Bertrand and Joseph are both hardworking, middle-aged Cameroonians with busy lives. Both are married with children; one is an auto mechanic and the other a taxi driver. They both completed primary school, and Bertrand went on to technical school while Joseph got his Ordinary-levels certificate. Both live in humble circumstances but manage to put food on the table and sponsor the education of their children, and neither has ever depended upon the state or government services to do so. Both were born in Cameroon in the late 1970s and therefore have only known one president – and one ruling party – their entire adult lives. Yet Bertrand considers himself a militant of the ruling party, while Joseph feels close to the largest opposition party. As a result, Bertrand believes that the ruling party wins in free and fair elections because it is genuinely popular and there is no other party that could better govern Cameroon. He is especially supportive of the ninety-year-old president, who he believes has the interests of ordinary Cameroonians at heart. As an opposition partisan, Joseph could not disagree more vehemently. When asked about his feelings toward the president he replied, “I don’t know if that man is an evil spirit or a human being.” He bemoans the country’s authoritarianism, economic malaise, and mismanagement of the secessionist crisis occurring in the western part of the country. Yet, despite the fact that elections in Cameroon have never resulted in a change in power – and despite their profoundly different views of Cameroonian politics – both men vote in elections and support their different parties.

What explains the radical political differences between these two men? What has driven them to such drastically different understandings of their own country? Some might argue that what differentiates them is their education or socioeconomic status. Citizens with higher levels of education are more inclined to support democracy and therefore would be more likely to reject the regime’s propaganda and support an opposition party. Alternatively, someone with low socioeconomic status might be likely to support the ruling party because they are more vulnerable to the vote-buying campaigns and social spending programs controlled by the party in power. Or perhaps, inversely, citizens with higher levels of socioeconomic status would support the ruling party because their livelihoods are more likely to be tied to public sector employment or clientelist linkages to the regime. Yet Bertrand and Joseph – like millions of other Cameroonians – have very similar levels of education and socioeconomic status. Neither of them works for or benefits directly from the state. And neither of them is from an ethnic group that is connected to the president.

This book will argue that by looking at these two men as individuals – by analyzing their ethnicity, their level of education, and their socioeconomic status – the political science literature on authoritarianism has missed the most important predictor of their political behavior: their social lives. Bertrand told me that he supported the ruling party because it provided peace and development. But as I dug deeper into his past and his childhood, he began to explain how the grandfather who raised him – a man who had served in the French colonial army during WWII and then spent much of his adult life as an active member of the ruling party – had taught him about the party and encouraged him to join. When I asked Bertrand to name the people in his life with whom he is closest today, he named ten people – his brothers, friends, a neighbor, a colleague, and a cousin – none of whom felt close to any opposition party. Half of them, in fact, were ruling party partisans like Bertrand. But this should not surprise us too much because Bertrand is from Bafia, a ruling party stronghold that has very little opposition presence. Born in a ruling party stronghold, Bertrand was socialized from an early age to support the ruling party and became grounded in a social network that reinforced his political beliefs.

The life of Joseph, the opposition partisan, mirrors that of Bertrand. When I asked Joseph why he supported an opposition party, he discussed his desire for change and hope for democracy. But as he described his childhood and adolescence, it became clear that he was raised in a very different social environment than Bertrand. Born in the opposition stronghold of Kumbo, Joseph was raised by two different uncles who were both supporters of Cameroon’s largest opposition party at the time. They frequently discussed politics in the household, and when Joseph left his childhood home to start his own family, there was no question which party he would support. Further, when I asked Joseph about who he spent time with today, he told me about thirteen social contacts, only one of whom supported the ruling party. Seven of his friends and family members, including his wife, who was the most important person in his life, supported opposition parties. In direct contrast to Bertrand, Joseph had been born in an opposition stronghold, socialized to support the opposition, and was now surrounded by a social network that reflected his political beliefs. Yet, if we had looked at these men in isolation, their politics would have remained a puzzle.

Broadly speaking, because we believe that support for democracy (or, alternatively, authoritarianism) is important for understanding the durability of autocratic regimes – as well as the likelihood of democratic consolidation following a political transition – political scientists have developed a number of theories to explain why some citizens living under authoritarianism support the regime in power. Most studies take a top-down approach to explaining political preferences in autocratic regimes, showing how citizens react to different government strategies of vote-buying (Blaydes Reference Blaydes2011), social spending (Magaloni Reference Magaloni2006), clientelism (Lust-Okar Reference Lust-Okar2006), patronage (Rosenfeld Reference Rosenfeld2020), repression (Frye, Reuter, and Szakonyi Reference Frye, Reuter and Szakonyi2014; Nugent Reference Nugent2020), or propaganda (Chapman Reference Chapman2021; Croke et al. Reference Croke, Grossman, Larreguy and Marshall2016). In contrast, the theory presented in this book will take a bottom-up approach to understand regime support, focusing on something that has been overlooked by political scientists working on autocratic politics: partisanship.

While partisanship in democratic regimes has been the object of intense study since the invention of mass public opinion polling, political scientists have largely ignored partisanship in the context of autocratic regimes. Yet, since the end of the Cold War and the rise of the electoral autocracy – the modal type of autocratic regime in the modern world (Levitsky and Way Reference Levitsky and Way2010; Morse Reference Morse2012) – millions of citizens living around the globe profess to support a political party, despite the fact that elections in their countries rarely (if ever) result in political change. Despite the general futility of elections, however, this book will argue that partisanship is crucial for understanding politics in these types of regimes. It is the foundation for understanding the legitimacy of the regime, support for democracy, participation in elections, and potentially a much wider set of outcomes such as political protest, violence, and perhaps even regime durability.

