The Centrifugal State: In Search of an Explanation of Why Countries Decentralise

The nation-state, that proud invention of modernity, finds itself caught in a vise. As global forces have often pulled countries closer together in processes of economic and/or political integration (witness the European Union) or collaboration, internal pressures push them towards greater fragmentation. Since the 1970s a strong centrifugal force within the nation-state—manifesting as decentralisation—has become one of the defining features of contemporary governance, reshaping polities from Britain to Brazil, Spain to China. The phenomenon transcends the simple devolution of power; it represents a fundamental reimagining of how modern states organise themselves in response to evolving demands and pressures.

Why is this global drive towards decentralisation happening globally? And why is it happening now? Three grand narratives have traditionally explained this way of transfers of authority from central to subnational tiers of government: the democratic imperative, the pursuit of economic efficiency, and the persistence of regional identity. These arguments have not merely described the decentralisation wave but have actively shaped it, interweaving with concurrent developments in democratic expansion (Treisman 2007, Beramendi 2012, Faguet 2014), structural economic transformation (Rodríguez-Pose, Ezcurra 2010, Hooghe, Marks 2013), and the resurgence of regional consciousness (McGarry, O’Leary 2009, Keating 2013). Yet, beneath these explanations lies a more complex reality: the relationship between rhetoric and implementation of decentralisation remains persistently opaque, with appeals to democratic principles, public welfare, or ethnic identity often serving as convenient justifications for policies implemented for altogether different reasons.

The central question, then, is whether this global push towards decentralisation represents a considered response to changing governance needs or merely an expedient reaction to mounting pressures. This distinction matters precisely because the supposedly immutable division between federal and unitary states has given way to a more fluid reality, where degrees of centralisation fluctuate in response to a myriad of different forces (Tiebout 1956, Kyriacou, Roca-Sagalés 2011). Understanding these dynamics is crucial for both theoretical discourse and practical governance, as it sheds light not just how states evolve but why they choose particular paths of institutional transformation.

The quest to understand what drives the global phenomenon of decentralisation has produced not just competing answers but also divergent analytical frameworks. The complexity of the phenomenon demands multiple perspectives, with two approaches emerging particularly strongly. The first, grounded in welfare economics, scrutinises how different tiers of government achieve optimal service delivery, seeking explanations in variables such as economic scale and territorial expanse. The second, rooted in political economy, rejects such technocratic determinism. Instead, it views decentralisation through the prism of competing interests, where institutional design emerges from the complex interplay of various actors’ ambitions and constraints. Each framework offers distinct insights into the mechanics of governmental reorganisation, warranting careful examination of their respective contributions to our understanding.

In addition, the determinants of optimal decentralisation fall into two broad categories: static and dynamic. Static factors—including national size, population distribution (Panizza 1999, Treisman 2007), ethnic composition (Alesina et al. 2003, Hooghe et al. 2010), and constitutional traditions (Treisman 2007, Stegarescu 2009)—help explain why federal India differs so markedly from more centralised France. The static nature of these variables offers relatively convincing explanations for the geographic variations observed in initial levels of decentralisation across different national contexts. Dynamic factors, by contrast, encompass evolving conditions: evolving security threats (Beramendi 2012); widening regional disparities (Rodríguez-Pose, Ezcurra 2010); economic integration (Marks et al. 2008, Kyriacou, Roca-Sagalés 2011); and the rise of the knowledge economy (McCann, Ortega-Argilés 2015). These shifting, yet overlapping explanations help illustrate the time dimension of decentralisation, explaining not just why countries differ but how they change over time in the decentralisation process.

While existing literature extensively explores the motivations behind decentralisation—whether democratic aspirations, economic imperatives, or regional identity—these narratives are often too broad to capture the subtle but significant shifts occurring in the governance of modern states (Treisman 2002, Dardanelli et al. 2019). This article dissects the oft-overlooked factors that weave these broader forces into a fabric that varies strikingly across time and place. In the paper, I do not settle for the usual binaries but engage with decentralisation as a dynamic spectrum influenced by both static factors, like constitutional heritage and population size, and dynamic forces, such as economic integration, shifting political winds, and technological advancements (World Bank 2000, Fisman, Gatti 2002). By positioning these factors within a political economy framework, the discussion goes beyond the typical cause-effect analyses, suggesting a more layered interaction of factors that redefines how we understand state transformation.

