De/Centralisation in Federations

Paolo Dardanelli
Paolo Dardanelli is Reader in Comparative Politics at the University of Kent. His research explores the causes and consequences of the territorial structure of political systems, from a comparative perspective across space and time. More information on his work can be found at and by following him on Twitter @PaoloDardanelli.


How powers are distributed between the federal government and the constituent units of a federation, or de/centralisation, is at the heart of federalism. There has recently been renewed interest in studying de/centralisation in federations, with several works addressing conceptualisation, measurement, theorisation, and causal analysis. This piece takes stock of this literature and discusses its contributions.



In essence, federalism is a constitutional device to divide powers between at least two levels of government. How powers are distributed and how the distribution changes over time is thus a, if not the, crucial question in the study of federalism. As Livingston (1956: 10) observed, “That the real key to the nature of the federation is in the distribution of powers seems to be agreed upon by nearly every writer who addresses himself to the question.”

Such a distribution is generally approached in terms of de/centralisation, whereby a system that attributes more power to the central government than to the constituent units is considered more centralised – or less decentralised – than a system in which the bulk of power belongs to the constituent units. Different forms and degrees of de/centralisation between federations and changes over time matter because they affect federations’ ability to perform effectively and efficiently, and deliver the purported benefits of a federal political system.


Conceptualising De/Centralisation in Federations

How has de/centralisation in federations been conceptualised? Three questions have attracted most debate. The first question, which touches upon the very essence of federalism, concerns whether federations can be distinguished from unitary systems by merely their degree of de/centralisation. It has been answered in widely divergent ways. Kelsen ([1945] 1961: 316) stated: “Only the degree of decentralization distinguishes a unitary State divided into autonomous provinces from a federal State.” King forcefully rejected this conceptualization, claiming instead that the defining feature of a federation is the “representation of regional units in the national legislature”, not the degree of de/centralisation (King 1982: 19, also 77, 146). Elazar (1971: 98-9) also rejected conflating federalism and ‘decentralisation’ but on the ground that federations, notably the United States, are characterized by noncentralisation rather than ‘decentralisation’. From a Kelsenian perspective, Dardanelli (2019) argued that the federal form of state can be conceptualized as a section on a de/centralisation continuum, marked by differences of kind, rather than merely of degree, with unitary states.

The second question is whether de/centralisation should be seen as a dichotomy or a continuum. Although this question has rarely been addressed explicitly, one conceptualisation or the other tends to be implicitly assumed. Kelsen’s ([1945] 1961: 316) view of de/centralisation as a continuum transpires clearly from his formulation of the essence of federalism. Elazar’s (1971: 98-9) notion of noncentralisation, by contrast, suggests a dichotomous conception. Dardanelli (2019), as seen, explicitly conceptualises it as a continuum divided by thresholds.

The third question concerns how to conceptualise de/centralisation itself. As shown by reviews such as Dubois and Fattore (2009), authors have conceptualized de/centralisation in a range of different ways. Although a degree of convergence has emerged, one important aspect is still unsettled: whether ‘shared rule’ is a dimension of de/centralisation. Hooghe et al. (2016: 19) treat ‘shared rule’ – i.e. the degree to which constituent units have a say in decision-making at the centre – as one of the two broad dimensions of de/centralisation, the other being ‘self rule’, i.e. “the authority that a subnational government exercises in its own territory.” (Hooghe et al. 2016: 23). By contrast, Dardanelli (2019) argues that ‘shared rule’ – unless bilateral (that is, enjoyed by an individual constituent unit in a bilateral relationship with the centre) – is not a dimension of de/centralisation; it should be treated as a separate concept and measured accordingly.


Measuring De/Centralisation

While studying de/centralisation dynamics has always been prominent in the scholarship on federalism, measuring it accurately and validly across federations and over time has long faced methodological and empirical problems. As a result, de/centralisation in federations had not been, until recently, systematically investigated across its different dimensions. Qualitative works, on the one hand, studied two or three cases (e.g., Esman 1984; Chhibber and Kollman 2004; Döring and Schnellenbach 2011). Quantitative scholars, on the other hand, covered a larger number of cases but relied solely on fiscal variables such as the proportion of total government revenues or expenditures accounted for by the constituent units as proxies for de/centralisation (e.g., Pommerehne 1977; Krane [1982] 1988). More recently, the Regional Authority Index (Hooghe et al. 2016) measured the authority regional governments – including the constituent units of federations – possess in terms of “self rule” (i.e., the powers they exercise vis-à-vis their own population) and “shared rule” (i.e., their power to influence national policies), but did not fully capture important de/centralisation dynamics in federal systems occurring in the policy and fiscal spheres.