So why do people in autocratic regimes support political parties? This book argues that, despite all the irrationality and obstacles to participating in autocratic politics, people in these regimes are socialized into becoming partisans by their partisan friends and family. Citizens choose to get involved in politics when the people who are important to them – their parents, cousins, friends, neighbors, and colleagues – are involved in politics (Han Reference Han2009), and they almost always support the party that the majority of their closest friends and family support (Huckfeldt and Sprague Reference Huckfeldt and Sprague1995; Lazarsfeld Reference Lazarsfeld1948). A growing literature based on democratic regimes has documented the social foundations of partisanship (Green, Palmquist, and Schickler Reference Green, Palmquist and Schickler2002) and the immense power that social networks play in influencing political behavior (Sinclair Reference Sinclair2012). Even things such as happiness, obesity, and smoking (Christakis and Fowler Reference Christakis and Fowler2007, Reference Christakis2008; Fowler and Christakis Reference Fowler and Christakis2008) have been shown to be contagious through social networks. In the context of developing countries, social networks facilitate clientelism (Cruz, Labonne, and Querubín Reference Cruz, Labonne and QuerubÍn2017; Ravanilla, Haim, and Hicken Reference Ravanilla, Haim and Hicken2022), the provision of public goods (Wilfahrt Reference Wilfahrt2021), the adoption of new technologies (Ferrali et al. Reference Ferrali, Grossman, Platas and Rodden2020), and even rebellion (Lewis Reference Lewis2020; Parkinson Reference Parkinson2022). Recent work in autocratic contexts has shown that the act of voting itself can be explained by social networks (Eubank et al. Reference Eubank, Grossman, Platas and Rodden2021). This book argues that partisanship under authoritarianism can likewise be explained by the political socialization that occurs within social networks.

Despite growing evidence from the developing world that social relationships are key to explaining a host of other political and economic outcomes, we have relatively little empirical work on the importance of political socialization, particularly in electoral autocracies. Nearly every study of political behavior in dictatorships is premised on a materialist framework. The extant literature largely frames political behavior as an economic cost–benefit analysis, in which citizens choose to support a party if and when they receive an economic reward (or stand to lose economic privileges) for their partisanship. This book will show that a materialist approach to political behavior in electoral autocracies can only weakly explain political beliefs, identities, and behaviors and oftentimes leads to conflicting expectations about these different outcomes. I will argue that a framework premised in social identity theory can provide us with a unified theory of partisanship that can better predict who becomes a partisan, which political party they support, and what they choose to believe about politics in their country.

1.1 Why Partisanship?

In autocratic regimes, where elections do not substantively affect the balance of power, why should we care about partisanship? Apart from any consequentialist outcomes, I begin by proposing that we should care about partisanship in such contexts for normative reasons. Millions of people around the world actively undermine their own democratic rights by supporting autocratic ruling parties. Alternatively, millions of others risk life and limb to support opposition parties in the context of repression. Political scientists should be able to provide a unified theory of these political choices, even apart from considering the consequences of such beliefs and behaviors for broader political outcomes. Table 1.1 presents a global list of contemporary electoral autocracies. It includes any country that has been consistently coded as an “electoral autocracy” by V-Dem’s “Regimes of the World” variable (Coppedge et al. Reference Coppedge, Gerring and Knutsen2023)Footnote ¹ for the past ten years (2012–22). Because this is a study of partisanship, I exclude regimes with no formal ruling party or where parties do not last more than one election cycle,Footnote ² as well as those that have experienced regime change within the past ten years.Footnote ³ The table lists the name of the country’s ruling party, the year the current regime first began holding multiparty elections, the percentage of citizens reporting themselves to be ruling party partisans,Footnote ⁴ the percentage of the voting-age population who voted for the ruling party in the last presidential election,Footnote ⁵ and the total turnout of the voting age population.Footnote ⁶

From Algeria to Zimbabwe, there are thirty-three electoral autocracies with ruling parties around the world that are at least a decade old. They span the continents of Africa, Europe, the Middle East, Asia, Central and South America, though by far the most – nineteen in total – are found in Africa. Perhaps surprisingly, for most electoral autocracies, the percentage of citizens who reported feeling close to the ruling party in public opinion surveys is quite close to the percentage of votes from the voting aged population that the ruling party received in the last election – across all electoral autocracies with data, the average for the former is 40.8 percent and the latter is 39.3 percent (with a correlation of 0.68). Nonetheless, there is considerable variation in ruling party partisanship across the universe of electoral autocracies, ranging from feeble, elite-based parties like the Parti démocratique gabonais in Gabon (13.5 percent) to much more deeply rooted parties, such as Chama Cha Mapinduzi in Tanzania (58.1 percent). However, ruling party partisanship tends to be highest in the most repressive regimes, such as Rwanda (67.7 percent) or Tajikistan (84.6 percent), reflecting the problem of interpreting such data at face value.

