NIE approaches to governance have shaped the policy agenda of donors and IFIs since the 1990s. NIE’s influence was evident in the World Bank’s Good Governance approach, which advocated for ‘getting the institutions right’ (including promoting market-led reforms, democracy promotion and transparency and accountability measures) (Rodrik Reference Rodrik2006). PSA, originally popularised by the work of Khan (Reference Khan, Harriss, Hunter and Lewis1995, Reference Khan2010), was a direct critique of NIE approaches to governance. Khan (Reference Khan2007) argued that East Asian countries achieved rapid growth and structural transformation not through enacting market-led reforms but through doing the exact opposite. East Asian governments actively intervened in their economies to support latecomer firms to invest in technological capabilities, contributing to structural transformation.

PSA pushes us beyond a narrow focus on formal institutions. Instead, PSA examines how distributions of power among organised groups shape how institutions operate. Clearly, a great deal of research within political science and political economy has focused on the importance of analysing power. The novelty of PSA has been ‘to precisely delineate difference between political settlements … that result from an interaction between global capitalism and unique distributions of power at local, national and regional levels that emerged from particular pre-colonial, colonial and post-colonial histories’ (Gray Reference Gray, Cheeseman, Bernard and Husaini2019, p. 1).

A remarkable feature of PSA’s rapidly growing influence has been its widespread proliferation in academic and policy circles in different disciplines, countries and donor agencies (Behuria et al. Reference Behuria, Buur and Gray2017). However, as scholars from varied disciplines combined the basic units of analysis of power and institutions with different theories and concepts, PSA was operationalised very differently (Gray Reference Gray, Cheeseman, Bernard and Husaini2019). PSA’s widespread usage was also a product of the influence the concept has had in donor programming, particularly in the UK government (but also in other European countries). Within donor agencies, ‘political settlements’ became a shorthand for analysing power relations between different stakeholders. Usually, the political economy analysis that followed simply mapped relationships between key elites and had very little to do with the economy.

Khan’s (Khan Reference Khan, Harriss, Hunter and Lewis1995; Khan & Blankenburg Reference Khan, Blankenburg, Dosi, Cimoli and Stiglitz2009) initial formulation of PSA explained the divergent economic trajectories of countries in Asia, Africa and Latin America. Though some of his work was comparative across cases, the Marxian and historical materialist roots of Khan’s PSA work were most evident in his analysis of specific cases, particularly Khan’s work on his own country, Bangladesh (Khan Reference Khan2004). The Crisis States Research Programme Consortium (CSRP), based at the London School of Economics and Political Science, extended Khan’s PSA, which they referred to as ‘the balance or distribution of power between contending groups and social classes, on which any state is based’ (Di John & Putzel Reference Di John and Putzel2009, p. 4). The CSRP and another UK aid-funded research centre, the Political Settlements Research Programme, based at Edinburgh, focused on analysing conflict and state-building. Between 2011 and 2019, the Department for International Development (DFID) funded the Effective States and Inclusive Development (ESID) Research Centre, which both used Khan’s (Reference Khan2010) approach and extended the influence of PSA analysis in academic and policy circles, including developing competing variants of PSA frameworks (Kelsall et al. Reference Kelsall, Schulz, Ferguson, Vom Hau, Hickey and Levy2022; Levy Reference Levy2014). In the last phase of ESID research, Kelsall et al. (Reference Kelsall, Schulz, Ferguson, Vom Hau, Hickey and Levy2022) embarked on a mapping and coding exercise of different political settlement types.

There is growing confusion and inconsistency in the definition of political settlements, its application, its analyses of power and the methodologies used to conduct PSA (Behuria Reference Behuria2025). The continued confusion about what constitutes PSA has led to disagreements among leading scholars that remain unresolved (Kelsall Reference Kelsall2018; Khan Reference Khan2018a, Reference Khan2018b). Kelsall et al. (Reference Kelsall, Schulz, Ferguson, Vom Hau, Hickey and Levy2022, p. 4/5) describe at least three different origin stories for PSA and highlight three different kinds of PSA: ‘a process of settling a conflict’, ‘a new and transformed political order’ and ‘the way in which different societies solved the political problems generated by economic growth’. Kelsall et al. (Reference Kelsall, Schulz, Ferguson, Vom Hau, Hickey and Levy2022, p. 21) define political settlements as ‘an ongoing agreement among a society’s most powerful groups over a set of political and economic institutions expected to generate for them a minimally acceptable level of benefits, which thereby ends or prevents generalised civil war and/or political and economic disorder’. However, both Kelsall et al.’s initial characterisation of three definitions of PSAs and their definition of political settlements misdiagnose the central line of distinction that divides the PSA literature: rival understandings of capitalist transformation processes.