The De/Centralisation Dataset (DcD)

The De/Centralisation Dataset (Dardanelli et al. 2019c) measures legislative and administrative de/centralisation in 22 policy fields and fiscal de/centralisation in five categories in Australia, Canada, Germany, India, Switzerland, and the United States for each decade from their foundation to 2010.

The DcD distinguishes between static and dynamic de/centralisation. Static de/centralisation is the distribution of powers between the central and constituent governments of a federation at any given time. Dynamic de/centralisation is the change in the distribution over time, toward either centralisation or decentralisation. Constituent-unit autonomy in continuously democratic federations is conceptualised as having two main dimensions: policy, itself divided into legislative and administrative, and fiscal autonomy. Legislative autonomy refers to a constituent unit’s control of primary legislative powers in a policy field. Administrative autonomy concerns the degree to which a constituent unit implements central government, as well as its own, legislation. Fiscal autonomy relates to its ability to obtain financial resources through its own tax and borrowing powers, and to allocate such resources as it pleases (Dardanelli et al. 2019a, 7-10).

Dynamic de/centralisation is conceptualized as having five main properties: (1) direction: whether change is toward centralisation or decentralisation; (2) magnitude: the magnitude of the change; (3) tempo: the frequency, pace, timing and sequence of change; (4) form: whether change occurs in the legislative, administrative or fiscal dimensions; and (5) instruments: the instruments through which change occurs, such as constitutional amendment, court rulings or conditional grants (Dardanelli et al. 2019a, 10-13). Policy de/centralisation is measured on 7-point scales, ranging from 7 (exclusive control by each constituent unit) to 1 (exclusive control by the federal government). Fiscal de/centralisation is measured on 7-point scales based on numerical indicators or qualitative assessment, ranging from 7 (maximal autonomy for each constituent unit) to 1 (minimal autonomy).


Dynamic De/Centralisation across Federations

The DcD data and the publications based on them (Dardanelli et al. 2019a,b) help us map dynamic de/centralisation across federations. The central finding is that most democratic federations have become more centralised over time, though neither uniformly nor invariably so. Centralisation has occurred primarily in the legislative sphere but less so in the administrative and fiscal spheres. Canada has followed a distinct path; it was significantly more centralised at the outset in 1867 but experienced a mix of centralisation and decentralisation over time, resulting in little net change overall. Given that the other federations have experienced considerable centralisation, Canada is once again an outlier today in being significantly less centralised than the other major democratic federations. In contrast with its peers, Canada has also experienced growing asymmetry whereby some (French-speaking Quebec, in particular) but not all provinces have developed their own policies in several fields, ranging from pensions to immigration. Centralisation has been less deep in the administrative sphere, particularly in Germany and Switzerland, thus leading to greater interdependence between levels in policy legislation and administration. In other words, most federations, notably the United States, have become much less ‘dual’ and much more ‘administrative’ in nature. There has also been much less centralisation in the fiscal sphere, particularly in Canada, Switzerland, and the United States, which underscores the limitations of using fiscal data to capture de/centralisation dynamics.

Dynamic de/centralisation is determined by a complex interaction of factors operating in a manner reminiscent of a “funnel of causality” (Campbell et al. 1960, 24-32). Broadly, socio-economic and socio-cultural change, occasionally reinforced by economic and security shocks, produce pressures in most federations to expand the scope and reach of the central government at the expense of the autonomy of the constituent units. These largely common forces interact, however, with the widely different structural features of each federation and are thus refracted in different ways in different contexts. Prominent among those structural features are the degree of economic integration and the relative strength of citizen identification with the constituent units compared with the federation as a whole. These interactions shape collective attitudes towards the federal balance and generate incentives and/or constraints on political actors. High economic integration and strong identification with the federation tend to foster centralisation; where these conditions are weaker, public attitudes tend to resist centralisation and even favour decentralisation. Political actors react to these incentives and/or constraints within the institutional framework of each federation (Dardanelli et al. 2019b, 16).



Measuring changes in the distribution of powers between the central government and the constituent governments of a federation over time is crucial to the comparative study of federalism. The De/Centralisation Dataset is a recent attempt to do so across federations and over time. Compared to other datasets, the DcD offers three main advantages: (a) detailed measures of legislative and administrative de/centralisation for 22 public policy fields, ranging from agriculture to transportation; (b) measures that capture the fiscal autonomy of the constituent units as opposed to their fiscal capacity; and (c) measures for the entire life of each federation. It can thus be a stepping stone to further research on many aspects of de/centralisation in federations.


