political revue


Year 1990, XXXII - Number 2 - Page 174



Which decentralization?
The concept of decentralization, like that of federalism, itself can take on extremely variable meanings according to the authors who employ them and the contexts in which they are studied or applied.
Thus, for example, Karl Deutsch in his article “Toward a Rational Theory of Decentralization: Some Implications of a Mathematical Approach”,[1] shows to what extent the logistics of large scale organizations suffer from lack of adequate communications. Yet, in the industrialized states, new technologies in the area of transport and communications in general should allow the attainment of an optimal level of decentralization, whether it be for a public organization or for a large private corporation. Deutsch underlines the need to assure more independent initiatives and more active popular participation by intensifying the feedback between administrators and citizens. Yet, the process of adjustment to new problems brought about by the modernization of communications, for which he believes there are clearly solutions based on improved communications, requires an ever greater degree of decentralization.
Michel Crozier, among many others, came to the same conclusion in The Bureaucratic Phenomenon when he wrote that “decentralization now appears to the more enlightened observers of the American business community as the necessary condition of any growth beyond a certain threshold”.[2] At the same time, although many authors appear favourable to greater decentralization for reasons of efficiency and democratization (close to the people), there are some reservations about this, notably in the area of individual rights and liberties (another aspect of democracy). Certain authors, such as Duguit in France, while favourable in principle to decentralization, consider that the protection of interests and individual liberties stems from judicial review of administrative acts rather than from local self government.[3] This theoretical attitude reveals a certain mistrust of political interference in local administration.[4]
In real terms, what kind of centralization is involved here? Let us proceed by a process of elimination: in this paper we will not consider “administrative decentralization” in which tasks are by definition assigned to administrative units strictly under the control of central (or national) authorities. In this case, a certain number of choices, adjustments, and thus decisions, are carried out by agents of the central government posted in the regions. This system presents several advantages with respect to decentralization, above all if these agents are “natives” of the region in question. They are still, however, representatives of this central government from which they draw their legitimacy. We find this kind of model in federal states, but most of all in unitary systems.
We are interested, to be sure, in administrative decentralization, taken as the situation where the above mentioned administrative tasks are entrusted to agents and governmental services not attached to the central government, but rather to governments representing citizens of a given region (province, canton, Land, state, etc.). In this case, it is at the same time a matter of territorial decentralization, implying the existence of a government in the wide sense of the word, that having one or more elected legislative assemblies (bicameral in all American States except Nebraska), an executive, an administration; courts of justice and eventually a constitution (as with all American States, but not the Canadian provinces).
What is called decentralization in the Federal Republic of Germany and into a lesser degree in the United States and Canada often takes this form. Thus, for example, the German Länder have more personnel and spend more money (all together) than the central government, but these are above all applied under a framework of decisions taken by the central federal institutions. These top-level decisions are taken in collaboration, at least to a certain extent, with the executive representatives of the Länder. In this case we may speak of co-operative federalism. This kind of federalism has many advantages, above all for societies characterized by a high level of political consensus and where the economic, cultural and social cleavages are consequently reduced or do not have cumulative disintegrative effects for the whole of the system.
For many theoreticians of federalism, such as Carl Friedrich or Ivo Duchacek for example, genuine federalism requires, however, more than administrative decentralization as just described. According to these authors, if we limit ourselves to this type of federalism, it is possible to slide irreversibly from the status of a federal state to that of a unitary one. The central government would tend to make more and more of the important decisions (notably in the economic and foreign policy spheres) and the federated entities would then content themselves with adjusting and implementing the “framework legislation” – all the more so since the technocrats of the central administration play a major role in their elaboration.
A vital objective for federalism: political decentralization.
In the federal perspective of authors such as Friedrich or Duchacek, and notably most of francophone Canadian federalists, the problem of the division of powers between the two levels of government cannot be avoided.