Gray (Reference Gray, Cheeseman, Bernard and Husaini2019) makes a useful distinction between the two distinct schools of ‘as process’ and ‘as action’ PSA research. ‘As process’ PSA literature (or those elaborating the structuralist, Marxian roots of PSA) conceptualises PSA as a political order that has not been consciously planned by different social groups. PSA of this kind has generally focused on the contestation over the distribution of the rents, which has shaped the unstable nature of capitalist accumulation in late-developing countries. ‘As action’ PSA literature, on the other hand, generally focuses on the role of agreements made by powerful elites. This literature employs PSA much more inconsistently and generally does not prioritise analysis of capitalist accumulation, focusing instead on social policy, conflict and state-building. This chapter (and the book) contributes to elaborating the structuralist, Marxian and historical materialist roots of PSA. In doing so, it builds on the work of Mushtaq Khan and others (Gray Reference Gray2018; Whitfield Reference Whitfield2018; Whitfield et al. Reference Whitfield, Therkildsen, Buur and Kjaer2015). It highlights how PSA can be used to help understand the contemporary transnational nature of vulnerabilities shaping late-development challenges. In the remainder of this chapter, the PSA approach is elaborated, highlighting the differences and misconceptions that exist about PSA and how it can be applied to help understand power dynamics characterising contemporary late development.

The Evolution of Political Settlements Analysis

Though Khan has often defined political settlements in slightly different ways, in its most basic form, political settlements are ‘a description of the distribution of power across organizations that are relevant for analysing a specific institutional or policy problem’ (Khan Reference Khan2018a, p. 5). This is in line with Migdal’s (Reference Migdal1988) work, which showed that strong states were relatively rare because they can persist only where there is a concentration of ‘social control’ and that can occur only where older forms of social control are undermined. Migdal’s work, like Khan’s PSA, highlights that where power is not centralised, ruling elites are forced to accommodate competing interests and thus are absorbed in a ‘politics of survival’ (Whitfield & Buur Reference Whitfield and Buur2014).

Khan’s (Reference Khan, Harriss, Hunter and Lewis1995) initial approach to political settlements had its roots in historical materialism, with the foundational understanding that ‘power is rooted in history’ (Behuria et al. Reference Behuria, Buur and Gray2017, p. 512). Under this historical materialist view of PS, to survive, governments mediate inter-class conflicts while also promoting the expansion of capitalist accumulation. This view is shared by older literature that has highlighted how the distribution of land was central to maintaining political authority in rural societies (Huntington Reference Huntington1968; Moore Jr. Reference Moore1967). The conflictual aspects of capitalist accumulation have accompanied both agrarian and structural transformation, often requiring ‘a decisive split between the ruling and landed elite’ (Lavers Reference Lavers2023, p. 24). Agrarian landholders have often proved to be obstacles to exchange rate and trade policies required for infant industry protection (Haggard Reference Haggard1990; Kay Reference Kay2002). Countries like Korea and Taiwan were particularly successful in breaking the power of landowners and introducing land reform (Amsden Reference Amsden1990; Kay Reference Kay2002). In India and Latin America, the opposite happened, with landowners resisting most moves to constrain their power (Kay Reference Kay2002).

Capitalist accumulation is always contested because it results in a distribution of resources and benefits to certain groups over others. Thus, power relations are always evolving rather than being firmly established under an agreement or formal settlement among groups in society. ‘Structuralist’ versions of PSA argue that rents are necessary and can be productive during processes of capitalist accumulation. They argue that violence and political contestation are intrinsic to processes of capitalist accumulation. Ideology and ideas are directly related to one another and shape capitalist accumulation processes.

‘Structuralist’ PSA also places emphasis on the colonial roots of how transnational coalitions shape domestic capitalist accumulation processes while centring analysis of dependencies within the global political economy. Some scholars have mistakenly highlighted Khan’s formulation as methodologically nationalist (Hickey Reference Hickey2013). This misunderstanding has been repeated in several publications using PSA (Hickey et al. Reference Hickey, Bukenya and Sen2014; Kelsall Reference Kelsall2018; O’Rourke Reference O’Rourke2017). Some of this is because of the influence of one diagram in Khan’s influential (Reference Khan2010) working paper. In this diagram, Khan (Reference Khan2010, p. 65) distinguished between four types of clientelist political settlements in a 2 × 2 matrix. Khan’s typology is based on analysis of the organisation of the ruling coalition, the time horizons they may develop and their implementation capabilities. Most PSA is based on analysing only two dimensions of the power Khan originally highlighted: the horizontal and vertical distribution of power. The horizontal distribution of power, which refers to the relative strength of the ruling coalition compared to competing elite factions, determines the time horizon of the ruling coalition. The vertical distribution of power, which represents the ruling coalition’s power relative to lower-level factions within their own coalition, shapes the ruling coalition’s enforcement capabilities. Potential developmental coalitions are characterised by their superior horizontal and vertical power compared to rival groups, resulting in long-horizon development orientations and high enforcement capabilities. Vulnerable authoritarian coalitions, while lacking in horizontal power, possess significant vertical power, resulting in a short-term orientation coupled with high implementation capabilities. Weak dominant coalitions, on the other hand, have greater horizontal power but are limited in their vertical power, leading to longer time horizons but reduced implementation capabilities. Finally, competitive clientelist coalitions have limited vertical and horizontal power, resulting in ruling coalitions that have short-term horizons and meagre implementation capacities.