To explain why decentralisation happens and happens in the way it does, I will first provide an overview of the global drive towards decentralisation. In the successive sections I lay out the theoretical and empirical foundations aiming to explain this phenomenon by examining both the static determinants and the dynamic influences that propel decentralisation in a variety of directions across different national contexts. The latter sections apply these insights within a political economy lens, dissecting how decentralisation is not only an outcome of welfare or efficiency considerations but also a calculated strategy shaped by political actors operating within complex power dynamics (Dardanelli et al. 2019). The conclusion synthesises these findings, challenging the reader to reconsider decentralisation not as a fixed endpoint but as an evolving dialogue between governance demands and the ambitions of those steering the state.

The world has become increasingly decentralised. Drawing from the authoritative data on decentralisation compiled by Shair-Rosenfield et al. (2021) and Hooghe et al. (2016), which covers 95 countries for the period between 1950 and 2018¹, we can paint a vivid picture of the extent to which power and resources have shifted from central governments to subnational tiers across the globe.

The numbers are eye-catching. The top 10 most decentralised countries in 2018, as measured by their Regional Authority Index (RAI) scores, span a diverse range of political systems and economic structures. Germany leads the pack with a RAI of 37.67, followed closely by Bosnia and Herzegovina at 36.34. India and Spain share the third spot with scores of 35.60, whilst Belgium rounds out the top five with 33.88 (Fig. 1).

Fig. 1.

The degree of decentralisation, measured by the Regional Authority Index in 2018, elaborated with data from Shair-Rosenfield et al. (2021) & Hooghe et al. (2016), created with Datawrapper.

The United States—often perceived as a paragon of federalism—ranks sixth with a score of 29.61. Pakistan, Canada, Switzerland, and Italy complete the top 10, with RAI scores ranging from 28.67 to 25.95. This list challenges preconceptions about which nations might be most decentralised, highlighting the complex interplay of historical, political, and economic factors that shape governance structures.

Perhaps most striking is the revelation that approximately 52.85% of the population in the represented countries lives in highly decentralised nations (those with an RAI score above 20). This statistic underscores the global shift towards more distributed forms of governance, a trend with profound implications for policy-making, economic development, and citizen engagement.

What explains this level of decentralisation? Table 1 looks at the population, GDP and GDP per capita means and medians for countries with high—defined as those with a RAI score of 25 or above—and low—below five in the RAI— degrees of decentralisation. It paints a picture where countries with higher RAI scores tend to boast larger populations, greater GDP, and higher per capita income. However, this relationship is far from straightforward, as evidenced by the considerable variations among similar countries in their levels of decentralisation.

For instance, the mean population for countries with RAI scores between 0 and 5 in 2018 was a modest 936,621, compared to 32,100,000 for those with RAI scores above 25. The median population was 390,353 and 25 million, respectively. Similarly, the average GDP for the less decentralised group stood at $7317.6 million, whilst their more decentralised counterparts boasted a mean GDP of $1,529,600 million. The disparity in GDP per capita is equally acute, with figures of $12,340.72 and $40,207.92 for the low and high RAI groups, respectively. The gap in wealth between the median highly decentralised and highly centralised country was given greater, as highly centralised countries had a median GDP per capita of just above $5300.

Yet, these averages mask significant outliers and variations. The group of highly centralised countries includes both small island nations (e.g. Bahamas, Barbados, Iceland, Malta) and larger developing (e.g. Cambodia, El Salvador, Guatemala, Laos, Mongolia) and developed (e.g. Estonia, Lithuania, Luxembourg, Singapore, Slovenia) economies. Thailand is the largest country by population in this group. The decentralised cohort spans federal giants like India and the United States alongside smaller, but highly devolved states like Belgium or Switzerland.

Figure 2 offers a compelling visualisation of decentralisation’s evolution over time. After a worldwide decline in decentralisation between 1950 and 1970, the median RAI score across the sample countries considered started a growth trajectory that has continued until the present. The gradual and gentle increase in transfers of powers to subnational tiers of government of the 1970s and 1980s turned into a surge between 1990 and 2000, coinciding with the end of the Cold War, the fall of the Berlin Wall, and the subsequent rush towards market economies. The global wave of democratisation and economic liberalisation in developing and emerging countries was also a powerful source of decentralisation processes (Rodríguez-Pose, Ezcurra 2010). Since then, the pace of transformation has eased, but continues to move upwards.

Fig. 2.