In this way, we can speak of dynamics, interaction, with federalism thus being considered as a process, a perpetual movement oscillating between diversity and unity, centralization and decentralization. In such a system, it is essential that decentralization (the one where the distribution of powers is modified, or political decentralization) should not be static. There is an oscillation between the two poles formed by the two levels of government just mentioned. We should note that all decentralization of powers is of a nature to engender tensions and conflicts at the same time as co-operation but this, it appears, is the price to be paid to avoid deeper and destabilizing conflicts (system conflicts).
Federalism in the sense of decentralization (of powers) thus becomes a process where the communities (regions, provinces) interact as autonomous units. As for Carl Friedrich, he goes still further when he states that in a real federal system there can be no sovereign. In such an order, autonomy and sovereignty are mutually exclusive. In this way, we must reject the concept of a unified and indivisible sovereignty underpinning public law in a unitary state such as France, from Bodin and Rousseau to the modern day. When decentralization in this country and for those who are inspired by this model is discussed, it is thus according to very different premisses.
Yet, while many contemporary federalists still speak of autonomy and division of powers, we should emphasize that in many federal states decentralization is above all envisaged as administrative or functional.
In reality, in the majority of present day federal states, it has become almost impossible to admit that the notion of pact or contract (between the two levels of government or between the constituent states) is inherent to federalism. If there was once a contract, it was not in any case a bilateral contract (because there would be a confederation and not a federation). The problem in the case of Canadian federalism is that there is still some ambiguity about this issue. For some, it implies autonomy and exclusive powers for the provinces. For others, favourable to greater national unity, the central government is the only depository of the national will and the only legitimate sovereign, irrespective of the importance of the regions and sub-nations making up the whole. In such a framework, decentralization obviously has a much more limited scope and excludes the essential powers since these concern the whole of the country. At the same time, emphasis is placed on co-operation and co-ordination resulting from the interdependence of the different levels of government. However, it is still possible in a federal state that a region, a province or another locally organized political entity could be concerned about better protecting its specific interests. It is then inevitable that it would attempt to expand to the utmost its maneuvering space and, sooner or later, question its participation in power, and far from being a purely academic question, as many often lead us to believe in certain cases it might be a matter of the only instrument available to a region whose survival is more or less threatened. This is where the tendency to reinforce the “provincial state” arises. It is from this perspective that the problem of federalism is envisaged by certain regions of Canada (particularly Quebec and Alberta).
Some people affirm, however, that at the level of the central power, the choice decision of orientations and resource allocation can be made according to a confederal type process because of the regional demands, actors and resources would thus predominate at the centre. Decentralization would thus be assessed less in terms of statutory constructions, administrative regulations and resource distribution than with reference to the interaction of the various actors. Such an approach of course, a counter-weight to the liberal and Marxist obsession with property but it risks falling in another trap by underestimating the sources of conflicts and the inequalities that concepts such as interdependence, interaction, or co-operation barely hide. It is partially the classical liberal myth that the federal interests are the sum of the regional interests, as formulated in the central institutions.
Four federalist prerequisites, or decentralization and related questions.
A) Local affairs and intermediate governments. The constitutions of the federal states in industrialized countries such as the United States, Canada, Switzerland, the Federal Republic of Germany, etc., clearly distinguish central governments (in the wide sense) from intermediary governments (provinces, states, cantons, Länder, etc.). On the bottom rung are local governments that normally come under the responsibility of the intermediary governments. This is the case in Canada where the provinces have the exclusive power to legislate in the area of municipal institutions and purely local matters. In the United States this power is indirectly conferred on the states through their residual power (what the constitution does not attribute to the central government).
The principal problem here arises from the difficulty in specifying what is a local affair in matters often involving the three levels of government, each within its own area of competence (the classic case is that of landing rights for the Concorde in New York in 1976).
Furthermore, the large metropolitan areas often transcend the limits of several states and require the creation of a fourth level of government, capable of offering services that can neither be offered by local government nor by the states (or provinces). But under whose jurisdiction does this type of government come?
In the Federal Republic of Germany this problem seems to be resolved in the sense that by virtue of concurrent powers, the central government has the means to legislate in local affairs with the Länder and its laws predominate in case of conflict.