This typology has been used widely, with the assumption that it can help predict certain outcomes depending on how a political settlement is classified. Yet, building on another matrix, which is in the same paper (Khan Reference Khan2010, p. 71), other work (Whitfield et al. Reference Whitfield, Therkildsen, Buur and Kjaer2015) has shown that countries with similar structures of ruling coalitions have very different outcomes in the implementation of industrial policy within the same institutional context. Even within the same country, outcomes in different sectors may be very different (Kjaer Reference Kjaer2015). In this other matrix, Khan (Reference Khan2010, p. 71) analyses patron–client structures and the organisational power of emerging capitalists. He categorises four kinds of patron–client structures, highlighting how likely they are to lead to relationships that contribute to investments in technological capability acquisition among domestic firms. In this way, he introduces a multi-scalar focus to PSA by analysing domestic firms in relation to their technological capabilities with lead firms globally. First, in the most constrained situation, firms have moderate to low capabilities and are politically weak. While potential developmental coalitions could drive early accumulation and could discipline such firms, ruling coalitions with shorter time horizons would have little power over such firms. Second, when firms have high capabilities but are politically weak, outcomes would vary depending on whether the ruling coalition can take a longer-term view or not. Khan (Reference Khan2010) highlights the case of South Korea, where ruling elites were able to take a longer-term view and develop effective state–business relationships. Khan (Reference Khan2010) contrasts the effective South Korean case with Thailand under Thaksin Shinawatra, where the dominant party retained a short-term view, leading to adverse outcomes.

Third, when firms have moderate to low capability but are powerfully networked, their political power can drive some limited accumulation, but governments will find it difficult to discipline firms in line with learning and the acquisition of technological capabilities. Fourth, when firms have high capabilities and are powerfully networked, powerful firms can drive substantial accumulation but they will limit entry of new capitalists. Governments would also find it difficult to discipline these capitalists. Some examples of such patron–client relations include the Indian government’s relationship with the country’s largest conglomerates. These different categories provide us with initial hypotheses regarding the variety of political constraints that governments face depending on the power and existing technological capabilities of powerful firms.

The bulk of PSA research, however, has neglected this 2 × 2 matrix. As such, the structuralist and multi-scalar focus of PSA is often downplayed. One of the book’s main intended contributions to the PSA literature on PSA is to elaborate how a focus on the historical evolution of transnational social relations is shaped by evolving economic structures and external relationships as national economic development trajectories take shape within the global economy. Since there are few firms that have significant technological capabilities across most sectors in Rwanda, this 2 × 2 matrix is not directly employed. Instead, the book employs PSA to analyse the significance of multi-scalar processes, particularly transnational dynamics within which internal and external factors combine under contemporary globalisation.

What Is Structuralist about PSA?

Structuralist PSA analysis focuses attention on the economic structures that newly independent countries in Asia, Africa and Latin America inherited after colonial rule. PSA begins with the assumption that the predominance of patron–client relations in former colonies has partly emerged from colonial relations, with European colonial administrations leaving their former colonies with undiversified economic structures, as well as dependence on primary commodity exports. Managing the export of cash crops and minerals was central to colonial economic policies and investment (Frankema et al. Reference Frankema, Williamson and Woltjer2018; Hill Reference Hill1963; Hopkins Reference Hopkins1973; Rodney Reference Rodney1972). Cash crop agriculture was central to organising political, social and economic change during the colonial era.

In the 1950s, in thirty-eight African countries, 65 per cent of total exports were accounted for by agricultural commodities, three-quarters of which were coffee, cocoa, cotton, groundnuts and palm products (Hance et al. Reference Hance, Kotschar and Peterec1961). At independence, commodity-dependent economies produced a low surplus, constrained by the limited diversification and productive capacities in their economies. These structural characteristics explain why clientelism is prevalent in low-income countries with undiversified economies. This contrasts with neopatrimonialism-based explanations of (African) underdevelopment, which argued that ‘the drivers of neopatrimonialism are located in social relations and in a set of assumptions about the pre-modern basis of authority on which African states are constructed’ (Behuria et al. Reference Behuria, Buur and Gray2017, p. 521). Some political settlements literature (Booth & Golooba-Mutebi Reference Booth and Golooba-Mutebi2012; Kelsall Reference Kelsall2013) accepted culturalist neopatrimonialism-based explanations of African underdevelopment, which argue that there is a modal pattern of rent-seeking across the continent. Neopatrimonialism-based arguments have been widely criticised (Gray & Whitfield Reference Gray and Whitfield2014; Mkandawire Reference Mkandawire2015). Among its many faults, neopatrimonialism-based explanations of African underdevelopment have contributed to the ‘excessive levelling of African political and economic landscapes’ (Mkandawire Reference Mkandawire2001, p. 289). Gray & Whitfield (Reference Whitfield and Buur2014, p. 7) highlight that such views are problematic given that they highlight uniformity across late-development experiences, while Khan’s original PSA highlights ‘diversity among African countries in terms of the distribution of power in society, with important implications for rent creation and management’. PSA highlights that the drivers of clientelism are not cultural but instead a result of undiversified economies and small formal sectors.