Measuring opposition partisanship cross-nationally is even trickier,Footnote ⁷ but definitionally, ruling party partisans outnumber opposition partisans. Nonetheless, there is huge variation in the extent to which the opposition wins votes in these types of regimes. Some electoral autocracies remain extremely competitive, such as Turkey, Malaysia, or Honduras, teetering on the line between autocracy and democracy. Others, such as Djibouti, Rwanda, and Equatorial Guinea, hardly have any independent opposition parties, teetering on the line between electoral autocracies and closed dictatorships. Regardless, the data in Table 1.1 make clear that political participation in electoral autocracies – even longstanding ones such as Singapore, Algeria, or Cameroon, where ruling parties have been in power for decades – is hardly meaningless. Millions of citizens in these regimes, oftentimes a majority of citizens, profess their support for political parties and go out to vote in elections.

Nonetheless, partisanship in autocratic regimes, particularly ruling party partisanship, is often assumed to have little meaning, apart from the economic advantages it can bestow. In fact, the literature on political behavior in autocratic regimes has rarely even considered partisanship directly, focusing instead on proximate measures of regime support, like voter turnout or approval of the president. Traditionally, the literature has mostly argued that autocratic political parties are ideologically empty,Footnote ⁸ and therefore partisan identities are often assumed to be transitory and superficial. Where partisan attachments are taken seriously, they are usually understood to be built primarily upon materialist calculations, such as access to clientelism, or relatedly, other social identities (primarily ethnicity), premised within a materialist framework. What do we gain by taking partisanship seriously as a social identity?

Although the focus of this book will be on explaining the origins of partisan attachments, understanding what differentiates ruling party partisans from opposition partisans has important implications for a potentially large host of political outcomes in electoral autocracies. First and foremost, this book will explicitly argue that partisanship is a strong predictor of political beliefs, namely, the perceived legitimacy of the regime as well as beliefs about democracy. Ruling party partisans are considerably more likely than opposition partisans to trust the various institutions of state and view their country as democratic. However, we might think that partisan divides may predict other aspects of public opinion, such as positions on public policy, beliefs about international actors (e.g., China or the United States), or even more fundamental issues, like attachment to a national identity or trust in other citizens.

Further, if partisanship predicts trust in the autocratic regime and beliefs about democracy, then it may be central to understanding important political behaviors. The most obvious of these is voter turnout. The existing literature has overwhelmingly argued that people vote for autocratic parties because they have received (or expect to receive) an economic reward for doing so. And although many people in these regimes vote exclusively for economic reasons, this explanation is obviously partial. How could Recep Tayyip Erdogan’s Justice and Development Party have paid off 29 million Turks in the 2023 election? In a country such as Burundi, one of the poorest in the entire world, how could the regime incentivize 3 million Burundians – more than half of Burundi’s voting aged population – to go to the polls? There must be something else that is leading people to participate without receiving a reward, and this book will argue that this “something else” is driven by social considerations. People identify with political parties because they have been socialized to do so, and when these attachments are strong enough, people will dedicate their scarce time and resources to support these parties, even when they expect nothing in return.

Understanding partisanship as a social identity may help to explain why millions of people receive gifts during elections but then decide not to vote or to vote instead for the opposition (Greene Reference Greene2022). It can also explain why people living in opposition strongholds – areas of the country that everyone knows will be “punished” by the regime through weaker social spending – will still vote for the ruling party. Ruling party social networks, isolated as they may be in an opposition stronghold, help sustain ruling party support in the absence of economic motivations – and vice versa. In general, understanding partisanship as a social identity can help to explain high voter turnout for ruling parties in some of the poorest, most corrupt countries in the world, like Chad, the Republic of Congo, Equatorial Guinea, Togo, Uganda, or Cameroon. Better understanding the origins of these identities and how they change may be central to explaining variation in political participation over time, between different regions, and across different countries.

Thinking of partisanship as a social identity may help to explain other forms of political behavior as well. For example, research has shown that citizens in autocracies who believe that elections are free and fair are less likely to protest, even when they disapprove of the regime’s economic performance (Williamson Reference Williamson2021). Viewed through the framework of partisanship presented in this book, this may help to explain why many electoral autocracies, particularly those in sub-Saharan Africa, have weathered decades of economic decline and worsening corruption. The legitimating ideology of ruling party partisanship may tamp down protest and stave off regime change “from below” (Matovski Reference Matovski2021; Sharafutdinova Reference Sharafutdinova2020). Relatedly, understanding the social underpinnings of opposition support may help to explain continued defiance and protest in the face or repression (Pan and Siegel Reference Pan and Siegel2020). LeBas (Reference LeBas2011) has convincingly shown that opposition parties which can successfully create cross-cutting partisan identities are less prone to fragmentation. Thus, understanding the origins of these identities might be key to understanding when and why opposition parties, on rare occasion, actually defeat incumbents.

Partisan divides may even explain outcomes following a regime transition (Wittenberg Reference Wittenberg2006). There is a growing literature that is increasingly finding that autocratic regimes leave a lasting legacy on political beliefs and behaviors long after they have fallen (Neundorf and Pop-Eleches Reference Neundorf and Pop-Eleches2020). These include less support for and satisfaction with democracy (Neundorf Reference Neundorf2010; Pop-Eleches and Tucker Reference Pop-Eleches and Tucker2017), higher levels of support for the previous regime (Mishler and Rose Reference Mishler2007), the emergence of political trust (Mishler and Rose Reference Mishler and Rose2001), as well as lower levels of civic and political participation (Bernhard and Karakoç Reference Bernhard and Karakoç2007; Ekiert and Kubik Reference Ekiert and Kubik2014). Given the theoretical framework presented in this book, many of these findings may be conditional upon partisanship during the autocratic period. Indeed, Neundorf, Gerschewski, and Olar (Reference Neundorf, Gerschewski and Olar2020) find that it is only citizens who were part of the “winning coalition” under the former autocratic regime who were more critical of democracy following a regime transition. In sum, although levels of partisanship may not change election outcomes in autocratic regimes, they may predict things that are arguably just as important, including beliefs about democracy, participation in protest, regime durability, and perhaps even democratic consolidation in the wake of regime change.