B) Elected governments and final legitimacy. For political decentralization to exist there must be intermediary government legislative assemblies that are democratically elected in the same manner as the national parliament. This democratic element is indispensable for a federal society of the type just mentioned, but it is a double-edged weapon, since the elected representatives of a province sitting in the national institutions claim to represent both all the country and their province of origin. In case of conflict, the problem of the principal allegiance might well come up, as was the case in Canada at the time of the repatriation of the constitution, in 1980-1981. The Canadian members of parliament from Quebec voted almost unanimously in favour of a new division of powers while the elected majority of the Quebec assembly resolutely opposed it. This is a classic example of the profound dissent that can exist between assemblies of different levels on vital issues.
C) Grey zones, concurrent and residual powers. Logically, we cannot speak of political decentralization without mentioning the grey zones and concurrent powers. Frequently, when there is a conflict between the two levels of elected government that concern us here, a struggle takes place in the area of concurrent power and in the grey zones that each of the two levels of government would like to clarify to their own advantage. The fact that certain constitutions (United States, but not Canada) accord residual powers to the intermediary government constitutes, at first sight, a favourable factor for decentralization, even though we should then clarify what in fact is left to the states, provinces, etc., in terms of real powers.
We may furthermore observe that central governments, in the name of the national good, quietly infiltrate this no-man’s land thanks to their financial powers and via joint programs covering a growing number of subjects.
For an author like Wildavsky, such a system implies tensions ,and conflicts but they probably generate real co-operation, while forced (co-operative-coercitive) co-operation would risk accelerating the polarization between the two levels of government by radicalizing the conflicts.
In this case, federalism implies concurrent powers and thus co-operation, but in various contexts and with various meanings.
D) Financial autonomy. All decentralization worthy of the concept implies a certain financial autonomy, albeit to different degrees and actual usage. For example, provinces or states can thus spend much money, but in the framework (and with the conditions) imposed on them by the central government. This money is, besides, spent in relatively less important areas (health, public assistance, primary education) while the central government keeps the essential governmental functions for itself (notably those of an economic nature).
Financial autonomy is at the same time a condition and an indicator of decentralization, but in order to evaluate its significance, it is necessary to study the various sources and methods of financing, and indeed, to define in exactly what areas the spending occurred.
Finally, there remains the question of determining how, and with what conditions (loose or restrictive), the laws were developed, as well as the decisions made about subsidies, tax transfer points and equalizing payments that are attributed to provinces and other entities of this nature.
Can the degree of decentralization be evaluated?
A) Limits to the quantitative approach. In the case of micro-analysis, it appears possible to construct indexes of decentralization when the number of key variables are few and quantifiable. K. Deutsch (quoted above) and many theoreticians of organizations have conducted interesting experiments which are limited in this respect.[5]
Stephen Ross[6] established a centralization continuum (the opposite is also possible) from a series of factual data: firstly, the distribution of financial responsibilities by level of government; and by the distribution of powers. He considered fifteen factors, established an index of services, and calculated the percentage of moneys spent by the governments concerned. Finally, the author studied the quantitative evolution of the public work force employed by the different levels of government and by area of activity. This enabled him to draw up a composite index of centralization.
This interesting experiment has many limits, including those mentioned above concerning financial autonomy, and in particular the limits imposed by conditional financial transfers, by cost-shared programs, and even by equalization payments (at first view, an unconditional transfer).
Richard Bird, in a quantitative study entitled The Growth of Government Spending in Canada,[7] also warns against such an approach.[8] He states that the concept of financial decentralization is not clear and that the study of transfers, for example, whatever method is used, is not based on a solid conceptual or empirical base. He himself, however, draws rather clear conclusions concerning the evolution of financial centralization.
B) Constitutional criteria. The constitutions of federal states contain some particularly revealing provisions regarding the centralization of some of the basic powers. A reading of article 1section 8 of the American constitution, of article 91 of the Canadian constitution or the economic articles of the 1947 Swiss constitution is highly significant illuminations in this respect.