Structuralist PSA is inherently multi-scalar. It highlights how ruling coalitions manage political order while investing in varied development strategies. Ruling coalitions in former colonies heavily depended on capitalists and landowners in the agriculture and minerals sectors. Ruling coalitions rely on relationships with buyers of commodities, as well as governments that host the multinational companies that buy a country’s minerals or exports. Export diversification results from industrial policies, with rents distributed to support investments in new export discoveries. This also reduces dependence on domestic firms and farmers that export commodities, as well as the buyers of those products. Even the process of export diversification faces ‘transition costs’, that is, the ability of powerful groups to block policies ranging from demonstrations, strikes to outright violence (Khan Reference Khan2010). In some Asian and Latin American countries, agricultural elites blocked attempts at diversification, fearing this would reduce their political power (Khan & Blankenburg Reference Khan, Blankenburg, Dosi, Cimoli and Stiglitz2009). In other countries, landed elites were incentivised to invest in industrialisation, or ‘transition costs’ were overcome when such elites were marginalised (Slater Reference Slater2010).

Political settlements analysis requires documenting power struggles over the flow of resources between groups. Power struggles have become increasingly multi-scalar, as African economies have been increasingly integrated into the global economy and global financial sectors. Many African government officials are just as invested in neoliberal ideologies or market-led reforms as donors. Since the 1970s, the socialisation of African bureaucracies to act in line with neoliberal best practice has been incentivised as African economies have been more integrated into the global economy. Donors have supported training initiatives for African economists who are trained in neoclassical economics through initiatives like the African Economic Research Consortium rather than the more traditionally heterodox Council for the Development of Social Science Research in Africa (Stein Reference Stein2021). Crucially, even as aid dependence may have reduced in many countries, African governments remain incentivised to adopt best practices in market-led reforms to maintain high ratings at credit agencies so they can secure loans at better rates.

Some would argue that there are cracks appearing in this neoliberal consensus because of the influence of alternative donors such as China. To a degree, this is correct. Ruling coalitions increase their autonomy by diversifying their external relationships, including through working with new donors, new domestic capitalists or foreign firms. This creates opportunities for inventive governments that use their agency to achieve more policy autonomy. However, PSA enables analysis of how these new bilateral relationships also come with structural dependencies on new actors (whether multinational firms or Chinese, Russian or Middle Eastern governments). In this way, foreign investors and donors are always integral to the political settlement because investments are central to how political settlements are financed.

Under contemporary globalisation, the strategies of ruling coalitions also inevitably require accessing financial support from outside (which may come with conditions) or ensuring that external support is not withdrawn (as has been the case when aid has been cut for various reasons across the continent). Within political settlements research, the primary instruments highlighted as crucial to managing political order during capitalist transformation are rents, ideas/ideologies and violence/coercion. The remainder of the chapter introduces the concept of holding power and then discusses three of the four determinants of holding power: rents, ideas and ideology, and violence and conflict. The salience of economic structure, the fourth determinant of holding power, has already been illustrated in this section.

Holding Power

In the PSA literature, the distribution of power across organisations in society describes the likelihood that specific organisations ‘hold out’ in the conflicts inevitably characterising capitalist accumulation (Khan Reference Khan2018a).Footnote ¹ Holding power refers to ‘the capability of an individual or group to engage and survive in conflicts’ and to impose costs on others while absorbing costs inflicted on them (Khan Reference Khan2010, p. 6). Policies are more easily enforced if ruling coalitions have more relative holding power than groups that are opposed to such policies. Conversely, it may be difficult to enforce policies if groups opposed to those policies have more relative holding power than ruling coalitions.

Holding power has been analysed in different ways within the PSA literature. First, holding power has been analysed by tracing how rents are allocated between groups and to individuals outside ruling coalitions (North et al. Reference North, Wallis and Weingast2009). Second, scholars have identified how elite access to influential positions are distributed across the government’s political and military apparatus, usually distinguishing members of coalitions based on ethnic identity, locational background or religion (Abdulai & Hickey Reference Abdulai and Hickey2016; Lindemann Reference Lindemann2011). Third, holding power is examined through studying the historical distribution of power in societies and relating it to sectoral outcomes. One example of this approach is Croese’s (Reference Croese2017) work on Angola’s housing sector.

