenges of global times. Its low degree of legislation and the lack of strong core institutions have let the authoritarian branch to gain ground (Flonk et al., 2020). Whereas China and Russia have enacted more elaborated systems and legal frameworks that fit their perspectives on internet governance, countries with less resources and institutions might find effective, less subtle ways to control social media and prevent internet-based communications. In this way, internet shutdowns, social media blocking, and surveillance are increasingly being used as political intimidation mechanisms in Africa.

In this study, we identified the group of countries that are undertaking digital authoritarian practices and the group that is not. As stated at the beginning, this thesis aimed to pinpoint the underlying conditions that make sub-Saharan free countries more prone to undertake digital authoritarianism practices by answering the following research question:

Under what conditions do free countries in sub-Saharan Africa undertake digital authoritarian practices?

We argued that any attempt to subvert freedom in the public sphere threatens democracy and, thus, is an authoritarian act. In the digital sphere, these restrictive actions are internet shutdowns, social media blocking and surveillance because the three of them curtail freedom of expression and of association, and access to information.

To assess that “such threats can be produced and diffused in transnational and multi-actor configurations” (Glasius & Michaelsen, 2018, p. 3795), a qualitative comparative analysis was conducted. The QCA solution fulfilled the initial hypothesis, which claimed that countries are adopting digital authoritarian practices in SSA where social mobilisation is high and Chinese influence is strong, regardless of the levels of internet penetration. Besides, by answering the sub-questions, it was possible to understand why cutting off the internet, blocking social media or surveilling social media are authoritarian practices that can be perpetrated by any government, even when it is not formally an autocracy. Furthermore, through the qualitative comparative analysis, it was found that Chinese soft power over some states is a constant condition in “almost” all of the cases that presented digital authoritarian outcomes. These countries could be taking the alternate sovereigntist approach, spearheaded by China, to make decisions over how to treat the internet domestically. The combination of this condition with protests and the degree of internet usage in those countries resulted sufficient for some countries to block the internet fully or partially and to prosecute inconvenient users.

QCA helped to find these combinations of factors that possibly lead to the outcome. However, it should be kept in mind that other factors, which are not contemplated might also play a part and that the absence of the outcome could be explained differently.

While the results cannot be generalised, nor is it possible to say that this is the only recipe for a state to behave in a digital authoritarian manner, a pattern can be seen within this configuration that becomes an interesting point of analysis for future research.

The results of this thesis show a relation between the increasing Chinese soft power and digital authoritarianism in the region. China’s influence and its interest on expanding their internet governance perspective should be investigated thoroughly. Finally, this thesis reinforces the need for an internet governance consensus and the recognition of digital rights in order to protect democracy.

All actors in society have gained an advantage with ICTs and the internet. Civil society has been able to organise, albeit so have their opponents (van de Donk, 2004). The power and counterpower of networked societies make it difficult to determine the final outcome of the use of technologies. However, the result will depend on our decisions since the internet is a tool that can be used for both purposes (Castells, 2017; Shirky, 2011). Digital democracy, then, will be built upon both its opportunities and its risks (Bennett, 2012; Shirky, 2011).

This thesis adopts the position that shutdowns, content blocking and social media surveillance restrict freedom of expression, freedom of association, and access to information. Thus, these practices are considered authoritarian regardless of the type of regime that adopted them. In the following sub-section, some public policy recommendations that might hinder the expansion of digital authoritarianism in SSA will be drawn.

7.1 Policy recommendations

- In the last Internet Governance Forum (IGF) celebrated in November 2019 in Berlin, the central focus of regulation was private firms and the norms around artificial intelligence, whilst the state was mostly treated as the primary regulator. In this way, there was little attention to its role—other than providing infrastructure for universal accessibility—as potential perpetrators of criminal acts themselves. Therefore, we suggest to include in the conversation, in the IGF and the WSIS, the much-needed boundaries for states to control and surveil the internet.
- Formulate regional and local legal frameworks to guarantee free connectivity to the internet. The inclusion of third-party organisations such as ECOWAS or the African Union could help to discourage such actions.
- Deter the laws that criminalise fake news and disinformation. Misinformation should be tackled with digital literacy or other systems of authentication. Leaving the power to the state or a private firm to decide what is true or not, right or wrong, might endanger democracy and justice.
- Acknowledge internet shutdowns, social media blocking and social media surveillance as digital authoritarian practices.
- Recognise digital rights, such as the freedom to access the internet and neutrality, as human rights.

References

Access Now (2019). KeepItOn Report 2018: The state of internet shutdowns around the world.

