Conference aim

Conference report

Programme

Workshop programme
Including
Full papers

Information

See paper in PDF

Citizen involvement and new roles for local politicians in decentralised city government: The Case of Stockholm

Henry Bäck
School of Public Administration
Göteborg University
P O Box 712
SE-405 30 Göteborg
henry.back@spa.gu.se

Folke Johansson
Department of Political Science
Göteborg University
P O Box 711
SE-405 30 Göteborg
folke.johansson@pol.gu.se

Paper submitted to the European Urban Research Association and Danish Building Research Institute conference "Area based initiatives in contemporary urban policy – innovations in city governance" Copenhagen 17-19 May 2001

Background

In 1996 the City of Stockholm introduced a new organisation with 24 neighbourhood committees¹. Beginning in 1997 the neighbourhood committees took over the responsibility for the provision of a number of services, mainly what could be labelled welfare services but also some technical services. The reform decision had been preceded by a field trial in three neighbourhoods in the years 1990-95 (Premfors m fl 1994). Albeit the idea of sub-local decentralisation in the city Stockholm could be traced as far back as to the 1940s (Sandqvist 2000), it is remarkable that no reform decision was taken until the interest for neighbourhood committees since long had waned in Swedish local authorities. In a way the traditional image of diffusion processes is turned upside down when innovations are spread from the periphery to the centre.

1) After the 1998 local election a coalition of Conservatives, Liberals, Christian Democrats and a local party took power in the city. The Conservatives preferably would abolish neighbourhood committees, while the other three supported the reform. The solution to the problem was a reduction of the number of committees to 18. It is interesting to note that this is the exact solution to the equation X = (3*24 + 1*0)/4 where X is the number of committees in the compromise. The assumption of power is described by Hanna Bäck (2000).

An interpretation that we have tried to launch (Bäck and Johansson 2000) implies that the institutional history differs between the big cities and the rest of the local authorities; The representative system with a council and central sector committees was well established in the cities long before the democratic breakthrough in local government. In other municipalities this system was slowly and sometimes reluctantly erected during the years 1920-50, only to be torn down as a consequence of merger reforms in 1952-74, leading to a situation where new institutions had to be rebuilt. Neighbourhood committees came to be one reaction to the centralism which these new bigger municipalities were perceived to be characterised by. The big cities, and especially Stockholm went unaffected through the institutional crisis implied by the merger reforms.

On the other hand the big cities as well as other municipalities were hit by the crisis of the welfare state in the late 80s and early 90s. In this new crisis they groped for new institutional arrangements and like other municipalities they found the by market liberalism inspired "New Public Management" solutions. But also neighbourhood committees, until now untried in the big city, were found. We thus suggest that neighbourhood decentralisation that in other municipalities was seen as a remedy to perceived democratic problems in the wake of the merger reforms in Stockholm came to function as a reaction to the crisis of the welfare municipality. This might seem paradoxical as many authorities where neighbourhood committees had been established already before the end of the 1980s now turned to centralisation and abolition of the neighbourhood committees as a response to the same economic crisis of the welfare municipality (cf e g the abolition of neighbourhood committees in the municipality of Ale described in Bäck 2000).

This alternative interpretation of the neighbourhood reform is very loosely if at all coupled to perceived problems in the local democracy. Nevertheless, it should be observed that as in other similar reforms, the objective of improving local democracy is in focus also in the Stockholm reform. The direct initiative to the reform was the so called "seven party address" (the Conservatives were against and did not sign) of June 8^th 1994, which explicitly stated as the first-mentioned goal of the reform: "The organisation is intended to enhance local democracy and to vitalise civil involvement." As it happens, most follow-ups and evaluation studies of neighbourhood reforms in Swedish local authorities have concluded that no such effects regarding citizen participation and involvement have been achieved (c.f. Montin 1989, Jönsson et al 1995, 1997a, 1997b). However, effects on the role of local politicians have been observed. Politicians who are members of neighbourhood committees are more closely in touch with neighbourhood communities and have gained a stronger position in relation to the professional administrative and service providing organisation (Jönsson m fl 1995, Premfors m fl 1994, Solli 1988).

In this paper we are going to approach the question whether such effects that have been observed elsewhere also are traceable in Stockholm. Are effects implying improved civil involvement and participation as absent as in other similar reforms, and could the same changes in the role of local politicians as in other municipalities also be observed.

Citizen involvement and participation

The concept ‘political participation’ does not have a clear connotation. It is rather one of the most debated concepts within political science. The existence of many views depends to a great extent on the fact that political participation is always a central component in discussing democracy. There are different views on the importance of participation when judging the functioning of democracy. There is however total agreement that when discussing democracy you must make up your mind about the role of political participation in your definition of democracy. (cf Milbrath 1972)

A first step is to consider whether political participation should be seen as instrumental or not. Should we perhaps reserve this concept for activities, which in one way or another are intended to influence the political system and the decisions taken by different political actors? The question of instrumentality is thus directly related to the effects that participation can have or is intended to have on the political system. Another central form of classification builds upon the possible effects of participation for the individual citizen participating. These theoretical notions have been summarised under the headings communitarian participation, educative theories and expressive participation. (Parry, Moyser and Day 1992)

If you see the communitarian part as important you imagine that all or at least many of the citizen should participate and also that the participation means that the decisions can be taken together, preferably in harmony and that the participation creates a spirit of community. Jean-Jacques Rousseau was one of the leading advocates of this view. There are obvious traces of a communitarian view in the prerequisites that we start from here. It is evident from the emphasis on the importance of the neighbourhood committees in creating a local community.

Within another theoretical notion the educational effects of participation are the most important. The citizens become better citizens by participating. They gain understanding also for other parts of society than the ones that they are normally involved in and they also gain confidence in the political system. The educational effects have been put forward by John Stuart Mill and more recently among others by Carole Pateman (Pateman 1980).