While there is much to learn from these approaches, holding power ‘depends not just on the resources the organization can deploy but also on its power to mobilize support’, or, more precisely, it ‘is partly based on income and wealth but also on historically rooted capacities of different groups to organize’ (Khan Reference Khan2010, p. 1). Holding power depends on the organisational capabilities and the capacity of leaders to mobilise support for their policies. As Khan (Reference Khan2018a, p. 640) notes, PSA ‘requires a deep understanding of the history, sociology and ideological and identity cleavages in a country, and how they have overlapped with and been used to mobilise around resource issues’. Assessments of holding power are as ‘important as it is difficult, both for external analysts and the protagonists themselves’ (Khan Reference Khan2018a, p. 643). This book attempts to evaluate the determinants of holding power in Rwanda. In doing so, it highlights how ‘structuralist’ PSA involves contestation among transnational coalitions comprised of varied actors. The next section distinguishes between how different PSA approaches have understood key determinants of holding power: rents, ideas and ideology, and violence and conflict.

Rents and PSA

There is significant debate about how to define rents and their role in late development within PSA. Structuralists would argue that for structural transformation to be achieved, industrial policy is necessary. Rents – defined as ‘excess incomes, which do not exist in perfect markets’ (Khan Reference Khan and Sundaram2000, p. 21) – are common instruments (tariffs, subsidies) in industrial policy. Thus, rents – though colloquially often associated with corruption or bribery – can often be productive. ‘Rents have a dual function – as well as acting as an incentive system, they also serve as a source of accumulation for investment’ (Gray Reference Gray2018, p. 66). This dual function, which is also aligned more broadly with evolutionary economics, is different from the work of public choice theory scholars such as Bates (Reference Bates1981). Bates’ work, which continues to significantly influence North American political science research on Africa, argues that the state is an arena where politicians and bureaucrats are motivated by rational self-interest.Footnote ² Similarly, most neoclassical economics scholarship has ignored the possibly productive and always conflictual aspects of rent distribution. Influential neoclassical scholarship (Krueger Reference Krueger1974; Posner Reference Posner1975) has argued that when states adopt industrial policies, firm managers are encouraged to establish and maintain policy-created rents rather than invest their time in managing production. Scholarship (Little et al. Reference Little, Scitovsky and Scott1970) has often argued that industrial policy has been associated with a wasteful allocation of government resources and corruption. This has led most neoclassical economists to view the distribution of rents as unproductive and undesirable. This remains the case with work on New Structuralist Economics that has been central to the revival of industrial policy discussions among IFIs and bilateral donors (Lin Reference Lin2011). Within such approaches, state–business relations are always perceived to be characterised by rent-seeking, which is always viewed as unproductive.

PSA acknowledges that there are different forms of rents. Some rents are undesirable for both political stability and structural transformation (like petty corruption). Some rents are necessary for political stability to maintain elite cohesion and disincentivise rebellion (like providing protected benefits to military elites). Other rents can be both necessary and productive (like tariffs) (Khan Reference Khan1996a, Reference Khan1996b). Thus, PSA argues that the distribution of rents is not undesirable or antithetical to achieving development. This also helps explain why corruption has coexisted alongside structural transformation in East Asia (Ang Reference Ang2020). Instead, PSA approaches argue that the provision and disciplining of rents in line with technological capability acquisition is central to achieving structural transformation (Whitfield et al. Reference Whitfield, Therkildsen, Buur and Kjaer2015).

Large segments of PSA research, especially the literature that does not focus on the economy, have broadened the conception of rents to include any kind of income flows (whether legal or not). In some PSA research, there is a tendency to simply assume that pay-offs or incomes would either negate conflict or encourage support for certain policies. In this way, non-structuralist PSA research regularly aligns itself with the public choice theory of Bates and moves away from the evolutionary focus on the necessity of the distribution of rents for structural transformation. This is a crucial misunderstanding within the PSA literature and also observers of the PSA literature who disregard the significance of varied approaches to analysing rents. There are also other misunderstandings relating to rents, particularly how the concept relates to ideas and ideologies, as well as violence and conflict, which will be discussed in the sections that follow.

The Role of Ideas and Ideologies

Social sciences – from political science to economics – have been increasingly concerned with examining the role of ideas in shaping policy choice, particularly in reference to the relationship of ideas and interests (Rodrik Reference Rodrik2014). Very broadly speaking, there are two fundamentally different ways of conceptualising ideology (Thompson Reference Thompson1984). The first is a neutral conception of ideology – described as ‘a system of beliefs’ or ‘symbolic practices’, which refer to social actions or political projects. This definition has ‘remained constant in political science over time’ (Knight Reference Knight2006, p. 625). The second conception links ideology to the process of sustaining asymmetrical relations of power or maintaining systems of domination. Within this conception, ideology is not restrictively understood as a set of beliefs. Instead, as Marx views it, ideology works as an inversion of reality (Cramer & Richards Reference Cramer and Richards2011). In this vein, Geertz (Reference Geertz1973, p. 220) defines ideologies as ‘maps of problematic social reality and matrices for the creation of collective conscience’.

