DIRECTLY ELECTED MAYORS -
DEMOCRATIC RENEWAL OR CONSTITUTIONAL CONFUSION?
Helge O. Larsen
Department of Political Science
University of Tromsø, Norway
Paper presented at Convegno Annuale, Società Italiana di Scienza Politica, Università di Napoli, 28-30 Settembre 2000
Panel 2: La personalizzazione della politica italiana.
I. Electing mayors by popular vote - a fashionable idea.
Never before have so many people in Europe been able to select their foremost local political leader by way of direct election as today. When Londoners take to the polling stations on May 8th this spring, the number will increase significantly. Last autumn also saw something as previously unheard of as 19 directly elected mayors emerging on the political scene in Norway. In Germany, the idea of letting people elect their mayor by direct vote seems to be spreading from the southern Länder (where it has been practised for a considerable amount of time) to the north and east. In 1993, Italy decided to move from a system of indirect to direct election of mayors. If we stretch the notion of Europe somewhat, Israel was first, introducing direct elections of mayors in 1975. In short, the idea of having direct election of mayors seems to be a popular one, and spreading.
Directly elected mayors is not a new invention, of course. In Europe, this arrangement has been found in both Bavaria and Baden- Würtenberg ( the two southernmost German Länder (states)) since the end of World War II, but with historical roots streching well behind the efforts of the Hitler regime to streamline German local government. Besides, these two states belonged to the American zone of post-war Germany, and thus the idea of having directly elected mayors was not an alien thought to the Americans. According to a survey in 1987, 90 percent of mayors in US cities with more than 100 000 inhabitants were directly elected.[1]
While some form of local government is found in more or less every nation-state in the modern world, that is not the case for the position of a mayor as the clear-cut political leader of an authority, and particularly not if this position is associated with executive responsibilities. In many European countries, if not to say the most, executive responsibilities lies in a collective political body, or alternatively with an appointed administrative executive. This is a question not only of functions and responsibilities, but of language and terminology. To a Norwegian writer, the term «mayor» comes easily when describing the elected political leader in Norwegian local government; the «ordförer». The term itself, however, litterally means «the person who speaks» (on behalf of others), and gives association to the function of a chairman rather than an executive. Also from a legal point of view, the Norwegian mayor is a sort of primus-inter-pares role; i.e. the first among equals. A Swedish colleague observed that in his terminology, there was probably no such thing as a Swedish «mayor». The difference between these two countries, with local government systems in many ways so similar, would be that the positions as leader of the council and leader of the executive body are separated in Sweden, and not in Norway.
This point is also observed by Clarke and colleagues, in that
«…there is confusion arising from the language itself. This is a particular problem for
British audiences; in England and Wales the title of mayor suggests a ceremonial and social role but not one which involves exercising political authority. In many systems,…the title «mayor» is used to describe what is in effect the political leader of the authority.»[2]
To the extent there is another internationally used term (at least to someone with a Western World bias), it is probably the German term «Bürgermeister», which literally means «the master of citizens», and is a term which at least in my thinking gives associations to a position equipped with both ceremonial and executive functions.
Having observed this, I will go on using the term «mayor» in the remainder of the paper, unless it makes better sense applying another term. By mayor, then, I mean the position in a local authority formally regarded as the highest, which most often implies chairing the supreme political body (the council), and representing the municipality both legally and ceremonially. I recognise that such a definition is not without its problems, which I shall return to. Taking Norway as an example again, the mayor is in all (435 municipalities and 19 counties) but three cases the single most important political figure, also generally in terms of political influence. But three authorities have adopted a system of local or municipal parliamentarism, which implies that the executive functions on the one hand and those having to do with chairing the council and representing the locality have been split on two positions, resembling those of Speaker of the House and Prime Minister at a national level.