A third way of looking at participation is stressing the expressive side. Those participating are supposed to try to mark their belonging to a certain group or a certain political outlook. One aspect of this can be that you participate because you want to feel that you are part of this group or this outlook. You can also within this notion participate in order to express your own opinion without any immediate wish to influence a certain decision. Something that is not incoherent with the possibility that one or several political decisions may have initiated the activity.

How to explain participation

With different ways of looking at participating follows hypothetically also different models of explanation. If you for example stress a communitarian aspect, you can hardly expect personal characteristics to be important. It is rather the shaping of the system that is most central. With an instrumental point of view, which in this case also can be linked to a democratic demand for equal possibility to influence, it seems reasonable to concentrate on the importance of personal efficacy for the participation. What you study is really more the prerequisites for participation than the direct cause for the activity.

From an evaluation point of view the structural explanations are of most interest. When you create a neighbourhood committee system, you change the political structures. Of course you can not analyse such changes separately but you have to consider several of the factors that have been mentioned here. They can to a great extent determine how an individual reacts to different political systems and different forms of organisation of the political decision making. Comparison over time is of course a basic method in evaluation studies.

Another important factor in judging the functioning of political systems is size. That is the number of citizens or the number of voters in the neighbourhood committee (or whatever unit you are studying). Traditionally you can see many democratic advantages with smaller units (c.f. Dahl and Tufte 1973).

Belief in the possibility to influence

The answers to a open-ended question concerning which expectations you did have on the neighbourhood committee reform showed that a lot of people thought that the distance between citizens and politicians would be shortened. Such a gap closing might of course have different positive consequences. The most important may be the view that the citizens have of this distance and its possible shortening. We have tried to measure this with different questions that are intended to give room for a more concrete picture. One of these concerns the possibility to influence politics. In the year 1966 more than half of the respondents (58 percent) totally or partly concurs to a statement saying that The introduction of neighbourhood committees make it easier for people in general to influence politics. Afterwards in 1999 a somewhat lower percentage, 41, concurs in the same way. The expectations are high on these committees. It should first be emphasised that the second result also must be seen as somewhat of a success. It is in fact four out of ten persons who think that it is easier to influence decisions. On the other hand it is always negative to be disappointed.

The percentage that thinks that it will be easier to influence diminishes from 58 to 41 percent. Men, elderly people, people with low education and workers are clearly overrepresented among those who have had to adapt their earlier expectations. Their general belief in the system has gone down relatively much. The percentage that think that it has become easier to influence is about one third also in the groups that are generally critical to the new system. We can note that people below 30 years, self-employed persons and those with a strong attachment to a certain political party tend to keep their belief in the positive effects of the system while those with no party attachment are more disappointed.

Political activity

The first form of political activity that we show is discussion of local political issues. The percentage, which claims to have discussed local political issues with people in their surrounding, has not changed from 1996 to 1999. Somewhat more than three out of four report having such discussions frequently and less than one fifth report not having such discussions at all. Most respondents chose the alternative Occasionally, which can include everything from once and up to the point where the individual himself see it as Often.

Table 1 Discussion of local political issues

	1996	1999
Often	27	26
Occasionally	56	56
Never	18	18
Sum	100	100
Respondents	1710	1576

Question: How often do you discuss local political issues with people in your surrounding?

Discussion of local political issues can be important for the individual and may also lead to efforts to influence political decisions. We now turn to the question whether you have in fact tried to influence a local political decision. We asked if you had tried at all and if so in what way? The results are that the percentage that has tried has not changed when you compare the situation beforehand 1996 (22 percent) with the situation after the reform 1999 (23 percent). The introduction of neighbourhood committees did not make an immediate difference. Rather the most striking in these results is the stability.

Table 2 Efforts to influence local political decisions

	1996	1999
Tried in any way	22	23
Influnce through (contact with): Civil servant in neighbourhood committee	7	5
Other civil servant		3
Neighbourhood committee politician	7	4
Other politician		2
Active in party organisation	2	1
Action group	3	3
Other organisation	2	2
Writing in newspaper	2	3
Petition/demonstration	13	15
Other way	2	2

The fact that the total percentage is the same shows directly that there are no changes. With regard to the different ways of influencing you can only note that their correspondents in the neighbourhood committees have substituted central civil servants and politicians. A change that is necessary as the issues are to a great extent handled only there. In that way the neighbourhood committees have been successful. All other kinds of political activity remains unchanged and the pattern of activity also corresponds fairly well with what we can see in the average local authority in Sweden. The most notable differences are that petitions/demonstrations are somewhat more common in Stockholm while activity in a party organisation is less important than in the average local authority. This is most certainly dependent on size. The political party is a more easily accessible channel in smaller communes.

The main pattern was that changes in activity were small. This goes also for different subgroups of respondents. With regard to sex there are no differences but age groups show varying behaviour. Those less than 30 years discuss local politics somewhat more 1999 than earlier while the percentage discussing is diminished at about the same rate among the middle aged. The older group, from 56 years and older, tries to influence political decisions somewhat less 1999 than 1996. Those self-employed also seem to be somewhat less active 1999 than 1996. The respondents with a strong party attachment both discuss local political issues more and try to influence local political decisions more 1999 than 1996. The idea that contacts through the neighbourhood committees would be more socially equal than other ways gets very little support in this analysis. Those groups that are normally politically active in society are more active with regard to neighbourhood committees as well as with regard to other local political institutions. The only exception being occupational groups where contacts with neighbourhood committees are equally frequent in all groups while other types of contacts follow the more normal patterns with higher frequency for self-employed and higher frequency for civil servants/salaried employees than for workers.

Participation – a central democratic object

The most frequently mentioned goals in relation to neighbourhood committees are increased political involvement and increased political participation. The result here shows that the hope of reaching these goals is not fulfilled. There is no increase in the number trying to influence political decisions and not in the number discussing politics. On the other hand there are quite a few who thin that the neighbourhood committees make it easier to influence local political decisions. The picture is not totally clear. There seems to be some contradictions. On the other hand there are rather simple explanations to these contradictions.