Access Now (2020a). #KeepItOn. Retrieved from https://www.accessnow.Org/keepiton/#coalition Access Now (2020b). KeepItOn Report 2019: Targeted, cut off, and left in the dark. Retrieved from https://www.accessnow.org/keepiton/

Adler, E., & Pouliot, V. (2011). International practices. International Theory, 5(1), 1-36. https://doi.org/10.1017/S175297191000031X (N.d.a). Authoritarianism: Encyclopeadia Britannica. Retrieved from https://www.britannica.com/topic/authoritarianism

Azmeh, S., Foster, C., & Echavarri, J. (2019). The International Trade Regime and the Quest for Free Digital Trade. International Studies Review, 0, 1-22. https://doi.org/10.1093/isr/viz033 Beetham, D. (2004). Freedom as the Foundation. Journal of Democracy, 15(4), 61-75. https://doi.org/10.1353/jod.2004.0057

Bennett, W. L. (2012). The Personalization of Politics. The ANNALS of the American Academy of Political and Social Science, 644(1), 20-39. https://doi.org/10.1177/0002716212451428 Bertelsmann Stiftung (Ed.) (2012). Transformation Index. Transformation Index BTI2012: Regional Findings West and Central Africa (1. Auflage). Gütersloh: Verlag Bertelsmann Stiftung.

Bigo, D., & Tsoukala, A. (Eds.) (2009). Routledge studies in liberty and security. Terror, insecurity and liberty: Illiberal practices of liberal regimes after 9/11. Milton Park, Abingdon, Oxon, New York: Routledge.

Boas, T. C. The Dictator's Dilemma? The Internet and U.S. Policy toward Cuba - SUMMER 2000. Retrieved from http://people.bu.edu/tboas/dictators_dilemma.pdf

Bosch, T. (2019). Social media and protest movements in south Africa #FeesMustFall and #ZumaMustFall. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 6683). London: ZED.

Brooker, P. (2008). Authoritarian Regimes. In D. Caramani (Ed.), Comparative Politics (2nd ed., pp. 103-116).

New York: Oxford University Press.

Castells, M. (2008). The New Public Sphere: Global Civil Society, Communication Networks, and Global Governance. The ANNALS of the American Academy of Political and Social Science, 616(1), 78-93. https://doi.org/10.1177/0002716207311877

Castells, M. (2017, December). Power and counter-power in the digital society. Making sense of the digital society, Berlin. Retrieved from https://www.hiig.de/en/events/power-counter-power-in-digital-society/

Cela, E. (2015). Social media as a new form of public sphere. European Journal of Social Sciences, 2(3). Retrieved from http://journals.euser.org/files/articles/ejser_may_aug_15/Erlis.pdf Chen Weiss, J. (2019). A world safe for autocracy?: China’s Rise and the Future of Global Politics. Foreign Affairs, July/August, 92-102.

Chen Weiss, J. (2020). Understanding and Rolling Back Digital Authoritarianism. Retrieved from https://warontherocks.com/2020/02/understanding-and-rolling-back-digital-authoritarianism/

China Africa Research initiative, CARI-SAIS (2020). Data. Retrieved from http://www.sais-cari.org/other-data Courtemanche, C. (2011, February 25). Internet, the engine room of revolution. The Sydney Morning Herald. Retrieved from https://www.smh.com.au/politics/federal/internet-the-engine-room-of-revolution-20110224- 1b71v.html

Cross, C. (2019). Cybercrime and the policing of politics in Tanzania. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 195-213). London: ZED.

Diamond, L. J., & Plattner, M. F. (Eds.) (2012). A journal of democracy book. Liberation technology: Social media and the struggle for democracy. Baltimore, Md.: Johns Hopkins University Press.

Diepeveen, S. (2019). A familiar refrain: political discourse and Facebook use in Mombasa, Kenya. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 214-235). London: ZED.

Dwyer, M., Hitchen, J., & Molony, T. (2019). Between Excitement and Scepticism. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 105-128). London: ZED. Dwyer, M., & Molony, T. (2019). Mapping the study of politics and social media use in Africa. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 1-19). London: ZED.

Ekine, S. (Ed.) (2010). SMS uprising: Mobile phone activism in Africa. Cape Town, South Africa: Pambazuka Press.

