1.1 Privileging the State

That conventional approaches to politics and development privilege the state is perhaps not surprising. The scholars, practitioners, and policymakers aiming to shape the distribution of power and resources in a manner that enhances human welfare – that is, to engage in the essence of politics and development – are closely linked to the state. They often sit in or hail from departments of government, politics, or economics, where – at least since Reference Gerschenkron and HoselitzGerschenkron (1952) – the state is considered to be the driving force behind development. They work with or in the World Bank, the United Nations, and other multilateral organizations, for whom states are both the major funders and primary interlocutors. They use available official (i.e., state-based) statistics, gathered by the state’s machinery in the interest of legibility, to implement research, pinpoint citizens’ needs, and assess policy impacts.

These scholars, practitioners, and policymakers generally focus on the strength of the state or the nature of its institutions, and they seek development solutions through state-building or institutional engineering. Even when they take social and economic contexts into account, considering how social identities or endowments affect outcomes, they largely overlook the variation in social institutions that compete or intersect with political institutions to shape individuals’ actions. In the state-centric perspective, outcomes that deviate from expectations are ‘failures’: problems of ‘corruption’ or ‘clientelism’ to be solved through the strengthening and reform of state institutions.

The dominant state-centric perspective impedes our efforts to bring non-state arenas of authority and social institutions fully into the study of, and programming around, politics and development. First, it portrays actors and institutions as either state or non-state, ignoring differences among non-state arenas and institutions. Empirically there are multiple arenas outside the state, based on different notions of community (e.g., religion, ethnicity, locality, economy), with different distributions of power and institutional arrangements. Yet, scholars, practitioners, and policymakers too frequently lump these together as an undifferentiated ‘residual’, focusing on the absence of ‘healthy’ state institutions rather than the presence of alternatives that drive outcomes. Other times, they focus on a single arena – for instance, the relationship between tribe and state, religion and politics, or ethnicity and service provision. In doing so, they often overlook important questions about the extent to which different arenas and institutions drive development.

Second, the focus on the state diverts the efforts of political scientists and development specialists away from developing a comprehensive and coherent framework for understanding non-state arenas of authority and related social institutions. Contrast the conceptualization and study of the state and its institutions with that of other arenas of authority and associated social institutions. Particularly since Reference Skocpol, Evans, Rueschemeyer and SkocpolSkocpol (1985) called on scholars to ‘Bring the State Back In’ and Reference March and OlsenMarch and Olsen (1984) (re)turned our attention to institutions, the state and its institutions have been a centre of attention. It is an entity with notions of communities and belonging (nations), regimes and authority, which shape the distribution of power and constrain members in ways that go beyond the sum of its institutions. The state is also recognized as having independent interests and more or less autonomy from societal actors (i.e., strength). Moreover, its institutions are distinguished in terms of (relatively) well-developed conceptual categories (e.g., democracies and autocracies, centralized and decentralized administration, proportional representation or majoritarian electoral systems). Explicitly recognizing variations in state strength and institutional arrangements facilitates theory testing, helps clarify scope conditions within which theories should hold, allows for distinguishing between institutional and contextual factors, and provides a scaffolding on which to place new findings.

The study of non-state arenas and social institutions lacks such crisp, well-established conceptual categories and frameworks of study. Political scientists recognize arenas of authority outside the state and related social institutions, but the language and frameworks they employ are less fully elaborated than those used to study the state’s role in governance and development. Researchers and practitioners recognize the importance of different non-state authorities, but they often focus on specific authorities (e.g., traditional authorities, gang leaders, warlords) and thus leave open questions of when and why different authorities have influence.Footnote ² So too, they use the term ‘social institutions’ to denote very different concepts, ranging from organizations that are either non- or semi-stateFootnote ³ to informal rules followed by state actors (e.g., Reference Helmke and LevitskyHelmke and Levitsky, 2004; Reference Helmke and Levitsky2006). Finally, researchers highlight different aspects of non-state arenas and social institutions, focusing on networks or specific rules.Footnote ⁴ However, they do not place these components in a broader framework. Without a unified language and framework of study, it is difficult to compare or reconcile diverse findings, accumulate knowledge, and achieve theoretical advances and practical insights for programming.

Third, and somewhat ironically, a state-centric approach impedes the study of the state. Assuming state predominance and under-theorizing social institutions not only precludes a full understanding of how social authorities and institutions affect governance and development outcomes, but it also distorts the view of the state. Reducing social institutions to context and under-specifying their variation makes it difficult to understand the role that state institutions truly play. Where the state is weak and social authorities are readily visible, their presence is viewed as a problem to be solved rather than forces to be understood. Where the state is apparently strong, social authorities and institutions are viewed as ineffectual and unimportant, even though they may be critical in shaping governance. Even multi- and bilateral development agents for whom state actors will remain the primary interlocutors need a clearer understanding of non-state arenas of authority and social institutions in order to be effective.

1.2 Competing Claims and Individuals’ Choices

The perspective I present here is not simply that non-state arenas of authority and associated social institutions shape individuals’ choices but that multiple communities often vie for control over their actions. The Ghanaian bureaucrat, Yemeni policeman, Jordanian voter, and American citizen presented earlier do not respond only to their kin, tribe, or religious arenas of authority any more than they respond solely to the state. The strengths of arenas vary across space and time, as well as for different individuals, depending on their position within the community (e.g., leader versus follower, elder versus youth). Often, the inelasticity of social demands becomes all too evident; particularly within development settings, institutions outside the state that shape actions lead to outcomes contra state and programming objectives. At other times, the state may wrest control from even powerful non-state authorities, either by acting alone or in conjunction with other arenas.

The social institutions within arenas of authority also vary, creating different incentives for members within them. Expectations differ across even seemingly identical arenas of authority. Take, for example, local and international religious arenas – even of the same denomination. At times, these make competing demands on members, forcing members to choose between them. This was evident over the backlash to Pope Francis’ 2016 Amoris Laetitia. Papal supremacy called for holistic sexual education for children, reintegration of divorced and remarried Catholics into the church, and respect for LGBTQ individuals, but many local dioceses – and a great number of individual Catholics – chose to ignore this newest apostolic exhortation. Therefore, in voting for the ‘bathroom bills’, proposed in US state legislatures between 2017 and 2019 as a way to limit accommodation of non-cisgender individuals, the arena of authority governing Catholic voters’ decisions could be either aligned with the Papal seat (which would be against the restrictive legislation) or with the competing religious interpretation within their local community (thereby validating discriminatory practices against LGBTQ individuals). When multiple arenas of authority make contradictory demands, individuals are forced to respond to some arenas over others.

Thus, understanding political behaviour and development outcomes requires that one recognizes both the importance of competing arenas of authority and the nature of the social institutions within them. Rather than view citizens and state officials as relating their actions solely, or even primarily, to the state, I argue that we need to start by considering how individuals – voters, public service providers, bureaucrats, politicians, and others – understand the acts in which they engage. Actions such as voting, dispute resolution, and public service provision are engagements with the state, but they are also, critically, engagements in multiple other arenas of authority. So too, actors are not ‘either’ state or non-state, political or social, but often acting simultaneously as players in multiple arenas. The apt question is not ‘is the service provision, election, or political behaviour in question located within the state realm?’ – and thus shaped by state institutions – but rather, ‘when individuals engage in these actions, what meaning do they attach to the actions?’ Which arenas of authority make demands upon them, what do they believe is expected of them in each, and with what consequences?

1.3 Plan of the Element

In this Element, I aim to overcome problems of the dominant state-centric perspective by setting state and non-state authorities and institutions on equal intellectual footing, providing a structure for accumulating knowledge about how these competing arenas and social institutions influence politics and development, and reconsidering the state. To do so, I take a ‘bottom-up’ approach that focuses on the perspective of individuals – voters, public service providers, bureaucrats, politicians, and others – and considers how their simultaneous membership in various arenas of authority shapes their choices and, ultimately, governance and development. I focus on how the everyday choices before individuals may take multiple meanings, provide guidance on how to understand the extent to which different arenas of authority influence actions, and illustrate how differences in social institutions affect individual choices and outcomes.

This Element is intended for two audiences. For scholars, I aim to bring together currently disparate findings from extant research, highlight general themes found in empirically rich but contextually specific (and less accessible) regional studies, provide a new perspective on governance and development, and pose questions for future research. For practitioners, I hope to help develop a structure for programming that is less prone to problems of isomorphism and state centricity than conventional approaches, and yet also less indeterminant than some of the existing alternative approaches. The goal is to provide a structure and language that allow scholars, practitioners, and others interested in politics and development to make sense of the many compelling studies to date, to structure research moving forward, and to design programmes that take these findings into account.

The Element proceeds as follows. Section 2 reviews the literature. It finds that state-centric and institutional approaches have dominated the study of and programming around political behaviour and development, while alternative approaches that highlight the complexity of development make it difficult for scholars and programmers to build on past experiences. Section 3 presents arenas of authority and social institutions, the building blocks of the framework. Sections 4–6 examine how arenas of authority and the social institutions within them affect political behaviour and development. Section 4 provides guidance on how to determine the extent to which competing arenas of authority influence decisions at the core of politics and development, while Section 5 turns attention to how variations in the social institutions within these arenas shape outcomes. Section 6 examines how social institutions within the arenas of authority outside the state affect state institutions. Section 7 concludes by considering how we can use the approach presented here in future research and programming and exploring unanswered questions.

Before proceeding, two caveats are in order. First, although I adopt the language of state and non-state institutions, I am uneasy with the distinction. As argued in Section 6, the distinction between state and non-state institutions is often overdrawn, and the boundary itself may be a useful subject of enquiry (see Reference MitchellMitchell, 1991). It is perhaps more apt to speak in terms of the basis of authority on which actors and institutions rest. The second caveat relates to the scope of this Element. I focus on the Global South throughout much of this text, but the social institutions discussed are not limited to what Western readers may think of as ‘those places’. I emphasize the Global South because that is where the vast majority of programming is implemented, but the issues raised very much describe life in Gothenburg, Sweden; New Haven, Connecticut; Marshall, Michigan; and elsewhere as well.

2.1 Conventional State-Centric Approach

Most political scientists and development practitioners privilege the state. Early modernization theorists (e.g., Reference LernerLerner, 1958; Reference LipsetLipset, 1959; Reference Smith and WaldnerRustow, 1970) presumed the state seeks, and ultimately will achieve and maintain, the monopoly over the legitimate use of force in a given territory, providing security and welfare to people therein.Footnote ⁵ The state is the locus of participation and representation, the engine of economic growth and development (Reference Gerschenkron and HoselitzGerschenkron, 1952). In general, these scholars argued that individuals in ‘traditional’ societies held values that constrained their demands on authorities and the state (Reference Almond and VerbaAlmond and Verba, 1963; Reference LernerLerner, 1958). Economic development – including the spread of roads, radios, and other aspects of modernization – would lead to greater mobility, expectations, and demands for democracy, which in turn would foster development. There would be hurdles. Reference HuntingtonHuntington (1968) famously argued that the strength of state institutions must keep pace with the level of social mobilization in order to avoid political decay and disorder. Yet, in general, development and democracy went hand in hand, and the state and its institutions were key. The state had the inherent ability to be more organized, technologically savvy, and capable of extending its power than social counterparts, putting them on the defensive. Where the state was not yet dominant, it would – or at least should – be so in the future. The question was when.

Contemporary scholars largely reject modernization theory’s teleological perspective, and yet many continue to privilege the state. It is by now well-recognized that the state extends power unevenly and often fails to act as early scholars predicted (e.g., Reference MigdalMigdal, 1988; Reference Falleti and CunialScott, 1972). Nevertheless, many view a high-capacity state as vital for economic growth and human development (see Reference Cho, Crenshaw and McCallCingolani, 2018 for a review). State institutions are also key: regime types affect political stability, economic growth, and human development; electoral systems shape voter behaviour, representation, policymaking, and economic welfare; administrative arrangements affect service delivery.Footnote ⁶ Research in this tradition has led to important insights about the logic of institutional arrangements that can be extended to other arenas of authority as well. Yet, these lessons are often overlooked because scholars in this tradition tend to view forces outside the state as disruptive. They label their impact ‘corruption’, ‘clientelism’, or ‘low quality government’,Footnote ⁷ and invest their time and energy into determining how state institutions (through the implementation of gender quotas, civil service exams, etc.) can overcome such forces.

Development practitioners also focus on state institutions. Particularly in the early 2000s, many explicitly called for state-building interventions. A report prepared for the UK’s Department for International Development noted, ‘The need to better understand state-building is not an academic exercise; states are crucially important to the future of those who live under their jurisdiction’ (Reference McGuireWhaites, 2008, p. 3). Prior to becoming president of Afghanistan, but after serving as finance minister and in the World Bank, Ashraf Ghani and his co-author, Clare Lockhart, wrote a book calling for greater attention to state-building (Reference Ghani and LockhartGhani and Lockhart, 2009). In light of spectacular failures in Afghanistan, Iraq, and elsewhere, the notion that external actors could successfully build states came under attack (e.g., Reference KrasnerKrasner, 2011), and ‘state-building’ became a dirty word. Nevertheless, practitioners continued to see the strengthening of ‘core government functions’ and other aspects of the state as the key to development, particularly in fragile and conflict-affected states (Reference LustUNDP and World Bank, 2017).