First the idea that neighbourhood committees make it easier to influence political decisions. Close to one half of the responding citizens share this opinion when the reform is introduced and more than one third still share it after three years. There is a rather large group, which displays what from a democratic view must be seen as a positive reaction here. However such a view does not necessarily lead to political activity. It is at the same time quite important as a factor of security in case one should feel the need to influence a decision or take some other form of contact.

The central question is; why does the political participation not increase? First we should remember that the personal characteristics of the citizens are the same and that the social structure is the same. Basically also the service provided by the neighbourhood committee is the same as the service earlier provided by the local authority centrally. Is it reasonable to assume that such a change directly or even in a middle term perspective should affect political participation? It is not likely, especially not if you have an instrumental view of participation. You are active with the purpose of attaining something – a change in policies a change in a more specific decision. The personal background may increase the probability that you do something but a direct cause is also a necessary condition. There is no indication or any reason to expect that such causes have become more common. It would have meant a serious failure if the reform did have such consequences. In the longer run political participation may increase somewhat because a lot of people now think that it is easier to contact politicians or civil servants. The threshold value for activity has become lower. An increase in activity as a result of this must however be seen as a long-term process. A process that has started already which can be seen by the fact that people now contact the neighbourhood representatives more than those representing the central parts of the local authority.

Changing roles for local politicians

The roles of local politicians can fruitfully be ordered into two dimensions. One of these concerns the role as a representative, and the other the intra-organisational role as decision-maker and governor (cf Young 1990, Nilsson 1993, Bäck 2000). In this paper we will discuss only the role as a representative. The questions to be posed to our empirical material can be summarised:

– How do politicians relate to the contradiction between territorially defined particular interests (the neighbourhood interest) and the common interest of the whole city? Has it been an effect of the neighbourhood reform that territorial interests have gained more legitimacy? How does such a change, if any, relate to how other particular interests (e g sectorally defined) are perceived?

– How have the politicians’ contacts with citizens and their associations developed? Has the neighbourhood reform implied increasing frequencies of contact with citizens with regard to questions concerning individual citizens as well as more political issues? Have the politicians become more deeply rooted in local associations?

The empirical material for the analysis is provided by two mail surveys carried out in the autumn 1996 and the autumn 1999 respectively with all directly and indirectly elected local politicians in the city of Stockholm. The 1996 survey was returned by 81 percent of the total of 832 politicians. The 1999 survey was distributed to 636 persons of whom 73 percent replied. The final report of the evaluation study will appear in 2001.

The number of politicians

Seen in a representation perspective it has been regarded important and advantageous to have a large number of elected officials. If elected office is regarded as a kind of civil involvement and participation a larger number of elected automatically implies a higher level of involvement than a smaller number. It has also been shown that a decreasing number of elected is something that tends to impair the group representativeness of the system. Social biases in representation are reinforced (cf Bäck 1993, 1996). Furthermore it should be pointed out the fact that having more politicians implies more interfaces between elected and electors than having a smaller number. There are many scholars who have expressed their worries for a decreasing number of politicians (Larsen & Offerdal 1992, SOU 2000:1, Rothstein et al 1995). John Stewart writes:

There are dangers in any reduction in the number of councillors if importance is attached to the representative role of councillors (Clarke & Stewart 1989).

A peculiarity of neighbourhood committee reforms in comparison with many other institutional reforms in local government’s political organisation is that they tend to increase instead of decreasing the number of politicians. This is also the case in Stockholm. It was noted above that there were in 1996 832 people holding directly or indirectly elected political office in Stockholm. This was an especially high number due to the fact that the city in 1996 actually ran two organisations in parallel. At the same time as the older organisation was still in place with a relatively large number of central sectoral committees and a local organisation with social welfare district committees, the new organisation with 24 neighbourhood committees was installed. If the persons only holding office in the social welfare district committees are deducted from the total number we are still left with 734 politicians – still a rather impressing number. We have not especially investigated the situation prior to the neighbourhood reform, but if we instead deduct those persons who in 1996 only held office in the new neighbourhood committees from the total number 458 would remain. The immediate effect of the neighbourhood reform on the number of politicians thus is an increase with 60 (from 458 to 734) to 82 (from 458 to 832) percent.

When neighbourhood committees were definitely installed in 1997 the social welfare district committees as well as some of the central sectoral committees disappeared. The shift of power in 1998 from a social democratic government in the city hall to a coalition of non-socialist parties carried with it a reduction in the number of neighbourhood committees, partly counteracted by a partial restoration of the central committee apparatus. The net effect of these changes was that we could count 636 politicians in 1999. The increase in relation to the initial position then halts at 39 percent.

Figur 1 The number of local politicians

These increases may seem considerable when expressed as percentages, but seen in relation to the population of the city (700,000) they are of course less impressing. At the starting point in 1995 there were some 1,500 Stockholmers per politician. At the peak in 1998 the corresponding number was 840 and 1999 each politician had 1,100 citizens to keep in touch with. A comparison could be made with the average conditions in the country in the mid 1990s. In the 1993 local councillor survey of the research programme "Democracy in Transition" (Bäck 2000) there were in 32 surveyed local authorities (of which four were county councils) 6,236 politicians, or slightly under 200 per authority. Many Stockholm neighbourhoods were of the same population size as an average municipality. Both the record figure of 1996 and the slightly lower figure of 1999 implies 35 politicians per neighbourhood.

Party structure and office structure

In the analysis of the 1996 politician survey (Bäck & Johansson 1997) we emphasised the importance of the combination of public offices which is held by the individual politicians for their attitudes and behaviour. More than half (52 percent) of the Stockholm politicians only held office in the new neighbourhood committees, one fifth (21 percent) held office only at the central city level (city council, executive committee, the remaining central committees). The rest was distributed between those holding office at both tiers (15 percent) and those with positions only in the doomed social welfare district committees (12 percent).