Elkins, Z., & Simmons, B. (2005). On Waves, Clusters, and Diffusion: A Conceptual Framework. The ANNALS of the American Academy of Political and Social Science, 598(1), 33-51. https://doi.org/10.1177/0002716204272516

Flonk, D., Obendiek, A., & Jachtenfuchs, M. (2020). Authority conflicts in internet governance: Liberals vs. sovereigntists? Global Constitutionalism, 9(2), 364-386. https://doi.org/10.1017/S2045381720000167 Freyburg, T., & Garbe, L. (2018). Blocking the Bottleneck: Internet Shutdowns and Ownership at Election Times in Sub-Saharan Africa. International Journal of Communication. (12), 3896-3916.

Frontline Defenders (2019, July 24). Nigeria: Ibrahim Garba Wala forcibly moved to a remote prison: Urgent appeal. Retrieved from https://www.frontlinedefenders.org/sites/default/files/ua_nigeria_ig_wala_240719_en.pdf Fuchs, C. (2014). Social media and the public sphere. TripleC, 12(1), 57-101.

Gerbaudo, P., & Treré, E. (2015). In search of the ‘we’ of social media activism: introduction to the special issue on social media and protest identities. Information, Communication & Society, 18(8), 865-871. https://doi.org/10.1080/1369118X.2015.1043319

Gladwell, M. (Ed.) (2010). Small Change [Special issue]. Annals of Innovation. (October 4, 2010): The New Yorker. Retrieved from https://www.newyorker.com/magazine/2010/10/04/small-change-malcolm-gladwell Glasius, M., & Michaelsen, M. (2018a). Authoritarian Practices in the Digital Age | Introduction. International Journal of Communication, 12, 3788-3794.

Glasius, M., & Michaelsen, M. (2018b). Illiberal and Authoritarian Practices in the Digital Sphere — Prologue, 12(ISSN 1932-8036), 3795-3813. Retrieved from https://ijoc.org/index.php/ijoc/article/view/8899/2459 Göbel, C. (2013). The information dilemma: How ICT strengthen or weaken authoritarian rule. Statvetenkaplig Tidskrift, 115(4), 385-402.

Grofman, B., & Schneider, C. Q. (2009, December). An Introduction to Crisp Set QCA, with a Comparison to Binary Logistic Regression. Political Research Quarterly, 62(4), 662-672.

Habermas, J. (1962). The structural transformation of the public sphere. Cambridge: MIT Press.

Hindman, M. (2008). The Myth of Digital Democracy. Princeton NJ: Princeton University Press.

Hoffman, S., & Taylor, E. (2019, November 14). EU-USRelations on Internet Governance.

Hofmann, J., Katzenbach, C., & Gollatz, K. (2017). Between coordination and regulation: Finding the governance in Internet governance. New Media & Society, 19(9), 1406-1423. https://doi.org/10.1177/1461444816639975 How WhatsApp is used and misused in Africa (2019, July 18). The Economist. Retrieved from https://www.economist.com/middle-east-and-africa/2019/07/18/how-whatsapp-is-used-and-misused-in-africa Howard, P. N., Agarwal, S. D., & Hussain, M. M. (2011). When Do States Disconnect Their Digital Networks? Regime Responses to the Political Uses of Social Media. The Communication Review, 14(3), 216-232. https://doi.org/10.1080/10714421.2011.597254

Huang, C.-H. (2008). China’s Renewed Partnership with Africa: Implications for the United States. In R. I. Rotberg (Ed.), China into Africa: Trade, aid, and influence / Robert I. Rotberg, editor (pp. 296-313). Washington, D.C.: Brookings Institution Press.

IGF (n.d.). About IGF. Retrieved from https://www.intgovforum.org/multilingual/content/about-igf-faqs Internet Society (2015, October 30). Policy Brief: Network Neutrality. Retrieved from https://www.internetsociety.org/policybriefs/networkneutrality/

Internet Society (2017, March 7). Perspectives on Internet Content Blocking: An Overview. Retrieved from ISOC website: https://www.internetsociety.org/resources/doc/2017/internet-content-blocking/

Internet Society (2019, December 18). Policy Brief: Internet Shutdowns. Retrieved from https://www.internetsociety.org/policybriefs/internet-shutdowns Iosifidis, P., & Wheeler, M. (Eds.) (2016). Palgrave global media policy and business series. Public spheres and mediated social networks in the western context and beyond. London: Palgrave Macmillan.

ITU (2018a). Core indicators on access to and use of ICT by households and individuals. Retrieved from https://www.itu.int/en/ITU-D/Statistics/Pages/stat/default.aspx ITU (2018b). Percentage of Individuals using the Internet. Retrieved from https://www.itu.int/en/ITU- D/Statistics/Pages/stat/default.aspx

Johannessen, M. R. (2013). Social media as public sphere: An exploratory study of online political communication in Norway (Doctoral dissertation). University of Agder, Kristiansand, Norway.