Consequently, most programming is centred around the state. Major multilateral and bilateral development organizations, themselves instruments of states, often focus their programming on state organs, implementing projects around budget support, administrative strengthening, or public infrastructure. Smaller development organizations, too, often partner with government agencies. Thus, even when these organizations engage business, NGOs, or other elements of civil society, they tend to do so privileging the state’s perspective.Footnote ⁸ Indeed, strengthening the capacity of these actors and organizations to engage the state is often a fundamental goal.

The instruments used to measure governance and development are also state-centric. The Fragile States Index, developed by the Fund for Peace, seeks to measure state capacity. Extant indicators of governance and service delivery (e.g., World Governance Indicators, Quality of Government) focus primarily on participation, transparency, accountability, and other dimensions of governance with respect to the state. Participation in elections or local council meetings is measured, whereas participation in tribal primariesFootnote ⁹ or non-state councils is not. Moreover, most indicators are at the national level, assuming that the important variation is to be found in national-level state institutions but not in local-level social institutions. Some may view the primacy given to national-level state indicators as reflecting the ease of using available data. Yet not all conventional measures are state collected, and alternatives can exist (see Appendix A for one such alternative). Moreover, measuring governance and development with regard to state institutions not only reflects the privileging of the state, but also contributes to it. Thus, while measures of state capacity and institutions are important, it is also necessary to correct the imbalance between measures of the state, of which there is abundance, and measurements of non-state arenas of authority, which are largely absent.

2.2 Conventional Society-Oriented Approach

The conventional society-oriented approach focuses on the roles of identity and social capital on political behaviour and outcomes. Scholars working in this tradition ask such questions as how such social factors shape voting, representation, and even democratization (Reference ChandraChandra, 2007; Reference Michael Albertus and SlaterPosner, 2005; Reference Putnam, Leonardi and NanettiPutnam et al., 1994; Reference HunterSzwarcberg, 2012). In answering these questions, they often point to the importance of competing authorities and social institutions. However, they privilege the state and do not fully theorize competing arenas of authority or the social institutions within them.

Scholars implicitly recognize that individuals are members of multiple communities, although they do not theorize about these in terms of competing arenas of authority. For instance, Reference HaleHenry Hale (2004, p. 480) argues that individuals have multiple dimensions of their identity and that they use them as a ‘sonar radar’ to navigate their world. Similarly, Reference PosnerDaniel Posner (2005) recognized that individuals have multiple dimensions of identity (e.g., religious, race, gender), and that they select from their identity repertoires when making choices. Yet, in exploring which identity structures individuals’ choices when they engage in (state-based) politics, these scholars focus primarily on the arrangement of state institutions and ethnic cleavages (see also Reference McCauley and PosnerMcCauley and Posner, 2019; Reference McLaughlinMcLaughlin, 2007). They pay far less attention to other factors that affect the strength of different arenas of authority. That is, they privilege the state.

Similarly, scholars taking this approach acknowledge the importance of rules. For instance, scholars of ethnicity emphasize rules of reciprocity, wherein they are expected to favour coethnics over non-coethnics (Reference Björkman and SvenssonBjörkman and Svensson, 2010; Reference Cammett and MacLeanCammett and MacLean, 2014; Reference CorstangeCorstange, 2016; Reference Habyarimana, Humphreys, Posner and WeinsteinHabyarimana et al., 2007; Reference Miguel and GugertyMiguel and Gugerty, 2005); those of gender focus on rules of patriarchy and homophily that at least partly explain why women are less likely to stand for office, vote in elections, or seek assistance from elected officials (Reference BensteadBenstead, 2016; Reference BjarnegårdBjarnegård, 2013). So too, scholars of social capital argue that engagement fosters the development of social organizations (or communities) in which there are ‘networks, norms and trust that facilitate coordination and cooperation for mutual benefit’ (Reference KohliPutnam, 1993, p. 35). Some recognize that the extent to which members identify with these groups or are constrained by their social institutions may vary (Reference Dulani, Harris, Horowitz and KayuniDulani et al., 2021; Reference HarrisHarris, 2022). However, the vast majority of this scholarship has not focused on differences in social institutions, and instead has assumed them to be uniform.

Yet, both the strength of competing arenas of authority and the content of institutions within them can vary significantly. This is true even within similar types of communities (e.g., ethnic, religious, geographic) within the same countries. For example, the Local Governance Performance Index (LGPI) surveys fielded in Kenya, Malawi, and Zambia in 2019 find a great deal of variation in the expectations regarding reciprocity among coethnics, across both subnational geographic locations and ethnic groups (Reference Lust, Kao and LandryLust et al., 2019). A framework that takes competing arenas of authority into account thus needs to consider variation in their social institutions, not to assume them to be constant.

2.3 Focus on Social Institutions

A third strand of literature highlights the importance of social institutions and increasingly recognizes variations across them. Scholars taking this approach have uncovered a number of important findings that I build on in the pages that follow. Yet, a review of this literature finds that these scholars, too, generally privilege the state and fail to consider competing, non-state arenas of authority.Footnote ¹⁰ Moreover, they lack an overarching schema that allows them to be fully in conversation with each other and to make sense of the increasingly vast literature.

Scholars taking this approach are primarily interested in how such institutions reinforce, complement, or replace formal state institutions, and they pay little attention to the competition across different arenas of authority outside the state. Reference Helmke and LevitskyHelmke and Levitsky (2004, p. 727; see also Reference Helmke and Levitsky2006) define informal institutions as ‘socially shared rules, usually unwritten, that are created, communicated and enforced outside of officially sanctioned channels’ (italics mine). That is, they portray social institutions in residual terms – as non-state.

Even Elinor Ostrom, undoubtedly the most influential scholar to highlight institutions outside the state, takes this approach. In Governing the Commons (Reference Ostrom1990), she argued that communities can devise rules to solve collective action problems and manage community resources in the absence of the state. She highlights the variation in non-state institutions, drawing attention to the importance of community boundaries, participatory decision-making, rules, enforcement, and dispute resolution mechanisms for community governance where the state is absent.Footnote ¹¹ Yet, even in the world she describes, where local communities have agency, the community is a single arena of authority. Furthermore, the state is still central: local solutions are particularly effective when the state recognizes them as legitimate. Viewing the state as the ultimate authority juxtaposed against a single local community ultimately excludes questions about competing arenas of authority outside the state.Footnote ¹²

Similar tendencies are found in the writings of development specialists. Many recognize the importance of non-state actors and social institutions, but they still privilege the state and under-theorize competing arenas of authority and social institutions. The Overseas Development Institute’s (ODI) Africa Power and Politics Programme (APPP) promoted the notion of ‘Going with the Grain’ (Reference KelsallKelsall, 2012), arguing that traditional authorities and local customs such as witchcraft or polygamy shape political outcomes and thus must be taken into account in development efforts (Reference Booth and Golooba-MutebiBooth and Golooba-Mutebi, 2012; Reference KelsallKelsall, 2008; Reference Kelsall2012). Reference LevyBrian Levy (2014), then at the World Bank, argued that we could better determine which levers could help establish ‘islands of effectiveness’ and eventually economic growth and better governance outcomes if we take into account whether polities are dominant or competitive, and governance impartial or personalized. More recently, Reference Andrews, Pritchett and WoolcockMatt Andrews, Lant Pritchett, and Michael Woolcock (2017) pushed even more strongly for an incremental approach to development that would consider contextual factors.

Yet even advocates of context-sensitive programming present typologies based on distinctions across state institutions – the mechanisms for centralizing rents and policy orientation for ODI’s APPP and competitiveness of politics and impartiality for Levy. They question the dominance of the Weberian arrangements but nevertheless view social institutions and actors as inferior, at best, and disruptive, at worst. They explain differences between outcomes that ‘should be’ according to state organizational rules and those that ‘are’ in terms of organizational capacity of state institutions (Reference Andrews, Pritchett and WoolcockAndrews et al., 2017, pp. 47–8). In doing so, they are unable to give guidance on how to determine which arenas of authority have weight and the impact that variations in their social institutions may have.

These gaps are evident when scholars and practitioners partner with each other, attempting to understand why interventions are more effective in some communities than others. Micro-level approaches that focus on institutions lend themselves to increasingly sophisticated, systematic studies of their effects, particularly through randomized control trials (RCTs). Despite enormous investments of time, money, and intellect, however, the main takeaway of efforts to date has been that development programmes often travel poorly across time and space (Reference Andrews, Pritchett and WoolcockAndrews et al., 2017; Reference Dunning and GrossmanDunning et al., 2019; Reference Michael Albertus and SlaterRaffler et al., 2020).

Efforts to understand why have yielded a vibrant, cross-disciplinary literature examining various aspects of social institutions. Some point to networks and social ties, considering how the strength of social ties in communities affect political participation, accountability, and service provision. Others examine the institutional rules that shape interactions between groups: altruism and reciprocity, group boundaries, lineage practices, or rules with regard to specific outcomes, such as conflict mediation.Footnote ¹³ Still others focus on how different types of authorities influence voting, representation, or service provision (Reference BaldwinBaldwin, 2016; Reference Díaz-Cayeros, Magaloni and Ruiz-EulerDíaz-Cayeros et al., 2014).

These studies provide important insights into arenas of authority and social institutions, but they are not unified by a common theoretical framework and language. This makes it remarkably difficult to compare findings across studies or draw together lessons offered by various scholars. Studies of networks and social density, for instance, tend to assume rather than interrogate the form of institutional rules that shape behaviour. Individuals from families, ethnic groups, or villages are uniformly expected to reciprocate with in-group members, or to view each other as sharing preferences and thus be more likely to coordinate. Those focused on differences in rules (e.g., reciprocity, inheritance) overlook differences in the strength of institutions. Moreover, these studies generally focus on one set of institutions, practices, or outcomes, failing to take into account how individuals may be subject to more than one arena of authority. They call for guidance on which arenas gain primacy and how social institutions within them interact to shape outcomes.

2.4 Focus on Membership in Multiple Communities

A fourth approach steps away from the state-centric, single-arena approach that underpins most work on governance and development. Both scholars working on hybridity of governance and those studying intersectionality recognize that individuals act within various, competing arenas of authority, and they pay great attention to how these arenas affect the distribution of power across actors. These scholars thus provide an important critique of the dominant approaches and a basis for advocacy. However, they do not offer guidance on how competing arenas of authority or the institutions within them affect the choices individuals make and the outcomes that ensue.

Most theorists of hybridity reject binaries of state and non-state actors, traditional and modern, local and international. They examine a range of governance processes, from dispute resolution and peacebuilding to welfare provision. Importantly, they argue that governance outcomes are ‘never simply a mix of two otherwise pure forms, but are perennially ongoing processes of amalgamation and dissolution’ (Reference Albrecht and Wiuff MoeAlbrecht and Wiuff Moe, 2015, p. 5). They produce what Reference GalvanGalvan (1997) calls ‘institutional syncretism’, in which structures from different arenas (as I call them) are transformed and combined to create new institutions. This view recognizes the importance of various arenas, but it sees them in terms of synthesis, rather than competition. Other scholars recognize the competition among various arenas. As Reference Cleaver, Franks, Maganga and HallCleaver, Franks, Maganga, and Hall (2013, p. 167) argue, ‘“Real governance” is formed, negotiated, and contested in the street, the clinic, the market and the press as well as in the formalized public decision-making arenas of community and local government.’ This view of governance as ‘institutional bricolage’ sees governance as shaped by multiple arenas and related institutions, but it does not seek to explore the impact of arenas and related institutions, systematically.

The study of intersectionality also acknowledges the importance of multiple communities which place demands upon and shape individuals’ actions. Intersectionality grew out of feminist theories of gender and race, bringing to the fore the insight that women (and others) are members of multiple communities and subject to various power structures. As Kimberlé Reference CrenshawCrenshaw (1990) argued, the dominant conceptions of discrimination and inequality portrayed disadvantage as located on a single axis (e.g., class, race, or gender) and thus ignored important intra-group differences. African American and Caucasian women are all women, but African American women are doubly disadvantaged, subordinate in both class and race. The power dynamics of gender, class, race, and other communities that come into play yield experiences that are distinct – not simply the additive experience of various discriminations (Reference Cho, Crenshaw and McCallCho et al., 2013; Reference McGuireSettles and Buchanan, 2014). Scholars of intersectionality thus emphasize lived experiences and explore how individuals’ multiple identities affect their experience with discrimination and inequality. Competing arenas of authority are at work in many of these analyses, although not explicitly recognized as such. Reference AckerAcker (2012), for instance, gives the example of hiring practices which have assumptions that may be based on race, class, and gender; Reference AnthiasAnthias (2013) points to societal arenas, considering inequalities of women within the public domain, as well as within their own families.