1.2 The Theory in Summary

The core theoretical insight of this book is that we can better understand the political beliefs and actions of ordinary citizens in electoral autocracies if we think of partisanship as a social identity as opposed to only a rational response to economic incentives provided by the regime. Drawing most explicitly on recent work by Keith Weghorst (Reference Weghorst2022) in Tanzania, but also the broader understanding of social networks elucidated by Sarah Parkinson (Reference Parkinson2022) in the context of Lebanon, I argue that while most existing frameworks of political behavior focus on “snapshot” moments before elections to explain the choices of political actors, we must instead look to the broader social contexts of people to fully understand their choices. Using this framework for understanding partisanship, the book makes three interrelated arguments.

First, building on recent work by Laebens and Öztürk (Reference Laebens and Öztürk2021) in Turkey, I show that ruling party partisanship and opposition partisanship are unique social identities that can be described by specific sets of beliefs, practices, and shared in-group and out-group perceptions. These identities are built upon the nature of political parties in electoral autocracies. Although no two parties are exactly alike, I argue that ruling parties across all autocracies possess a defining set of characteristics that produce a common identity of “ruling party partisan” in these regimes. Similarly, opposition parties in such regimes are also unified by a specific set of characteristics that produce a common identity of “opposition partisan.” Specifically, ruling parties campaign primarily on issues of economic development and security (Matovski Reference Matovski2021), while opposition parties overwhelmingly campaign on the issue of democracy (Letsa Reference Letsa2019; Paget Reference Paget2023). Independent opposition parties operate on the premise that the regime is autocratic while ruling parties contend that elections are free and fair, the regime is democratic, and the opposition always loses simply because it is unpopular. As a result, the defining characteristic of partisan identities in electoral autocracies – and what makes them unique from partisan identities in more democratic regimes – is polarized beliefs about democracy. While both groups profess to prefer democracy to authoritarianism as generic forms of governance, opposition partisans are far more likely to think that the current regime is autocratic while ruling party partisans believe the opposite.

Although affective polarization is currently on the rise in democratic regimes (Iyengar et al. Reference Iyengar, Lelkes, Levendusky, Malhotra and Westwood2019; Wagner Reference Wagner2021), leading “out-of-power partisans” to increasingly believe that the other party won through undemocratic means (Gidengil, Stolle, and Bergeron‐Boutin Reference Gidengil, Stolle and Bergeron‐Boutin2022; Kingzette et al. Reference Kingzette, Druckman, Klar, Krupnikov, Levendusky and Ryan2021; Orhan Reference Orhan2022; Svolik Reference Svolik2020), the democracy cleavage in democratic regimes is not universal in the way that it is in autocratic regimes, where there is rarely, if ever, a change in power between parties. Thus, opposition and ruling party partisan identities in electoral autocracies are common across the universe of cases because of their distinct and shared sets of beliefs as well as their in-group and out-group perceptions. I argue that the partisan-motivated reasoning uncovered by Elizabeth Carlson (Reference Carlson2016) in Uganda and the political polarization described by Adrienne LeBas (Reference LeBas2011) in Zimbabwe are structural features of partisanship in electoral autocracies.

Second, I argue that people are actively socialized by the people in their lives to adopt these partisan identities. Citizens learn partisan political beliefs from their friends and family: The key to explaining partisanship in electoral autocracies is the political socialization that occurs through social networks. Social networks are built from childhood and include family, friends, religious communities, and work colleagues. Existing work has shown in democracies that citizens who have social networks active in politics are more likely to join themselves (Sinclair Reference Sinclair2012) and that citizens hold more politically congruent opinions when surrounded by their co-partisans (Klar Reference Klar2014). Building on these findings, I argue that whether you come to support the ruling party, an opposition party or no party at all depends decisively upon whether the people in your life support that party as well.

As a consequence of this argument, I propose three different mechanisms of socialization. First, ordinary people are more likely to be socialized by close contacts rather than weak ties (Harmon-Jones et al. Reference Harmon-Jones, Greenberg, Solomon and Simon1996; Jones and Volpe Reference Jones and Volpe2011). As a result, I argue that childhood and adolescent socialization within family homes should play an important role in the transmission and adoption of partisan identities (Jennings and Niemi Reference Jennings and Niemi1968). Children raised in partisan homes should be considerably more likely to adopt the partisan identities of their parents than children raised in nonpartisan homes. However, political socialization is a life-long process, and many people – particularly (though not exclusively) those raised by nonpartisans or weak partisans – end up being socialized to support a party in adulthood. I propose, therefore, that a second key condition in the transmission of partisanship is the strength of the partisan identities of those around you. Building on work by Baldassarri and Grossman (Reference Baldassarri and Grossman2013), who find that strong group attachment leads to higher levels of prosocial behavior, I argue that ordinary citizens are more likely to be socialized by strong partisans than weak partisans. As a result, I propose that party activists should be key points of socialization – these strong partisans should be much more politically influential within their social circles than ordinary partisans. This argument echoes the work of Baldwin (Reference Baldwin2016), who convincingly shows that opinion leaders like traditional chiefs can play an outsized role on vote choice within their communities.