The relatively small group overlapping the central and local levels turned out to be a group with strategic properties:

– They had higher hierarchical positions within the politician collective (38 percent held chairmanship positions as compared to 10 percent of the "pure" neighbourhood politicians and 13 percent of the "pure" central level politicians)

– They distinguished themselves as advocates of the local neighbourhood interest by more often than the others asserting that municipal service had been distributed in a less appropriate way over the city surface. Eight percent spontaneously claimed that they regarded themselves as representatives of their own neighbourhood, a response given by a mere four percent of the "pure" neighbourhood politicians and five percent of the "pure " central level politicians. In good accordance with their view of themselves as neighbourhood representatives they also displayed an especially positive attitude towards the neighbourhood reform.

– However, at the same time they had an especially good understanding of the common city interest. If this common interest should enter into conflict with the interests of the neighbourhood "only" 70 percent were prepared to give priority to the local interest (as compared with 81 and 76 percent for the other categories respectively). As a consequence they were less enthusiastic than the others to proposals to introduce direct neighbourhood-wide elections, an arrangement that could weaken the integration of the city. Only 54 percent compared to 59 and 63 percent respectively in the other two groups were positive to the idea of direct elections.

– Their public spirit also was expressed in that especially few (37 percent) would give priority to the interests of a service sector in a hypothesised conflict between a sector and the common interest of the city (to be compared with 61 and 45 percent respectively).

– The relation to the local community is also upheld in their deeds. In this group 83 percent (as compared to 56 and 65 percent respectively) had during the last year had contacts with associations or firms. A larger proportion than in the other two groups reported contacts with individual citizens in "personal" as well as more "political" matters. In this group 59 percent rather often or practically daily discussed local government matters with other citizens than their own family members, colleagues, friends and neighbours. This figure could be compared with 42 and 44 percent in the other two groups.

To summarise: The politicians who held office on both tiers in the city were the goodies of the reform: Active, influential, keeping contact with the local community acting as its representative still understanding the need to keep the city together. These two-level politicians act as an integrating link in the political system, thereby facilitating the successful implementation of the reform. In the absence of such integrating links the reform could either route towards the disintegration and decomposition of the city or towards increasing centralisation, which could endanger the credibility of the reform as a democratisation project.

When discussing the modus operandi of a political organisation, it is not sufficient only to consider formal structures – what formal bodies and offices are in place, which are their rights and obligations. One must also consider the party political structure that cuts across the formal structure. Even though the formal structure might seem decentralised, even fragmented, uniform action can be maintained via intra-party steering, parties often being considerably more hierarchical and centralised organisations. Legendary Chicago mayor Richard Daley’s control chanelled through the Cook County Democratic Party (cf Banfield 1961) and the Communist Party strict control of the according to the letter of the constitution federal and democratic Soviet Union could be two examples.

With these considerations in mind we will in the following analyses use a model that could be simply summarised:

Y = f(O,P,T)

Y is the attitudes and behaviour of the politicians: Here the following aspects will be scrutinised:

Attitudes

– The feeling of belonging to a local neighbourhood
– Assessment of the geographical distribution of municipal services
– Territorial particular interests versus the common good
– Service sector particular interests versus the common good
– Spending preferences

Behaviour

– Citizen contacts
– Discussion of local politics with surrounding persons
– Involvement in associations

O is the pattern of offices held by the respondent. This pattern is expressed in the terms discussed above, i e the degree of overlapping between the central and local levels. Two dummies will be used in the analyses: Overlapping offices and Central level offices respectively. The technical implication of this is that regression coefficients represent differences in comparison to those only holding office at the local neighbourhood level, i e how much larger or less will Y be in the category under consideration compared to "pure" neighbourhood politicians.

P is a vector indicating position in the party political structure. Three variables are included in the analyses:

– Majority will be 1 for respondents belonging to the ruling majority of the council (Social Democrats, Left Party and the Green Party in 1996 and Moderates (Conservatives), Christian Democrats, Peoples Party (Liberal) and the Stockholm Party² in 1999) 0 for respondents affiliated with the opposition parties.

2) A local party in the city of Stockholm, originally an offspring of the environmentalist movement in the 1970s but now clearly distinct from the Green Party taking an explicitly neutral position between the socialist and non-socialist blocs.

– Moderate will be 1 for Moderate Party members and 0 for all others. A reason for distinguishing the Moderates is that their party actually never supported the neighbourhood reform, while all other parties are initiators and supporters of the reform.

– Leftwing is coded 1 for Social Democrats and members of the Left and Green parties. All others are coded 0. This variable intends to catch the effect of the left-right dimension traditional in local politics in Stockholm.³

3) About the dominance of the left-right dimension in Stockholm city politics see Hanna Bäck 2000

T, finally, is time. This variable is 0 for 1996 and 1 for 1999. If there are detected significant effects of this variable on Y, we at least know that something happened between the two points in time that can not be explained by changes in the pattern of offices held by local politicians, the change of regime in 1998 and changes in the party political composition of the council. It is among these observed significant time-effects that consequences of the neighbourhood reform are to be sought. It could, however, still be the case that such observed changes are dependent on other events between 1996 and 1999 not controlled for.

Furthermore we mean that significant effects of O also indicate possible effects of the neighbourhood reform as the very categories defining O are constituted by the neighbourhood reform.

Local affinity and territorial particular interests

The distribution with regard to the feeling of affinity with the neighbourhood of residence hardly displays any change over time at all, which also is supported by the regression analysis. The regression equation explains (R²_adj) 5.5 percent of the variance and the only significant slope coefficient is with the variable "central level office".⁴

4) The unstandardised regression coefficient is -.493, meaning that central level politicians feel almost one half point (on a five points scale) less affinity than pure neighbourhood politicians. The standardised regression coefficient is -.243.