These approaches yield important lessons, but they stop short of providing guidance on how to determine which arenas of authority are salient, and to what effect. Scholars of hybridity and intersectionality point to how individuals’ membership in different communities affects their power and experience. However, by emphasizing the complexity of institutions, they do not provide guidance on how to determine which arenas of authority (or communities) matter, when. Nor do they pay explicit attention to the social institutions within these arenas. Moreover, many working in this vein resist the development of a positivist framework. For some, the goal is descriptive, not prescriptive. It is to understand the hybrid nature of governance, the ways in which power is embedded within it, and how it is shaped and reshaped. They provide an important tool for greater understanding and advocacy (e.g., Reference Cleaver, Franks, Maganga and HallCleaver et al., 2013; Reference Forsyth, Kent, Dinnen, Wallis and BoseForsyth et al., 2017; Ginty and Richmond, 2016) but not a framework that helps reconcile extant findings, guide future research, and inform development actors.

2.5 Conclusion

The perspective that I present in this Element – that politics and development outcomes are the result of individuals’ everyday choices, shaped by multiple, competing arenas of authority and the institutions within them – draws upon and extends decades of scholarship on politics and development. However, I make three conceptual moves that, taken together, distinguish this from previous frameworks. First, I argue that the seemingly political decisions that individuals make often have multiple meanings (Reference Falleti and CunialWedeen, 2002) and that individuals associate their choices with arenas of authority and attendant communities. Second, I set arenas of authority associated with different communities and the state on equal theoretical footing. Third, I emphasize the importance of institutional variation in both the state and various non-state arenas of authority. By doing so, I develop a framework that brings disparate literature into conversation, fosters knowledge accumulation, and facilitates a new approach to the study of, and programming around, politics and development.

3.1 Arenas of Authority

An arena of authority is a sphere of activity with clear membership, goals, and institutions. Visualize an arena of authority as a physical arena. It has boundaries, which distinguish members of the community (inside the arena) from outsiders. Membership in the community may be based on such foundations as ethnicity, tribe, or religion, but regardless of the foundation, the community seeks to propagate itself beyond the current generation. This common goal does not imply equality or a lack of conflict. The community may be highly differentiated, with leaders and followers, masters and slaves. Indeed, members need not necessarily have joined the community by volition. Members may also contest the rules or compete with each other over resources. Ultimately, however, those within the arena of authority are engaged, more or less consciously, in a grand project of sustaining the group. They are thus mutually interdependent, with each member’s welfare tied to that of others within the arena.

Like the state,Footnote ¹⁴ social arenas of authority govern much of individuals’ lives, from cradle to grave. They regulate marriage and biological reproduction, raise resources for community goods, provide welfare, and resolve disputes. Indeed, in much of the world, the vast majority of individuals take civil and criminal cases before customary venues or nowhere at all, and public confidence is often higher in such venues than in their state counterparts (Reference Gutmann and VoigtGutmann and Voigt, 2020; Reference LoganLogan, 2013). Some striking statistics: rural Liberians took only about 4 per cent of cases involving economic disputes and 8 per cent of those involving violent crimes to the state courts, compared to 36 per cent and 40 per cent, respectively, to customary courts (Reference KohliSandefur and Siddiqi, 2013, p. 42);Footnote ¹⁵ Indonesians were equally likely to take cases to customary or state courts, but they have much higher confidence in the customary courts (Reference HarperHarper, 2011); and increasing numbers of Muslims in Germany opt to register marriage, divorce, and solve disputes in sharia’ courts unrecognized by the state (Reference JarabaJaraba, 2020). Much of the world’s population also relies on social arenas for access to land. Indeed, in 2017, the Reference KohliWorld Bank (2017) reported that only about 30 per cent of the world’s population had state titles to their land. This is despite decades-long global efforts at state land titling, in a state-centric attempt to increase legibility, extend state control, and (arguably) foster economic growth and equity.Footnote ¹⁶

Arenas of authority have boundaries. Boundaries separate those who are within the community from outsiders, and they are more or less porous. Some arenas tightly restrict who may enter or leave the community. Children born to a Jewish mother are Jewish, while those born to a Muslim father are Muslim – or at least they should be according to membership rules in Judaism and Islam. Ethnic groups can be similar, particularly if one defines ethnicity as Reference ChandraKanchan Chandra (2006, p. 398) does: ‘a subset of identity categories in which eligibility for membership is determined by attributes associated with, or believed to be associated with, descent (described here simply as descent-based attributes)’. In the middle of the spectrum are groups in which not all individuals are eligible for entry and exit may be difficult, but not impossible. Consider members of campus Greek societies and inner-city gangs. Fraternities admit students but not ‘townies’, and gangs often admit individuals from certain areas and demographic groups but not others. At the other extreme are arenas with porous boundaries. Take for example some neighbourhoods or open religious congregations. One might become a resident in a neighbourhood simply by moving into a community, a congregational member by converting.

It may be tempting to associate the porousness of boundaries with the type of community – to view geographic communities as relatively open, ethnic or sectarian groups as closed. Yet, there is considerable variation in membership rules within the same type of arena, across both space and time. For instance, upper-class, white Americans may face few barriers to accessing any neighbourhood, at least financially; however, the same is not true for their less-wealthy counterparts or for African Americans in the 1950s, for whom redlining practices limited their ability to enter certain housing markets.Footnote ¹⁷ It is also not true elsewhere in the world. In parts of Ghana, Oman, and Zambia, for instance, permission of traditional authorities is often required for one to obtain land in their area. So too of religious communities. In some cases, one can change religion as easily as clothing styles, leaving one arena to join another. Elsewhere, this is not the case. One does not simply convert from Sunni Islam to Shi’ism, yet alone to Christianity, in much of the Middle East, nor revoke membership within Hindu or Muslim communities in much of India and Pakistan. The costs of attempting to exit the arena are high, bearing at times even the penalty of death, and the ability to enter other arenas is limited.

Finally, social arenas, like states, seek a degree of sovereignty. Sovereignty is generally associated with states, defined as the supreme authority over a polity within an established territory. Yet, substantively, the key to sovereignty is the supreme legitimate authority over a polity. As Reference AgnewAgnew (2005, p. 441) argues, such ‘effective sovereignty is not necessarily predicated on and defined by the strict and fixed boundaries of individual states’. Indeed, political authority is neither exclusively territorial nor restricted to states. Arenas of authority seek ultimate, legitimate authority over activities that are key to their community’s survival. As examples, Lebanese religious sects seek to maintain control over marriage, divorce, and other family matters; Malawian ethnic groups aim to control rules governing inheritance, authority, and land rights.

Both states and arenas of authority seek sovereignty, but the nature of sovereignty the state requires may differ from that sought by social arenas. Here it is useful to recall Reference Jackson and RosbergJackson and Rosberg’s (1984) distinction between empirical and juridical statehood in sub-Saharan Africa. Empirical statehood is based on the exercise of power – the monopoly of the legitimate use of force on populations within a territorial entity – and it exercises power with an eye inward, towards the national population. Juridical statehood is based on international recognition, and its aim is to be included in the international society of states. In the modern era, states maintain themselves by obtaining de jure recognition. Arenas of authority neither necessarily require nor receive such recognition. They seek de facto power over the issues that are critical to their community (a point we will return to in Section 4), and they vary in the extent to which this requires territorial control.

These competing aspirations for sovereignty open a range of possibilities for the relations between states and other arenas. At times, conflict is inevitable. Individuals located simultaneously in these overlapping arenas can comply with one authority only if they fail to comply with another. ISIS and states in the West provide a striking example of this. The estimated 20,000 foreigners who fought on behalf of ISIS in Syria and Iraq could only respond to ISIS’ authority by opposing that of their home states, effectively committing treason. At other times, states can cede authority over individuals, or over some spheres of their social and economic activities. Reference MigdalJoel Migdal (1988) highlighted the fact that states often do so when they are incapable of maintaining rule or unwilling to expend resources necessary to control populations or territory. This occurs not only when states are weak but also when they are strong – as seen in governments’ hesitance to enter inner-city ‘ghettos’ of the United States (Reference LustSampson et al., 1999) and marginalized suburbs in Sweden (Reference Esaiasson and SohlbergEsaiasson and Sohlberg, 2020).

Before turning to the institutions within these arenas, note that not all entities outside the state are social arenas. Non-profit organizations and corporations often have more limited goals. They aim to produce widgets, garner profits, and maybe even ensure the longevity of the enterprise. They can benefit from and thus may seek to develop the community’s skills. But they do not aim to govern the community. Reebok may care what its employees wear on their feet and seek customer loyalty over Adidas, but their board – most likely – cares little about who they marry and how they live. Where they do (e.g., family businesses, company towns), the economic system is better understood as an arena of authority than an organization. Indeed, markets and class-based systems are arenas of authority. However, organizations, as entities, that work unproblematically (save, at least, for illegal enterprises) with and under the state as well as other arenas of authority, and do not compete over individuals’ allegiances, are not. (See Table 1 for a summary.)

3.2 Defining Social Institutions

Social institutions structure engagement within arenas of authority. Sociologists debate the definition of social institutions perhaps as much as political scientists tangle over how to conceptualize the state (Reference MillerMiller, 2019). Some scholars define social institutions narrowly, such that they refer to the rules that govern actions and the rewards or punishments associated with compliance – a view of institutions in accordance with that of the Nobel Prize-winning economist Reference NorthDouglas North (1990), among others. Yet, rules invoked are often role-dependent, varying according to the position that one holds in the community. Moreover, even when enforced, the impact of various rules depends on the magnitude of costs or benefits associated with compliance. To highlight this, I define social institutions as the roles, rules, and rewards that structure activities within a community as it attempts to govern and ensure its survival.Footnote ¹⁸ Social institutions determine individuals’ positions within a community, the actions available to them and others, and the consequences thereof. In doing so, they affect the distribution of power in the community, members’ expectations of others’ responses, and consequently, individuals’ decisions. (See Table 2 for a summary.)

Because arenas of authority are multifaceted, the social institutions within them regulate a range of needs. These extend beyond the issue around which the community is formed. For instance, social institutions within ethnic arenas of authority determine not only who belongs (kinship), but also who has access to land and other resources (the economy), who has the right to rule the community (government), and what knowledge is passed down, when, and to whom (education). Institutions within religious communities regulate not only the relationship between members and their god (religion), but also tithing and charity (the economy), marriage and reproduction (kinship).

Within each arena, individuals assume roles. These are the positions that members of the community can hold, distinct from the individuals who hold them. Roles are associated with rights and responsibilities, and the rewards that follow. To simplify, there are leaders and members. In religious or kinship systems, for instance, these are priests and parishioners, imams and followers, chiefs and subjects, elders and youths. The ability to access different positions may depend on gender, age, bloodline, or other qualities. Where these characteristics shape roles within arenas of authority, we can understand them as gendered, ageist, or ethnic, respectively. Yet, such characteristics as gender, age, or ethnicity are not arenas of authority in and of themselves. Arenas of authority are associated with an identity group that defines the community, but not every identity group constitutes a community that defines an arena of authority.

The roles that men and women, the elderly and young, or those of different bloodlines are allowed to occupy vary across time, and across arenas. Again, this is true even in communities founded on similar bases. Consider the relationship between gender and leadership roles in religious arenas. In most religions, women are subordinate to men, barred from taking the highest leadership roles. Yet this varies across religions, and even within them. For instance, while most Protestant Christian religions allow the ordination of women, the Southern Baptist Convention, which constitutes the largest protestant denomination in the United States, bars women from the highest positions and promotes distinct gender roles in the household, encouraging women to submit to their husbands’ leadership (Reference CraryCrary, 2019). Gender roles also vary across time. The Catholic Church, for example, prohibited women from any formal leadership roles until January 2021, when Pope Francis amended Canon Law to allow women to act in leadership roles played by lay leaders (e.g., altar servers or readers) (Reference HunterPovoledo, 2021).

Rules determine who may assume certain roles, as well as who can enter the community and obligations of those within it. Think of rules as statements that determine who is permitted or required to/not to take an action, within specific conditions, and with expected sanctions or rewards for compliance/non-compliance.Footnote ¹⁹ Rules set forth expectations and implications, but they allow individual choice. It is true that an individual does not get to choose ‘on what’ they act (i.e., whether or not there is a rule regarding the choice at hand) or the consequences of their actions, but they decide what action to take.

Finally, social institutions entail the rewards and consequences of actions. John Harsanyi writes that ‘People’s behavior can be largely explained in terms of two dominant interests: economic gain and social acceptance’ (Reference Harsanyi1969, p. 524). Both come in the forms of carrots and sticks. Individuals may receive bonuses or economic rewards, or face fines or material losses. They may enjoy respect and moral standing or face social shaming and ostracization.