Finally, drawing most explicitly on Americanist literature (Huckfeldt and Sprague Reference Huckfeldt and Sprague1995; Sinclair Reference Sinclair2012), I propose that partisan homogeneity within social networks should also be important to explaining individual-level partisanship. In other words, one of the best predictors of whether someone supports a political party, and further, whether that party is the ruling party or the opposition, will be whether someone spends their time with mostly nonpartisans, mostly ruling party partisans, or mostly opposition partisans. I further show that, on average, due to the nature of autocratic politics, ruling party networks tend to be much more politically homogenous than opposition networks, and thus opposition partisans face higher levels of cross-partisan influence. Finally, I show that the partisan homogeneity of networks itself is not driven by homophily – collective levels of education, ethnicity, or access to the regime – that might undergird partisan homogeneity. By measuring the key demographic features of both ordinary citizens and the people in their social networks, I show that the primary sources of homophily do not explain the strong correlation between co-partisanship within networks. I also contend that partisan homogeneity is not itself entirely ecological. By leveraging data based on life timelines, I show that people don’t primarily self-select into partisan networks because of their preexisting partisan preferences. Instead, I argue that people are influenced by their networks and that political contagion has the ability to turn people into partisans.

Building on this core framework, the final part of the theory focuses on the orientations of partisanship within social networks. Given that social networks can determine individual-level partisanship, what explains why some social networks are made up of nonpartisans while others contain mostly ruling party partisans or opposition partisans? Simply put, I argue that it depends heavily on where you live: Structural homophily plays an outsized role on partisan orientations in electoral autocracies (Blau Reference Blau1977; Feld Reference Feld1982). Drawing on recent work by Michael Wahman (Reference Wahman2023) in Southern and Eastern Africa as well as Zambo (Reference Zambo2004) in Cameroon, I show how party politics are strongly territorial in nature, and this territoriality sharply circumscribes the choices of partisanship available to ordinary citizens. In order to explain the divergence between the political orientation of networks, it is therefore crucial to understand the unique political geography of electoral autocracies. Unlike democracies and full autocracies, access to political information is not spatially homogenous in these types of regimes. Over decades of undemocratic electoral competition, opposition parties are only able to compete consistently in a handful of constituencies, while the ruling party dominates in much of the rest of the country. In turn, this geography of “strongholds” restricts the ability of ordinary citizens to affiliate with opposition parties (Wahman Reference Wahman2023; Zambo Reference Zambo2004). Because of the lack of a free media and the geographic isolation of the opposition, partisan identity becomes quarantined to its geographic strongholds (Paget Reference Paget2022). As a result, partisanship is largely a function of geographic proximity to opposition parties, with social networks acting as the primary mechanism linking parties to citizens. Electoral districts that are proximate to opposition activities will produce opposition networks, whereas districts core to the ruling party (with little to no opposition activity) will heavily feature ruling party networks.

Finally, coming full circle, I show that this political geography has broader effects on society beyond the production of partisan social networks. Not only do political parties produce a unique set of political views that are adopted by their partisan adherents, but these views also leech out into geographic space, affecting nonpartisans and even cross-partisans in party strongholds. In other words, I show that even nonpartisans and ruling party partisans in opposition strongholds hold much more skeptical views of the government than do their counterparts who live in ruling party strongholds (and vice versa).

1.3 Research Design and Methodology

1.3.2 Data Collection

Thus, I test the theory of this book in Cameroon because it is a stable, long-standing electoral autocracy with active, independent opposition parties and a relatively permissive political environment for gathering public opinion data. Before discussing data collection, I note that my theory was developed during fieldwork for a separate research project, during which I lived in Cameroon for a year (2014–15), personally interviewing 576 ordinary Cameroonians as part of 2,400-respondent public opinion survey implemented across 7 of Cameroon’s 10 regions, including the Anglophone Northwest and Southwest regions. Quotes from these survey interviews are sprinkled throughout the book, as well as quotes from party elites whom I interviewed during this time – including the late John Fru Ndi and Adamaou Ndam Njoya, two of the most respected opposition leaders in Cameroon. However, I do not draw from the survey data itself to test my theory, as that year of fieldwork and the data that I collected laid the groundwork for generating the theory presented in this book.

Instead, I use different sources of data collected in 2021 and 2022 to explain the social origins of partisanship. In order to understand and elucidate the process of political socialization, I conducted life history interviews with, and then surveyed the full social networks of twelve ordinary Cameroonians from three separate research cites: Boumnyebel (a rural opposition stronghold), Bafia (a rural ruling party stronghold), and Yaoundé, the capital of Cameroon, which is politically mixed but historically votes for the ruling party. These sites were selected to represent a range on the spectrum of political geography to elucidate the potentially different political dynamics in strongholds versus more competitive districts. The twelve subjects themselves also represent a range of political and social backgrounds. Five are ruling party partisans, four support independent opposition parties, one supports a co-opted opposition party, and two are non-partisans. Their ages range from 24 to 56, and while some never finished primary school, others have multiple postgraduate degrees.