Joseph, S. (2012). Social Media, Political Change, and Human Rights. Boston College International and Comparative Law Review, 85(1), 145-188.

Kalyango, Y., & Adu-Kumi, B. (2013). Impact of Social Media on Political Mobilization in East and West Africa. Global Media Journal, 12(22), 1-20. Retrieved from http://www.globalmediajournal.com/open-access/impact-of- social-media-on-political-mobilization-in-east-and-west-africa.pdf

Karekwaivanane, G., & Mare, A. (2019). "We are not just voters, we are citizens": social media, the #ThisFlag campaign and insurgent citizenship in Zimbabwe. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 43-65). London: ZED.

Keller, J. (2011, September 4). The Middle Eastern Nation Most Vulnerable to Internet Shutdown. The Atlantic. Retrieved from https://www.theatlantic.com/technology/archive/2011/09/the-middle-eastern-nation-most- vulnerable-to-internet- shutdown/244507/

Kietzmann, J. H., Hermkens, K., McCarthy, I. P., & Silvestre, B. S. (2011). Social media? Get serious! Understanding the functional building blocks of social media. Business Horizons, 54(3), 241-251. https://doi.org/10.1016/j.bushor.2011.01.005

King, K. (2013). China's aid and soft power in Africa: The case of education and training / Kenneth King. African issues. Oxford: James Currey.

Krapiva, N. (2020). ECOWAS Togo court decision: Internet access is a right that requires protection of the law. Retrieved from https://www.accessnow.org/ecowas-togo-court-decision/

Linz, J. J. (2000). Totalitarian and authoritarian regimes. Boulder, Colo., London: Lynne Rienner.

Logan, C. (2008). Traditional Leaders In Modern Africa: Can Democracy And The Chief Co-Exist? (Working Paper no. 93). Cape Town, South Africa.

Lyon, D. (2007). Surveillance Studies: An Overview, Polity: Cambridge.

Masilo, B., & Seabo, B. (2015, December). Facebook: Revolutionising electora campaign in Botswana? Journal of African Elections, 14(2), 110-129. Retrieved from https://journals.co.za/content/eisa_jae/14/2/EJC183722

Meissner, M., & Wübbeke, J. (June 2016). IT-backed authoritarianism: Information technology enhances central authority and control capacity under Xi Jinping. China's Core Executive. Leadership styles, structures and processes under Xi Jinping (No. 1). Berlin. Retrieved from Mercator Institute for China Studies website: https://www.merics.org/sites/default/files/2018-07/MPOC_ChinasCoreExecutive_web.pdf

Mello, P. (2020). Qualitative Comparative Analysis: Research Design and Application. Washington, DC: Georgetown University Press (forthcoming).

Morozov, E. (2011). The net delusion: The dark side of Internet freedom. New York: Public Affairs.

Ngwainmbi, K. E. (1995). Communication efficiency and rural development in Africa: The case of Cameroon. Lanham: University Press of America.

Nyabola, N. (2018). Digital Democracy, Analogue Politics: How the Internet Era is Transforming Politics in Kenya. London: ZED.

Nyambura, H. (2018, December 1). Activist Alenga Torosterdt charged with publishing false information. Retrieved from #iFreeKe website: https://www.ifree.co.ke/2018/12/activist-alenga-torosterdt-charged-with-publishing-false- information/

Nye Jr., J. S. (2004). Soft Power: the means to success in world politics. New York: Public Affairs.

Orji, N. (2019). Social media and elections in Nigeria: Digital influence on election observation, campaigns and administration. In M. Dwyer & T. Molony (Eds.), Social Media and Politics in Africa: Democracy, Censorship and Security (pp. 152-172). London: ZED.

Papacharissi, Z., & Fatima Oliveira, M. de (2012). Affective News and Networked Publics: The Rhythms of News Storytelling on #Egypt. Journal of Communication, 62(2), 266-282. https://doi.org/10.1111/j.1460- 2466.2012.01630.x

Paparachissi, Z. (2009). The virtual Sphere 2.0: the Internet, the public sphere and beyond. In A. Chadwick & P. Howard (Eds.), Routledge Handbook of Internet Politics. London and New York: Routledge.