Table 3 Local affinity (percent)

	1996	1999	Change
Not at all	1	1	0
Small	4	3	-1
Somewhat	10	13	+3
Rather much	37	36	-1
Very much	49	48	-1

If the distributions in the different office-holding categories are scrutinised it appears that the group with overlapping offices distinguish themselves at both points in time, and also that there is a considerable increase in this very group. This cannot be seen in the regression analysis as this change in regression analysis terms in an interaction effect between the variables O and T, which has not been specified in the model. If such an interaction term (between "holding overlapping offices" and time) is added to the regression equation an almost significant slope coefficient appears (p=.049).

The appearing differences between the different office holding categories, with especially low values for those only holding central city offices, in combination with the increase in local affinity in the group holding overlapping offices are facts that support a conclusion that the neighbourhood reform has had an effect. The reform has created a group of local politicians with a strong and increasing feeling of affinity with the local territory.

Figure 2 Percent high or very high local affinity (percent)

The introduction of neighbourhood committees implied that the responsibility for service provision was organised territorially. There are as well ideas that local neighbourhoods constitute local communities where people are brought together and where living in the same locality produces a ground for common interests. To the neighbourhood committee as a political organisation then is allocated the role of representing the locality and its interests towards the surrounding. It therefore ought to be expected that the reorganisation of the city should lead to an increased saliency for the territorial dimension in local politics, and that politicians, now to a large extent organised territorially should become more attentive to the geographical distribution in the city. To approach this issue we posed a survey question about how the distribution of local government services is assessed: is the distribution appropriate, and if not what parts of the city are favoured and what parts are treated unfairly.

Table 4 The geographical distribution of local government service (percent)

	1996	1999	Change
1 Appropriate	69	52	-17
2 Own neighbourhood favoured	11	18	+7
3 Own neighbourhood treated unfairly	21	30	+9
Balance of opinion (2-3)	-10	-12	-2

It is obvious that there is a marked increase in the proportion holding the opinion that local service is distributed in a less than appropriate way over the surface of the city. On the other hand there is no especially pronounced increase in discontent with how one’s own neighbourhood is treated. The results indicate that, as expected, the attention to and the importance of the territorial dimension in politics increases after the neighbourhood reform.

This being an assessment of the services delivered we ought to check for the effects of party affiliation and whether one’s party is in office or in opposition, as these factors can be expected to affect the evaluation of the services provided by the local authority. Such a control is performed by applying the regression model. This analysis show that the conclusion remains. The variables "majority" (B=.101), "leftwing" (B=-.198) and "time" (B=-0.180) all have significant effects on the assessment of the appropriateness of the service distribution. Those belonging to parties in office thus tend to have a more positive evaluation than opposition party members. Members of the left-wing parties tend to be more dissatisfied than members of the right-wing, and the proportion of contended respondents decreases from 1996 to 1999. The categories defined by holding office at the local and/or the central level, however, do not differ significantly from each other⁵. The conclusion that the territorial dimension in local politics has gained in saliency seems to hold for this closer scrutiny.

5) The regression model explains 6.4 percent of the variance

Figure 3 The distribution of local service (percent "appropriate")

Whom a representative considers himself to represent is known in the literature as representation focus (Eulau et al 1959, Whalke et al 1962). We have asked our respondents if there are some group or groups or interests that they, beside from being party representatives, consider themselves to be the representatives of. Around half of the respondents in the 1999 survey say that they have one or more such representation foci. The most usual responses can be interpreted in terms of groups more or less dependent on local government welfare services: Children, young people, parents, and families (12 percent), weak groups, the sick, disabled persons (8 percent). After that comes "my own neighbourhood" with 7 percent. This represents a slight increase from the 1996 survey, in which 6 percent gave this answer. This small difference nevertheless results in a significant regression coefficient (B=.027; p=.039). The fact that the direct effect of time passing appears larger than when controls are not applied could be explained by the existence of indirect causal chains with a sign opposite to the sign of the direct effect. Categories in which such a territorial representation focus is less common have increased their proportion of the population studied (Moderates, "pure" central level politicians) while categories where a territorial representation focus is more common have decreased relatively (Social Democrats, local level politicians).⁶

6) The pure central level politicians have increased their share of the whole group (excluding Social Welfare District Committees) from 24 to 28 percent.

From the central city level perspective the territorially defined neighbourhood interests can be considered as particular interests. Regarding the interests of particular groups as not legitimate is typical of Swedish political debate in recent years. In a study of values evidently underlying organisational reforms in Swedish local government under the first half of the1990s a strong repudiation of the idea that particular group interests have democratic legitimacy was observed (Bäck 2000). The objective of democracy is to solve common problems and there is no room for conflicts of interest, especially not for asymmetric conflicts of interest. Leif Lewin (1998) means in the book "Bråka inte!⁷" that this consensual democracy actually is typical of Swedish politics in the whole 20^th century.

7) "Don’t make trouble!"

In this perspective there are remarkably many of our respondents who claim that a member of a neighbourhood committee primarily should be guided by the needs of the neighbourhood if these should come into conflict with the general interest of the whole city. Almost 90 percent hold such a view. Sectorally defined particular interests, however, have a lower degree of legitimacy, which is more expected. Only half of the respondents hold the opinion that a member of a service sector committee primarily should see to the needs of the sector.