Evolution of the median level of decentralisation in the sample countries between 1950 and 2018, elaborated with data from Shair-Rosenfield et al. (2021) and Hooghe et al. (2016).

Figure 3, plotting countries’ initial RAI scores in 1990 against their change in RAI 28 years later, by 2018, further complicates the picture. The scattered distribution and near-flat regression line indicate that a country’s starting point in 1990 bears little relation to its subsequent trajectory of decentralisation. This finding undermines any notion of path dependency in decentralisation processes and highlights the importance of country-specific factors in shaping governance structures.

Fig. 3.

Changes in decentralisation relative to a country’s starting point on the decentralisation scale (1990–2018).

As we ponder these trends, while the data shows that we are living in an increasingly centrifugal world, it becomes clear that decentralisation is not a one-size-fits-all phenomenon. The trends raise questions about the drivers of decentralisation. Is it primarily a response to economic imperatives, a reflection of changing political ideologies, or perhaps a combination of both? The data suggests a complex interplay of factors, challenging simplistic explanations and inviting deeper analysis. Its causes and consequences vary widely across different contexts, influenced by a country’s history, political culture, economic structure, and myriad other factors. The data presented here offers a springboard for further investigation, inviting scholars and policymakers alike to investigate the differences and nuances of decentralisation and its implications for governance in the 21st century.

Many theories have been used to try to explain why some countries decentralise in the way they do and why others decentralise much less or not at all.

Perhaps the dominant theory has been the ‘decentralisation theorem’ (Oates 1972, 2005). This theory represents a distillation of earlier insights of the so-called ‘efficiency approach’ proposed by Samuelson (1954, 1955), Tiebout (1956), Musgrave (1959), and Arrow (1969). This framework matches governmental competencies to their optimal territorial scale. It weighs the economies of scale inherent in centralisation against the benefits of tailoring services to local preferences, offering a sophisticated calculus for determining the optimal distribution of governmental powers and to increase efficiency in service delivery and, ultimately, in economic performance.

The theorem’s logic is as persuasive as it is straightforward. Some governmental functions—national defence being the archetypal example—benefit from economies of scale that make centralisation inevitable. Others, such as waste collection or employment training, gain more from local customisation than they lose in efficiency. The theorem posits that, under specific conditions, varied local provision proves Pareto superior to uniform central delivery across all jurisdictions (Oates 2005).

While traditionally cast as normative rather than explanatory, the theorem assumes new significance when viewed through the lens of voter and investor preferences (Rodden, Wibbels 2010) and policymakers’ pursuit of public welfare (Weingast 2014). This reframing transforms it from mere prescription to explanation, suggesting that decentralisation occurs when either local services command a larger share of national resources or when previously centralised functions prove better suited to local delivery.

However, the decentralisation theorem, for all its intellectual sophistication, often falters on the jagged rocks of political reality. Its critics take particular aim at its central conceit: that decisions about governance levels emerge from dispassionate cost-benefit analyses rather than the murky waters of self- or party-political interest. The underlying assumption that politicians, parties, and pressure groups act as dutiful servants of optimal policy design remains a highly challenged one, particularly in places with weak institutions. Indeed, Rodden (2006) dismisses the idea that decentralisation flows naturally from welfare economics or public choice systems as fundamentally divorced from observable patterns. Whether sparked by the tides of globalisation or the winds of democratisation, the final shape of decentralisation emerges from an intricate negotiation among politicians, bureaucrats, and interest groups. As Garman et al. (2001: 209) observe, understanding this choreography demands intimate knowledge of the actors’ preferences, capabilities, and the institutional stage upon which they perform.

Treisman (2007) strikes at the heart of the matter: the theoretical efficiency of different decentralisation levels bears little relation to policymakers’ actual decisions. He notes that “few economists—and still fewer political scientists—believe that actual arrangements are always the most efficient ones. So even if fiscal decentralisation is more efficient in ethnically divided societies, that does not imply that ethnically divided societies would tend to be more fiscally decentralised” (2007: 15).

Even Oates (2005), the theorem’s architect, already bent to these winds of criticism, acknowledging the need for a ‘second-generation’ theory of fiscal federalism. Yet the marriage of economic determinism and political contingency remains an uneasy one. The universal principles of efficiency maximisation sit uncomfortably alongside the idiosyncratic nature of political reality. While economists pursue the holy grail of universal causation through cross-country analyses, political scientists, geographers, and historians adopt what might be seen as a more down-to-earth view when dealing with decentralised authority. However, they often remain anchored in the granular details of specific cases, creating a methodological divide as deep as it is persistent and which has significant implications for the capacity of different theories to explain the complex reality on the ground.