According to Lijphart, the relationship between the legislative and the executive branch of government in different national political systems may broadly be classified into two dominant types, the parliamentarian and the presidential model,[3] and with the latter resembling most the classical Montesquieuan ideas of division of powers among the bodies of the state. In local government, the relationship between the council and the executive body is not necessarily based on a majoritarian principle, but perhaps, more often on proportional representation. Using Norway as case again, the mayor is elected by and from the council as leader of both the council and the executive body, and his or her party may or may not enjoy a majority in the council. Most often these days the mayor has to draw support from a coalition of parties in the council, and is not guaranteed that this will be a stable coalition.[4]
Directly elected mayors, particularly when coupled with executive functions, would in principle pull the system in a presidential direction. It is not a necessary precondition for the existence of some sort of municipal presidential regime, as is shown by Hoffman-Martinot in the case of France:
«Mais l’election au suffrage universel direct n’est pas la condition indispensable à la constitution d’un regime présidentiel municipal, comme le montre parfaitement le cas francais. En France en effet, seuls les conseillers municipaux sont élus par les citoyens, avant de choisir parmi eux le maire. Mais celui-ci domine la scène politique municipale, après avoir été activement propulsé et soutenu au XIXe siècle par l’Etat comme relais électoral et administratif essentiel de ministères parisiens ..»[5]
What is clear, then, is a need to distinguish between the quest for a more executive political position in local government - an executive mayor - and the the electoral arrangement for selecting a person for a mayoral position, be it executive or not. As I shall return to, these two concerns have most often been coupled, and in England as well as in Italy and Germany, the emphasis has been on the executive dimension. This is in contrast to Norway, where the current experiment is carried out without virtually any change in the formal authority of the position. As far as I can see, New Zealand represents a case where the stress is on the electoral reform, and with a deliberate desire not to equip the popularly elected mayors with real political powers, but instead to underline the symbolic, representative and possibly integrative aspects of the role as mayor.
This may seem a little odd, at least at first glance. Whether the focus is on the electoral or the executive aspect, or both, one would think that they both would pull in the same direction as far as mayoral authority and influence is concerned. It seems like a reasonable assumption that a mayor elected directly by the citizens could enjoy and would make claims to a different and stronger legitimacy than one indirectly elected by the councillors. In such a perspective, also the Norwegian experiment may be questioned, since it seems to assume that it is unproblematic to change the way mayors are elected, but not the formal powers associated with this position. A Norwegian political scientist has provokingly asked if one really wants to create a lot on Norwegian «local banana republics»[6] Before returning to the Norwegian as well as reforms in other European countries, it may therefore be appropriate to take a closer theoretical look at the role(s) of mayor, and the way political legitimacy is linked to variations in formal authority as well as actual behaviour in political leadership positions such as the mayorship.
II. Political leadership: Authority de jure or de facto?
In any organisation or political system, roles may vary with respect to how well defined they are by formal institutional norms and prescriptions. The behaviour of persons in organisational roles may also vary with context and situation. One among several ways of seeing political leadership, is to regard it as a position within an institution, or an organisational function (Seligman 1950)[7]. When leadership is positionally defined, it is primarily associated with rights, duties and obligations associated with a level or a position in a hierarchial structure, or an informal, stratified collective (Edinger 1975).[8] One could say that leadership in this perspective refers to the authority vested in a certain position, and independent of who
is the incumbent.
A problem with such an approach is it's static character, which may cause problems if we want to explain changes in influence or authority associated with a position. Here, leadership is not seen as a quality of communication and interaction, where the source of authority rests on voluntary consent from followers.
Another difficulty is, of course, that an a priori connection between positions, authority and influence may lead us astray from real loci of influence. A third difficulty is that a purely positional view excludes the possibility of seeing authority and influence in non-stratified situations as leadership. Identification is made easier with a positional approach, but at
the risk of overlooking phenomena one would otherwise call leadership (Bacharach and Baratz 1963,1970).[9]
Leadership may refer to formal roles or positions, but also to types of activities or actions, or to the result of such actions (Olsen 1980).[10] Even if almost everyone studying the phenomenon has his or her definition, Edinger groups the efforts into two categories, based on either position or behaviour:
"When leadership is behaviourally defined it is identified with persons who shape up the actions of other persons. In that sense, it is associated primarily with followership and processes of interpersonal relations, rather than with interrelated positions
in a hierarchial structure" (Edinger 1975:256).
A problem with a purely behavioural approach may be that such a multitude of relations are called leadership that the concept is completely diluted and void of analytical value. This problem is present also when studying political leadership, although the question of mobilisation and of building and maintaining a following is not restricted to leaders already occupying formal positions.
The design of a study, or in which way the arrows should point, is, of course, a matter to be related to the analytical intentions of a study. When studying political leadership, there
are, in my view, some advantages and good reasons for treating role as independent variable. For one thing, the identification of leaders is made easier. This may help in comparative research, since more or less similar leadership positions are often found across systems (e.g. that of prime minister or mayor). Secondly, political scientists are most often concerned with analysing political processes either played out within or with reference to formal political institutions.
By taking a formal position as the point of departure, one may learn more not only about that particular position, but also about the system it is a part of. Leadership may thus be associated with the execution of particular functions within a political system; e.g. co-ordinating functions of an instrumental as well as an expressive (symbolic) character. Formal structure or authority may significantly affect leadership legitimacy. Whether a mayor is directly or indirectly elected may be quite important for the legitimacy of the position and in turn the
actions of the incumbent.