Table 5 Territorially and sectorally defined particular interests (percent preferring the interests of the neighbourhood or the service sector to the common interest of the city)

	1996	1999	Change
Neighbourhood interests	89	88	-1
Service sector interests	54	49	-5

One may wonder what makes this difference between a neighbourhood and a service sector. One possible explanation could be found in the rhetoric surrounding the neighbourhood reform, where territorial community has been stressed. The case simply has been put so, that it is the duty of a neighbourhood politician to represent the neighbourhood and its interests. Looking at the local community in the neighbourhood renders another aspect. Stressing the common interests of the neighbourhood implies that conflicts of interests within the neighbourhood are of less importance and/or less legitimate. The neighbourhood interest and the city interest in a way belong to the same family. In both cases it is a territorially defined interest that takes precedence and is regarded as more important and more legitimate than other contrasts within the neighbourhood or within the city, e g between classes, genders or ethnic groups. There is a slight change over time in how the importance of the neighbourhood interest is viewed. Nor does any significant time effect appear in the regression analysis. On the other hand both party political and office pattern variables have significant effects. The Moderates are less willing to advocate the neighbourhood interest than adherents of the other parties (B=-.185) and politicians with overlapping offices (B=-.157) or only holding office at the central level (B=-.147) also more seldom are supporters for the local interest.⁸

8) The proportion of variance explained is 5,4 percent

Non-territorial particular interests

The respondent politicians have been asked how they look upon local government services decentralised to the neighbourhood committees: Should the authority make more or less efforts or is the present situation satisfactory? There are shifts in priorities between the two points in time. There are especially marked increases in the number who think that the authority should increase its efforts in welfare, child care, cultural activities and libraries. This implies some changes in the ranking of different service areas: Primary education, child care and welfare is given a higher ranking at the expense of care of the elderly and streets and roads management.

Tabell 6 Spending preferences (Opinion balance)⁹

	1996	1999	Change
Primary education	+76	+75	-1
Care of the elderly	+81	+74	-7
Recreation	+65	+71	+6
Welfare	+40	+55	+15
Child care	+32	+47	+15
Streets and roads	+54	+46	-8
Cultural activities	+19	+27	+8
Libraries	+16	+26	+10
Consumer advisory service	+11	+9	-2

^{9) The opinion balance measure is the difference between the percentage wanting to increase efforts and the percentage advocating a decrease.}

It can, however, be concluded that the responding politicians on the average are prepared to increase spending in more policy areas than they are willing to decrease spending in. A summarising measure on spending preference can be an arithmetic mean over the nine service areas. This measure is for the year 1996 +47, and increases in three years one meagre unit to +48. The willingness to spend thus has increased, but very marginally. This, however, is a case where it could be warranted to suspect that an increasing trend could have been concealed by the party political shifts. The essence of what happened in the 1998 election was that the more spending happy left wing was reduced in favour of the more parsimonious right wing parties. Therefore the party political changes could be expected to carry with them decreasing spending preferences that could hide an underlying increasing trend.

Table 7 Spending preference. Regression analysis.

	B	b	Significance
Majority	-.089	-.114	.000
Moderate	-.332	-.336	.000
Left-wing	.231	.295	.000
Overlapping offices	-.093	-.089	.002
Only central level	-.082	-.086	.003
Time	.056	.071	.012

R²_adj = 0,365

The regression analysis strengthens these suspicions. The expected party political effects do appear. Incumbent parties and Moderates are less willing to spend than opposition parties and left-wingers. The time effect is positive and statistically significant.

It could thus be established that the preparedness to increase spending rises between the years 1996 and 1999 after controlling for party political circumstances and the structure of elected offices. On the other hand, it is more in doubt whether this increased spending willingness can be coupled with the neighbourhood reform. A fact to be held against such a conclusion is that neighbourhood politicians as a group do not differ systematically from central level politicians. In comparison with "pure" neighbourhood politicians both "pure" central level politicians and politicians holding offices in both tiers display a lower level of spending willingness. Furthermore, (not demonstrated in this table) the direction of movement without controlling for party political conditions is upwards for the "pure" neighbourhood and central level politicians and downwards for those with overlapping offices. Those holding office on both levels thus would like to cut spending while those holding office on only one of the levels prefer increases.

This seems to be a situation, where we have been able to establish a significant change depending on something else than the neighbourhood reform occurring between the two points in time. Our mind goes to cut-back policies following the bourgeois assumption of power in 1998. These policies may have brought about a preference for increased spending irrespective of party allegiance.

Politicians and citizens

There is a marked increase in contacts between citizens and politicians. In matters only concerning the contact seeking citizen herself or her next-of-kin the proportion of politicians which on the whole can report contacts of that kind during the last twelve months increases by eight percentage points. The percentage reporting more than five such contacts increases with nine points. Considering contacts in more general political matters the number having any contacts to report increases with nine percentage points and those with more frequent contacts with 13 percentage points. It may surprise that these rising contact frequencies do not leave any imprints in the measurements we have at the level of citizens. It should, however, be kept in mind that the two populations surveyed are dramatically different in magnitude. If one hundred politicians earlier not reporting any citizen contacts begin to report contacts, the percentage being contacted rises with 14 percentage points. if on the other hand one hundred citizens not having reported any contacts with politicians do report such contacts, they represent only a 0.01 percentage increase.

Table 8 Citizen contacts (percent)

	Never	1-5 times	6-10 times	More than10 times
Personal matters 1996	28	44	11	17
1999	20	43	15	22
Change	-8	-1	+4	+5
Political matters 1996	19	41	16	25
1999	10	36	21	33
Change	-9	-5	+5	+8

This contact activity is related to the structure of offices held. The "two-level" politicians are especially active. "Pure" neighbourhood politicians, however, account for the biggest increases.

Table 9 Citizen contacts per office category (percent)

	Neighbourhood	Overlapping	Central level	All
Personal matters 1996	67	84	67	72
1999	81	92	72	80
Change	+12	+8	+6	+8
Political matters 1996	82	91	77	82
1999	91	97	85	90
Change	+9	+6	+8	+9

The regression analyses show that there are significant positive effects of belonging to the group with overlapping offices as well as a positive significant effect of time. Belonging to the group with only central level office has a positive effect on contacts in political matters but not in personal matters. Party affiliation, on the other hand, as well as incumbency or opposition has no importance.¹⁰

10) The explained proportion of the variance is 5.6 percent for personal matters and 7 percent for political matters.