Decentralisation is, in reality, mostly driven by a constellation of very diverse factors, which intermingle in a web of unique relationships in every space that gives the subject its variety and richness. Two crucial observations merit attention. First, decentralisation’s various forms likely spring from different sources. This explains apparent paradoxes, such as the simultaneous flowering of nation-states in interwar years of the 20th century, as well as after the collapse of the Berlin Wall (Alesina, Spolaore 2003) alongside the centralising forces of the welfare state. Similarly, we witness the spectacle of policy responsibility flowing outward while control, through targets and monitoring, tightens at the centre.

Second, centralisation and decentralisation coexist in an eternal dance, their steps traced across countries, regions, and policy domains. The decentralisation pendulum keeps on swinging constantly. While recent decades have seen decentralisation on the ascendancy across most parts of the world, other periods have bowed to centralising forces. As Blankart (2000: 32) suggests, decentralisation often emerges as a reaction to ‘over-centralisation’, making it impossible to understand one without the other. It also arises when the ‘centralised state’ faces severe crises and decentralisation is viewed as a potential, often desperate solution for states that are on the brink of collapse. In essence, centralisation and decentralisation are two faces of the same political coin.

In the following sections, I examine the static and dynamic determinants of decentralisation, picking Oates’ (1972) decentralisation theorem as the starting point to steer through different political economy explanations before exploring historically specific causes. This implies grappling with the tension between case studies and quantitative methodologies, seeking harmony between these seemingly discordant approaches to understanding decentralisation’s origins.

In the taxonomy of causation, determinants reveal themselves in both static and dynamic forms, like photographs and films of the same landscape. Wealth, territorial expanse, population, and ethnic composition present themselves as the static counterparts to growth, shifting boundaries, demographic change, and cultural evolution. This distinction informs about the dichotomy between spatial variation and temporal transformation. Before I chart how the currents of economic, political, and technological change shape decentralisation’s evolution, we must first understand how their static manifestations position nations along the spectrum of decentralisation. Some indicators, like territorial dimensions, resist rapid change and thus speak more eloquently in their static form. Others, such as wealth and growth, find their voice in the language of dynamic transformation.

Constitutions are the architectural blueprints of governance. However, most of them change at a glacial pace and thus stand as static pillars of decentralisation (Escobar-Lemmon 2003). One might assume that federal states like the United States would invariably demonstrate greater decentralisation than unitary ones such as Spain, Greece, or Hungary. After all, federal constitutions enshrine subnational autonomy beyond the reach of governmental whim (Treisman 2007). Yet reality proves more nuanced. Two complications muddy these constitutional waters. First, the binary opposition of federal and unitary systems dissolves into a continuum upon closer inspection. As Keating (2013: 60) observes, “the distinction between federalism and the stronger forms of regionalism is becoming ever more difficult to make.” Second, even this spectrum proves an unreliable guide. Germany, despite its federal framework, can, in some areas, display greater centralisation than France, often viewed as an archetype of unitary governance. Yet, France has spent most of this century cautiously and almost stealthily embracing greater decentralisation (Jeffery 2015). Nevertheless, empirical investigation confirms that federal constitutions generally herald higher levels of decentralisation, though Treisman (2007) suggests this correlation might better be understood through the prism of economic development.

Colonial heritage, too, casts long shadows across modern governance structures. British colonial administration, with its tradition of local autonomy and institutional architecture, planted seeds of decentralisation that outlived the empire itself (Hadiz 2010). In contrast, some of Spain’s Latin American domains remain monuments to centralisation. Even in federal states, such as Argentina, Mexico, or Venezuela, their federal constitutions have oftentimes been mere facades obscuring deeply centralised realities (Falleti 2010). Treisman’s (2007) analysis suggests that colonial legacy, alongside territorial scale, explains much of the variation in contemporary decentralisation, though this relationship may be part of a more intricate universe in which economic development holds the dominant thread (Rodríguez-Pose, Sandall 2008).

The decentralisation theorem’s concept of ‘preference heterogeneity’ provides another lens through which to view static determinants. The theory of fiscal federalism suggests that decentralisation flourishes where voter preferences regarding taxation and service provision show marked regional variation. Assuming policy efficiency influences implementation, greater preference heterogeneity should herald more decentralised structures (Hooghe et al. 2010). This framework invites us to examine how territorial scale, population density, ethnic composition, and regional income disparities shape decentralisation processes across the world.

