The Role of State Capacity in Development Studies Free

Abstract

The role that state capacity plays in development has been the subject of growing cross-disciplinary research. The state capacity concept has faced criticism on numerous fronts, with scholars recently devoting serious efforts to overcome it. This article offers a rich overview of the concepts’ journey in terms of its theoretical roots, its uses by different disciplines, its definitions and measurements, and the empirical evidence around the impact of state capacity on development outcomes. It argues in favor of an optimistic view of the field’s advances by showing: (a) convergence in conceptualizations and (b) high-quality innovations in measurements. The article also contributes to the field by offering a definition of state capacity that consolidates all elements of the dominant strand and further tackles a number of issues that remain weak and should take priority in a future agenda.

1 State Capacity: An Encounter of Two Worlds

As in many other domains of life, the 1970s gave way to intense intellectual debates in the literature on state capacity. Against the backdrop of a highly divisive context, economic historians and political sociologists would document and exchange stories about the role of the state, choosing to follow the neo-Marxist view equating states to superstructures mirroring societal interests, or instead conform to Weber’s notion of state autonomy and neutral competence. Theoretical discussions took place around the drivers of state formation, state strength, and the dynamics between agency and structure in the shaping of statehood. Yet, from both an ontological and methodological perspective, these debates pertained to a rather compact world, with all strands living in relative harmony. The state was (again) portrayed as a centralized single entity whose capacity could be captured and summarized, and therefore qualify within the explanations behind different historical paths toward development.

While this sense of clarity was being enjoyed by some disciplines, political scientists first voiced some discomfort at the way the state was being treated as a macro explanatory variable for development. As succinctly put by Barbara Geddes (1996), after being “brought back” to academia, the state had turned into a “great clumsy creature that no one quite knows what to do with” (1).

Almost twenty years passed until Francis Fukuyama’s article The Strange Absence of the State in Political Science (2012) and subsequent publication (2013), which did not seem to do much to lure political scientists into the charms of bureaucracies as study subjects. Fukuyama’s points managed to introduce a number of reconfigurations to research on state capacity and governance by sparking a prolific discussion on why so many scholars are keen on studying power-limiting institutions while few focus on the way states accumulate and exercise power through their bureaucracies. Part of the explanation, Fukuyama argues, is that the predatory reputation of states among political economy scholars has turned the attention toward power limitation. Or perhaps, less overtly, that assessing checks is empirically easier than assessing execution power.

Such a remark by Fukuyama must have felt like friendly fire to public administration scholars across the world, who for decades had been dissecting bureaucratic virtues and vices. Yet, the dividing line between public administration and political science continues to generate controversy. In this sense, Fukuyama’s article made transparent the tensions lying at the crossroads of different disciplines interested in the state, no longer able to ignore each other. Advancing the literature on state capacity would now take the form of an encounter between two worlds. First, the world of capacity, dominated by practitioners and public administration scholars focused on organizational studies, mostly interested in understanding the organizational drivers of successful policy implementation in remote places (e.g., OECD 2006; Peters 2015; Wu, Ramesh, and Howlett 2015). Second, the world of the state, an umbrella construct overly simplified by certain disciplines in order to fit it into models as one (or very few) informative variables for large-N statistical analyses.

This cohabitation proved difficult but necessary. Organization scholars could achieve greater generalization for their studies at the expense of giving up excessively granular approaches to bureaucracies, while comparative scholars could finally include a missing link between inputs (e.g., rules and resources) and the resulting developmental performance (at the expense of admitting the state is not monolithic).

The greater diversity in the field can be observed on different fronts. First, in the growing number of empirical efforts documenting the effects of state capacity on development. State capacity is thought to make a difference for growth and long-run performance (Dincecco and Katz 2014; Evans and Rauch 1999; Hamm and King 2010; Hamm, King, and Stuckler 2012), industrialization (Evans 1995), innovation (Weiss 1998), health and social outcomes (Cingolani, Thomsson, and de Crombrugghe 2015; Scartascini, Stein, and Tommasi 2008), and more generally to policymaking quality (Peters 2015; Rothstein 2005; Scartascini, Stein, and Tommasi 2013; Weaver and Rockman 1993) and policymaking freedom (Franco-Chuaire, Scartascini, and Tommasi 2017). In terms of international affairs, state capacity has been linked to the prevention of civil wars and conflict (DeRouen et al. 2010; DeRouen and Sobek 2004; Fearon 2005; Fearon and Laitin 2003), reduced war contagion (Braithwaite 2010), the implementation of peace agreements (DeRouen et al. 2010), and reduced human rights violations (Englehart 2009), among many others.

At the same time, state capacity gained terrain as a comparative discipline with a global view. Scholars from multiple latitudes have offered systematic comparative efforts on state capacity (Centeno, Kohli, Yashar, and Mistree 2017) and refined methodological frameworks tailored to practitioners on-site have been further supported by international organizations such as the World Bank (Andrews, Pritchett, and Woolcock 2017). The field also gave rise to renewed debates on the proximate and historical determinants of state capacity (and its erosion), many times challenging conventional wisdom (e.g., Gennaioli and Voth 2015; Reinsberg, Kentikelenis, Stubbs, and King 2019).

As Fukuyama’s critique came to light, a number of simultaneous voices echoed similar concerns in the form of comprehensive reviews of the concept (Cingolani 2013; Hanson and Sigman 2011; Savoia and Sen 2015) or critical notes on its use (Centeno et al. 2017; Enriquez and Centeno 2012; Williams 2018). Most of these acquired a negative tone and issued warnings about concept stretching, tautologies, and construct validity, even suggesting an end to the concept’s use altogether (Williams 2018).

The objective of the present review is to counteract some of these negative views by highlighting the advances, innovations, and increasing convergence in the field of state capacity for development in recent years. It will examine trajectories and trends from a number of angles: the concept’s theoretical grounding, its demonstrated impacts on development, its conceptualizations and operationalizations. It will extend a definition of state capacity that consolidates a large number of converging views. All in all, the article is intended to offer a critical analysis of the strengths and weaknesses associated with the concept and draw an agenda for future contributions.