Roles, rules, and rewards combine to shape the choices and actions that are the core of politics and development. Coercive power has long been held as the key motivation for compliance with state authority, spanning from Hobbes and Mill to modern theorists, such as Reference Smith and WaldnerTyler (2006). Scholars recognize that coercive power, or carrots and sticks, may motivate actors with regard to non-state leaders as well (see Reference Falleti and CunialScott, 1972, or Reference Mares and YoungMares and Young, 2016). But desire to do the right thing may as well. Individuals may respond to calls for action because they believe that it is right and proper to do so (Reference LipsetLipset, 1959, p. 71). In this case, they contribute regardless of other’s abilities to reward or sanction (Reference BeethamBeetham, 1991; Reference Kelman and HamiltonKelman and Hamilton, 1989; Reference HaggardSparks et al., 1996; Reference Falleti and CunialSunshine and Tyler, 2003; Reference McGuireTyler, 1990).

Before discussing the stability of arenas of authority and social institutions, it is important to clarify what social institutions are not. Social institutions are not culture – if culture is defined as beliefs and preferences – even if culture includes preferences over political outcomes or policies (e.g., Reference Almond and VerbaAlmond and Verba, 1965; Reference Inglehart and WelzelInglehart and Welzel, 2005).Footnote ²⁰ Expressed preferences may reflect social institutions, as individuals may develop or choose to reveal preferences that are in accordance with social rules in order to avoid repercussions. However, preferences should not be confused with the rules that shape them. Social institutions, like social norms, are also not simply reactive instincts or habits, like quieting a crying baby or writing a diary (Reference Bell and CoxBell and Cox, 2015; Reference Legros and CislaghiLegros and Cislaghi, 2020). Social institutions entail a set of obligations and consequences. In this respect, social institutions are similar to injunctive norms, although not to descriptive ones. However, as Reference Legros and CislaghiLegros and Cislaghi (2020) show, there are multiple, and at times incompatible, understandings of ‘social norms’. Thus, for clarity, I focus on social institutions defined in terms of roles, rules, and rewards.

Finally, and importantly, social institutions are not simply informal institutions. Scholars and practitioners tend to equate state rules with formal institutions, and those outside the state with informal ones. Yet the rules in social institutions can be formal (where formal is understood as explicitly stated, parchment institutions) or even change in their degree of formality over time. Moreover, as Samuel Reference BowlesBowles (2016) argues, such codification may, itself, influence the outcomes. Only by distinguishing between formal and informal rules in social institutions can we explore the effect of formalization.Footnote ²¹

3.3 The Stability of Arenas of Authority and Social Institutions

Arenas of authority and social institutions within them need to be relatively stable and not simply reflect other underlying factors if they are to form the basis of a useful approach to understanding political behaviour and development outcomes. Communities develop social institutions as they seek to govern themselves, and they do so in response to specific challenges their context raises. Consequently, one of the thorniest issues in the study of authority and institutions is separating their effects from the factors that give rise to them. If social institutions simply reflect existing conditions, then one cannot assess their impact independently from these conditions (e.g., address endogeneity issues). However, there is good reason to believe that although arenas and social institutions are neither exogenously determined nor static, they are relatively stable and, over time, may become divorced from their initial purpose. They thus provide useful leverage for research and programming.

Evolutionary biologists, cultural anthropologists, economists, and psychologists have argued that norms develop as a mechanism to ensure a group’s survival (Reference Bowles, Choi and HopfensitzBowles et al., 2003). These may be shaped partly by ecological conditions. For instance, Reference HaggardHarry Triandis (1989, p. 510), a pioneer in cross-cultural psychology, argues that thinly populated areas were more likely to develop individualist cultures; since ‘[o]ne can scarcely reject a neighbour if one has only one neighbour’, those living in sparsely populated areas came to accept a great deal of diversity. Economic institutions may also influence social institutions. Reference Henrich, Boyd and BowlesJoseph Henrich and colleagues’ (2001) masterful study of reciprocity in fifteen small communities – across twelve countries on five continents – found that how much people rely on market exchange in their daily lives and the level of cooperation required for economic production affected the levels of cooperation in lab-in-the-field games. People who were nearly self-sufficient and engaged in small-scale production, such as those living on family-based farms in Peru or Bolivia, were much less likely to cooperate. In contrast, Indonesians who relied primarily on large-scale whale fishing – where more than a dozen men may set out in a large canoe to hunt whale – were much more likely to cooperate. These findings are in line with Reference MacLeanLauren MacLean’s (2010, chapter 6) study of reciprocity in Ghana and Cote d’Ivoire. She finds that the shift from cocoa to tomato farming expanded the scope of social ties and made relations more diffuse in Ghana, while the failure to make such a shift led to reciprocal relations that were stronger but also more concentrated among a smaller family group in Cote d’Ivoire.

Institutions are continually contested and evolving, but they are also sticky. Forces such as urbanization, technological changes, ageing populations, and efforts aimed at shaping societies through rules aimed at ending ethnic, sexual, or gender discrimination are all catalysts for change. Nevertheless, institutions outlast the conditions that gave rise to them, despite constant pressures (Reference Bowles, Choi and HopfensitzBowles et al., 2003). Moreover, the constellation of roles and rules that define social institutions shape the distribution of power. Thus, a change in a single rule that may disrupt the balance of power is often countered by other rules. Reference Michael Albertus and SlaterRubie Watson (1990, p. 241) illustrates this with regard to inheritance laws. She argues that the change to partible inheritance, or the division of land equally among offspring in the family, led to downward mobility in China but not in Europe because partible inheritance in Europe was coupled with rules that promoted late marriage, close-kin marriage, and marriage within class. In short, institutions are sticky.

More importantly for the purpose here, individuals making decisions – whether to run in elections and whom to support, whether to contribute to community funds, or whom to turn to in resolving disputes – generally make their choices in the context of relatively legible and stable institutions. When they do not – when technological, environmental, medical, or other changes heighten uncertainty – it is possible to consider the range of outcomes they expect, still taking into account their perception of social institutions. Political scientists should be concerned about endogeneity issues, cautioned against attributing causal influence to institutions or assuming that thorny political problems can be solved through (re)designing state institutions. However, there is much to be gained by considering arenas of authority and social institutions at a given point in time. To understand decisions and subsequent outcomes from the perspective of everyday people engaged in making choices, the origins of different authority structures are of little consequence. What matters are the communities they belong to, their roles within these communities, the attendant rules and associated rewards, and the extent to which these affect the decision at hand.

4.1 Salience of Arenas

The first challenge is to determine which arenas of authority are associated with the issues in question. Two factors come into play. First is the extent to which the issue is salient to fundamental tenants of the community, and thus salient to the community’s imperative to maintain itself. Second is the extent to which elites within an arena use the issue to extend their power vis-à-vis competing elites from other arenas.

To understand whether individuals see the issue at hand as relevant to the community, Kate Baldwin, Kristen Kao, and I find value in the notion of domain congruence.Footnote ²² Domain congruence reflects the extent to which an issue or activity and the arena of authority are related. If individuals believe that the community cares deeply about the issue, they are more likely to expect that leaders and community members will expend resources to sanction noncompliance. Non-coercive power is also likely to be higher when there is greater domain congruency. Citizens may view it as more normatively appropriate to comply with demands of authorities and other community members when the activities called for match their domain of power, or where they have greater expertise (see Reference McGuirePresthus, 1960, p. 195 for the classic statement of this). Individuals thus should be more likely to respond to demands made when they are associated with arenas in which there is greater domain congruence.

We find evidence of this in a study of authority in Kenya, Malawi, and Zambia. The study employed a survey experiment aimed at understanding how citizens respond to requests when asked to do so by different authorities.Footnote ²³ It focused on three activities (voting for a presidential candidate, contributing to a school fund, and contributing to a burial fund) and sought to understand the extent to which state and traditional leaders at local and supra-local levelsFootnote ²⁴ mobilize support. We find that the relative power of leaders depends on the match between their domain (or arena of authority) and that of the activity. For instance, the MP, as a supra-local, state leader, was better placed to mobilize support for the presidential candidate than to mobilize contributions to a burial or school fund. In contrast, the local councillor had greater influence on contributions to the school fund, which can be understood as a local, state-oriented activity.

Individuals view different arenas as more or less appropriate to issue areas even when they are located at the same level (e.g., are local or supra-local) or are of the same type. For instance, in a study examining when individuals comply with directives to take Covid-19-related precautionary measures, my colleagues and I found that Malawians viewed village heads as more legitimate in issuing such directives than local religious authorities. They appear to view advocating precautionary health measures as fitting within the village head’s purview of maintaining community welfare – an association that may be reinforced by the role these village heads play in organizing immunization drives and other health initiatives (Reference Kao, Lust and DulaniKao et al., 2021).Footnote ²⁵

Issues may also be salient because they are part of a broader struggle between communities in different arenas. Elites and other community members in arenas of authority can compete with those of other arenas to gain or maintain their power. For example, they can use action-oriented issue framing to shape members’ understanding of the activity, the choices at hand, and their implications. Take, for instance, dispute resolution in Jordan. Jordanian tribal codes generally demand that parties to a conflict not only accept the solution but also view it as a final reconciliation. Thus, as the head of public relations for the Jordanian Public Safety Department explains, ‘Tribal customs preserve the restoration of social balance after a crime has occurred’, whereas when state directives are issued, tribal processes are still required to prevent continued conflict (Reference HaggardWatkins, 2014, p. 40). The rules guiding the tribal process imbue the outcome with a fundamentally different meaning from one reached within the state courts alone, and they help to maintain the tribal arena.

Similar dynamics are found in the West. In his study of a ranching community in northern California, Reference EllicksonEllickson (1994, p. 60) highlights how the community maintained social order apart from the state. In part, this was done by rules that defined ‘being a good neighbour’ as one who sought to solve such issues as cattle poaching and breach of agreements within the community, rather than to take disputes to state courts. Thus, when deciding whether to take an issue to the state court or address it locally, the rancher not only views the choice as one over the issue itself, but also as one that signals their adherence to the ranching community.

Simply framing an issue as part of a larger goal does not necessarily ensure that members view it as salient. Social movement theorists argue that the success of such framing depends on the consistency of the message, its empirical credibility, and the credibility of its messengers (Reference Benford and SnowBenford and Snow, 2000). Success also depends upon the strength of competing arenas. Reference KohliThachil (2014), for instance, finds that the elite-led Bharatiya Janata Party (BJP) used a combination of Hindu nationalist rhetoric and privately provided services to mobilize support from lower castes in Chhattisgarh. However, its strategy met strong resistance in Uttar Pradesh, where the Bahujan Samaj Party (BSP) had successfully won over the lower castes, thus polarizing the class divides. Even when communities use similar strategies to influence individuals’ decisions, different arenas of authority may not have equal sway.

4.2 Strength of Arenas

The influence of different arenas of authority over individuals depends on the strength of arenas. This is determined, in part, by the nature of the arena: the level of solidarity among its members, the range of aspects within individuals’ lives that the arena touches, and the extent to which the community and its leadership can monitor and sanction members. Other factors are specific to individual members: the extent to which they are beholden to the community and the exit options they enjoy.

First, social solidarity strengthens arenas. Members are more likely to view obligations as legitimate when they feel themselves to be part of the community. In such cases, they are also more likely to feel they will reap benefits from their contributions when they act together. This is particularly true in small communities, where the fact that each contribution counts can help overcome collective action problems (Reference OlsonOlson, 1965).

Two studies of India highlight the importance of solidarity. India’s caste system is a well-known example of a closed social system, with caste-specific roles, rules, and rewards, and clear discrimination of the lowest cast, the Dalit or ‘untouchables’. Often, there is solidarity within the castes but not across them. Reference WaringTimothy Waring’s (2011) research on irrigation in villages of Tamil Nadu illustrates how individuals resist demands to contribute when solidarity is absent. He finds an association between the size of one’s caste in a community and how much labour people volunteered to cooperative irrigation systems: individuals volunteered 77 per cent more days, on average, if the share of their caste rose from 0 to 50 per cent of the village. Moreover, lab-in-the-field experiments revealed that Dalits in Dalit-non-Dalit groups were less likely to contribute, and Dalit collaborators confirmed that ‘Dalits often reduce their cooperation when called upon by high caste people, because they assume that they will not benefit from any project they are asked to support’ (Reference Waring and BellWaring and Bell, 2013, pp. 402–3). Yet, Reference Smith and WaldnerPrerna Singh (2011; Reference Lust2015) argues that such solidarity, or what she called ‘We-ness’, can develop in communities even when they are divided by religion or caste. Examining Kerala, India, she argues that Keralites established a ‘We-ness’ based on a common language and shared Malayali culture. This allowed them to bridge religious and caste divides, support community services, and even engage effectively in the battle against the coronavirus pandemic (Reference KohliSingh, 2020).

Studies of race and voting also suggest that individuals respond more to demands made in arenas of authority where they feel themselves to be full members of the community. Reference DawsonMichael Dawson (1995) argued that African Americans complied more with the demands of their racial groups than class groupings because experiences with racial discrimination led them to view their personal interest as closely tied to that of other blacks, or to have what he called ‘linked fate’. Importantly, in this view, social institutions that guide the behaviour of groups in other arenas of authority (i.e., the rules leading non-African Americans to discriminate against out-group members) may foster solidarity within a community. Such feelings of solidarity affect the extent to which individuals vote in accordance with social institutions in racial arenas of authority. In the United States, black and Latin American respondents who expressed sentiments of linked fate were more likely to support minority candidates (Reference Bejarano, Brown, Allen Gershon and MontoyaBejarano et al., 2021). In contrast, in South Africa, individuals whose appearance led them to be singled out as different, and thus experience fewer feelings of linked fate, were less likely to vote with their group (Reference HarrisHarris, 2022).