Similarly, analysis of the decisions of Supreme Courts and Constitutional Courts, over a long period, allow us to draw a certain number of “assumptions” concerning the evolution of the centralizing/decentralizing process, insofar as we prudently use such indicators in conjunction with others, however open to criticism or incomplete they may be.
C) Distribution of functions. In his book Federalism,[9] William Riker classified “the degree of centralization” of American federalism into nineteen categories of activities (or functions). He distinguished five degrees: maximum decentralization, predominate decentralization, equality, predominate centralization, and maximum decentralization. He proposed this in the perspective of a “zero sum game” where what one loses is automatically won by the other. He goes on to conclude that in this country there are almost no more functions exclusively carried out by the governments of the states, while in 1790 the states’ rule was the common order of things, with the exception of foreign affairs and defence.
Such a method, when complemented by the study of the evolution of central political institutions, thus allows us, according to him, to bring out certain general tendencies, but this method has some fundamental draw backs. The first stems from the fact that it starts from a questionable assumption where the notion of conflict or competition prevails over that of co-operation or interdependence. It leads to precise quantitative results, whereas the evaluations are more impressionistic than mathematical, in an area where the use of statistics cannot be used in a systematic and complete way. Furthermore, it is difficult to judge the relative importance of each function. For example, how much we assess the political importance of one area such as foreign affairs with respect to that of economic development? This is all the more so since both are nowadays closely related, above all in the United States.
In Canada, too, many authors have attempted similar studies, notably Claude Morin[10] when he writes about the magnetic pendulum, or the power of centripetal forces.
D) The evolution of institutions. The study of the evolution and the creation of the new political and administrative institutions, at the different levels of government, can provide partial solutions. Thus, for example, in the United States and in Canada, during the last decades many states and provinces have equipped themselves with an increasingly sophisticated and modernized bureaucratic machine, especially in the industrialized regions.[11] In the case of Canada, we sometimes even see a direct confrontation between the federal and provincial governments. At other moments co-operation carries the day, but it occurs more and more between the various departments of each government. On the other hand, in Canada as well, the creation and growing use of federal-provincial conferences constitutes another useful indicator. In this case, it is a matter of a new co-operative institution of the various executives (prime ministers) of the provinces and the federal government. In practice, however, it is very difficult to evaluate it on a centralization/decentralization continuum.
We should also note the importance of provincial parties which are independent (and the degree to which they are independent) from the national parties, since this could constitute an interesting indicator to add in a more general study of this subject. Similarly, this could include the study of national parties as efficient (or inefficient) transmission paths of regional demands. When efficient, they contribute to national integration while reinforcing the machinery of national political parties.
E) The spirit of institutions and concluding remarks. In conclusion, we may note that many constitutionalists stress the “spirit” of institutions, as part of an overall vision. At first sight, this also might throw some light on our own concerns. But this would take us into an area even more difficult to analyse from the viewpoint of methods of observation.
Thus, for example, for Aaron Wildavsky, real federalism implies at one and at the same time conflict and co-operation, or centralization and decentralization, but as part of a balanced consensus.
As for Vile, he considers that constitutional, legal, political, administrative and financial techniques contribute, at least in the United States, to maintaining or eroding what he calls “the equilibrium between independence and the mutual interdependence of the different levels of government”.[12]
These authors, as well as Carl Friedrich,[13] thus favour the notion of equilibrium or federalist spirit. In the constitutional practice of the Federal Republic of Germany (see the decisions of the Constitutional Court in Karlsruhe) the concept of Bundestreue is invoked (federal loyalty or equilibrium).[14] It is a concept that is of greater practical importance than we would be inclined to imagine and thus deserves special attention.
There is still the question of how such a concept is rendered in concrete reality.[15] Thus, the notion of federal equilibrium can be interpreted in various ways from country to country, from one period to another, notably according to whether we favour a model of political decentralization or of administrative decentralization, to take these two points of reference.