It has been observed that cut-back policies tend to lead to increased levels of citizens’ activity, and it has been somewhat acidly suggested that if high frequencies of contacts between citizens and politicians are a criterion of good democracy, democracy will be best enhanced by cutting public services to zero (Rubenowitz 1994). At the same time it has been expected that the neighbourhood reform by way of closing the gap between electors and elected would result in more contacts. What can now be concluded from these results looked upon in these conflicting perspectives? First it could be assumed that a closer relation between neighbourhood politicians and the citizens hardly can be expected to imply more contacts with those politicians only holding office on the central level. In this perspective it should be noted that we only have observed a significant effect of being a "pure" central level politician on contacts in more political matters. It could be that civil discontent in political matters results in citizens contacting politicians in city hall, and that the reported contact frequencies in political matters can be related to cut-back policies. In those matters concerning only oneself and one’s next-of-kin it is more natural to turn to the local representatives in the neighbourhood committee. From table 7 it can be concluded that the reported increase is greatest for neighbourhood politicians with regard to these "personal" matters, but that they are of approximately of the same size for all office categories if we look at the "political" matters.

In terms of the explicit goals of the neighbourhood reform this interpretation can be regarded positive. At the same time the interpretation is disturbing in relation to the role of the neighbourhood politicians. Members of the public who want to influence public policies turn to centrally located politicians, while neighbourhood politicians come to exercise the role of "the guy who can fix things" in the administration. It is the image of a patronage system that lurks. The position of the neighbourhood politician is dependent on the ability to fix solutions for his clients by-passing established administrative routines. In interviews with neighbourhood politicians in other municipalities than Stockholm (Bäck 2000), it appeared that many held exactly such an ideal. It was the archaic parish trustee that constituted the ideal image.

In evaluations of neighbourhood committee systems there often emerges what could be called "the queue syndrome": In queues in the supermarket, in the liquor store or at the bus stop committee members get into conversation with private neighbourhood residents concerning authority business (see e g Westerståhl 1995). In order to approach the question if the queue syndrome is present also in Stockholm’s neighbourhood committees we have asked our politician respondents how frequently they discuss local authority matters with people in their surrounding – in their family, at their workplace, with their neighbours and with other local residents. This very last alternative is intended to catch "the queue syndrome."

The category "other local residents" ranks relatively low compared to other more intimate groups. 75-83 percent of the respondents say that they very often or almost daily discuss local authority matters with their family, their colleagues or their friends. One such close category that also ranks low is "neighbours" with whom the respondents seldom discuss authority matters with. Why such discussions with neighbours are avoided of course is an interesting question that we, however, will not dwell upon any more. It can however be concluded that about half of the politicians discuss with people "they meet in town" and that this proportion is increasing between our two points of measurement. This can be taken as evidence of the Stockholm neighbourhood reform carrying with it a "queue syndrome."

Table 10 Discussions about local authority matters (per cent fairly often or daily)

	1996	1999	Change
Family	83	83	0
Colleagues	80	75	-5
Friends	79	80	+1
Other local residents	45	54	+9
Neighbours	34	37	+3

There also emerges a difference between the three office categories. On both occasions the "double politicians" holding office at the central as well as at the neighbourhood level display the highest activity levels, while the other two groups – the pure local level and the pure central level politicians – are less active.

These observations from the tables are corroborated by the regression analysis which displays significant positive effects from holding overlapping offices and from time. Altogether we think that these observations indicate an increasing discussion activity between the more leading neighbourhood politicians who also have elected offices in city hall and "ordinary" citizens thereby giving evidence for the Stockholm neighbourhood reform having "queue-syndrome effects" like those observed in other municipalities.¹¹

^{11) Perhaps it should be added that also affiliation to the left-wing parties has a significant positive effect on discussions with "other local residents." The regression model explains 3.9 percent of the variance.}

The politicians and association activities

It is not surprising that the respondents to the 1999 politician survey on the average are members of as much as 3.6 voluntary organisations¹². They constitute a group where also association activities should be expected. Holding elected posts in associations is somewhat more rare: Measured in the same way, the Stockholm local politician is a member of the board in 0.6 associations. It also is relatively unusual that those associations that one is a member of are local in the sense that they recruit their members and confine their activities to the very neighbourhood of residence. Here the average number is 0.8.

12) More accurately they are members of at least 3.6 associations, because the sum refers to the number of positive responses to membership in a number of categories of associations specified in the question.

The most usual associations for the politicians to be members of are organisations relating to residence (e g tenant-owners’ societies, house-owners’ associations or tenants’ associations), trade unions, consumer co-ops and sport clubs. If we instead look at the associations where our respondents are members of the board consumer-coops disappear while action groups and the like are added to the list. The profile of the local neighbourhood associations that the politicians engage in, on the other hand, is rather different. Associations related to residence also now rank high, but they are followed by action groups, parents associations and sports clubs.

Table 11 Association activities (percent)

	Members	Members of the board	Members of local neighbourhood associations
Residence associations	71,3	12,3	31,3
Trade unions	67,9	9,9	3,2
Consumer co-ops	58,4	1,5	2,4
Sports clubs	32,8	3,7	4,4
Action groups	25,2	4,3	10,6
Parents associations	13,4	1,7	4,7
Pensioners’ clubs	9,5	2,4	3,7
Business associations	7,3	1,5	0,4
Women’s organisations	6,7	0,9	0,7
Other organisations	64,7	17,0	15,5

If the neighbourhood reform should be expected to have some consequences at all in this respect it ought to concern activities in the association life in the local neighbourhood. It could be expected a deeper involvement in local associations among neighbourhood politicians than among central level city hall politicians. One could also expect an increase over time in this respect.

The regression analysis is only partly in the expected direction. The slope coefficients for "Moderate" and for "Central level office" are both significant and negative, i e Moderate party members are more seldom than supporters of the other parties members of local associations and politicians on the central city level are more seldom than politicians on the neighbourhood level involved in local neighbourhood organisations. On the other hand there emerges no significant time effect.¹³

13) Explained variance is 2.2 per cent.