Several conclusions arise from this examination. The first is the growing diversity in terms of how disciplines and methodologies account for the role of state capacity in development, with an increasing trend toward incorporating more aspects into existing models. The second is the notable growth in efforts to extend and improve empirical works measuring the effects of state capacity for development. Third, the current overview makes way for an optimistic view in terms of methodological maturity, by showing that while the concept has been stretched, a clear dominant definition has emerged, and that promising empirical innovations are on a good path to addressing previous limitations. Yet, important challenges remain in terms of causality and contextualization.

The article is outlined in the following way. The next section reviews four theoretical strands that have advanced the use of the concept in different disciplines. It is meant to create bridges for readers interested in developmental states who are coming from different disciplines. Section 3 provides a comprehensive analysis of the existing evidence on the effects of state capacity on development, clustered by themes. Section 4 provides a systematic overview of the conceptualizations and measurements in the literature, and proposes an umbrella definition of state capacity in line with the dominant strand. Section 5 critically reflects on the advances and pending challenges in the field, while the final section provides a few concluding thoughts.

2 Four Theories of State Capacity and Development

This section provides an overview of four interrelated theoretical streams highlighting the role of state capacity in development, namely the bellicist, Weberian, relational, and political economy approaches. While the boundaries between them are not clear cut, there are nuances in the way they understand the dynamics of capacity building. While both the bellicist and Weberian traditions are state-centric, they focus on different Weberian attributes: the first on the centralization of coercive power, and the second on the insulation of modern bureaucracies. At the same time, while the relational tradition acknowledges the role of wars and fiscal extraction in statebuilding (as in the bellicist tradition), it is deemed more socio-centric, with state capacity shaped by the reinforcing interactions between state and society. Finally, the political economy tradition is much more recent and differs methodologically by placing a greater emphasis on agency versus structure, and incorporating more hybrid elements in its models of capacity building. Only a selection of representative works is included, although each strand has been prolific and complemented by multiple empirical works (only a few referenced here).

2.1 The Bellicist Tradition

The pioneer work of this tradition is the highly detailed historical account of Charles Tilly (1975, 1992), a collection of essays that looks into the role of war in the formation of national states in Western Europe. These narrate the evolution of such states going from a fragmented network of feuds into centralized absolutist regimes able to claim the monopoly of coercive power, the most primary element of stateness. The pivotal objective of Tilly’s work is to show that variations in state formation processes respond to different countries’ military needs for warmaking (mainly in France, England, and Germany), by means of securing an efficient and centralized revenue extracting apparatus. In Tilly’s work, stateness in Western Europe is defined by formal autonomy, differentiation from nongovernmental organizations, centralization, and internal coordination (1975, 35). The level of stateness is in turn determined by the state’s capacity to raise tax, that is, the capacity to “build an apparatus which effectively drew the necessary resources from the local population and checked the population’s efforts to resist the extraction of resources” (40).¹ Tilly’s work had multiple followers and influences on the literature on state formation and fiscal capacity in Western Europe (e.g., Downing 1993; Ertman 1997).

Yet, the bellicist theory proved insightful not only in the Western European context. The work of Herbst (1990), for example, uses Tilly’s framework in Africa and East Asia as two contrasting cases that grant support to the bellicist theory. In Africa, Herbst boldly predicts that states are unlikely to build strong capacities, since the fundamental structural changes needed for state formation are not present in the absence of external war. In Taiwan and South Korea, in contrast, extended periods of warfare fostered nationalism and allowed states to extract substantial resources from the population, favoring the control of dissident groups.

Similarly, the work by Centeno (2002) seeks to unravel the impact of internal and external war in state effectiveness in Latin America. In contrast to the European case, organized violence in the form of state war is a rare occurrence in the continent, presenting an interesting complement to Tilly’s theory. State capacity here is understood as “the ability of the relevant political authority to enforce its wishes and implement policies” (3). The book documents how and why the lack of this type of violence and low state capacity reinforce each other over time. To answer this, Centeno points to a number of starting conditions: (a) Latin America fought “limited wars” instead of mass wars; (b) it minimized border conflict by accepting colonially imposed borders; (c) external powers mediated in wars; (d) domestic elites were more divided than in Europe; (e) more caste divisions existed, and (f) administrative chaos followed decolonization (22–23). After this initial independence, context path–dependency followed, and the lack of centralized violence led to weak statebuilding. The bellicist theory suggests several channels through which wars build states: by centralizing power and revenue extraction, by creating a higher level of national belonging that encourages cooperation toward the state, through war success as a way to grant state legitimacy, and through mass conscription, which improves organizational capacity and further centralization. Centeno explains why these channels are broken in Latin America. While administrative capacity barely existed at the moment of independence, partial wars did not encourage mass mobilization of armies, therefore discouraging its professionalization. Also, social divisions were above national sentiment and the results of war were mostly negative: debt, breakdown, and chaos. Yet, within this universe, he finds variations in the continent: the centralizing power of war was stronger in Chile, Mexico, and Argentina. In concluding, Centeno suggests thinking about treating Europe as an exception instead of a rule.

The bellicist thesis influenced advances in the study of state capacity in other subdisciplines, particularly in the more recent game-theoretic literature on investments in state capacity, as will be shown.

2.2 The Weberian Tradition

The Weberian tradition is also considered state-centric, but focuses on unpacking the role played by the key emerging actor mediating between conflict, resource extraction, and stateness: the modern bureaucracy.

A classic early work along these lines is that of Theda Skocpol (1979) States and Social Revolutions: a Comparative Analysis of France, Russia, and China. It analyzes how the state plays a role in the outbreak of revolutions in the three countries, by comparatively and historically assessing the interactions between the state apparatus and the ruling elites, and between bureaucratic actors themselves. In her model, the nature of these interactions determines the state’s autonomic power. As hinted at earlier, while Skocpol focuses on the administrative capacity to carry out central policies, she acknowledges other dimensions, too, such as coercive and extractive capacity: “Obviously, sheer sovereign integrity and the stable administrative-military control of a given territory are preconditions for any state’s ability to implement policies. Beyond this, loyal and skilled officials and plentiful financial resources are basic to state effectiveness in attaining all sorts of goals” (Skocpol 1979, 16). The work shows that state structures affect the nature of revolutions, and these, in turn, shape historical processes of state formation: “strengthened states—more centralized, bureaucratic and autonomously powerful—emerged from all three Revolutions” (285). Her work is ground-breaking in the study of revolutions and state formation and has multiple derivations.