The data collected from these twelve subjects were meant to test the plausibility of the idea that opposition and ruling party partisanship are unique social identities, as well as to investigate and illustrate the process of political socialization in Cameroon. In the first part of data collection, I conducted full life histories of each subject, focusing on their childhoods and adolescence and the moment in which they became politically conscious. In the second part of the data collection, I asked each subject to name every person in their current social network, posed a series of survey questions about each contact, and then reached out to the contacts to survey them with a mirroring set of questions. This second set of data was intended to capture how political influence works within the social networks of adult citizens who have already formed their partisan attachments. Thus, the first set of data – the life history interviews – looks at formative political socialization, while the second set of data – the full network two-way surveys – is designed to understand contemporary political influence and social processes. Together, these two sets of data are not meant to formally test the theory presented in this book but instead to provide a richer picture of the nature and texture of partisanship and political socialization in an electoral autocracy.

The third original data set comes from a 1,200-respondent public opinion survey that I fielded in four of Cameroon’s ten regions (the Centre, Littoral, Ouest, and Sud) in February 2021. Within these four regions, enumeration areas (EAs) were selected based on two primary criteria. First, whether the EA was in an electoral district that voted consistently for the ruling party, voted consistently for the opposition, or was competitive between the ruling party and the opposition, representing a full range of competitiveness.Footnote ¹³ Second, EAs were selected to produce variation within these three categories between rural and urban areas. A map of the survey sites is presented in Figure 1.1, and the details of the enumeration procedure can be found in Appendix 1.

Figure 1.1 Survey sampling map of Cameroon

The purpose of fielding an original survey was to be able to test three key features of my theory: (1) to show that ruling party partisans and opposition partisans hold distinct political beliefs as well as different in-group and out-group perceptions, (2) to estimate the relationship between various features of social networks and individual-level partisanship, and (3) to test the relationship between social network characteristics and political geography. Thus, a core feature of the survey was its capacity to measure the social networks of a large, diverse group of citizens in an autocratic regime. In order to do this, the survey asked each respondent to name all the people with whom they have regular conversations about life and current events. The respondent was then asked a series of questions about each of the people named, including their ethnicity, education, and partisanship. This original social networks data, described in more detail in Chapters 4, 7, and 8, forms the basis of my empirical testing of the theory presented in this book.

The mixed-method approach used in this book allows for a more thorough and richer understanding of how ordinary citizens think about politics in autocratic regimes. By relying on both qualitative and quantitative data, I am able to show the chaotic, nonlinear, and muddled ways in which ordinary people learn about politics, while at the same time simplifying such complexities to test broader truths. The messy nature of the qualitative data provides a deeper and more convincing picture of political life in an autocratic regime, while the survey data allows us to generalize away from specific experiences to understand politics in electoral autocracies more broadly.

1.4 What Do We Gain by Framing Partisanship as a Social Identity?

The foundation of this book is to use social identity theory to better understand the nature of partisanship in electoral autocracies. There are a number of reasons why this framework provides more traction than existing theories, which are almost entirely premised on the broad assumption that people support various political parties for economic reasons. First, the existing approach produces a number of indeterminant hypotheses about how education, ethnicity, and socioeconomic status will produce support for the ruling party or, alternatively, the opposition. Should someone with high socioeconomic status support the ruling party because she is most likely to have clientelist linkages to the regime (Miguel, Jamal, and Tessler Reference Miguel, Jamal and Tessler2015) or the opposition because her high level of education leads her to reject government propaganda (Croke et al. Reference Croke, Grossman, Larreguy and Marshall2016; Geddes and Zaller Reference Geddes and Zaller1989)? Should someone with low socioeconomic status support the ruling party because he is more sensitive to vote buying and social spending (Blaydes Reference Blaydes2011; Magaloni Reference Magaloni2006), or should he support the opposition because he stands to gain the most, economically, from a transition to democracy (Acemoglu and Robinson Reference Acemoglu and Robinson2006; Boix Reference Boix2003)? Without a unified theory of partisanship, we are left with piecemeal and oftentimes contradictory hypotheses. I argue that if we treat partisanship instead as a social identity, we gain far more analytical purchase on predicting individual-level political support under authoritarianism.

In addition, the theory in this book seeks to push back on the overwhelmingly structural, top-down nature of political science theory in the study of autocratic regimes. In general, the institutionalist literature in the study of authoritarianism is far more developed than work focused on political behavior (e.g., Gandhi Reference Gandhi2008; Geddes Reference Geddes2003; Svolik Reference Svolik2012). This is a natural reaction to the dual realities that (1) microlevel data – particularly public opinion data – is generally difficult to collect in such regimes and (2) political behavior “matters” less in autocratic regimes. When elections do not result in political change, we naturally care less about how people participate in them. However, both of these conditions are changing. On the one hand, public opinion data from electoral autocracies has become increasingly more common, thanks in large part to the incredible work of the various Barometer survey organizations as well as the WVS, among others. On the other hand, with the Arab Spring and the global trend of democratic backsliding, scholars are increasingly aware that understanding public opinion in autocratic contexts is actually quite important. Nonetheless, analysis of political behavior in these regimes remains quite structural in nature: Much of the theorizing in this literature starts from the vantagepoint of the state, showing how people react to regime strategies, such as vote buying (Blaydes Reference Blaydes2011; Kramon Reference Kramon2017), public spending (Magaloni Reference Magaloni2006), and repression (Nugent Reference Nugent2020; Young Reference Young2019). This study starts, theoretically, from the vantagepoint of the citizen. It focuses on the grassroots level, trying to understand what is happening around the kitchen table between grandmothers and their grandchildren and in local bars between friends. Thus, this book seeks to flip the way in which we typically approach studying autocratic politics.