Lavers (Reference Lavers2018) provides a useful emphasis on ‘paradigmatic ideas’ as the fundamental basis of ideologies, binding elite cohesion and mobilising popular support for political parties as they target a specific goal. To illustrate the significance of a focus on ‘paradigmatic ideas’, Lavers (Reference Lavers2018, p. 12) argues that in both Ethiopia and Rwanda, a paradigmatic idea for the ruling parties in these countries was the ‘need for rapid socioeconomic development as a means of enduring peace and stability’. In other countries such as Bolivia, Ghana, Peru and Zambia, resource nationalism was instrumentalised as a paradigmatic idea.

Within PSA debates, analysis of the relationship between ideas and interests has been vibrant. A common criticism of Khan’s (Reference Khan2010) elaboration of PSA was that it did not discuss the role of ideas sufficiently (Hickey Reference Hickey2013). By the mid-2010s, the PSA literature had conceptualised the role of ideas in terms of understanding legitimisation of development strategies and mobilising support (Behuria et al. Reference Behuria, Buur and Gray2017). Different groups of scholars, who were using PSA, analysed the role of ideas in varied ways. Most often, ideas and ideologies were simply being brought in to include the influence of external actors. Sometimes, PSA analysis also focused on how political parties and rivals employed ideas and ideologies to mobilise support for their policies. Most authors were analysing ideas and material incentives as separate from one another in explaining the motivations of different actors (Abdulai & Hickey Reference Abdulai and Hickey2016; Hickey & Izama Reference Hickey and Izama2017; Yanguas Reference Yanguas2017).

Scholars elaborating the structuralist roots of PSA rely on historical materialist understandings of ideology. Gray (Reference Gray2018), for example, argues that political commitments to political ideologies can structure institutions and the distribution of power by influencing perceptions of the legitimacy of claims of different groups. Non-structuralist PSA approaches generally assumed ideas to be neutral in terms of their relationship with capitalist transformation, rarely centring analysis of Marxian or historical materialist understandings of ideas. Lavers’ (Reference Lavers2018) work on this takes a less clear stance, highlighting both the existence of historical materialist understandings of ideas and ideologies but also more neutral views of the roles of ideas/ideologies in mobilising support and incentivising the behaviour of actors.

The split within PSA scholarship is again evident in how different groups view the relationship between ideas/ideology and material incentives/capitalist accumulation. Observers may argue that whether ideas or material benefits take priority depends on the specific case. However, this can often come down to prior assumptions, particularly in relation to how the ideational and the material relate to one another in shaping capitalist accumulation processes. Khan’s PSA retains capitalist accumulation as a focal point of analysis. Thus, ideas and ideologies are perceived to be explicitly linked to mobilise support, and rivals can be punished for working against ideological goals. Drawing on ideas and ideology provides the moral legitimacy to act and to mobilise support (which would not be possible through the distribution of material benefits alone).

Centring analysis of capitalist accumulation requires PSA to highlight the inescapable relationship between ideologies and interests. Within Marx’s writings, however, there is no uniform theory of ideology (Larrain Reference Larrain1979). Althusser (Reference Althusser1971) built on Marx’s work to argue that ideology represented the imaginary relationship of individuals with their real conditions of existence. Ideology does not represent the real relations that govern the existence of individuals. For Althusser, ideology is a ‘lived experience’ where subjects are constituted in and through ideology (Purvis & Hunt Reference Purvis and Hunt1993). He uses the concept of ‘interpellation’ to show how ideology constitutes people as subjects and situates subjects within specific discursive contexts. Human beings relate to their material conditions through ideology. Ideology legitimises the exploitation and unfairness associated with ‘real’ conditions of capitalist accumulation. Exploited groups accept such conditions as opportunities for their advancement, whereas capitalists, as well as the ruling coalition they finance, mask their domination by justifying exploitation through the attainment of collective achievements. To legitimise inequality, exploitation should not be perceived as negative but as ‘a directly relational factor that is at the same time a form of inclusion in a given social system’ (Cramer Reference Cramer2006, p. 112). Thus, ideology is how human beings relate to their world. It does not simply distort real relations but connects individuals with material conditions.

Althusser (Reference Althusser1971) argues that ideology organises the behaviour of individuals in society. Maps of social realities are challenged through the contestation of memories and the roles of individuals who occupy positions within those memories. Workers or those who support dominant classes (in political causes, rebellion or war) are not simply obedient followers. Those who resist dominant ideologies may create their own ideologies. Hirschman (Reference Hirschman1971) argues that such social formations require ideologies even more than dominant classes. However, even the ideologies of non-ruling organised groups have a relationship to material interests. Non-ruling organised groups either have been excluded from controlling systems of exploitation or are exploited within them or were never included in the first place. Gaining support among those who resist dominant ideologies is essential for rival ideologies to be relevant.