A similar perspective is presented in Bringing the State Back In by Evans, Rueschemayer, and Skocpol (1985), a compilation aimed at showing how the autonomous power of the state is built and operates. In the introduction, Skocpol (1985) defines state capacity as whether a state is able to “implement official goals, especially over the actual or potential opposition of powerful social groups or in the face of recalcitrant socioeconomic circumstances” (9). This capacity is achieved when “[state] organizations claiming control over territories and people may formulate and pursue goals that are not simply reflective of the demands or interests of social groups, classes or society” (9). The authors call for an inductive approach to the study of the state, which they contend is the best way to capture the complexity of state–society relations. In this book, Rueschemeyer and Evans (1985) place the effectiveness of state intervention in the economy at the center. Extending Weber’s notions, they argue that a balanced combination of professional cadres in the bureaucracy and proper levels of intra-state agency coordination are the key to economic transformative capacity, an idea that would be more extensively explored in Evans (1995).

More recent scholarship has sought to expand on the historical processes of capacity building in modern European states, emphasizing the mechanisms of bureaucratic capacity building (Van de Walle and Scott 2009) or the historical determinants of merit-oriented versus patrimonial civil services (Charron, Dahlström, and Lapuente 2012).

In a similar vein, the book by Barbara Geddes, Politican’s Dilemma (1996), is a salient initiator of debates on how political institutions shape politicians’ incentives to build administrative competence in the more developmental context of Latin America. Her model lies somewhere in between intentional and structural approaches, where certain institutions constrain politicians’ choices regarding administrative reform, and these choices, in turn, have an impact on the structure of the politics–bureaucracy interaction in the future. Geddes considers state capacity to be equated to the implementation power of the state, a task that falls inherently under the bureaucracy and is as important as decision making. This implementation power depends on the possibility to count on a politically insulated bureaucracy through advancements in terms of merit-oriented administrative reform: “the capacity to implement state-initiated policies depends on the ability to tax, coerce, shape the incentives facing private actors, and make effective bureaucratic decisions during the course of implementation. All of these abilities depend in turn on the existence of effective bureaucratic organizations” (14). In Geddes’s mindset, the state is not an actor with relatively autonomous preferences, but an entity consisting of self-interested individuals of four different types (presidents, legislators, party leaders, and bureaucrats) who seek to maximize their career opportunities but face different behavioral incentives. She introduces the idea of the politician’s dilemma: a president faces a tradeoff between appointing competent state mangers who increase the chances of fostering growth and development, and appointing partisan managers to reassure their support. Both dimensions contribute to their career prospects. Administrative reform is therefore seen as a collective action problem of the type iterated in the prisoner’s dilemma, where cooperation to reduce patronage through reform happens only if others are willing to cooperate as well.

Politicians who might otherwise consider offering reforms as a strategy for attracting support will not be able to afford the cost in lost political resources as long as they compete with others able to use such resources in the struggle for votes. This is the politician’s dilemma. A politician might in some circumstances, however, be willing to give up this resource if everyone else were also willing (42).

Geddes’s work builds a bridge between the Weberian origins of modern professional bureaucracies and its correlates in neopatrimonial contexts. Another prolific contribution along these lines is the work of Peter Evans on “pockets of efficiency” in national bureaucracies (Evans 1989); that is, administrative strongholds where bureaucratic action is successfully isolated from electoral or patronage politics and reproduces itself over time on the basis of rational–legal rule (see, for example, Hout 2007 or Roll 2011). At the same time, the concept of neopatrimonialism is often closely related and intertwined with patron–client networks, patronage, and the notion of “rentier states” in political economy studies.² This literature, however, is more linked to the next strand, where state structures mirror societal interests as opposed to acting autonomously.

2.3 The Relational Tradition

The relational tradition is less state-centric in the sense that while acknowledging the importance of warfare and other drivers, it points at the specific interaction between society and the state as the ultimate determinant of state formation processes. Two proponents, Joel Migdal and Michael Mann have persuasively framed discussions in this direction.

The book by Migdal, Strong Societies and Weak States (1988) is illuminating in terms of how state capacities are configured in developing countries. Migdal sees the development of the state apparatus as determined by the social structures it seeks to regulate. He contends that the state-centric perspective is not instructive in developing countries, as autonomy and centralization are rare occurrences. Migdal defines state capacity as “the abilities of state leaders to use the agencies of the state to get the people in the society to do what they want them to do” (1988, prologue), or otherwise to “achieve the kinds of changes in society that their leaders sought through state planning, policies and actions” (4). In this context, his perspective prioritizes the capacities to extract resources, regulate social relationships, and penetrate the territory. He resorts to the dichotomous typology of weak and strong states, the former having high capabilities to complete these tasks, the latter unable.³ The main question addressed is why some countries were able to establish effective behavioral rules for societies, while others failed. He finds that the key to explaining state capacity is the distribution of social control among many organizations, including the state. In order to build a strong state (for example, Israel), it is necessary for colonial powers to centralize social control and resources within a governmental organization instead of diffusing control through local chiefs. The latter option favors strong societies and weak states, and facilitates state capture. While strong societies are conducive to democracy, they hinder “stateness,” and therefore, the formation of strong states.

In The Sources of Social Power (1986 and 1993), Michael Mann introduces some innovative concepts in the field of state capacities. He studies power relations in terms of multiple and intersecting networks of influence, which represent organizational means of obtaining human goals. He defines four so-called sources of social power: ideological, economic, military, and political, the latter being the state’s domain. He contends, in turn, that states exert two types of power: despotic and infrastructural. Despotic power is the state’s power to impose mandates over society, or “the range of actions the state elite is empowered to undertake without routine negotiation with civil society groups” (Mann 1988, 59), while infrastructural power refers to “the institutional capacity of a central state, despotic or not, to penetrate its territories and logistically implement decisions” (59). The distinction introduced by Mann led to new debates, as despotic power was until then the most frequent interpretation of state power.⁴

Relational approaches to state capacity gained important terrain in moving beyond organizational views of the state, which in turn fostered theoretical discussions about governance as a broader concept than government (see Robinson 2008 for an example of state capacity analyzed in terms of socio-economic engagement).