When leadership or leadership roles are treated as independent variable, the analytical intentions are often to assess the relative importance of a particular position for decision
outcomes or policy formation. Taking a structural approach, variation in outcomes may be related to variation in formal authority among the different actors or players. Formal structure
both enables and restricts the actors.
Characterising behaviour: Role as dependent variable.
Organisational roles may be more or less prescribed by formal norms and rules. A general assumption may be that the closer one moves to the top of the organization, the less restricted are roles by formal rules. The requirements facing leaders are often spelled out in vague terms. This may particularly be so if we talk about political leadership roles. If this is the case, one
may reasonably expect variation in the way different incumbents perform in their position. This is confirmed by one of the leading scholars in the field:
"The main empirical finding to date in political science role studies related to political leadership is that for any formal institutional role, different styles of incumbent performance are possible" [11].
When the analytical interest is focused on behaviour and the way incumbents in certain positions act, it has been quite common to try to develop typologies of such behaviour. This is the case also for political leadership positions, and a number of such categories or typologies also pertaining to the position of mayor have been suggested, mainly in studies of US mayors.[12] I will later in this section propose one possible scheme for mapping such mayoral behaviour.
In political science leadership studies of the institutional kind, formal authority has probably been the most common indicator to select, particularly when asked: does leadership matter?This is not sufficient, in my view (a view I do not at all claim to be holding alone).[13]Basically, there are two reasons for this; one more theoretical-substantial, and one of a more methodological kind. I will discuss these briefly.
The influence and the legitimacy of a political leader rests not only upon the amount of formal power (authority de jure) associated with a position, but also on the way the incumbent
manages[14] this authority (authority de facto). Basically, leadership is a relation between leader and followers, and a meaningful application of the concept requires both to be present. It is a relation of power, or rather authority .[15] The relation is often conceived and expressed in causal terms, picturing the leader as a person of power or influence, and thus able to
persuade others to change their preferences or actions.[16] As Lasswell would have it:
"This term is intended to designate the public leader, the one who plays an influential, conspicuous part in public affairs. Leadership goes beyond advice to commitment; it goes beyond management to goal-setting and high-level integration. The full-time public figure or active politician comes in this category."[17]
The methodological point has to do with the ways we proceed in order to arrive at good and fruitful understanding of how (political)leaders think about and behave in their roles. I have
argued for the need to take behavioural aspects into account, also including the motivations and reasons of actual persons in leadership positions that Searing calls for. I believe this to
be in line with what Edinger recommended already a couple of decades ago:
"As for now the most promising approach to comparative leadership analysis appears to be deductive-inductive analysis directed toward ever broader empirical generalisations. On the one hand, this calls for bringing theoretical abstractions down to the level of the "real world" of concrete human relationships. On the other hand, it means raising particularistic phenomena to a level of generality where patterns of variance may be identified, compared, and, to some degree et least, explained.".[18]
In order to understand political leadership, and the way political influence associated with a particular position varies across systems, it is, of course, necessary to look at the formal institutional framework. This is not sufficient, however, since the actual influence may vary despite institutional similarities, and according to situational factors. By the way leaders act in positions, they actually manages authority. Using «Parsonian» terms, one may talk of institutionalised as well as non-institutionalised behaviour.[19] Such an approach corresponds with Searing’s «motivational approach», which calls for a move in the direction of frameworks that includes rather than excludes informal norms and behaviour, even if this sometimes may come with the cost of conceptual difficulties. Thus, it makes sense to map how political leaders themselves perceives of their role(s), agreeing with Searing that «..the roles of politicians are dynamic and adaptive patterns of goals, attitudes and behaviors».[20]
In my efforts to understand the role of mayor in Norway, I have taken as a point of departure a scheme with two dimensions, generating four different leadership styles among mayors. The dimensions - and the types - have been developed more or less in the way Edinger recommends; as a type of pendulum movement between theoretical assumptions and empirical observations.
In our interviews with mayors, we tried to find out something about how mayors perceive of their own role, and how they go about filling it. Among a lot of other questions, we asked the
simple and straightforward (and normative) question: "What characterises a good mayor in your opinion?" In another question, we asked the mayors to give their impression as to what were the expectations of the inhabitants to their mayor.
To guide my thinking about this, I found it useful to conceive of four different ways mayors create or maintain their authority. To use the language of Parsons, this may be four ways mayoral authority seems to be institutionalised.[21] They are generated from two dimensions, a universalistic - particularistic, and an Instrumental - expressive dimension. At the poles of the horizontal dimension, I would put the instrumental aspect at the one end, and the expressive at the other. Along the vertical axis, the opposing aspects would be the universal versus the
particularistic one. It is then possible to distinguish between four distinct types of mayoral styles, which in turn may influence their authority. It is, in my opinion, important to be