Polyakova, A., & Meserole, C. (August, 2019). Exporting digital authoritarianism. Washington, DC. Retrieved from Brookings website: https://www.brookings.edu/research/exporting-digital-authoritarianism/ (N.d.b). Protest: Merriam Webster Dictionary. Retrieved from https://www.merriam-webster.com/dictionary/protest

Ragin, C., & Rihoux, B. (Eds.) (2008). Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques: SAGE Publications.

Reporters Without Borders (2020). Africa. Retrieved from https://rsf.org/en/africa

Rights Activist, IG Wala Who Was Sent To Prison Over Facebook Post Receives Presidential Pardon (2020, April 21). Sahara Reporters. Retrieved from http://saharareporters.com/2020/04/21/just-rights-activist-ig-wala-who- was-sent-prison-over-facebook-post-receives-presidential

Rihoux, B., & Meur, G. de (2008). Crisp-Set Qualitative Comparative Analysis (csQCA). In C. Ragin & B. Rihoux (Eds.), Configurational Comparative Methods: Qualitative Comparative Analysis (QCA) and Related Techniques. SAGE Publications.

Rotberg, R. I. (Ed.) (2008). China into Africa: Trade, aid, and influence / Robert I. Rotberg, editor. Washington, D.C.: Brookings Institution Press.

RSF (2018, May 18). Kenya ’s journalists can now get ten years in prison for defamation. Retrieved from https://rsf.org/en/news/kenyas-journalists-can-now-get-ten-years-prison-defamation Runciman, C. (2017, May 22). SA is protest capital of the world. Independent Online. Retrieved from https://www.iol.co.za/pretoria-news/sa-is-protest-capital-of-the-world-9279206 Schneider, C. Q., & Wagemann, C. (2010). Standards of Good Practice in Qualitative Comparative Analysis (QCA) and Fuzzy-Sets. Comparative Sociology, 9(3), 397-418. https://doi.org/10.1163/156913210X12493538729793 Sehring, J., Korhonen-Kurki, K., & Brockhaus, M. (2013). Qualitative Comparative Analysis (QCA): An application to compare national REDD+policy processes (Working Paper 121). Bogor, Indonesia.

Shahbaz, A. (2018). The Rise of Digital Authoritarianism. Retrieved from https://freedomhouse.org/report/freedom- net/2018/rise-digital-authoritarianism

Shahbaz, A., & Funk, A. (2019). Social Media Surveillance. Retrieved from https://freedomhouse.org/report/freedom-on-the-net/2019/the-crisis-of-social-media/social-media-surveillance Shapiro, A. (2000). The Control Revolution: How the Internet is Putting Individuals in Charge and Changing the World We Know. New York: Public Affairs.

Shirky, C. (2011). The political power of social media. Foreign Affairs, 90(1), 28-41. Retrieved from https://www.j stor.org/stable/25800379 (N.d.c). Slacktivism: Lexico by Oxford Dictionary. Retrieved from https://www.lexico.com/en/definition/slacktivism Sunstein, C. (2008). Infotopia: How Many Minds Produce Knowledge. Oxford, New York: Oxford University Press. Taye, B. (2020, March 6). Interview by L. Mühleneimer [Netzpolitik]. By email.

Tonwe, D. A., & Osemwota, O. (2014). Traditional rulers and local government in Nigeria: a pathway to resolving the challenge. Commonwealth Journal of Local Governance, 128-140. https://doi.org/10.5130/cjlg.v0i13/14.3728 Trines, S. (2019, April 30). New Benefactors? How China and India are Influencing Education in Africa. Retrieved from https://wenr.wes.org/2019/04/how-china-and-india-are-influencing-education-in-africa Tufekci, Z. (2017). Twitter and tear gas: The power and fragility of networked protest. New Haven: Yale University Press.

USAID (2018). Senegal democracy and governance: Fact sheet. Retrieved from https://www.usaid.gov/senegal/newsroom/fact-sheets/senegal-democracy-and-governance-fact-sheet Van de Donk, W. (2004). Introduction: Social movements and ICTs. In W. van de Donk, B. D. Loader, P. G. Nixon, & D. Rucht (Eds.), Cyberprotest: New Media, Citizens and Social Movements. London: Routledge.

Walker, C., & Ludwig, J. (December, 2017). Sharp Power. Rising Authoritarian Influence. International Forum for Democratic Studies. Washington, DC. Retrieved from National Endowment for Democracy website: https://www.ned.org/wp-content/uploads/2017/12/Introduction-Sharp-Power-Rising-Authoritarian-Influence.pdf We are Social (2019). Digital 2019. Global Digital Overview. Retrieved from https://datareportal.com/

West, D. M. (October 2016). Internet shutdowns cost countries $2.4 billion last year. Washington, D.C.