2.4 The Political Economy Tradition: Capacity as Investments

A more recent and fast-growing strand of game-theoretic literature has emerged with a focus on the incentives actors face to actively invest in state capacity. Much of this literature converges in purpose with economic history works on state formation (e.g., Cárdenas 2010), while it is conceptually embedded in the “limited state” scholarship (North 1990). In general, it is assumed that contracting institutions are essential in order to provide the adequate environment for investment and innovation. The enforcement of property rights requires a strong legal system that constrains the ruling elite’s expropriation potential through predictability and rule of law. In this respect, state capacity is understood mainly as legal capacity, and it is associated with less intervention potential.

For Margaret Levi (1988), state capacity is the state’s ability to provide collective goods by raising revenue. She follows the line of Tilly by studying the determinants of fiscal capacity, but focuses on internal political conditions. Levi contends that leaders are self-interested and revenue maximizers for the whole of their tenure, as revenue “en-hances the ability to extend rule” (2). The main hypothesis is that the chances to raise revenue and the organizational structures favored are affected by three main constraints. First, the insecurity of the leader’s mandate duration. Leaders discount the future based on their expectations to remain in power, and normally, political rivalries and insecurity lead to high discount rates. In these settings, extraction policy tends to be less concerned with damaging economic performance. Second, the transaction costs associated with revenue raising. It is assumed that agents will break contracts whenever convenient, depending on the costs associated with the changes. Policymakers need to lower transaction costs, as a tradeoff exists between revenue extraction and transaction costs. The latter are defined as “the costs of measuring, monitoring, creating and enforcing compliance” (12). A third constraint is the bargaining power of rulers. Rulers have more bargaining power if they can monopolize coercive, economic, and political resources. Given these factors, creating compliance is a continuous interaction between rulers and citizens in the form of a collective action problem. Levi’s main hypotheses find support through an empirical strategy that explores the variation in major tax policy choices as a consequence of changes in the three political constraints mentioned above. The historical case studies are chosen according to different levels of state development but undergoing historical change: Ancient Rome, medieval England, and France, Britain during the industrial revolution and post–World War II Australia.

Acemoglu (2005) presents a model seeking to outline the effects of state strength on the economy. The article attempts to fill the gap regarding the tradeoff between the distortions posed by self-interested rulers with taxation power, and the inefficiencies that arise from the lack of state authority. For this, it builds on the typology by Migdal (1988) on weak and strong states, and the tradeoff between the power of the state versus the power of society. A state can be weak or strong both politically and economically: in politically weak states, rulers can be replaced easily, while economically weak states have low capacity to tax. Rulers need to choose between using public resources for their own consumption or devoting them to the provision of public goods. In this model, too weak and too strong states (both economically and politically) create distortions. In weak states rulers cannot derive high future benefits from public investments, and therefore choose to underinvest in public goods and state capacity. When states are too strong, they can impose such high taxes that the economy can lose vigor. The ideal scenario is somewhere in the middle, where rulers can tax and have incentives to invest in public goods. An equilibrium named consensually strong state exists if the state is politically weak but high taxes are imposed. This state lends support to the inverse relationship commonly observed between taxing capacity and executive constraints, and it emerges when decreases in political power raise investments in public goods, and as a consequence, citizens agree to pay higher taxes.

Cárdenas (2010) offers further theoretical insights on the determinants of state capacity building in Latin America. In line with Centeno (2002) he argues that the Latin American state has been “extremely weak in terms of the most basic capacity measures” (2). Although initially capacity is understood as the “professionalization of the state bureaucracy, its ability to protect property rights and make credible commitments to private investors, as well as its ability to raise revenue from the society” (2), later, fiscal capacity takes the main role: “state capacity is defined as the state’s ability to generate tax revenue from the public” (15). The article presents a game-theoretic framework based on Cárdenas and Tuzemen (2011) that models the impact of four factors in statebuilding: two traditional ones in bellicist accounts like internal and external war, and two novel ones, political and economic inequality. Capacity accumulation is modeled as investments under uncertainty, in the same vein as Besley and Persson (2009, and others). In a nutshell, whoever is in power (the citizens or the elite) has to make decisions on how much to invest in state capacity for the future, how much to tax, and the amount dedicated to providing public goods. Investments in state capacity are risky because they take away resources from private consumption, and can be used by the opposition at a later point in time for redistribution. There is uncertainty about how much society will value public goods in the second period, because it is determined by exogenous factors (e.g., internal or external wars). Each sector derives utility depending on the value of public goods and the resources left for private consumption after tax. The government makes the three decisions based on the maximization of the sum of weighted utilities of the two groups (the weight depends on whether the government is utilitarian or group-biased) plus the expected payoff for the second period, for the group who is in power in the first period. The decision on how much to invest in state capacity for the second period depends on two unknown variables: the future value of public goods, and who will hold power, in addition to the expected net payoff. The result is that the optimal level of investment depends on the value of public goods in both periods, as well as the level of group-bias (the weight given to the value of private consumption for the ruling group). While conflict affects the first, political and economic inequality affect the second. The results show that political and economic inequalities lower the incentives to invest in state capacity.

Acemoglu, Ticchi, and Vindigni (2011) present a model showing some mechanisms by which inefficient states arise and persist as a consequence of patronage politics. They consider that “societies with limited state capacity are those that invest relatively little in public goods and do not adopt policies that redistribute resources to the poor” (1). State capacity and state efficiency are somewhat equated and involve the abilities of a central authority to monitor bureaucrats. This capacity is in turn dependent on previous investments. Starting from a nondemocratic regime, the article claims that a transition to democracy might foster the selection of an inefficient state by the rich elite in control. This happens as the elite captures democratic demands through patronage, and avoids redistribution demands. The use of patronage can become self-reinforcing and thus explain the persistence of inefficient states. After democratization, bureaucrats can vote for either the pro-poor or the pro-elite party. The electoral support of the bureaucracy is key to winning elections. Although the bureaucrats are by assumption poor, they find it more convenient to vote for the pro-elite party, as they foresee that the poor will demand changes in the bureaucracy in order to better redistribute rents through taxing policy. If the elite wins, bureaucrats receive convenient rents from the elite that they would otherwise lose after efficiency-oriented reforms. The elite also gain from an inefficient state, as it avoids more taxes and redistribution. The emergence and persistence of inefficient states is more likely to occur when inequality between rich and poor is higher, because the rich elite is more prone to avoid redistribution and therefore allow higher rents for the bureaucracy. On the other hand, if rents are too high and become too costly for the elite, the latter might simply prefer to redistribute, pushing bureaucrats to adapt their demands accordingly.