Second, the more multifaceted the arena of authority – encompassing a wider range of behaviour and outcomes relevant to individuals’ welfare – the greater its influence over individuals’ choices.Footnote ²⁶ The rewards of complying with rules within an arena relate not only to the direct act at hand, but also to the indirect consequences that come from maintaining strong relations with other members of the community. In arenas associated with a wide range of activities and future benefits – finding a job, securing one’s property, obtaining help in old age – individuals have greater incentives to comply with the social institutions.

Examples abound. Reference SchafferFrederic Schaffer (1998) finds that individuals vote in accordance with male heads of households because the costs of transgressing patriarchs (i.e., social institutions in kinship arenas) go far beyond voting day. He explains:

Encompassing arenas are found outside the Global South as well. Studying the United States, psychologist Paul Piff and his colleagues (Reference Piff, Kraus, Côté, Cheng and KeltnerPiff et al., 2010; Reference Haggard2012; Reference Piff and RobinsonPiff and Robinson, 2017) find that the poor are more likely to respond to social obligations around family and friendship arenas. This may be because the poor rely on neighbours as their safety net, and thus they are much more likely to expect and give assistance to each other. The American poor act much like the Ghanaians in Reference PankaniWinifred Pankani’s (2014) study discussed at the outset of this Element. Similarly, studying poor communities in England, Reference JöstPrisca Jöst (2021) finds evidence that the poor’s voting behaviour is more closely associated with their beliefs over how others in their community vote than is the case for wealthier Brits. She argues that the poor have fewer links to individuals outside their community and thus fewer external options. Consequently, the local arena of authority is more encompassing, having a greater influence on the poor.

Encompassing arenas affect not only members, but leaders as well. Lily Tsai’s influential study of service provision in China demonstrates this nicely. Tsai centres her argument around solidary groups, in which members are ‘obligated to behave and judged according to the group’s standards of what constitutes a good person and a good member – “good” meaning not just what is good for the group but the goodness (versus badness) of human action and character’ (Reference Falleti and CunialTsai, 2007, p. 95). When solidary groups, which are akin to arenas of authority, encompassed all citizens in the local government jurisdiction and embedded local officials, both citizens and local officials were more responsive.

Third, arenas are stronger when leaders or other community members are more capable of monitoring and enforcement. Rules that enhance visibility foster compliance. This explains why many communities encourage contributions during weddings, public benefit auctions, and other ceremonies during which behaviour is publicly observed, emotions of solidarity run high, and social sanctions are particularly effective (Reference AmbecAmbec, 2008). Proximity facilitates monitoring as well, as it allows authorities and other community members a better vantage point for monitoring individuals. Finally, in arenas with dense social networks, in which people have a large number of ties with others in the community, information flows more swiftly through communities, and leaders and members are more likely to know of and sanction non-compliance. This role of networks may explain why individuals respond more to demands of coethnics than to non-coethnics (Reference Habyarimana, Humphreys, Posner and WeinsteinHabyarimana et al., 2007; Reference Miguel and GugertyMiguel and Gugerty, 2005), as well as Reference GranovetterMark Granovetter’s (1973; Reference Granovetter1983) finding that weak ties are more beneficial than strong ties in mobilization.

Prisca Jöst and I (2022) uncover evidence of the role of networks in a study of the poor in Kenya, Malawi, and Zambia. Using a hierarchical model that allows us to consider the relationship between community social ties and compliance, we analyse the survey experiment on authority that Kate Baldwin, Kristen Kao, and I employed (see Section 4.1). Prisca Jöst and I find that the poor are more likely to participate when asked by local traditional authorities and neighbours than they are when asked by more remote leaders, a finding that highlights the importance of proximity. Moreover, these local authorities had even greater influence over respondents living in communities with dense social ties. Higher expectations of social sanctioning in these communities appear to explain greater compliance, at least in part. Networks within communities – geographical or otherwise – facilitate monitoring and enforcement, strengthening the arena of authority.

Finally, individuals’ responsiveness to different arenas of authority depends not only on the nature of the arena, but also on individuals’ characteristics. To some degree, there are idiosyncratic differences in people that make them more or less likely to comply with authority – state or otherwise. Reference EllicksonEllickson (1994, Chapter 3) describes a rancher in Shasta, California, who seemed impervious to the rules and rewards of the ranching community, for no obvious reason. However, circumstances can also systematically affect the hold that arenas have over individuals. Arenas have less influence over those who can more easily opt-out of the community or weather the costs associated with reneging on obligations. As Reference HaggardTriandis (1989) notes, in general, more affluent and more mobile individuals can escape the watchful eye of their community; the poor and less mobile rely more on benefits accorded by compliance, and are less able to escape sanction.

This helps to explain why individuals who are disadvantaged within an arena may nevertheless turn to it, even when given the opportunity to venue shop. Reference KohliJustin Sandefur and Bilal Siddiqi (2013) demonstrate this in Liberia. There, women were far more likely to take their disputes to customary, rather than state, fora, even when they were aware of their alternative options, and despite the fact that customary tribunals disadvantage women. The exception was women in the process of suing men, who were more likely to file charges in state courts. However, this exception may prove the rule. In this context, it is extraordinary for women to sue men, a move afforded to those who are better able to refuse compliance with the social institutions. Expecting that alternative venues will free the disadvantaged from repressive relationships misses the fact that those most disadvantaged are often least well-positioned to escape the arena of authority.

4.3 Shaping the Decision Field: Institutional Congruence versus Identity Cleavages

The discussion so far has focused on the factors that strengthen arenas, at times leading one to dominate another. In reality, however, people are members of numerous arenas, some of which may dictate the same action when choices are set before them. Where this is the case, individuals may behave in accordance with social institutions of multiple arenas, none of which is – by itself – the strongest. Thus, what matters in determining how arenas come together to shape an individual’s choice is how social institutions within these arenas overlap, not how membership does.

Scholars of identity and politics have focused on the latter: the overlap in membership, or put differently, identity cleavage structures. Some emphasize nested groups, focusing on cases in which individuals in one community are subsumed in a larger community (e.g., members in a locality are all members of a larger region) and asking which groups individuals respond to most (Reference LawlerLawler, 1992; Reference Lawler, Thye, Yoon and AbrutynLawler et al., 2016). Others are particularly interested in how cleavage structures affect outcomes such as conflict (Reference Gordon, Landa and BihanGordon et al., 2015), democratization (Reference DeutschDeutsch, 1961; Reference LijphartLijphart, 1977; Reference LipsetLipset, 1959; Reference Lipset1960), and economic growth (Reference HunterSelway, 2011).

These scholars emphasize different social arrangements, but they largely agree that what matters is how these different identity-based communities influence individuals. Reference Smith and WaldnerTaylor and Rae (1969, p. 534) refer to the different ‘political norms’, associated with an individual’s membership in different groups, that lead the individual to be ‘cross-pressured’. Reference LipsetLipset (1960, p. 88) emphasizes the ‘multiple and politically inconsistent affiliations, loyalties and stimuli’ (my italics) that individuals face. Reference Lawler, Thye, Yoon and AbrutynLawler and colleagues (2016, p. 149) point to ‘nested group commitments’ which can at times be in conflict – as the example of Dalits in Tamil Nadu illustrated.

However, the cleavage structure of identity groups does not necessarily correspond to the overlap of social institutions within them. Identity groups may be nested and yet make competing demands on their members. At the same time, identity cleavages may be cross-cutting and yet require the member to make the same choice.

To understand this, consider the Ghanaian bureaucrat presented in the introduction. The bureaucrat had spied her chief standing in the office queue, waiting to process his papers, and led him to the front of the line. This response would be consistent with the constellation of rules in Scenario 1, presented in Table 3. In this case, the bureaucrat is a member of the state, her ethnic group, and her kinship group. She recognizes that the administrative rules require her to treat applicants in the order they arrive, but social institutions within her ethnic arena as well as those within her more immediate kin arena require her to serve the elder first. In the absence of strong state institutions, she does so. In the second scenario, however, she may make a very different choice. The institutions associated with the state and ethnic arenas remain the same. However, imagine she comes from a very nationalist family, so rules in her kin arena require her to uphold state rules. Now, even if the ethnic arena has the single largest hold (e.g., she anticipates good standing in her ethnic group affects her ability to get other jobs or marry well), she may still require the chief to wait his turn. She does so when there is a congruence of rules between the state and kinship arenas, and when they are jointly stronger than her ethnic arena.

4.4 Illustration: Subnational Variation in Jordanian Electoral Participation

Before turning to social institutions in greater detail, I want to illustrate the points set forth thus far by examining electoral participation in Jordan. The case demonstrates variation in the strength of arenas of authority based on kinship, religion, and state, and shows how differences in these arenas and individuals’ reliance shape the influence of different arenas over their electoral behavior. Moreover, the case illustrates how subnational variation in the strength of these arenas affects electoral participation – in particular, why Jordanians living in rural areas have been much more likely to vote than those living in the capitol city, Amman.

Four arenas of authority are relevant to voting in Jordan: the state, tribe, family, and religion. With regard to the state, individuals are expected to vote for those who will represent their interests in the parliament. With regard to the tribe, parliamentary elections are an opportunity to respond to obligations, to demonstrate tribal allegiances, and, if one is lucky enough to have a tribal member elected, to activate the representative’s obligations towards tribal members and secure access to resources. Tribal arenas thus promote voting for ‘their own’ as a way to seek and fulfil tribal obligations.Footnote ²⁷ With regard to family, close-kin ties add an additional layer of obligation that goes beyond that of tribe, and one that is activated for Jordanians who do not identify with a tribe.Footnote ²⁸ Finally, with regard to religion, Islamist communities promote voting for Islamist candidates as a demonstration of religious conviction.

Perhaps surprisingly, the state plays little direct role in Jordanians’ choices over candidates. Jordanians generally do not see casting ballots as a way to ensure effective political change. Since the parliament’s reopening in 1989, the legislative body has been increasingly impotent in the face of the monarchy, and citizens’ trust in it has only declined (from 46 per cent of Jordanians stating they had some or great trust in the parliament in 2010 to only 14 per cent stating this in 2018) (Reference BarometerArab Barometer, 2019, pp. 8–9). Even most elites do not view obligations to the state as a primary reason to vote. An elite survey published in 2012 found only 30 per cent of 185 respondents said they participated in the 2010 elections because it was their national duty do so or in order to enhance democracy – compared to over half who said they did so due to their tribal or social ties (Reference Al-AzzamAl-Azzam, 2012, p. 355). Moreover, Jordanians may rely on the state for many things, but they do not need to vote to receive services in exchange. Electoral rules affect the choices before Jordanians, as we will see in Section 6, but obligations to the state and the expectations of rewards in the state arena do not drive voting.

Tribal, family, and religious arenas have greater influence on voters’ decisions. For Jordanians of tribal origin, and particularly those living in rural areas, shaykhs play an important role in everyday life: they resolve disputes and help access services. At election time, many tribes use primaries to choose candidates and issue endorsements, sending clear messages that voters should support ‘their own’ (Reference KaoKao, 2015; Reference HaggardWatkins, 2014). Obligations around family are even more strongly held. Even Jordanians of East Bank origin, for whom tribe is important, seek to have close relatives and family members elected over other tribal members, and some tribes institute rotation agreements that allow each family to take a turn accessing parliament. Finally, Jordanians rely on their religious community for spiritual guidance and well-being as well as financial and material support (Reference ClarkClark, 2004; Reference Smith and WaldnerWiktorowicz, 2003). Thus, calls to support Islamist candidates have sway.

The extent to which these arenas influence electoral engagement depends on the electoral context and voters’ characteristics. I find evidence of this in an analysis of survey data collected with Lindsay Benstead and Kristen Kao in 2014 (see Appendix B for details). Members of tribes were more likely to feel obligated to support and vote for a candidate who is a tribal or family member. This was especially true of those with little education and those living in rural areas, who are more highly dependent on tribes for governance and welfare. Moreover, when a change in the electoral law (described in Section 6) forced voters to choose between supporting candidates on the basis of religion or tribe, most rural Jordanians of East Bank origin supported their tribal candidate (see Reference PatelPatel, 2015). In contrast, those living in cities, where tribal influence over governance is weaker, were more likely to vote for a candidate supported by their local religious leader.

Or, to put this in a more stylized fashion, consider a Jordanian voter who has three candidates before her: a State Candidate, who is a non-Islamist, loyalist to the regime with no tribal or family ties to the Voter; a Tribal Candidate, who is a pious member of the voter’s tribe but not her family; and an Islamist Candidate, who is a pious candidate but a vocal opponent of the monarchy. Note that the rules for the tribal and religious arenas are congruent if the candidate is both tribal and religious. If the voter makes choices based only on the state arena, they vote for the State Candidate, if only on the tribal arena, for the Tribal Candidate, and if on the religious arena, for the Islamist Candidate. In this example, taking the strength of the arena into account, they vote for the Tribal Candidate in rural areas and the Islamist Candidate in urban areas.