World Bank (2020). World Integrated Trade Solution. Retrieved from https://wits.worldbank.org/countrysnapshot/en/MOZ WSIS (2005, December 18). Tunis Agenda for Information Society. Retrieved from ITU website: https://www.itu.int/net/wsis/docs2/tunis/off/6rev1.html

Zhao, S. (2014). A Neo-Colonialist Predator or Development Partner? China's engagement and rebalance in Africa. Journal of Contemporary China, 23(90), 1033-1052. https://doi.org/10.1080/10670564.2014.898893

8 Appendices

8.1 Appendix 1: List of Sub-Saharan Africa (UNDP)

Table 14: Appendix 1. Sub- Saharan Africa (UNDP)

Abbildung in dieser Leseprobe nicht enthalten

Source: UNDP

8.2 Appendix 2: Most common triggers for internet blockages in SSA (2016-2019)

Table 15: Most common triggers for internet blockages in SSA (2016-2019)

Abbildung in dieser Leseprobe nicht enthalten

8.3 Appendix 3: Chinese Investment in SSA (2016-2019)

Table 16: Chinese Investment in SSA (2016-2019)

Abbildung in dieser Leseprobe nicht enthalten

Sources: UNCTAD, China Global Investment Tracker (2016-2019)

1 Chinese Investment is not reported annually, this figure can account for contracts with a duration of 2 or more years
2 Chinese Investment may correspond to investment larger than a year and the FDI reported by the countries may not include all Chinese contracts, then the estimated share can be greater than 100%
3 Number of Belt and Road Initiative projects

8.4 Appendix 4: Freedom in the World Index 2016-2019

Abbildung in dieser Leseprobe nicht enthalten

8.5 Appendix 5: Digital Authoritarian Practices

Disclaimer: The numbers in Table 18 and the List of events do not represent the total of events but the total of reports. The cases found in different sources and that the specific date could not be verified could be double counted.

Table 18: Summary - Digital authoritarian practices by country

Abbildung in dieser Leseprobe nicht enthalten

List of events (type, year, source, country)

Abbildung in dieser Leseprobe nicht enthalten

8.6 Appendix 6: Internet Penetration Index

Abbildung in dieser Leseprobe nicht enthalten

8.7 Appendix 7: Raw Data

Table 21: Raw Data

Abbildung in dieser Leseprobe nicht enthalten

8.8 Appendix 8: Script Code for R

Abbildung in dieser Leseprobe nicht enthalten

8.9 Appendix 9. Consistency and Coverage Formulae

Figure 9: Formulae of Consistency

Abbildung in dieser Leseprobe nicht enthalten

Figure 10: Formulae of Coverage

Abbildung in dieser Leseprobe nicht enthalten

[...]

---

¹ The QCA refers to the independent variables as conditions that may determine the outcome or phenomenon. Conditions have a membership to the sets, which is determined by thresholds established in the calibration phase, so that a qualitative comparison can be made. In Chapter 5, a more detailed explanation of this method will be provided.

² The practice of supporting a political or social cause by means such as social media or online petitions, characterized as involving very little effort or commitment (“Slacktivism,” n.d.c).

³ 2019 values are estimated.

⁴ There are three principal variants: crisp sets (csQCA), fuzzy sets (fsQCA) and multi-value (mvQCA).

⁵ A condition preceded by a tilde (~) refers to the absence or value 0 in a given dichotomous variable. In the example, [~c] is read as: condition C is absent or low. Likewise, value 1 can be read as: condition C is present or high.

⁶ Formulas to calculate these measures are available in Appendix 10.

⁷ A complete table showing this distribution is given in Appendix 1.

⁸ https://www.accessnow.org/keepiton/

⁹ https://rsf.org/en

¹⁰ https://www.aei.org/china-global-investment-tracker/

¹¹ https://unctadstat.unctad.org/wds/TableViewer/tableView.aspx

¹² The flows are the transactions made in one year, while the stock shows the total cumulative value of the FDI in the host country.

¹³ http://www.sais-cari.org/other-data

¹⁴ https://acleddata.com/#/dashboard

¹⁵ Some protests may have violent events provoked by different groups or the police, categorising protests as ‘riots’.

¹⁶ https://www.itu.int/en/ITU-D/Statistics/Pages/stat/default.aspx

¹⁷ Appendix 7 contains all the raw data used for the dichotomisation.