Besley and Persson (2007, 2008, 2009) do extensive work on unraveling the determinants of state capacity. In this setting, capacity building is seen as an investment made by incumbents as a function of future levels of social valuation for public goods. These levels are allowed to vary exogenously according to different political and historical circumstances, so that capacity building acquires the format of investments under uncertainty.

Besley and Persson (2008) analyze how self-interested incumbents decide to use part of the government’s revenue to invest in fiscal capacities that enable higher tax extraction from the society in the future. Raising public revenue is key to providing a higher level (or a better quality) of public goods, in this case exemplified by defense against external threats. They analyze how two types of exogenous conflicts, internal insurgency and external war, affect the value that society grants to public goods, and as a consequence, the relevance of investing in fiscal capacities. The model predicts that the expectation of future external conflict will raise today’s investments in fiscal capacity, and all sectors of society will be taxed equally, such that revenue raising is optimized. In contrast, the expectation of future internal conflict has the opposite effect: as all fiscal capacity is destroyed when internal conflict occurs, the value of public goods decreases, incumbents favor their own group in society with redistribution, and no public goods are provided.

An extended model (Besley and Persson 2007, 2009) also analyzes investments in legal capacity as an endogenous policy decision. Legal capacity is understood as the capacity to protect and enforce property rights, a market-supporting institution. Apart from conditioning investments on internal and external conflict, they also show that other political factors can change the value of public goods, such as the level of inclusiveness of institutions (capturing political polarization), the level of political stability, and whether the political regime resembles that of a utilitarian planner or a politically controlled one. Several derivations follow: investments in legal and fiscal capacity are complementary; wealthier countries choose higher levels of both types of capacity; greater political stability raises investments in state capacity; a more representative political system (meaning, closer to the utilitarian planner ideal) also raises investments in both capacities; and finally, the higher the wealth of the ruling group, the higher the investments in legal capacity, and the lower in fiscal capacity.

One of the novelties of Besley and Persson’s (2011) Pillars of Prosperity is that it incorporates the role of violence and seeks to interact more closely with the literature on conflict and fragile states. It further adds complexity on the conceptualization of state capacity, and makes a special effort to distinguish policies from institutions. State capacity is seen as capital investment, and it has four dimensions: (a) taxing; (b) contract enforcing (called productive capacity); (c) coercive (peace and order); and (d) regulatory/distributive (public goods provision). The authors point out that while most of the economic development literature takes for granted the ability of the state to perform these tasks, this is very much different for the development community in weak and fragile states. The book asks what determines the building of different capacities, and why are they complementary. In this framework, self-interested governments decide on policy depending on a set of constraints posed by the political institutions, and always seek to be re-elected. At the same time, their re-election depends on investments on violence, similar to those who are out of power and maximize their opportunities to regain it. The conclusions show that stable politics, along with consensual institutions and a higher demand for public goods, are more conducive to greater investments in state capacities. There is a particular emphasis on the importance of cohesive institutions. There are three resulting state types: (a) common interest states; (b) redistributive states; and (c) weak states. At the same time, the book explains the empirical pattern of clusterization of low income, fragile states, and violence by the existence of complementarities between state capacities. One very interesting angle of the book is that it explains how different types of foreign aid can affect statebuilding. For example, cash aid can help common-interest states, but has no effect when states are redistributive or weak. Infrastructure aid can raise private incomes and help stress the importance to build the state. Differently, military assistance can reinforce the power of the incumbent and increase violence.

3 The Empirical Evidence on the Role of State Capacity in Development

As suggested in the introduction, one salient aspect of the literature on state capacity is the growing empirical research on the effects of state capacity for different development outcomes. Three subdisciplines are the predominant contributors: conflict studies, development economics, and comparative public administration, while game-theoretic political economy cuts across all three (mostly the first two) and has helped in their convergence.

3.1 Conflict and Fragile States

Largely following the bellicist tradition, the conflict literature in international relations took a special interest in state capacity after numerous incidents of state collapse at the beginning of the twenty-first century. The work of Fearon and Laitin (2003), for example, shows that among the factors that facilitate guerrilla warfare and insurgency are low financial, organizational, and political state capacities. Similarly, Fearon (2005) complements Collier and Hoeffler’s argument (2004) on the importance of primary commodity exports as drivers of civil war, by showing that this link runs only through low state capacity. DeRouen and Sobek (2004), in turn, find that state capacity impacts civil war outcomes and duration directly. They argue, however, that while a more effective bureaucracy favors the government, higher coercive power does not necessarily have the same effect.

The relationship between conflict and state capacity is further and extensively explored in a special issue of the Journal of Peace Research. There, Sobek (2010) shows that more capable states provide lesser opportunities for rebels to initiate violence, as they are able to channel social demands in a way that limits the possibilities of rebellious collective action. In the event of violence, moreover, capable states are prone to choose bargaining as a way out. Braithwaite (2010) contends that lower state capacity gives a lesser chance to resist the spread of regional conflict. Thies (2010), in contrast, finds no direct effect from state capacity on conflict onset, but an inverse relationship. DeRouen et al. (2010), in turn, examine the importance of state capacity in the implementation of peace agreements and find an interaction between the two by which third-party intervention is more significant in low-capacity settings.

State capacity has also been analyzed as a commitment-enabling device. The work of McBride, Milante, and Skaperdas (2011) presents a dynamic game-theoretic approach where state capacity levels affect the likelihood for peace to prevail in conflict resolution processes as opposed to the continuation of conflict. State capacity is conceptualized as credible commitment to resolution enforcement and it affects decisions on future military spending, conditional on a number of context-specific factors. The derivations of the model show that only when state capacity is sufficiently large is peace self-reinforcing.