A perspective that pays attention to the strength of competing arenas of authority also helps explain variation in turnout. Voting is higher in rural areas, where tribal arenas are stronger. Voting is not purely voluntary, as tribal members sit outside polling stations to mark who has voted, sending out for those who fail to arrive.Footnote ²⁹ Nor do all tribal members benefit in the future. In 2013, Jordanian women, gathering under a tent near a polling station in Tafilah, where they had been bussed in to rally for their tribal candidate, complained that the candidate would ‘throw away his telephone’ as soon as the elections were over.Footnote ³⁰ Nevertheless, individuals vote, and they do so in response to the salience and strength the tribal arena has in elections. This explains why turnout in Jordan’s 2013 elections reached nearly 80 per cent in rural tribal areas compared to 34 per cent in Amman.Footnote ³¹

5.1 Group Membership and Boundary Maintenance Rules

As discussed in Section 3, arenas may have more or less porous borders, making it easier or more difficult for individuals to enter or exit the community. Boundary maintenance rules also establish how an individual should act in order to demonstrate allegiance to the community. This can be over appearance – the side curls of Hasidic Jewish men, the distinctive blue dresses and white bonnets of Amish women – or over actions, such as prohibitions against publicly airing views that oppose the group’s tenets. Rules governing group membership and boundary maintenance not only help to establish and signify membership in the group, but also influence a wide range of engagement and outcomes, from electoral competition and representation to citizenship and policymaking.

Where boundaries are relatively closed, competition and participation take on new meaning. Campaign events become theatres for reinforcing community solidarity. People attend rallies because they want to show their support for and strengthen ties with the candidate and community; many are family, friends, neighbours, or co-congregationalists.Footnote ³² Events are technically public, but they border on exclusivity. This is particularly true in constituencies with competing, tightly knit communities. In a campaign study I conducted during the 2010 Jordanian elections, for instance, one researcher was asked to leave a campaign tent because he was from another local family and thus suspected of being a spy. The candidate and his entourage did not believe the event could be used to sway the man’s vote. Similarly, surveys of individuals attending the Jordanian and Egyptian 2010 campaigns found that a large percentage of respondents saw campaign events as ‘closed’.Footnote ³³ Even in Tunisia’s 2011 campaign, which took place after the fall of Ben Ali and when public mobilization was especially high, 14 per cent of campaign attendees we surveyed viewed campaign events as exclusive.Footnote ³⁴ In such cases the locus of competition shifts to the nomination stage (much as it does in one-party states). Tribes, families, and religious congregations choose candidates.Footnote ³⁵ At times, they institute rules such as rotation of candidates across clans or families to reduce the potential for internal conflict, or even establish tribal primaries which are sometimes as formalized and hotly contested as general elections. Overall, the intention is to appear as a unified block, signalling to candidates outside the group that they are unlikely to succeed in garnering their votes.

A similar example is found in Libya. In 2012, a General National Congress (GNC) member described to me how rules aimed at maintaining the appearance of group coherence – or group boundary maintenance associated with the tribal arena – impacted elections. His small tribe had formed a coalition with two other tribes to run for a proportional representation (PR) seat in his district. They chose their coalition strategy by ‘counting heads’ of the different tribes, and they organized their campaign at weddings – a space in which their gatherings would be isolated from, and go undetected by, larger tribes that may have tried to undermine their efforts. Importantly, they knew that not every tribal member would support the selected candidate, but they nonetheless sought to maintain the appearance of cohesion. Fellow members could choose to not vote for their candidate, but to voice this intention publicly – and especially in favour of a candidate from another tribe – was unacceptable.Footnote ³⁶ Such restrictions not only affect the outcome of elections, but also the expectations of other voters and the competitiveness of campaigns. When no one stands up in dissent, it is easier to believe everyone is united behind the candidate.

Boundary rules also play a role in the integration of refugees. Take, for instance, Jordanians’ responses to refugees fleeing the Syrian civil war. In some cases, they treated refugees as outsiders whose presence they resented, but elsewhere, they admitted them into the community. In part, this was because boundary rules of tribal arenas of authority already designated them as part of the in-group. For instance, members of the Bani Khalid tribe, which spans Jordan, Lebanon, and Syria, were ‘welcomed and accepted into the Jordanian community because they carried the same tribal name’ (Reference Miettunen and ShunnaqMiettunen and Shunnaq, 2020, p. 11). They were integrated, at least in part, because they entered the community as a member of the same tribal arena – one which transcended state borders.

Boundary rules governing social arenas of authority affect relations with the state as well. Reference MacLeanLauren MacLean (2010) shows how group boundary rules of the Akan in Ghana and Cote d’Ivoire shaped individuals’ relationships with the state. In Ghana, boundary rules designated a wider range of individuals as community members, but the rules governing members’ commitments only required low levels of mutual support. In contrast, in Cote d’Ivoire, boundary rules restricted community entry to a close circle of family members but required they provide each other a higher level of assistance. She argues that this had important implications for how individuals engaged with the state. Ghanaians had a much closer relationship with the state, viewed local officials as key actors, and saw engagement with the state in terms of civic duty, while Ivoirians understood the state to be a remote entity and any interaction with it as the reserve of ‘big men’.

Finally, studies of mobilization around AIDS issues also demonstrate the importance of group boundaries on political engagement and policymaking. Reference CohenCathy Cohen’s (1999) study of the African American community’s response to AIDS found that the community initially viewed the epidemic as one afflicting gay, white men, and, for some in the community, God’s punishment for sinful activity. Consequently, the leadership shied away from pressing for responsive policies. Similarly, Evan Reference LiebermanLieberman’s (2009) study of AIDS policies in Africa found that where group boundaries are strong (i.e., less porous), there was less support for broad policy responses to the epidemic. Where the disease came to be associated with certain groups – as in South Africa, where whites saw it as a ‘black disease’, while blacks perceived it as a ‘white disease’ – the policy discussion was one of blame and stigmatization, with little support for broad-based policies.

In short, boundary rules influence individuals’ behaviour and, ultimately, politics and development outcomes. The extant literature suggests some hypotheses that require further investigation. Stronger, less porous boundaries appear linked to more encompassing prescriptions on the behaviours of members, the appearance of less competition, and more difficulty for elites in other arenas to try to mobilize members across boundaries. This can have implications for policymaking and implementation when issues are associated with certain communities. Where there are less porous boundaries, in-group members may express greater support for policy responses, but they may be less willing and able to reach across boundaries to other communities, and thus to gain their support.

5.2 Rules of Engagement: Individualism-Collectivism

A second set of rules revolve around engagement: who is expected to act, how, and with regard to what? Here, I focus on the distinction between rules of engagement based in individualism and those centred on collectivism, a distinction that Reference Michael Albertus and SlaterHarry Triandis (2001, p. 907) has called the single-most important cultural distinction.Footnote ³⁷ This distinction has significant implications for the likelihood that individuals contribute to public goods or participate in political action, and, consequently, welfare and development outcomes.

All social institutions govern communities and are, in that sense, communal, but they differ in the extent to which they frame action in terms of individual or collective goals. In individualism, social institutions aim to create and maintain a community in which individuals are autonomous and encouraged to self-actualize. In collectivism, individuals’ independent desires are subordinate to the collective, and the ‘good’ is determined by what benefits the collective, not the individual (see Reference Markus and KitayamaMarkus and Kitayama, 1991; Reference HaggardTriandis, 1989). Those transgressing a community norm – speaking out against established power, voting for an ‘unacceptable’ candidate, refusing to contribute to community initiatives – bring not only social shame or physical and material harm to themselves, but also to others in the community. So, too, those who improve their welfare raise the community’s status and benefit all.

Whether social institutions centre on individualism or collectivism influences individuals’ decisions to contribute to community development.Footnote ³⁸ The traditional statement of the collective action problem presumes individualism. Individuals want to benefit from the collective good, so if their own actions alone will not determine success and if they will benefit regardless of their own actions, they will choose not to contribute. They need selective incentives, or to be engaged in an iterated game (Reference AxelrodAxelrod, 1984), in order to do otherwise. In contrast, in collectivism, the same conditions apply: individuals’ actions alone will not determine success, and these individuals may benefit from success regardless of their contribution. But collectivism entails additional incentives to contribute. Identifying one’s self with the community or holding strong notions of solidarity may drive contributions (e.g., Reference LawlerLawler, 1992). It is not surprising, thus, that social movements often put forth frames intended to inculcate collectivism (Reference Benford and SnowBenford and Snow, 2000). Moreover, where social institutions require one to put the collective good above one’s own, failing to do so may incur sanctions. Collectivism is not to be equated with ‘warm and fuzzy’ feelings (Reference Liu, Morris, Talhelm and YangLiu et al., 2019). In-group vigilance and social sanctions associated with collectivism can be painful, particularly where networks are strong (Reference Hu, Lin and CuiHu et al., 2015).

Studying Ghanaian villages and their home town associations in the Netherlands, Reference Mazzucato and KabkiMazzucato and Kabki (2009) show how collectivist social institutions foster community contributions. They describe institutions that local villages designed to ensure funeral contributions. Funerals are perhaps the most important of ceremonies in Ghana, and holding a funeral in one’s hometown is an act that both represents belonging and affords an opportunity to fundraise for the bereaved, maintaining social insurance. In these villages, funerals are also an opportunity to collect funds for the village, as 10 per cent of the contributions go to village development funds and 90 per cent to the bereaved family. Importantly, traditional authorities require their subjects to contribute to these funds regardless of whether they are residing in the village or abroad, attending the funeral or not. Moreover, according to the social institutions described, failure to pay incurs collective punishment: a family is not allowed to hold a funeral anywhere on the village’s territory if past funeral dues are unpaid. Migrants’ families back home are shamed if they fail to pay, and migrants suffer the costs of this shame. Thus, diaspora’s contributions to funeral costs and other collective causes, which are critical to development outcomes, continue in part because family members back home constitute linkages with collectivist social institutions. Importantly, Reference Mazzucato and KabkiMazzucato and Kabki (2009) note that such institutions are particularly effective in small villages, where monitoring and enforcement are easier (i.e., local arenas are stronger).

Whether social institutions are individualist or collectivist also shapes political participation. Indeed, distinguishing between individualism and collectivism may help reconcile debates found in the literature on clientelism and voting. Many studies of vote-buying assume that voters act within individualist contexts and vote-buying is a one-shot exchange between a broker and voter. Scholars then ask how brokers and candidates ensure that voters support them after the handouts are given. They search for mechanisms that facilitate monitoring (Reference Michael Albertus and SlaterStokes, 2005; see also Reference Hicken and NathanHicken and Nathan, 2020 for a review), or less frequently, individual-level traits – such as a strong intrinsic sense of reciprocity – that explain why voters remain true to brokers (Reference Finan and SchechterFinan and Schechter, 2012; Reference Lawson and GreeneLawson and Greene, 2014).

Yet, a very different picture emerges if collectivist social institutions drive voters’ decisions. Citizens may support candidates who are members of their community (e.g., family or kin, coethnic, or from the same locality) at least in part because they feel duty-bound to do so (Reference CrenshawCruz, 2019; Reference KaoKao, 2015; Reference Falleti and CunialRavanilla et al., 2021; Reference SchafferSchaffer, 1998; Reference Haggard2014). They may also believe candidates from their community are equally duty-bound to help them meet their everyday needs, easing their way through a maze of sluggish, unresponsive bureaucracy (Reference Lust-OkarLust-Okar, 2006; Reference Lust-Okar2009). Voters thus support the candidate because they believe that the candidate they support is the one who is obligated to help them or their community, given collectivist institutions in arenas of authority outside the state. As one Jordanian put it, ‘Everyone in [Jordan] needs a VIP to solve his problems’ (Reference Lust, Hourani and El-MoumaniLust et al., 2011, p. 120). The question for scholars and practitioners, then, is not which version of clientelism is correct but rather which social institutions structure individuals’ choices, and thus, what assumptions should guide research and programming.

In sum, the individualist-collectivist distinction in social institutions appears to have important implications. Where social institutions are individualist, one can expect contributions are made along the logic of the collective action problem and selective incentives may be required to motivate contributions. However, where institutions are collectivist, selective incentives are likely to be less effective. Social sanctions meted out by the community when individuals place their interests over those of the community motivate compliance. In short, much as state institutions that promote patriotism can engender individuals’ sacrifice for sake of the community, social institutions based on collectivism can spur cooperation.

5.3 Rules Governing Roles: Leadership Selection and Tenure

The final set of rules regard the roles individuals hold. In this section, I focus on rules governing leadership selection and tenure. These influence the extent to which leaders have influence over other community members, and the ability of community members to hold leaders accountable.