Suggested Dardanelli, P. 2020. ‘De/Centralisation in Federations’. 50 Shades of Federalism. Available at:



Campbell, A., P. Converse, W. Miller and D. Stokes. 1960. The American Voter. New York, NY, USA: John Wiley.

Chhibber, P. and K. Kollman. 2004. The Formation of National Party Systems – Federalism and Party Competition in Canada, Great Britain, India, and the United States. Princeton, NJ, USA: Princeton University Press.

Dardanelli, P. 2019. Conceptualizing, Measuring, and Mapping State Structures – With an Application to Western Europe, 1950-2015. Publius: The Journal of Federalism 49/2: 271-98.

Dardanelli, P., J. Kincaid, A. Fenna, A. Kaiser, A. Lecours and A. K. Singh. 2019a. Conceptualizing, Measuring, and Theorizing Dynamic De/Centralisation in Federations. Publius: The Journal of Federalism 49/1: 1-29.

Dardanelli, P., J. Kincaid, A. Fenna, A. Kaiser, A. Lecours, A. K. Singh, S. Mueller and S. Vogel. 2019b. Dynamic De/Centralisation in Federations: Comparative Conclusions. Publius: The Journal of Federalism 49/1: 194-219.

Dardanelli, P., J. Kincaid, A. Fenna, A. Kaiser, A. Lecours, A. K. Singh, S. Mueller and S. Vogel. 2019c. De/Centralisation Dataset 1790-2010.  [Data Collection]. Colchester, Essex: UK Data Service,

Döring, T. and J. Schnellenbach. 2011. A Tale of Two Federalisms: Germany, the United States and the Ubiquity of Centralisation. Constitutional Political Economy 22/1: 83-102.

Dubois, H. and G. Fattore. 2009. Definitions and Typologies in Public Administration Research: The Case of Decentralisation’, International Journal of Public Administration 32/8: 704-27.

Elazar, D. J. 1971. Community Self-Government and the Crisis of American Politics. Ethics 81/2: 91-106.

Esman, M. 1984. Federalism and Modernization: Canada and the United States. Publius: The Journal of Federalism 14/1: 21-38.

Hooghe, L., G. Marks, A. H. Schakel, S. Chapman Osterkatz, S. Niedzwiecki and S. Shair-Rosenfield. 2016. Measuring Regional Authority – A Postfunctionalist Theory of Governance, vol. 1, Oxford: Oxford University Press.

Kelsen, H. (1945) 1961. General Theory of Law and State. New York, NY, USA: Russell & Russell.

King, P. 1982. Federalism and Federation. London: Cass.

Krane, D. (1982) 1988. The Evolutionary Patterns of Federal States. In C. Lloyd Brown-John (ed.), Centralizing and Decentralizing Trends in Federal States. Lanham, MD, USA: University Press of America, pp. 39-62.

Livingston, W. S. 1956. Federalism and Constitutional Change. Oxford: Clarendon Press.

Pommerehne, W. 1977. Quantitative Aspects of Federalism: a Study of Six Countries. In W. Oates (ed.), The Political Economy of Fiscal Federalism. Lexington, MA, USA: Heath, pp. 275-355.


Further Reading

Dardanelli, P. 2019. Conceptualizing, Measuring, and Mapping State Structures – With an Application to Western Europe, 1950-2015. Publius: The Journal of Federalism 49/2: 271-98.

Dardanelli, P. and J. Kincaid (eds). 2019. Dynamic De/Centralization in Federations. Publius: The Journal of Federalism 49/1: special issue.

Hooghe, L., G. Marks, A. H. Schakel, S. Chapman Osterkatz, S. Niedzwiecki and S. Shair-Rosenfield. 2016. Measuring Regional Authority – A Postfunctionalist Theory of Governance, vol. 1, Oxford: Oxford University Press.

Hooghe, L., G. Marks, A. H. Schakel, S. Chapman Osterkatz, S. Niedzwiecki and S. Shair-Rosenfield. 2016. Community, Scale, and Regional Governance – A Postfunctionalist Theory of Governance, vol. 2, Oxford: Oxford University Press.

Treisman, D. 2007. The Architecture of Government – Rethinking Political Decentralization. Cambridge: Cambridge University Press.

Leave a Reply