3.2 Innovation and Industrialization

In Embedded Autonomy (1995) Peter Evans analyzes the ways in which state behavior has the power to shape structural change and promote industrial growth through its transformative capacity. By focusing on newly industrializing countries during the 1970s and 1980s, he develops the idea of embedded autonomy, a combination of internal bureaucratic coherence within agencies and external connectedness with key industrial sectors. This embedded autonomy is behind the high levels of state capacity that led to the successful industrialization of some countries (most prominently Korea) over others (Brazil and India as cases of partial success). The historical process that leads to embedded autonomy is achieved through a combination of two types of state intervention in the IT industry in particular: a “midwifery” role of attracting capital to new sectors and a “husbandry” role of nurturing and developing these sectors. These two roles stand in contrast to other more “extreme” roles, namely that of “custodian” or general regulator only, and that of “demiurge,” providing goods and services directly.

Linda Weiss (1998) follows the line of Evans (1995) by exploring the transformative capacity of states against the backdrop of globalization, and challenging mainstream views about the demise of the state in this new type of world (e.g., Strange 1996). Similarly, she contends that the level of state intervention is not informative of its power, but transformative capacity is. The latter is conceived as “the ability [of the state] to coordinate industrial change to meet the changing context of international competition” (7). As a step further from Evans, Weiss studies the somewhat contradictory relationship between transformative and distributive capacity through five case studies: Taiwan, Japan, South Korea, Sweden, and Germany. In her cases, transformative capacity dominates the emerging economies of Taiwan and South Korea, whereas distributive capacity remains at the core of the Swedish economy. Germany and Japan, on the other hand, are two cases with a proper combination of both capacities. She finds that while distributive capacity grants high density to state–society relations, it hinders innovation within the industrial sector.

3.3 Growth

Following the line of state capacity as bureaucratic professionalization, Evans and Rauch (1999) represent a seminal empirical assessment of the impact of bureaucratic “Weberianness” on economic growth. By developing original survey data on bureaucratic features such as meritocratic recruitment, salary arrangements, and career paths, they present evidence showing strong associations between “Weberianness” and economic growth in emerging economies. Similarly, Bockstette, Chanda, and Putterman (2002) develop an index of State Antiquity that they later associate positively with GDP per capita levels and economic growth. The article shows and reinforces the idea that early state formation and institution building is an advantage in achieving higher performance for development.

The work by Hamm and King (2010) concentrates on the role of state capacity as a mediating variable between foreign direct investment (FDI) and development. It assesses the link between FDI and growth in a panel of thirty-one post-Soviet transition economies between 1989 and 2004, and finds that FDI levels have a positive effect on GDP per capita only contingent on high levels of state capacity. In fact, the authors find that the lack of a well-functioning state causes FDI impacts to be negative. Hamm, King, and Stuckler (2012) further show that mass privatization in post-communist countries hindered state capacity, and the latter affected growth negatively between 1990 and 2000.

The articles by Dincecco and Katz (2012, 2014) and Dincecco and Prado (2012) focus on the impacts of state capacity on long-run economic performance in Europe. Dincecco and Katz (2012) present evidence that regime type affects tax revenue extraction and spending quality, and this in turn affects long-run performance. Dincecco and Prado (2012) present an empirical strategy that uses exogenous variation in fiscal capacity in order to assess its effects on productivity. They argue that through war, states make fiscal innovations that persist in time and can explain permanent gains in GDP per worker.

The work of Knutsen (2013) links economic and political factors by presenting an empirical analysis of the separate and combined effects of regime type and state capacity levels on growth. It understands state capacity as the successful implementation of public policies through an efficient rule-following bureaucracy. Empirical evidence from a group of forty-five sub-Saharan countries suggests that democracy has a higher effect on growth in low-capacity countries than in dictatorships, whereas political regime has an insignificant effect in high-capacity states. These findings challenge common views pointing at the detrimental effects of democracy in low-capacity states.⁵ Knutsen argues that democracy limits survival-oriented behavior in low-capacity states, as democracy (vertical accountability) can act as a substitute for state capacity (horizontal accountability).

3.4 Policymaking Quality and Welfare

A number of empirical works have associated state capacity with general welfare and policymaking quality. The work of Henderson, Hulme, Jalilian, and Phillips (2007) demonstrates a tight association between the level of Weberianness in bureaucracies and poverty reduction. Similarly, Acemoglu, Garcia-Jimeno, and Robinson (2015) show that local state capacity has sizable spillover effects in neighboring municipalities in Colombia, and this affects poverty levels, quality of life, and public service provision. Scartascini et al. (2008), in turn, present evidence suggesting that health and education spending is more effective in countries with better state capacity and policymaking capabilities. Cingolani et al. (2015) show that greater Weberianness is linked to better health outcomes as determined by the Millenium Development Goals. Dahlström, Lindvall, and Rothstein (2013) show that bureaucratic capacity has a positive effect on programs that require discretion, but no effect on more standardized programs. Franco-Chuaire, Scartascini, and Tommasi (2017) show that state capacity affects the range of policies available to countries, for example, in the case of risk mitigation policies against trade openness.

4 Defining and Measuring State Capacity: Multidimensionality and Convergence

After establishing the theoretical roots and some of the effects of state capacity on development, this section aims to provide an overview of its precise conceptualizations and empirical measures. It does so with the particular purpose of demonstrating that while the concept remains versatile, it fulfills two desirable properties: a healthy level of definitional convergence, and an increasing innovation in terms of measurement. After observing a wide range of views in the literature, this section ventures a definition of state capacity that blends all elements of the dominant stream.

Figure 1 shows a chronological list of the most salient definitions of state capacity, while Figure 2 shows a selection of some of the most popular measures used for empirical work. It is possible to observe that some scholars acknowledge the existence of several aspects or dimensions of capacity (e.g., Besley and Persson 2011; Fortin 2010; Kocher 2010), while others aim to capture state capacity in a generic way (e.g., Bockstette et al. 2002; Fearon and Laitin 2003; Hamm and King 2010; Scartascini et al. 2008).

In general, the focus is on one or a combination of the following aspects of state capacity: (a) coercive or military (Fortin 2010; Tilly 1975); (b) fiscal (Besley and Persson 2007, 2009, 2011; Dincecco 2011; Levi 1988; Tilly 1975); (c) administrative (Fukuyama 2004, 2013; Geddes 1996; Skocpol 1985); (d) legal (Besley and Persson 2007, 2009); and (e) infrastructural (Fortin 2010; Mann 1986, 1993; Migdal 1988; Soifer 2008). While the concept presents itself as multidimensional, this does not pose threats to construct validity, as will be argued next.