Larger nations, by virtue of their scale and diversity, naturally incline toward decentralisation. Beyond the simple arithmetic of diversity, vast territories present practical challenges to centralisation, where distance itself becomes a barrier to unified governance (Lessmann, Markwardt 2010). Empirical evidence from Panizza (1999), Arzaghi and Henderson (2005), and Treisman (2007) supports this geographical determinism. Population density writes its own chapter in this story: sparse populations may demand decentralisation when distance renders central control impractical (Letelier-Saavedra, Sáez-Lozano 2015), while concentrated urban populations can achieve the economies of scale that make local governance viable (Arzaghi, Henderson 2005).

The relationship of ethnic diversity with decentralisation proves more complex than first appearances suggest. Spain’s mosaic of Catalan, Basque, and Galician identities might seem to explain its greater decentralisation compared to more homogeneous France (Keating 2013). However, empirical evidence paints a more ambiguous picture. Some states, facing secessionist pressures, employ decentralisation as a pressure valve for ethnic tensions (Brancati 2009). Yet, Treisman (2007) finds ethnic fractionalisation sometimes correlates with centralisation, perhaps because it impedes economic development. Arzaghi and Henderson (2005) suggest that ethnic diversity might actually encourage governments to restrict decentralisation to local levels, maintaining control over potentially restive minorities and as a way to potentially—although not always successfully—limit separatist tensions from emerging.

The relationship between wealth, inequality, and decentralisation also resists simple characterisation. Regional income disparities, while theoretically conducive to decentralisation by creating divergent preferences (Lessmann 2012), tend to strengthen the centre’s redistributive role (Kyriacou, Roca-Sagalés 2011). Wealthier regions frequently naturally seek autonomy to retain their resources, yet national prosperity presents a paradox: it might fund ambitious national projects (Letelier-Saavedra, Sáez-Lozano 2015) or empower citizens to demand greater local control (Treisman 2007). Empirical investigations by Panizza (1999) and others generally support a positive relationship between national wealth and decentralisation (Letelier-Saavedra, Sáez-Lozano 2015).

In brief, the evidence speaks a bit more clearly on certain points about how static factors determine the drive towards decentralisation: larger, more populous nations and wealthier nations (Table 1), as well as those with federal constitutions, tend towards transferring more power and, to a lesser extent, resources towards subnational tiers of government (Rodríguez-Pose, Vidal-Bover 2024). Other static factors— ethnic diversity, inequality, and wealth—tell more complex tales, their influence mediated by context and circumstance. Thus far, the causative elements of fiscal federalism theory find qualified support in the data, though reality proves, as ever, more nuanced than theory might suggest.

The march toward decentralisation in recent decades cannot, however, be simply understood by looking at the static factors. Alongside those static factors lie a whole range of dynamic determinants of decentralisation, which are often the cause but also the consequence of variations in the allocation of powers and resources between different tiers of government.

Two of those dynamic factors have been globalisation (Ezcurra, Rodríguez-Pose 2013) and democratisation (Pius Kulipossa 2004, Enikolopov, Zhuravskaya 2007), twin forces which have profoundly reshaped the political scenery. Decentralisation has also kept company with significant variations in economic growth, widening inequality, and the metamorphosis of industrial economies into knowledge-based ones (Rodríguez-Pose, Ezcurra 2010, 2011, Filippetti, Sacchi 2016, Ligthart, Van Oudheusden 2017). Add to this pageant the rise of cities, the emergence of supranational entities, and the novel policy challenges of our age, and one begins to glimpse the complexity of modern governance. Each thread in this weave demands careful examination, for in their interplay lies the story of how nations redistribute power.

The scholarly examination of decentralisation’s causes, while often focused on discrete factors, invariably kneels before economic integration (e.g. Rodríguez-Pose, Gill 2005, Kyriacou, Roca-Sagalés 2011, Beramendi 2012, Hooghe, Marks 2013). The basic arithmetic of this relationship appears straightforward: international market integration has rendered smallness in government less costly (Rodrik 2011). After all, optimal national scale is marked by a delicate equilibrium between administrative overhead and the economic integration. As past economic integration during the globalisation period dissolved trade barriers, smaller administrative units shed the bureaucratic burden of large nations without sacrificing commercial advantage. This theory, however, like many elegant theories, faces complications in the messy laboratory of reality.