Equilibrium – yes! – but generally with a predominant concern for central institutions (and in particular the supreme courts) of national integration and even unity, and thus the rejection of any form of decentralization when the decision makers concerned consider that it represents a centrifugal force capable of somewhat destabilizing the entire federal political system. In the name of equilibrium and integration, the national government would thus always have the last word in a confrontation with the intermediate governments, contrary to what Carl Friedrich writes: “In a federal system there cannot be a sovereign and no one has the last word”.[16]
That fact that the federal state is a state, with all of its attributes, in front of embryonic states that constitute the political entities such as provinces, sets insurmountable limits to the process and to policies of “decentralization”, even in societies where such political entities enjoy a relatively extensive degree of autonomy.
Just where can this process or these policies go without destabilizing the equilibrium and thus provoking disintegration? Here too, the answer depends on the specific needs of the societies in which the political framework is federalist. And, before attempting to evaluate its degree of decentralization, it is necessary to understand the particular context of the society in question, the absence of which would distort any interpretation, above all when it is matter of quantifiable data which at first glance are more precise than other sorts of data.
In spite of all their shortcomings (when taken separately), the criteria above mentioned, to the extent that they are combined together, are of such a nature as to allow us to finally develop a body of assumptions (in the absence of anything better) about the evolution of centralization, both as a process and as a deliberate or incidental policy.

[1]American Political Science Review, 63 (1969), pp. 734-749.
[2]Michel Crozier, Le phénomène bureaucratique, Paris, Editions du Seuil, 1963.
[3]Léon Duguit, L’Etat, Tome I et Traité de Droit constitutionnel, Tome III, Paris, 1923.
[4]Albeit in a completely different political and administrative context we find similar hesitations with regard to the administration of certain American states (above all those of the South). See our article “Droits de la personne et processus de centralisation: rôle de la Cour supreme des Etats-Unis”, Canadian Journal of Political Science / Revue canadienne de science politique, 20:4 (1987), pp. 711-729.
[5]Karl Deutsch, op. cit. (at note 1), p. 743.
[6]Stephen Ross, “State Centralization and the Erosion of Local Autonomy,” in Journal of Politics, 36 (1974).
[7]Richard Bird, The Growth of Government Spending in Canada, Toronto, Canadian Tax Foundation, 1970, p. 37.
[8]For another quantitative approach for the study of revenues and spending of the different levels of government in the federal states, see also Krane Dale, “The Evolutionary Patterns of Federal States,” in C. Lloyd Brown-John, Centralizing and Decentralizing Trends in Federal States, New York, University Press of America, 1988, pp. 39-62.
[9]William Riker, Federalism: Origin, Operation, Significance, Boston, Little Brown, 1964. See especially the typology of functions and the table “The Degree of Centralization by Substantive Functions”, pp. 82-83 in Chapter 3, “The Maintenance of Federalism: The Administrative Theory”.
[10]Claude Morin, Le Combat québécois, Montréal, Boréal Express, 1973, pp. 57-75.
[11]See the notion of “entrepreneurial regionalism” in Raymond Breton, Regionalism and Supranationalism, Montréal, Ed. Cameron, Institute for Public Policy Research, 1978.
[12]M.J.C. Vile, Politics in the USA, London, Hutchinson, 1987, p. 15.
[13]See Chapter 1of Carl Friedrich, “Théorie du fédéralisme en tant que processus”, in Tendances du fédéralisme en théorie et en pratique, New York, Praiger, 1968, p. 19.
[14]H.A. Schwartz-Liebennann von Wahlendorf, “Une notion capitale du droit constitutionnel allemand: la Bundestreue,” in Revue du droit public et de la science politique, 1973, pp. 769-792.
[15]For background information, see our article “La Cour constitutionnelle fédérale et l’autonomie des Länder en R.F.A.” in La revue juridique Thémis, 22:1 (1988), pp. 42,58-9.
[16]C. Friedrich, op. cit., p. 19.



il federalista logo trasparente

The Federalist / Le Fédéraliste / Il Federalista
Via Villa Glori, 8
I-27100 Pavia