The dominant class cannot solve problems associated with the inequalities of capitalist accumulation without ideology. Within some of Marx’s writings, ideology hides the true relations between classes and explains the relations of domination and subordination as harmonious so labour acquiesces to the needs of the dominant class. Thus, ideology legitimises the class structure and becomes indispensable for reproduction (Larrain Reference Larrain1979). Dominant classes in the industrialised West often highlight the capitalist market as a symbol of human choice and freedom, rather than acknowledging its imperatives and compulsions (Wood Reference Wood1994). Marx and Engels (Reference Marx and Engels1970) argue that dominant classes give ideas a form of universality, and represent them as the only rational, valid ones, thus presenting their interests as the common interests of all members of society.

The ideas that are presented as providing a form of ‘universality’ become foundational paradigmatic ideas to contribute to elite cohesion, mobilise support and demobilise dissent. Ideology provides a ‘moral explanation’, which can sustain individuals during chronic strain, either by denying it outright or by legitimising pain through the goal of achieving a higher end. It provides ‘solidarity explanations’ to knit a social group or class together. It also ‘claims complete and exclusive possession of political truth’ and ‘works towards a utopian culmination of history’ (Geertz Reference Geertz1973, p. 198). Ideological goals make ideological action attractive to loyal actors and force these actors to work in line with achieving a common purpose (Anselmi Reference Anselmi2013; March Reference March2003).

This book’s analysis shows how ideas are inseparable from interests and shape who gets benefits and who is excluded from benefits. It takes a Marxian and historical materialist approach to understand the relationship between ideas and interests. It shows how RPF ideology has created a ‘solidarity explanation’ that claims exclusive possession of the truth and is strengthened by the collective memory of a violent history. This will be discussed in more depth in Chapter 3.

The Role of Violence and Conflict

The insight that the distribution of rents in low-income countries is sometimes necessary has gradually gained more acknowledgement among multilateral institutions and development policymakers. Nobel Prize Winner Douglass North made waves in the 2010s after publishing a new book with his colleagues, Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History (North et al. Reference North, Wallis and Weingast2009). North et al. (Reference North, Wallis and Weingast2009) argued that this book broke with many of North’s long-held theoretical assumptions, implying that the book represented something of a paradigm shift. The book’s impact was evident immediately after its publication, influencing the World Bank’s (2011) World Development Report on Conflict, Security and Development.

There are reasons to assume that Violence and Social Orders and a subsequent edited collection (North et al. Reference North, Wallis, Webb and Weingast2011), which recognised the influence of Khan’s PSA, represented a shift in the thinking of influential economists in line with structuralist understandings of late development. North et al. (Reference North, Wallis and Weingast2009) acknowledged that the distribution of rents may be important in solving the ‘problem of violence’ common in late-developing countries. North and his colleagues went against decades of work, within IFIs and among prominent neoclassical economists, which argued against industrial policy and the role of state intervention in low-income countries. However, like other influential works (Acemoglu & Robinson Reference Acemoglu and Robinson2012; Grindle Reference Grindle2004; Rodrik Reference Rodrik2008b), which are celebrated for departing from dominant development thinking, North et al. (Reference North, Wallis and Weingast2009) remained committed to the desirability of dominant Good Governance and market-enhancing policies (securing property rights, transparency, reducing corruption and accountable political systems). In this way, North et al. (Reference North, Wallis and Weingast2009) remain committed to the neoclassical definition of capitalism that well-defined property rights reduce transaction costs and allow markets to allocate resources more efficiently. Instead, Khan’s (Reference Khan, Sundaram and Reinert2005, p. 71) approach rests on the Marxist definition of markets as a system of compulsion, developed by Wood (Reference Wood2002), which compelled capitalists and workers to continuously strive to improve their performance just to survive.

North et al. (Reference North, Wallis and Weingast2009) argue that two kinds of societies exist: Limited Access Order (LAO) societies and Open Access Order (OAO) societies. OAOs are depicted as providing for equality for all citizens and entry without restraint into all aspects of economic and political life. Key conditions to be characterised as an OAO include the rule of law for elites, perpetually lived organisations in the public and private arenas, consolidated civilian control of the military and other organisations, which have the potential to enact violence. North et al. (Reference North, Wallis and Weingast2009, p. 114) also argue that the life of all citizens should be characterised by ‘impersonal exchange’ and thereby suggest that an ideal society is one that is governed by market relations (in line with the Good Governance approach). However, in this book, North et al. (Reference North, Wallis and Weingast2009) begin by arguing that most societies – and all low-income countries, which they characterise as LAOs – must deal with the problem of containing violence. In contrast to OAOs, which, North et al. (Reference North, Wallis and Weingast2009) argue, have states that satisfy the Weberian assumption of a monopoly on the legitimate use of violence, LAOs are controlled by coalitions of military, political, religious and economic elites. In LAO societies, elites limit access to valuable political and economic functions to generate rents. In LAO societies, powerful individuals who receive rents are disincentivised from using violence and instead choose to cooperate with ruling coalitions to be peaceful. Thus, North et al.’s (2009) framework relies on methodological individualist assumptions, with individuals resorting to violence to further their self-interest (Gray Reference Gray2016). The only path to securing peace is presented as providing these powerful individuals with rents.