Because the present study is based empirically in Cameroon, it also builds upon the Africanist literature on partisanship (Bratton, Bhavnani, and Chen Reference Bratton, Bhavnani and Chen2012; Carlson Reference Carlson2016), though, strictly speaking, the scope conditions of the argument exclude Africa’s more democratic regimes. Nonetheless, the Africanist literature could also potentially benefit from viewing partisanship as a social identity. A social perspective can help to explain why many studies have found that citizens in democracies don’t punish corruption or poor governance (Dunning et al. Reference Dunning, Humphreys, Hyde, McIntosh and Nellis2020). When partisanship and political participation are rooted in social considerations as opposed to economic ones, it helps to explain the enormous gap we see between poor governance in much of sub-Saharan Africa and the political behavior that fails to punish such weak performance.

Echoing the research agendas of authors such as Adrienne LeBas (Reference LeBas2011), Regina Smyth (Reference Smyth2020), Eloïse Bertrand (Reference Bertrand2021), and Anders Sjögren (Reference Sjögren2023), the theory presented in this book also seeks to push for a renewed emphasis on the importance of opposition parties in autocratic regimes. However, while extant work on the topic tends to focus on the origins and functions of the parties themselves, this book instead investigates the effects of opposition parties on the beliefs and actions of ordinary citizens. Oftentimes overlooked as weak and incompetent, I argue that they actually provide a critical public service. Even when opposition parties are not particularly popular and provide little hope for electoral turnover, in many countries they are one of the only actors providing the vital function of openly and vocally calling out the autocratic abuses of the state. Critically, the argument made in Chapter 9 will show that such parties do not simply arise in areas where ordinary citizens already believe the state is autocratic, but instead, where opposition parties are established, they actually provide this information to citizens and actively change public opinion. Of course, opposition partisans are most likely to receive and internalize this information, but I also show that even apolitical citizens in proximity to opposition activities are affected by their messages. Thus, if we care about citizen perceptions of and support for democracy within the context of authoritarian regimes, we might recenter the importance of opposition parties in such contexts.

Another contribution of the theory outlined in this book is that it uncouples ethnic identity from partisan identity, which so far has been the only sociological lens used to understand political behavior in Africa (Harris Reference Harris2022). In an effort to push back on racist tropes about African political behavior, I would argue that the political science literature has overcorrected itself. Scholars have bent over backwards to provide evidence that African voters behave rationally and react to political incentives and stimuli in ways that are divorced from social considerations. A framework that models partisanship as a social identity in and of itself unburdens a sociological approach that has heretofore equated political choice to ethnic or racial identity. By showing that partisan identities are unique from ethnic ones, the argument made in this book opens up an entire area of study for Africanist political science on the origins, durability, and effects of partisan identities.

Finally, this book seeks to fill a gap in the literature that presents an inconsistent view of human motivations across the field of political behavior. While scholars recognize that citizens in democracies have dense social worlds that shape their political preferences, we assume that autocratic citizens are motivated to participate in politics almost exclusively by economic incentives. The overarching goal of this book is to de-exotify the everyday politics of autocratic regimes. Theories of political participation in such regimes, particularly those that rely on the explanations of electoral patronage and vote-buying, support a normative narrative that autocracies are illegitimate; they provide an explanation for why citizens would vote to support authoritarianism. Political scientists have inadvertently supported this exotification of autocratic politics by constructing a different approach for understanding political behavior in autocratic regimes (Przeworski Reference Przeworski2022). While it is now uncontested that citizens in consolidated democracies participate in politics for social reasons, we have barely just begun to imagine that ordinary citizens in autocracies participate for similar reasons and that this participation might constitute a source of pride for citizens of such regimes. Understanding the more banal and ordinary reasons why citizens participate is critical to de-exotifying politics in non-Western contexts.

1.5 Plan of the Book

Chapter 2 will outline this theory in more detail, focusing on what it means for partisanship in electoral autocracies to be a social identity. Within this theoretical framework of social identity theory, the chapter sets out expectations for how partisanship can best be explained by social and environmental factors, including social networks, political geography, and the process of partisan socialization. I argue that individual partisanship is best predicted by the characteristics of one’s social network, and that, in turn, the political characteristics of one’s network can largely be predicted by the unique political geography of electoral autocracies. Chapter 3 introduces the case of Cameroon, providing a brief history of Cameroonian politics, with a focus on the evolution of political parties and their geographic roots in the colonial period. The chapter concludes by discussing the ways in which Cameroon is both similar and different to other electoral autocracies, and what this means for the generalizability of the theory presented in this book.

Chapter 4 begins the empirical analysis of the theory. It starts by investigating what exactly it means to be an opposition or a ruling party partisan in the context of an electoral autocracy. Using my original survey data, I show that partisans in Cameroon are divided on the issues of regime trust and perceptions of democracy. I also investigate in-group and out-group perceptions. I then turn to WVS data from thirty-three different regimes around the world to show that partisans across a wide range of electoral autocracies are likewise divided on these issues, while their counterparts in democratic regimes are not.