The EU stands as perhaps the most forceful theatre for testing these hypotheses. Here, the dismantling of trade barriers linked to European integration has proceeded at unprecedented speed and scale, offering a natural experiment for changes in regional autonomy. Hooghe and Marks (2016) paint a picture of regions empowered, through a system of multilevel governance, by direct access to supranational institutions and resources. These regions, far from passive observers, have seized their moment, establishing Brussels outposts and forging alliances under banners like the ‘Europe of the Regions’ (Keating 2013). Their entrepreneurial spirit in carving out political space suggests a deep connection between economic integration and regional empowerment.

Yet empirical evidence proves stubbornly ambiguous. While some researchers uncover positive correlations between economic integration and decentralisation (Stegarescu 2009, Kyriacou, Roca-Sagalés 2011), others suggest globalisation’s risks demand stronger central coordination (Rodrik 2011). Diaz-Cayeros (2006) notably found trade openness walking hand in hand with centralisation, arguing that global market volatility requires steady central stewardship. The contradiction speaks to the complexity of modern governance, where competing imperatives often pull in opposite directions.

Democratisation, the second great force under examination, proves equally enigmatic. The temporal symmetry is striking: the World Bank (2012) documents democracy’s dramatic expansion between 1974 and 2012, a period coinciding with decentralisation’s global ascendancy. This democratic wave coincided with decentralisation in Spain (Colino 2013), Eastern Europe (Fritz et al. 2009), and Latin America (Falleti 2010), where centralised autocracies gave way to more distributed democratic systems. Yet, exceptions abound: Treisman (2007) notes Estonia’s post-democratic centralisation contrasting with Spain’s decentralisation, while Zimbabwe and Belarus chart opposite courses in their relationship with autocracy. These contradictions suggest that democracy’s relationship with decentralisation, while significant, resists simple characterisation.

The theoretical marriage of democratisation and decentralisation finds its strongest expression in fiscal federalism. Indeed, preference heterogeneity—the framework’s milestone—achieves its fullest meaning under democracy, where regional voter preferences can find authentic expression in policy. Without functioning democratic mechanisms, the varied wishes of different regions remain mere whispers in the wind. Faguet (2014) sees decentralisation as both democracy’s handmaiden and redemption song, offering central governments a path to restored legitimacy and citizens a voice in their local affairs.

A more prosaic explanation emerges from regional dynamics of resistance. Provincial areas, often home to those out of favour with authoritarian regimes, may prove more fertile ground for anti-dictatorial mobilisation than capitals, which tend to attract regime loyalists. Victory achieved, regional elites naturally seek to cement their influence through decentralised democratic constitutions (Brancati 2009). This is a pattern that was particularly in evidence in Brazil’s experience with decentralisation and centralisation (Montero 2007). This reading suggests decentralisation as less a product of democratic theory than of political pragmatism.

While some econometric studies, notably Enikolopov and Zhuravskaya (2007), support democracy’s role in fostering decentralisation, the direction of causation remains contested. Treisman’s (2007) finding of negligible links between democracy and decentralisation stands as a provocative challenge to conventional wisdom. The question of whether democracy breeds decentralisation or vice versa remains one of social science’s more intriguing chicken-and-egg problems.

The relationship of economic growth with decentralisation has primarily been examined in reverse. Most research focuses on decentralisation’s impact on growth. Analyses by Rodríguez-Pose and Ezcurra (2010), Lessmann (2012), and Letelier-Saavedra and Sáez-Lozano (2015) have begun to reverse this lens. The theoretical arguments echo earlier debates about wealth levels: growth might fuel centralisation through expanded welfare spending, or spark decentralisation by educating and empowering local electorates. Empirical studies generally favour the latter interpretation (e.g. Rodríguez-Pose, Ezcurra 2010, Lessmann, Markwardt 2010), though the picture remains far from definitive.

One might resolve this theoretical tension by examining modern economic transformations. The new economic geography highlights regions’ evolving economic roles, particularly through ‘untraded interdependencies’ (Storper 1997). As industries cluster to harvest proximity benefits, growth poles emerge, their distinct economic identities fuelling calls for regional autonomy (Keating 2013). This more nuanced reading suggests that perhaps it is not growth per se that drives decentralisation, but rather the particular character of modern economic development, with its emphasis on knowledge clusters and regional specialisation.