4.1 A Converging Concept

Four generalized approaches can be found among definitions of state capacity, depending on whether they focus on inputs, processes, goal content, or goal achievement.

One set of definitions focuses on specific attributes in place, either taking them as the inputs or endowments that are believed to be conducive to state capacity (or a sign of achieved capacity). While most works will measure capacity through the existence of certain attributes, this literature defines it as the mere presence of those attributes, such as the existence of a professional bureaucracy or military, or any similar infrastructure throughout its territory (e.g., Acemoglu et al. 2016; Bäck and Hadenius 2008; Evans and Rauch 1999).

Another more demanding conceptualization of state capacity refers to a number of process-related aspects in terms of the ways in which states are expected to act. These features are only observable in action. For example, Huntington (1968) expects capacity to be translated into “organizational and procedural durability, complexity, adaptability, coherence and autonomy” (12, in Hanson and Sigman 2011, 5). Similarly, other scholars expect specific features of the policymaking process, such as the capacity to set priorities, maintain both adaptability and stability, or reduce particularism (Rothstein 2005; Scartascini et al. 2008; Weaver and Rockman 1993).

A third type of definition is normative in terms of the nature of state goals, leading to ideological nuances. In this literature, for example, states are expected to provide public goods, enforce property rights, deliver benefits to households and firms, or promote change in order to converge with a competitive context (Besley and Persson 2007, 2009, 2011; Weiss 1998).

The dominant view, however, differs from input, process, or goal content-centered approaches. It focuses instead on the state’s ability to fulfill its goals, regardless of the goals’ nature. This dominant view is therefore non-normative in terms of both policy content and how policy execution should be carried out. In this dominant view, scholars tend to select different thematic dimensions through which states pursue their goals, for example, through coercion, bureaucratic competence, or fiscal extraction. These dimensions are used in order to narrow down the study of the mechanisms for capacity building, but they do not always represent necessary conditions for capacity building (e.g., in this dominant view, states may have high capacity without courts or rule of law).

Based on this dominant strand, state capacity can be generally defined as the state’s ability to credibly implement official goals against potential resistance, by means of taxation, the bureaucracy, the military, the legislature, and the courts, and throughout its entire domains. In pure form, this definition inevitably points in the direction of dissecting the multiple layers of official objectives, ranging from circumstantial desires of elected leaders to historical and path-dependent objectives enshrined in higher level legislation such as constitutions. This endeavor has a complexity of its own, since—as wisely high-lighted by Centeno et al. (2017)—we have not sufficiently acknowledged the tradeoffs between different goals and how achieving one can lead to hindering another.

4.2 The Measures: Continuous Innovation

While a long and diverse list of alternatives can be found in the realm of capacity measurements, a number of data sources can be spotted recurrently for each dimension. Coercive capacity has tended to resort to basic measures of military power such as military personnel or spending (e.g., from Correlates of War), or one of the many existing state fragility indices (e.g., State Fragility Index from Polity IV or the Index of State Weakness from the Brookings Institution). Administrative capacity, in turn, has been most prominently captured through the Index of Bureaucratic Quality from the International Country Risk Guide by the PRS Group. Fiscal capacity has been measured through different variants of tax revenue. Infrastructural capacity and the reach of the state have been proxied by aspects such as road density or census frequency.

Yet, important innovations are taking place with the aim of either achieving higher granularity or extending the geographical and temporal scope. In the field of administrative capacity, for example, the Quality of Government Institute has been releasing its expert survey capturing the professionalization, openness, and impartiality of bureaucracies, and intends to increase its scope over time (Dahlström et al. 2015).⁶ While expert surveys have limitations of their own, the possibility to assess bureaucracies from a number of key dimensions as opposed to only one is critical. Moreover, new empirical efforts are finally capturing bureaucratic information at the agency level through, for example, employee data (see Bersch et al. 2017 for Brazil or Rasul and Rogger 2018 for Nigeria). In terms of effective implementation, new technologies allow the measurement of precise patterns, as in the case of social policy targeting and implementation in one of India’s largest states (Muralidharan, Niehaus, and Suhktankar 2016).

In terms of fiscal capacity, for example, measures have tended to move from plain revenue figures to revenue-against-expected-revenue measures (more closely matching the idea of realized potential). Important innovations in this area have also allowed a much more precise measurement of the territorial reach of extraction, as the work of Harbers (2015) does for Ecuador by using satellite imagery to measure economic activity.

Connected to this, key innovations are also taking place in terms of the reach of the state and infrastructural power. Such an example is the work of D’Arcy and Nistotskaya (2017), who present a novel measure of capacity using cadaster data for a large number of countries and a long temporal scope, illuminating new insights in regard to the historical determinants of state capacity. While not comparative across countries, the measures used by Acemoglu et al. (2015) on the network of colonial roads in Colombia, or the historical opening of post offices in U.S. counties (Acemoglu, Moscona, and Robinson 2016) allow testing propositions about the spacial diffusion of capacity, a rather unexplored aspect of the phenomenon. And while optimizing both scope and granularity generally involves a tradeoff, the field is advancing as to allow their joint maximization.

5 Critical Reflections on Achievements and Challenges

The concept and its derivations have been subject to numerous critical assessments in recent years.⁷ The majority relate to how state capacity has been stretched, misconceptualized, or poorly measured. This section reflects on how the discipline has dealt with a number of topics, in descending ordered of (perceived) success.

5.1 Conceptual Stretching

Several scholars have argued that state capacity lacks conceptual clarity and may encompass too much (e.g., DeRouen and Sobek 2004, among many). This broad overview shows that while four main “visions” exist, there is in fact a great deal of convergence toward a neutral goal-oriented definition, where capacity is equated to the realization of official goals against potential resistance. The recent discussions taking place in high-ranking journals have established a narrower and more fluid dialogue, which lessen concept-stretching problems.

5.2 Normativity

As a consequence of this convergence toward a goal-achievement definition, the field has also taken steps to resolve problems emerging from normative definitions of state capacity. Various scholars have, for example, pointed at the opposing dynamics between legal (power-limiting) versus implementation (power-deploying) capacity (Cingolani 2013; Fukuyama 2013). The fact that these two capacities are exerted by different institutions and may be in direct tension with each other raises questions about how to account for both simultaneously within the same construct. The work of Fukuyama (2013) has advanced clear answers in this respect, differentiating capacity from other power-limiting concepts such as good governance, rule of law, or bureaucratic autonomy. While there is no full consensus in this respect, the field has clearly advanced in disambiguating such concepts.