Chapters 5 and 6 investigate the grassroots nature of political socialization. They rely on qualitative data that were collected in order to better understand the texture and fabric of social relationships and political influence among ordinary Cameroonians. Chapter 5 zeroes in on the life histories of twelve research subjects. Based on in-depth interviews, the chapter describes how each subject came to form their political beliefs and associate themselves with a political party (or not). Section 5.1 details the stories of three of the partisans – Bertrand, Joseph, and Smart – who were clearly “invited” to join the party in childhood or later in life by a party activist. Section 5.2 discusses the stories of Martin, Anita, and Jacques, whose process of political socialization reveals the ways in which political geography influences partisan socialization within an opposition stronghold. Section 5.3 highlights the lives of Mireille and George, whose stories illuminate the ways in which clientelist reasons for supporting political parties are highly reliant on the dynamics of one’s social network – or not, in the case of George, the one partisan in the sample who did not describe a process of political socialization.

Moving from early adolescent socialization to current forms of socialization, Chapter 6 explores the nature of political influence within the current social networks of the twelve research subjects. The chapter presents three core observations. First, the data seem to indicate that two or three of the research subjects – all party activists – are “influencers” within their social networks: A very high percentage of their social networks report that they have been politically influenced by the subject. Second, the four ruling party partisans have more politically homogenous social networks than either the opposition partisans or the nonpartisans and that, further, they also appear to be less politically influenced by opposition partisans than vice versa. Thus, it appears that ruling party partisans are far more politically “insulated” than opposition partisans are. Finally, I describe how the “inner circles” of the twelve subjects are a lot more politically homogenous than are the social contacts who have less intimate relationships with the subject. In other words, while the weak ties of social networks are relatively diverse, close “inner circles” tend to be much more political homogenous.

I test the core hypotheses related to political socialization in Chapter 7 using the large-N data from my original survey. I find support for the hypotheses that citizens exposed to politics in their childhood homes are more likely to become partisans; that party activists tend to be more influential in their social networks than regular partisans; and that political homogeneity in social networks is an extremely strong predictor of individual-level partisanship. Thus closeness, strength of identity, and homogeneity all appear to be important for understanding political socialization. The data also suggest that opposition partisans tend to be more influenced by ruling party partisans than vice versa. The final empirical test of Chapter 7 is designed to assess the possibility that partisan homogeneity within networks is being driven by choice homophily. The survey asked the respondents for how long they had known each of their discussion partners. In addition, each respondent who reported being a partisan was asked how old they were when they first started feeling close to their specific political party. Thus, the final empirical analysis reruns the statistical models including only discussion partners whose relationships predate the respondent’s partisanship. In other words, it eliminates the possibility that respondents chose discussion partners because of their partisan identities. The results are robust to these more narrow specifications, indicating that self-selection based on partisanship is not the primary driver of partisan homogeneity within social networks.

Chapter 8 then turns to the question of the origins of this partisan homogeneity: If the political nature of social networks is driving individual-level partisanship, what explains why some networks are oriented toward the ruling party while others lean toward the opposition? The chapter shows how the unique nature of political geography in electoral autocracies produces an extreme level of structural homophily that can predict the political nature of social networks. The first part of Chapter 8 defines political geography and outlines its measurement within Cameroon. I create a factor variable based on four distinct measures of opposition activity, including original electoral data collected from Cameroon’s electoral commission (ELECAM) as well as data from the original survey on local opposition offices and political campaigns. Second, the core test of the theory shows that, even controlling for a host of demographic and network-level variables, partisan homogeneity is strongly predicted by this political geography factor variable. I also show that partisan networks still drive partisan outcomes in cross-partisan strongholds, indicating that it is the networks themselves that matter for understanding partisanship, not just that there are more partisans in general in party strongholds. The final section of Chapter 8 zooms in on the 113 survey respondents who have moved between different party strongholds, showing how the partisan composition of social networks appears to change over time based on geography.

Chapter 9 circles back to the question of the content of partisan identities, connecting it to the role that political geography also plays in constructing these identities. Relying again on the original survey data, it first shows that all people who live in ruling party strongholds – not just ruling party partisans – perceive themselves in very different ways than do people living in opposition strongholds, again, regardless of partisan status. The second half of the chapter turns to Afrobarometer data from Cameroon, Tanzania, Togo, Uganda, and Zimbabwe (as well as, separately, Ghana) to make several interrelated points. First, the data shows that the effects of political geography hold outside of Cameroon, underlining the external validity of key parts of the theory presented in this book. Second, it reveals that political geography alone can predict political beliefs in electoral autocracies. Citizens living in opposition strongholds – even nonpartisans and ruling party partisans – are less trusting of their governments, less satisfied with democracy, and more likely to believe that the president ignores the law and goes above Parliament. Third, looking exclusively at Uganda, the data provides some evidence that it is political parties that actively shape these political beliefs in electoral autocracies. Finally, using Afrobarometer data from Ghana, the chapter concludes by showing that there is no relationship between political geography and political beliefs in an established democracy; this relationship is unique to electoral authoritarianism. Chapter 10 concludes by describing the contributions and implications of this theory of partisanship in electoral autocracies, as well as some of its limitations, and ends by discussing how it might travel outside the context of electoral autocracies.