Leadership selection rules determine who is eligible to rule and who has a voice in choosing the leader, what Bruce Bueno de Mesquita and his colleagues called the ‘selectorate’ (Reference Bueno de Mesquita, Smith, Siverson and MorrowBueno de Mesquita et al., 2003). These determine the level of competition, ranging from highly competitive contexts, where many are eligible for leadership and the selectorate is large, to less competitive environments, in which the number of contenders is small and the selectorate limited.

There is widespread evidence that leaders are more responsive to members when the level of competition over leadership is higher. Chieftaincies in Sierra Leonne are found to be less repressive and yield better development outcomes when leadership selection is more competitive (Reference Acemoglu, Reed and RobinsonAcemoglu et al., 2014). A similar association exists between higher competition and better public goods provision in the Philippines (Reference Cruz, Labonne and QuerubínCruz et al., 2020). So too, Jordanians enjoy better municipal services where there is greater competition over local elections (Reference GaoGao, 2016), while in India, slum dwellers benefit when party brokers compete with each other (Reference AuerbachAuerbach, 2019).

Leaders, like others, are members of multiple arenas of authority, and the extent to which they respond to demands made by those in different arenas depends, in part, on how critical these arenas are to their success. Consider, for instance, Lebanese politicians, who act as elites in both sectarian communities and political parties, as well as within their respective arenas of authority. Melani Reference CammettCammett (2014) finds that service providers connected to sectarian-based political parties in Lebanon favour in-group members in districts where their sect is large enough to win elections but that they do not do so in districts in which voters from their sect do not constitute the plurality. In deciding to whom they grant favour, these politicians respond to demands from their party (and competition for leadership within it) and demands from their sect. Their decision to serve outsiders, which can be understood as the fulfilment, or not, of obligations to favour their sectarian community, is based on the logic of political competition.

Rules over the length of leadership tenure affect leader responsiveness as well. When leaders’ tenure is long and their welfare dependent on the community (e.g., exit options are limited), they can personally benefit from the development of their local area. Drawing on Mancur Reference OlsonOlson’s (1993) concept of a stationary bandit, Kate Reference BaldwinBaldwin (2016) argues that leaders with longer time horizons are more likely to see the benefits of fostering development. Thus, traditional authorities in Zambia act as development brokers, and not simply vote brokers, because the traditional institution of life-long hereditary rule encourages them to invest in their area’s development.

Yet, to be a ‘stationary bandit’ requires that the leader enjoys security. Reference PallerJeffrey Paller’s (2019, p. 20) study of urban politics in Ghana points to how land security influences leader responsiveness. He distinguishes between three types of settlements: indigenous settlements (with leaders most akin to Baldwin’s stationary bandits), stranger settlements (with diverse communities but secure land tenure), and squatter settlements (newly emerging settlements with no sense of belonging and little land security). In the latter, he argues, leaders do not face periodic elections or other term limits, but they focus more on their position within other communities (and arenas of authority) than on the local settlement. He explains:

In short, the rules that govern leader selection and removal are likely to affect leaders’ responsiveness to members’ demands and, consequently, politics and development outcomes. Leaders are more likely to be responsive when they are chosen from members in the immediate arena of authority, when they expect to have a long tenure, and when their welfare is closely tied to that of the community. In this way, expectations again mirror many of those found regarding the state: states with more democratic, stable regimes are expected to be responsive, fostering economic growth and human development.

5.4 Illustration: Social Institutions, Land Titling, and Property Rights

To illustrate the insights from bringing arenas of authority and social institutions into analyses of politics and development, I return to the issue of state land titling. Recall from Section 3 that less than one-third of the global population had state-backed property titles in 2017, despite strong efforts to institute state land titles. Considering how assumptions underlying land titling do, or do not, comport with the lived experiences of land users thus sheds light on the barriers to titling and encourages thinking about alternative ways to enhance land security.

Efforts to advance state land titling are based on several fundamental assumptions. As Reference Bruce and Migot-AdhollaJohn Bruce and Shem Migot-Adholla (1994, p. 3) note, the notion that land tenure security is ‘the perceived right by the possessor of a land parcel to manage and use the parcel, dispose of its produce and engage in transactions, including temporary or permanent transfers, without hindrance or interference from any person or corporate entity’ is critical to both land registration and titling. In accordance with the state-centric approaches discussed in Section 2, the state is portrayed as the dominant, or at least potentially dominant, arena of authority, and an impartial actor – the protector of land users’ rights. Moreover, land users are expected to enjoy easy entry and exit into geographical communities and thus to be able to access land or dispose of it at will; and the rules of engagement are individualistic, allowing users to make decisions independent of community goals and considerations.

Yet, the state is often not the most salient arena for individuals seeking access to land. For many in the world, the state is a distant actor, largely absent from their lives, and when present, often not trusted.Footnote ⁴⁰ Local traditional leaders and other community members, in contrast, are very much present and more often trusted. They are the key to solving disputes, receiving social assistance, accessing services, and – importantly – securing land (Reference LoganLogan, 2013). Moreover, as we show in Malawi and Zambia, local traditional authorities are often more likely than the state to monitor and sanction non-compliers (Baldwin et al., 2021). Not surprisingly, in a study aimed at understanding the value of localizing state administrative services, Reference Dulani, Lust and SwilaBoniface Dulani, Hannah Swilah, and I (2016, p. 12) found that Malawians preferred titles from their local chiefs over state titles because, they argued, their property is insecure without chiefs’ support. Their remarks echoed the cautions from Reference Bruce and Migot-AdhollaBruce and Migot-Adholla (1994, p. 8): ‘a formal title certificate or other official document is, at best, merely an affirmation of this social guarantee; it does not create it’.

It is not just the relative strength of the state vis-à-vis customary arenas of authority that limits the advance of state titling, but also the disconnect between assumptions underlying state titling projects and the nature of social institutions that govern individuals’ experiences around land. In much of the world, communities are governed by collectivist institutions, and land is an important component of community identity. Access to land is understood as part of one’s right as a member of the community, as determined by ancestry and belonging. Individuals from outside the community do not have the right to simply acquire it, or those from inside the community to dispose of land as they wish. Outsiders may obtain land, but the community can, and does, expel them from the area if land pressures rise (Reference Berge, Kambewa, Munthali and WiigBerge et al., 2014; Reference McGuireTakane, 2008). That is, the boundaries to the local arena of authority are often largely impermeable, and collectivist institutions limit the ability of individuals to transfer property as they wish. Moreover, as Adam Harris and Lauren Honig (2022) show, many Malawians and Zambians have higher trust in neighbours who rely on customary land tenure, and they anticipate greater cooperation from them as well. The individualist underpinnings of state land titling are often at odds with the social institutions in the dominant, local arenas of authority.

This is not to say that concerns about land security are misplaced. Systems governing land rights outside the state often create insecurity, particularly for some groups. As noted earlier, outsiders often face greater land insecurity, as the community can invoke rules over boundaries at any time. Lineage-based inheritance laws and residence patterns disadvantage men or women, depending on whether they are matrilineal or patrilineal, respectively (Reference Berge, Kambewa, Munthali and WiigBerge et al., 2014; Reference Kutsoati and MorckKutsoati and Morck, 2016). But such insecurity and inequalities also exist with regard to the state. The 2019 LGPI survey reveals that 15 per cent of Malawians and 12 per cent of Zambians in the border region were worried that they may lose access to their land. Importantly, they see a range of threats, from relatives and customary leaders to commercial farmers, as well as the state.Footnote ⁴¹ Understanding which arenas of authority are most salient, and the institutions governing them, is key to determining which groups are most at risk.

Solutions to land security problems thus need to take into account how individuals experience the relative strength of different arenas of authority, and the social institutions within them. One approach is to formalize customary land tenure arrangements. In Malawi, Karen Ferree, Lauren Honig, Melanie Phillips, and I (2022) find that individuals prefer land with written documents regardless of whether they are offered by the state or a chief, and they do so as much to signal to other citizens in their community that they hold the land as to benefit from state-backed security. This is not surprising, given the fear many have that other community members may try to grab their land. Another option, for those who seek to strengthen the state, is to design solutions that incorporate social institutions. For instance, Reference Berge, Kambewa, Munthali and WiigBerge and colleagues (2014) point to Norway’s odel system as an example of how customary lineage rights can be incorporated into state law. The odel, which has been around for thousands of years, gives first rights of purchase to those in the kinship circle who have controlled the land for a fixed period. Incorporating such a system into a state titling scheme does not overcome the weakness of the state, but it can make state law compatible with social institutions.

6.1 Electoral Rules

Electoral rules are conventionally viewed as political institutions that influence engagement and representation. They are the outcomes of negotiations in which political actors are centre stage. Taking this approach, for instance, Reference BoixBoix (1999) argues that the relative size of political parties determines whether elites establish PR or majoritarian rules. Scholars and practitioners increasingly recognize that other forces play a role. Some note that electoral rules may not work as expected because context (e.g., social divisions and norms, weak rule of law and media systems, and electoral fraud and intimidation) mediates outcomes (Reference Ferree, Powell and ScheinerFerree et al., 2013; Reference Ferree, Powell and Scheiner2014; Reference Krook and MoserKrook and Moser, 2013), while others consider how electoral institutions activate fault lines, determining which identities are salient (Reference ChandraChandra, 2007; Reference Michael Albertus and SlaterPosner, 2005). These approaches are insightful but not fully sufficient. They explore how elites’ negotiation of electoral rules shape elections but fall short of comprehending how the same considerations impact other arenas of authority. A more complete understanding is essential as academic perspectives influence praxis; political scientists not only study electoral rules, they also engineer them (Reference Carey, Hix, Htun, Mozaffar, Powell and ReynoldsCarey et al., 2013; Reference Htun and PowellHtun and Powell, 2013).

The story of the Libyan 2012 GNC electoral law illustrates how arenas of authority outside the state respond to electoral rules, and reciprocally shape electoral institutions and outcomes. The electoral law, formulated after the downfall of Muammar Qhaddafi, was the subject of extensive international intervention. Electoral engineers prepared plans based on their best understanding of Libya’s conditions and electoral institutions. Yet, they failed to anticipate how elites in arenas of authority outside the state, intent on maintaining their communities’ social institutions, would respond and ultimately upend the electoral engineers’ best-laid plans.

The preferences of international experts and the larger political parties were constructed within the paradigm of elections as a state institution. International experts generally advocated a PR system with gender quotas; they believed these would strengthen political parties, which were nearly absent in the wake of the previous regime,Footnote ⁴² and promote the status of women in a society where most held conservative views on gender roles.Footnote ⁴³ Many elites returning from abroad (with few domestic bases of support) and the Muslim Brotherhood (the largest party) favoured a PR system with gender quotas for the same reasons. Given the nascent state of parties in Libya, however, the proposal was a two-tiered, mixed electoral system, with PR seats evenly distributed across the country in larger constituencies on one tier and majoritarian seats on a second tier.

However, local elites viewed electoral institutions not only as a state institution but also as one associated with other arenas of authority. Elites in parts of the country where boundaries around ethnic, tribal, or locality-based arenas of authority were strong – and where there were deep divisions between neighbouring villages – could not accept the possibility that a politician from a neighbouring area would be their representative. Moreover, social institutions within these arenas prohibited women from working outside the home or traveling without a male companion, yet alone going to Tripoli to represent their constituency. Consequently, these elites rejected the proposed system not only (or even primarily) because of its impact on electoral behaviour and representation but also because it would potentially, and fundamentally, undermine their arenas of authority and the social institutions within them.

Negotiations over districting and seat distribution regarding the division of proportional representation (list, or PR) and independent candidate (IC) seats ultimately resulted in a complicated, hybrid system – and one much different than the electoral engineers intended. As shown in Figure 1, areas such as Gheryan, where there were strong local leaders and distinct tribal identities, accepted only IC seats, and PR seats were redistributed to areas such as Tripoli, with a larger number of residents from other regions of the country and returnees from abroad. Altogether, fifty of the seventy districts had parallel voting (e.g., voting for both PR and IC seats), while nineteen districts had only IC seats and four districts only PR seats. Put simply, the result was a hodgepodge system that reflected compromises between elites aiming to accommodate demands emerging from different ethnic and local arenas of authority and the social institutions within them. The compromise averted conflicts (at the time) but also complicated election logistics and created inequalities in representation.Footnote ⁴⁴

Figure 1 Distribution of seats in Libya 2012 general national council elections, by district

Source: Reference CenterThe Carter Center, 2012b, Appendix G.

In contrast, electoral engineers in Jordan devised rules that worked in concert with elites in both tribal arenas and the state, achieving both parties’ desired outcomes. At the end of the 1980s, Jordan reinstated multiparty parliamentary elections and allowed political parties to mobilize. The goal was to reduce opposition around economic crises and ease Jordan’s entry into a peace treaty with Israel. But the parliament elected in 1989 was a nuisance at best and destabilizing at worst. Parliamentarians frequently opposed government policy, with the Islamic Action Front (IAF) leading the opposition. Consequently, the monarchy changed the electoral law in 1993, aiming to weaken the opposition. The new electoral law moved from the 1989 electoral law’s multi-member district (MMD)/multi-vote system to an MMD/single, non-transferable vote (SNTV) one.