This exploration of decentralisation’s dynamic drivers yields even less certainty than the earlier examination of static determinants. Integration, growth, and democratisation all present theoretical and empirical ambiguities that resist neat resolution. Yet, this might speak more to the challenge of tracking causation through time than to any fundamental weakness in the relationships. While these forces have undoubtedly shaped decentralisation’s recent trajectory, their precise mechanisms and relative importance remain tantalisingly elusive. The story of how nations redistribute power between different tiers of government, it seems, is written not in bold strokes but in subtle shades of grey.

To try to solve the riddle of what has triggered the global drive towards decentralisation, we also need to look elsewhere for other potential, complementary explanations. Political economy and history are, in particular, two fertile grounds where research on decentralisation has had less to say. Oates (2005) was already conscious that his fiscal federalism theory harboured a blind spot: its silence on the political machinations that may be an important driver of decentralisation.

To investigate this shadowy area of existing knowledge on decentralisation, one must pierce the veil of institutional arrangements and examine how the ambitions of politicians and other political and societal actors shape the destiny of devolved power. Some patterns emerge with the reliability of seasons—electoral cycles chief among them—while others prove as mercurial as local political cultures themselves. What binds these disparate threads is the notion that decentralisation policy emerges from an intricate interaction between central and local powers, a perpetual negotiation whose steps we are only beginning to map. In the theatre of decentralisation, political actors often seek to advance their interests while maintaining the delicate balance of governance.

Riker’s (1964) theory of federalism stands as perhaps the clearest exposition of decentralisation’s political arithmetic. His framework divides the political world into two camps. In one realm, party members and voters select regional leaders who then anoint their national leader. In the other, they choose the national ‘sovereign’ who then dispenses regional authority like feudal titles. The former breeds decentralisation, while the latter tends toward centralisation. This clean bifurcation flows from the relative heft of central and regional politicians in each scenario. When national leaders owe their position to regional barons, power and resources trickle outward; when, in contrast, regional authorities serve at the pleasure of the centre, their ability to demand autonomy withers (Rodríguez-Pose, Gill 2003).

The left-right divide can also be cast as a crucial determinant in this intricate political calculus. Left-wing parties are often wedded to the twin imperatives of macroeconomic management and redistribution, naturally gravitate toward centralisation. The right, by contrast, sees in decentralisation a mirror of its own principles: privatisation and liberation from state oversight. Yet political reality, with its fondness for irony, offers plenty of evidence of counter-narratives. In the British theatre, it was Labour that took up the standard of decentralisation, while the Conservatives—those supposed champions of devolved power—mostly have stood athwart the path of change.

The ebb and flow of electoral fortunes adds another layer to the complexity of the determinants of decentralisation. Keating (2013) observes that the ascension of a new government, particularly one that has built its power base in the periphery before conquering the centre, can create a brief window for radical reform. Yet power at the centre proves an intoxicating draught; unless change is swift, the appetite for decentralisation often fades. Opposition parties, meanwhile, finding themselves locked out of central authority, naturally seek to carve out influence in the regional and local spheres, those smaller stages where they might still command an audience.

The political economy of decentralisation manifests in myriad other forms, each adding its own hue to this rich political canvas. Modern governance faces challenges that seem tailor-made for local solutions—environmental concerns and social services among them—as John (2014) notes. The waning of national, class-based politics, coupled with the growing alignment between local business interests and regional politicians, further tilts the balance. Meanwhile, the Leviathan theory of government (Brennan, Buchanan 1980) offers a lens through which to view the eternal struggle between tiers of government, each seeking to maximise not public welfare but their own power and resources.

These insights from political economy fill a crucial void in fiscal federalism’s theoretical framework. They bring some light into those puzzling cases where the distribution of policy competencies appears to defy the conventional wisdom about economies of scale and preference heterogeneity, revealing the hidden hands that truly shape the architecture of governance. However, as their predecessors, political economy alone fails to provide a full proof theory of why decentralisation has become much more prominent and about its rise of recent.

As we have seen, explaining why decentralisation takes place—and why it is taking place at the rate it is—is proving far more intricate than any single theoretical framework might suggest. While we have dwelt upon the universal forces that shift the equilibria of optimal devolution, the contemporary drift toward decentralisation spans merely a few decades. Before this period, we witnessed simultaneous yet contradictory trends: centralisation as a corollary of the emerging welfare states, while empires’ slow dissolution yielded its own form of devolution. Such paradoxes mock any notion of immutable relationships between decentralisation and its supposed determinants. To untangle this knot, we must acknowledge a final category of cause: those singular moments in history that have shaped the landscape of governance.