5.3 Capacity to Travel in Time-Space

Several scholars have pointed at the deficiencies in terms of cross-national and time series indicators for state capacity (Brans 2003; Cingolani 2013; Dalström et al. 2010; Fukuyama 2013; Lapuente 2007), and in particular regarding subnational measures (Enriquez and Centeno 2012). While many of these concerns persist, the literature has moved in the right direction. As earlier illustrated through recent empirical works in Brazil, India, and Ecuador, subnational measures have made critical progress, and bear promises of scalability in order to achieve both intra- and inter-country comparability. In terms of time-series measures, much of the literature has simply focused on largely deterministic factors that by nature entail very low intertemporal variance. This is explained by the necessary interest in the structural aspects of state formation, but limits the possibilities of assessing the micro-determinants of capacity and their policy implications. Yet, good progress is taking place in this respect too, with novel indicators gaining speed, such as the Quality of Governance EU Regional and European Quality of Government Index (EQI) data, or the Sustainable Governance Indicators from the Bertelsmann Foundation containing indicators for policy performance. In this respect, the field continues to be indebted to developing economies, but advances in EU and OECD measurements are worth highlighting.

5.4 State Capacity for Different Stages of Development

While better capacity to travel in time-space is being achieved, the ambition of higher comparability should be balanced against the need to admit a certain degree of relativity in the concept of state capacity. This may be necessary in order to admit different state capacity visions for different stages of development. As Fortin (2010) points out, for example, a focus on fragility is helpful at the extreme of the spectrum where authority collapses, but of little everyday use for most polities (656). While countries like Sudan or Libya might be concerned with the centralization of coercive power, countries like Mexico or Argentina are more likely to focus on their fiscal or regulatory capacities. This might speak in favor of different “generations” of capacities, and therefore lower-level generalizations and empirical measures. This sides with the idea of Enriquez and Centeno (2012) and Williams (2018) that state capacity should be contextualized, although not to the point where the concept is diluted, losing comparative potential. This tradeoff has not been sufficiently brought to the fore.

5.5 The Observability Challenge

The overview of conceptualizations has shown that state capacity could well be understood as a system with different components. Certain scholars have focused on the presence of certain inputs believed to lead to the realization of the state’s potential, others preferred to focus on processes, while most of the literature focuses on whether the state’s outputs take the form of fulfilled goals. Focusing on any isolated component of capacity entails problems (Hanson and Sigman 2011; Lodge and Wegrich 2014). A focus on inputs does not necessarily guarantee a realization, as in the case of professional bureaucracies (Fukuyama 2013). Processes are difficult to crystallize in simple measures that enable valid inferences. Outputs may lead to circularity problems of the sort Kocher (2010) alludes to “Why is there high insurgency in this country? Because there is low state capacity. How do I know there is low state capacity? Because I observe high insurgency.” Similarly, Soifer (2008) highlights circularity problems in the measurement of state infrastructural power: “Scholars need to measure infrastructural power in a way that distinguishes power from its effects, and this is often difficult to do.” Moreover, “because we have no independent measure of the increased weight of the state apparatus from its effects, a convincing demonstration that the state has had an effect requires the elimination of alternative explanations for the observed effects” (247). Given these problems, the literature is left with the difficult task of measuring implementation capacity without easily distinguishing implementation capacity from all other external factors that determine implementation results (which are many). It follows that these circularity problems are particularly relevant in regard to empirical tests and causality, as when practice measures of capacity and performance conflate (Cárdenas 2010; Centeno et al. 2017).

5.6 Capacity to Travel Across Generalizability Levels

While the idea of state capacity likely operates at the highest level of abstraction, the literature talks about capacity at different levels, state being the most aggregated. The capacity concept has been extensively used in more specific public administration and public policy domains, in the form of policy capacity (Peters 2015; Wu, Ramesh, and Howlett 2015); administrative capacity (Cingolani 2019); managerial capacity (Andrews and Boyne 2011; Hammerschmid et al. 2014); connective, ambidextrous, and learning capacity (Gieske, van Buuren, and Bekkers 2016); among others. Hence, state capacity should allow the possibility of de-scalability while preserving some commonalities across the different realms. In the definition above, the idea of fulfillment of official goals conforms to this de-scalability principle, as official goals can range from those of a small organization (even a private one) to those operating at the national level. The connection to the institutions, rules, endowments, and procedures that realize capacity should, however, be adapted to the unit at hand. The dominant state capacity conceptualization converges, for example, with the definition of policy capacity: making policies perform as intended by the actors who designed them (Peters 2015). The dialogue between these different levels is not at all well established and deserves further efforts.

6 Concluding Remarks

The literature on state capacity has been insightful in explaining the historical determinants of today’s levels of development by dissecting the mechanisms by which central states manage to progressively enable higher welfare levels for their citizens, ranging from the provision of public order to redistribution and social policy.

There is extended empirical evidence suggesting that state capacity plays a critical role for development and it is therefore worth furthering all efforts to unpack and explore this broad and pluridimensional concept. Research on state capacity has fast flourished in recent years, aided by a series of lively debates among political scientists. These debates have helped more explicitly acknowledge the pitfalls of the state capacity literature and encouraged new initiatives to overcome them.

This article presents an overview of the trajectory of the state capacity concept and its uses in development studies, allowing a quick glance at the field’s diversity and trends. It has highlighted that while the variance in conceptualizations is high, there is a dominant stream that allows researchers to downplay the idea of concept stretching. This article, moreover, advances a definition of state capacity that integrates all elements of this dominant stream. It also highlights the recent innovations in the field of capacity measurement.

All in all, the article presents an optimistic view of the importance the concept can play in understanding the mechanisms behind development and that the field is active and constantly improving. Having partially addressed problems in the areas of comparability and construct validity, the concept continues to face challenges that should be addressed in a future agenda.

Notes

References

I am grateful to the organizers and participants of the Quality of Government Workshop on State Capacities (Gothenburg, June 2019) for the inspiring discussions on some of the topics presented here.