There is a tendency within economic models of conflict and some influential American political science literature to retain assumptions of methodological individualism and rational choice. In the economics of conflict literature, Paul Collier and Anke Hoeffler best characterise this approach. They (Collier & Hoeffler Reference Collier and Hoeffler2004) argue that ‘greed’ is the most significant motivation for individuals to engage in conflict. However, individuals and groups resort to violence not just for their own monetary gain or fear of being deprived of their assets but also for social, emotional and ideological reasons (Cramer Reference Cramer2006). In Khan’s (Reference Khan2010) work, the role of violence is under-explored. Yet PSA consistently highlighted that violence is intrinsic to processes of capitalist accumulation (Behuria et al. Reference Behuria, Buur and Gray2017; Gray Reference Gray2016). In contrast, non-structuralist PSA approaches (Kelsall et al. Reference Kelsall, Schulz, Ferguson, Vom Hau, Hickey and Levy2022) have aligned with North et al. (Reference North, Wallis and Weingast2009), arguing that productive outcomes are associated with limiting violence (in line with the liberal peacebuilding literature).

PSA has tended to be primarily focused on analysing violent processes rather than outright civil war. Non-structuralist PSA approaches focus on limiting violence or eliminating it. Thus, such approaches are incompatible with a Marxian characterisation of capitalist accumulation. The PSA approach employed in this book is used to capture processes of inevitable violence, contestation and inequality that characterise late development. Chapters 3–10 elaborate these ‘structuralist’ aspects of PSA.

Conclusion: The Potential of ‘Structuralist’ Political Settlements Analysis

Global development policymaking and academic work have long been criticised for ignoring the centrality of the politics in specific contexts when explaining why policies fail or succeed (Ferguson Reference Ferguson1994; Hickey Reference Hickey2008, Reference Hickey2009; Venugopal Reference Venugopal2022). Comparative politics, too, has tended to relegate analysis of power to the margins in favour of analysis of regime types of dictatorships or democracies. Structuralist PSA provides us with a way to sketch out how the evolution of transnational coalitions evolves as countries travel through their contested development trajectories. Traditional PSA, in its attempt to become more amenable to disciplines such as political science, has fallen into the trap of developing 2 × 2 typologies of different types of political settlements, which may predict certain socio-economic outcomes. However, such typologies limit the explanatory power of PSA, which is better elaborated through analysing the historical evolution of power relations across the macroeconomy, the financial sector, as well as across other key economic sectors. In the pages that follow, Rwanda’s political settlement will be sketched out in this way to describe how, since 1994, the RPF has attempted to maintain political order while investing in its ambitious goal of becoming a services-based hub.

RPF ideology is rooted in the specific histories of its members, as well as the lived experience of conflict in Rwandan society. It is based on the refugee identities of families of the ruling coalition who led the liberation in 1994. Most of these RPF cadres lived in neighbouring countries within East Africa for decades before returning to Rwanda. Elite cohesion is based on loyalty to an ideological goal of never returning to a haunting national past of violence, dislocation and genocide. A new national identity (Ndi Umunyurwanda – I am Rwandan) is portrayed as a departure from regressive ideologies of previous governments, based on divisive ethnic identities.

The PSA analysis presented in the next section describes how the RPF’s development strategy is predicated on the external financing of the political settlement. As a result, dissidents and critics of the government have focused their attention on discrediting Rwanda’s external reputation, as well as its domestic developmental achievements. Unlike most studies on Rwanda, this book shows how Rwanda’s main political opponents have been demobilised. The RPF has maintained political order through disrupting possible networks from rivals and marginalised segments of the population, as well as potential financiers. The RPF has been particularly successful at maintaining elite cohesion and demobilising threats even during times of economic weakness, which most of the previous independent Rwandan governments were unable to do. Transnational contestation in Rwanda’s political settlement, which the book argues is a characteristic of contemporary late development, is particularly extreme because of the RPF’s prioritisation to become a services hub.

Chapter 3 provides a snapshot of Rwanda’s economic rise since the 1994 genocide. It then illustrates the political economy of present-day Rwanda, including pinpointing the vulnerabilities characterising the external dependence of its services-first strategy.