The German narrative offers a particularly stark illustration of history’s heavy hand. The twin catastrophes of world wars carved deep channels of decentralisation, while the Nazi Party’s ascendance triggered a violent centralising spasm. These seismic shifts bear little kinship to the theoretical frameworks outlined earlier, yet their impact stands both undeniable and profound. From such cases emerge certain bitter truths: the drum-beat of war—or merely its distant echo—invariably summons centralising forces, as defence budgets swell and nationalist sentiment drowns out regional voices (Beramendi 2012).

Financial tempests, too, can serve as agents of transformation. Rodríguez-Pose and Gill (2005) trace how the economic storms of the 1970s birthed the centralising impulses of Thatcher’s Britain, as local authorities found their wings clipped. Yet one must resist the temptation to cast every historical moment as unique and unrepeatable. While election victories, supranational alliances, or fiscal crises may appear as singular catalysts of change, they often serve merely as visible markers in a deeper process. Like earthquakes that merely manifest deeper tectonic shifts, changes in government often simply crystallise gradual processes that align comfortably with fiscal federalism’s theoretical architecture.

Panizza (1999: 119–122) notes how history’s apparent impulses have led many scholars to despair of applying economic theory to cross-national variations in power distribution. The common wisdom holds that the backstage political theatre determining policy defies prediction and, therefore, theoretical capture. Yet Panizza hypothesises that, to test history’s role empirically, if fiscal federalism suggests an optimal level of decentralisation for each nation—one that democracy and rational choice should naturally seek—then deviations from this ideal must represent history’s interference: those unpredictable outcomes born of accident and personality. These ‘historical errors’ gradually yield to democratic rational choice’s corrective force, though not without resistance from political reality’s stubborn facts.

These theoretical frameworks stand alone in their ability to weave together econometric precision and historical narrative, harmoniously accommodating both universal and particular causes within a form of understanding. They offer a rare bridge between the quantifiable certainties of economic theory and the messy realities of historical circumstance.

As the twentieth century dawned, many researchers watched modernity’s march and concluded that centralisation was its inevitable companion. ‘Popitz’s law’ (Popitz 1927, 1932)—named for a Prussian finance minister of the 1930s—was formulated from observing the gradual centralisation of Switzerland, the US, Argentina, and Canada. Yet the latter half of the 20th century and the first quarter of the 21st century have witnessed a decentralising tide that would mock any such claims to timeless truth. Across the world we are witnessing the emergence of centrifugal states.

In this paper I have searched for an explanation of why countries decentralise and why they decentralise when they do. The tension between general and specific explanations remains central to this inquiry, a theoretical Gordian knot that continues to challenge our understanding. While this fundamental discord remains far from being resolved, I have attempted to impose order by categorising causes and examining their elaborate interplay. Indeed, the discussion presented in this article suggests that even modest generalisations about decentralisation stumble upon case-by-case variation, like universal rules shattering against the rocks of specific circumstances. Though I can offer no new empirical insights, I have brought together diverse indicators rarely considered in concert, creating a more nuanced picture of decentralisation’s complex reality and honest drivers.

Overall, the accumulated research points to constitutional federalism and country size as decentralisation’s most reliable predictors. Time’s influence proves more elusive, as even economic growth’s supposedly central role as a driver of decentralisation faces both theoretical and empirical challenges. Further complexity emerges from the causal web linking the major trends of recent decades: globalisation, growth, structural change, inequality, and democracy. Their temporal coincidence has made it tempting to select a few and build empirical cases for their primacy. Yet the persistence of contradictory cases, periods, and hypotheses demands intellectual humility.

Given the challenges of data coverage, reliability, and significance, coupled with the theoretical complexity of reconciling general and specific casual systems, the myriad of studies and the drivers of decentralisation can at best partially explain how decentralisation unravels and why it has taken so much of a hold in this historical period. Hence, researchers interested in decentralisation processes and their causes and consequences must continue to fill in blank spaces in existing knowledge while knowing full well that they will be uncapable to touch and adequately explain why decentralisation takes hold in some places and not in others. While this yields valuable insights and reliable conclusions, we remain far from complete understanding decentralisation. We must therefore continue with this type of research, as each study will add another important piece of this intricate puzzle of governance and power distribution.