The result was what both the state and tribal elites wanted: a significant weakening of the Islamists and reinforcement of tribal arenas of authority. Individuals previously could divide their votes, casting one vote for their tribal leader and a second vote for the candidate who represented their policy preferences. After 1993, they were limited to one vote. When required to choose between casting votes in response to social institutions in tribal arenas of authority and those in religious ones, most East Bank Jordanians supported their tribal authorities (recall reasons for this described in Section 4). Thus, Islamists (not all of whom were IAF) saw their share of Jordan’s eighty legislative seats fall from thirty-four seats in 1989 to twenty-two seats in the 1993 (Reference Michael Albertus and SlaterSchwedler, 2015). Islamists’ presence in parliament was further reduced through their boycott of the 1997 and 2010 elections. In contrast, seventy-five independents gained seats in the 1997 elections, including sixty-eight tribal chiefs (Inter-Parliamentary Union, 1997), and more than ninety tribes were represented after 2010 (Reference HanandahHanandah, 2010). The electoral rule’s alignment with tribal social institutions produced these outcomes.Footnote ⁴⁵

6.2 Administrative Boundary-Drawing

Arenas of authority outside the state, and social institutions within them, affect the development of administrative boundaries as well. Administrative boundaries are often drawn along tribal or sectarian lines, as a response to demands from social groups. When this fails to be the case, they are often the subject of contestation.

Reference GranovetterGrossman and Lewis (2014) find this in Uganda. They argue that ethnic groups in outlying areas of their district seek territorial secession when they are of a different group than the majority of those in the district centre. This is because these groups perceive themselves to be disadvantaged by the district centre and thus demand a district to gain greater spoils. Importantly, however, the disadvantage comes not from the logic of the state system but rather because resources are demanded and believed to flow along ethnic lines. That is, the ethnic arenas, and social institutions regarding responsiveness to in-group members, drive resource distribution and, hence, the demand for new districts.

State bureaucrats can face strong resistance when they attempt to ignore these arenas. Tunisian bureaucrats attempting to draw municipal boundaries as part of the decentralization process in post-revolutionary Tunisia learned this first hand (Reference KherigiKherigi, 2021b). Mokhtar Hammami, the head of the Ministry of Interior’s Department for Local Authorities (DGCPL) and bureaucrat who led the boundary-making process, took pride in the Ministry’s formulaic, ‘neutral’ process – or, put differently, a process that privileged the state. He explained:

The result, however, was that local communities contested the new boundaries. Like the Libyans described earlier, Tunisians living in areas where arenas based on local ties were strong were unwilling to be either joined with ‘strangers’ in the municipality or split from relatives by municipal boundaries. Local elites fought to ensure that state boundaries were consistent with, and reinforced, social institutions within their arenas of authority, and often they succeeded. Reference KherigiKherigi (2021b) reports that the Ministry of Local Affairs faced hundreds of complaints and was forced in many cases to modify the boundaries.

6.3 Political Parties

Arenas of authority and social institutions also affect the development of political parties and party systems. As with electoral rules, scholars and practitioners primarily view political parties and party systems in terms of the state- as organizations and systems that structure competition over political office. These scholars recognize that social context influences the development of parties and party systems (Reference BoixBoix, 1997; Reference Lipset and RokkanLipset and Rokkan, 1967), but do not explicitly consider arenas of authority and the social institutions within them. These authorities not only influence the development of parties and party systems, but also give them meaning that extends beyond their roles in the state.

Where the state is weak and other arenas of authority strong, there is little incentive to form political parties. Kuwait offers a striking example of how non-state arenas of authority can substitute for political parties. There, parties are illegal, although political ‘tendencies’ exist. More importantly, politically active elites appear satisfied with this arrangement. They rely instead on tribes and, to a lesser extent, business networks to hold gatherings (diwaniya) and debate political issues, to select electoral candidates, and to mobilize voters. Thus, although Kuwait has one of the most politically active and contentious societies in the Gulf, Kuwaitis have not rallied for the right to form political parties. However, when the Emir unexpectedly began to enforce a decade-old law prohibiting tribal primaries in 2008, Kuwaitis were enraged. Banning political parties was one thing, but eliminating tribal primaries completely unacceptable (Reference LustTetreault, 2014).

Jordan, too, illustrates how non-state arenas can undermine the development of political parties. Recall from the previous discussion that Jordan has strong kinship arenas of authority, with social institutions that require individuals to cast their ballots for co-tribal candidates. Tribes, then, often act as political parties – engaging in primaries and get-out-the-vote efforts. Even when they form political parties, often it is because, as one Jordanian tribal elite once told me, that’s how modern politics is played. The result is what Reference Mainwaring and ScullyMainwaring and Scully (1995) describe as inchoate party systems, where parties are dominated by individual personalities, have low organizational capacity, and develop shallow roots in society.

Political parties thus established are an example of isomorphism, the development of organizations that are political parties in form but not substance. They do not play the textbook roles: to develop policy positions to represent their group interests, recruit and nominate candidates, or educate and mobilize the citizens. Instead, they act as organizations (at times subsidized by state coffers or international donors) that require and reward members of non-state arenas of authority for demonstrating their allegiance. They are instruments of social arenas as well as the state.

This affects how citizens engage with the parties, as well as how parties evolve. In Jordan, for instance, citizens spend little time during the campaign period seeking information about candidates or their positions. Detailed assessments of candidates’ qualities and platforms are not necessary; they simply need to look at which candidate represents their community. Consequently, fragmentation has persisted. Jordanian candidates have won seats with as little as 2 per cent of the popular vote, the margin of votes between last winner and first loser is often in double digits, and the percentage of wasted votes reach as high as two-thirds of the total vote count; in the 2010 parliamentary elections, 12 seats (10 per cent of the total) saw fewer than 100 votes separating winners and losers. Such results leave would-be candidates in upcoming elections sufficient reason to believe they have a chance of winning, even with relatively few resources (National Democratic Institute, 2010, p. 36). Large numbers of candidates and low margins of victory thus both result from and promote high turnover in parliamentary seats. Not surprisingly, in 2010, Jordan saw nearly two-thirds turnover in parliamentary seats, and in 2007, three-fourths of the representatives entered for the first time (Reference Lust, Hourani and El-MoumaniLust et al., 2011, p. 127).

6.4 Implication: The Role and Development of (Non-)State Institutions

Electoral rules, administrative boundaries, and political parties are often viewed as state-based institutions; yet, as the aforementioned examples illustrate, they often are both shaped by and act as institutions in arenas outside the state. Electoral rules, for instance, can affect the extent to which individuals have incentives to demonstrate their allegiance to their tribe, locality, or other social group. Gender quotas can alter the rewards associated with following obligations for women to stay removed from the public sphere. So, too, actors and organizations associated with non-state arenas can act as substitutes for institutions associated with the state. The ability of tribal, religious, or other organizations to coordinate candidates and mobilize support undermines party development, but it also means that one finds the relevant institutions that shape electoral mobilization in tribes or other arenas, not in the state. Institutions associated with the state are often institutions of non-state arenas as well, and institutions in non-state arenas often shape engagement in the state. That is, institutions are not always either state or non-state institutions but may act in multiple arenas of authority at the same time.

This has important implications for development practitioners and policymakers. Consider democracy promotion, for instance. Political party subsidies may have little impact on competition if the institutions that mobilize support and shape representation are found in other institutions. So, too, trainings that focus on campaign messaging will have very little impact if voters make their decisions based on whether a candidate is from their community, rather than the policies they espouse. Finally, as the example from Libya shows all too well, electoral rules that are designed without taking into account how these rules are both state and non-state institutions can have unintended consequences. Actors intent on state-building need to take into account how institutions can be simultaneously institutions of state and non-state arenas.

7.1 Defining Arenas and Social Institutions

Taking the competing arenas of authority and social institutions that affect individuals’ decisions into account requires that they be well defined. These are not merely the residuals, defined by what is left over after one has taken the state and its institutions into account. Nor should they be understood only as disruption to the state – corruption, clientelism, or capture. Finally, they are not simply identity groups. Arenas may be based on ethnicity, religion, or other identities; however, it is not the nature of the identity but rather the characteristics of arenas and institutions that shape action. The goal is not just to recognize that ethnicity, religion, or other identities affect the choices made by voters and politicians, citizens, and service providers. It is to understand how differences in the nature of these arenas and institutions outside the state shape individuals’ engagement – regardless of the identity on which they are based.

Arenas of authority are spheres of engagement. They have a community which acts within its boundaries, distinguishing members from non-members outside. Unlike states, social arenas of authority need not be territorially bounded. However, individuals within arenas are bound together by a common goal to perpetuate the community. Individual members may have competing interests, but they are engaged in a common project. This requires governance over a range of actions, from how individuals enter or exit the community to what roles they may hold, and how they act with members and outsiders.

Social institutions delineate the roles, rules, and rewards that govern communities within the arena of authority. These determine who has power and how they maintain it, the actions that are permitted or prohibited, how adherence accords benefits and transgressions incur sanctions. Social institutions are not simply beliefs or practices (habits); in contrast to beliefs, they are rules about what one should and should not do, and unlike habits, they have an intentionality that is driven by anticipated costs and benefits. Nor are they social networks, social capital, social actors, or any of the other myriad ‘socials’ that are used, somewhat interchangeably, in the literature.

Social institutions vary significantly both across arenas and across time. It may be easier or harder for individuals to enter or exit the arena (e.g., creating boundaries that are more porous or impermeable), or for individuals to change their roles or be selected as leaders (e.g., social mobility). Rules governing how individuals engage with others in the community – whether women and men can debate issues together, whether youth can challenge their elders, whether in-group members can police inter- and intra-group conflicts – differ in important ways, even across arenas formed around the same identities.

7.2 Charting the Path Forward

A framework that takes arenas of authority and social institutions into account prompts analysts and practitioners to ask a number of questions: which arenas of authority matter for the issue at hand? How do the social institutions within the different arenas shape behaviour and development outcomes? (See Figure 2.) Moving forward requires fully interrogating expectations derived from the extant literature, answering new questions that emerge from the framework, and addressing methodological challenges.

Figure 2 Process of inquiry

7.3 Towards Effective Development

The approach outlined here benefits scholars and practitioners alike. For scholars, it provides new insights into studies of both the state and social forces. Studies that focus solely on the impact of political institutions on outcomes are likely to attribute far more import to the political institutions than is warranted. So too, efforts to understand the role of identity or to link identities and social conditions – such as economic inequalities or rural–urban divides – are incomplete unless they consider the relationship between such identities and overlapping arenas of authority.

This perspective also provides important insights for practitioners. It helps to explain why isomorphic mimicry, or the adoption of state organizational forms or policies that appear to be ‘modern’ or ‘best practices’, can result in zero improvement in state capability, or even decline (Reference Smith and WaldnerPritchett et al., 2012). Quite simply, individuals implementing policies or engaging within these organizations may do so on the basis of arenas and social institutions outside the state, or indeed, attach very different meanings to engagement than the international practitioners and their local allies intended. The everyday approach to politics and development I present here also helps explain why advocates may face resistance in implementation. Where programmes and policies threaten the interests of elites in powerful arenas of authority outside the state, as they did in boundary making in Tunisia’s decentralization process, for instance, they may not only face resistance but also spark unrest and even violence. Finally, the framework provides a basis for understanding why scholars find that informational interventions often to do not foster social accountability (Reference Dunning and GrossmanDunning et al., 2019). When individuals are subject to powerful incentives from arenas of authority outside the state, they may fail to hold officials and service providers accountable, even when they have full information.

The framework not only explains such potential outcomes but also provides a means for accumulating knowledge and improving programme design. Practitioners have increasingly recognized the importance of context, as it has become clear that programmes which worked splendidly in one area or country do poorly elsewhere. Indeed, even programmes that work well in one place at one time fail to do so in subsequent iterations. However, the response to date does not provide a strong foundation for learning from past experiences. For example, Reference Andrews, Pritchett and WoolcockMatt Andrews, Lant Pritchett, and Michael Woolcock (2017) propose a problem-driven iterative adaptation (PDIA) method, in which practitioners identify multiple ideas, trying them out in an experimental process that allows for the emergence of hybrid solutions. The approach put forth here goes beyond the goal of PDIA’s incremental and experimental approach. I anticipate that it is possible to develop a firmer understanding of politics and development such that lessons can be drawn and transferred to new places and times.

Advancing research and development programming requires revising insufficient and often state-centric theories, recognizing the ranges of authorities at play, and the nature of institutions associated with them. It starts by interrogating assumptions about how individuals understand the everyday choices before them and considering how the roles, rules, and rewards in multiple, competing arenas of authority influence their behaviour. If scholars and practitioners can identify systematic differences in the nature of arenas and social institutions that shape individuals’ choices and, ultimately, development outcomes, then they can begin to ask the right questions, design better research, and develop appropriate programmes from the outset. The framework I have presented here, and the process of inquiry that follows from it, provide a basis for such efforts.