Summary

I table 10 we have summarised our observations in terms of effects of "office structure" and "time". If there are any consequences of the neighbourhood reform for the studied aspects of the representative’s role the following patterns ought to emerge:

1. Change over time: We expect a change between the two points in time. It should, however, be kept in mind that the 1996 measuring is not in the strict sense an ex-ante measuring. The reform is already decided upon and to a certain extent running. It is above all when we think that how the reform is implemented is important that we should expect differences over time.
2. Systematic differences between the "office categories" are almost by definition consequences of the neighbourhood reform as the three categories are creatures of the reform. All such differences, however, cannot be seen as expected. We expect them to have a direction implying a closer relation between neighbourhood politicians and the local community.

With these considerations we have in the table shaded those effects that seem most consistent with the reform expectations.

Table 12 Summary of effects of office structure and time

	Office structure	Change over time
4 Contacts in political matters	Pure central level politicians most	Increase
1 Local affinity	Pure central level politicians especially weak
6 Local neighbourhood associations		No change
2 Contacts in personal matters The queue syndrome	Overlapping offices especially much	Increase
7 Legitimacy of local interest	Pure neighbourhood politicians especially much	No change
3 Willingness to spend		Increase
5 Dissatisfied with service distribution	No differences

Note: Numbers refer to discussion below

The following seven patterns emerge:

Change over time and difference between groups in the directions expected

1. The affinity of the politicians with their neighbourhoods is deepening and politicians only holding central level offices display an especially weak feeling of belonging. These results are in line with what may have been expected.
2. The number of politicians reporting contacts with ordinary citizens in matters only concerning the contacting citizen herself or her next-of-kin is increasing. Those politicians joining offices at the local and the central levels receive most of the contacts. What we termed "the queue syndrome" displays a similar pattern. Considering the important position the "two-level politicians" seem to have in the system, also these results are to be expected.
3. The willingness to spend more on the activities under the jurisdiction of the neighbourhood committees increases over time, and in this respect the "pure" neighbourhood politicians are the most ardent advocates.

Expected change over time, but no expected differences between groups appear

4. As expected the number of politicians being contacted by ordinary citizens in more political matters increases, but the politician category displaying the highest contact frequencies is the "pure" central-level politicians.
5. Also, as expected, the number claiming that local government services are unfairly distributed across the city’s territory increases. However, no significant differences emerge between the different politician categories.

Expected differences between groups but no change over time

6. Those politicians only holding office at the central city level are, as expected, the least involved in local associations in the neighbourhoods. On the other hand, there is no tendency that involvement in local associations is increasing.
7. The legitimacy of territorially defined particular interests does not increase over time, but "pure" neighbourhood politicians are eager advocates of the local interest against the common city interest.

Discussing changes between the two points of time when we did our measuring one should also consider other events in the city than the implementation of the neighbourhood reform. Are there other events that might have caused the observed changes in attitudes and behaviours? One such alternative explanation of course is the continued cut-down policies in social welfare, underlined by the shift of power from a Social Democratic to a bourgeois leadership dominated by the conservative Moderates in 1998. It seems reasonable to assume that cut-back policies rather than the neighbourhood reform are the causes of changes 3 (increased willingness to spend) and 4 (increased frequency of contacts in political matters). It is more difficult to make a credible argument for cut-back policies producing deepening local affinity (1), more citizen-politician contacts in "personal" matters and discussions between politicians and ordinary citizens when meeting "in queues" (2). In these cases the implementation of the neighbourhood reform seems to be a more credible explanation than cut-back policies. When it comes to the increasing number claiming local government services being less fairly distributed over the territory, it might be possible to find arguments for both standpoints.

A more far-reaching summary of the analyses and the interpretations made in this paper could be formulated into two propositions:

P1 The neighbourhood reform has carried with it an increased feeling of local belonging to the neighbourhoods among Stockholm’s local politicians. A local politician role is emerging, implying at least among the more leading neighbourhood politicians frequent discussions of local authority matters with ordinary citizens they happen to meet "in town" and frequent contacts with ordinary citizens seeking help to "fix things" for themselves. Neighbourhood politicians also are somewhat more involved in local associations than "pure" city hall politicians.

P2 The territorial dimension in local politics is becoming more important. The opinion that there are imbalances in the distribution of local services between the different parts of the city is more commonly held. Territorially defined particular interests become relatively legitimate, especially when considering politicians lacking a direct affiliation with city hall.

A development characterised in this way might point in the direction of an important transformation of the structure of the political system in the city. On the local neighbourhood tier there are traces to be found of a more old-fashioned – pre-modern – political system, where individual politicians play a role for the contacts between individual citizens and the welfare municipality. In city hall, on the other hand, the drama of modern politics is at stage. This play now and then takes its expression as cuts in the welfare production to be distributed at the local level. This will lead to tensions and a fermenting discontent can be discerned even among local politicians. Those local politicians holding offices on both levels fill a function as integrators in this increasingly dualistic system. Other integrating mechanisms not touched upon in this paper are the system of steering by objectives, the resource allocation system, and perhaps most important of them all the steering and co-ordination channelled through the political parties. This last form of steering and co-ordination is streamlined by the procedure of city-wide indirect elections, guaranteeing the same political majority in all neighbourhood committees as well as in the city council. The overlapping of public offices – in the French doctrine known as cumule des mandats – is still an integrating mechanism. In the same way as it is asserted that this system tends to fuse the levels of French government (cf Hague et al 1992), overlapping offices contribute to fusing the government of Stockholm into a system where it is not obvious that it is relevant to pose the question which level is steering.

References

Banfield E C1961 Political Influence: A New Theory Of Urban Politics New York: The Free Press

Bäck Hanna 2000 "En ny ’regering’ tillträder i Stockholm" in Henry Bäck & F Johansson (eds) Mellan samlingsregering och parlamentarism. Studier i genomförandet av begränsat majoritetsstyre i Stockholms stad Stockholm: IKE

Bäck Henry 1993 Hur många och hurudana? Om organisationsförändringar, politikerantal och representativitet i kommunerna Stockholm: Institutet för kommunal ekonomi

Bäck Henry 1996 "Decentralisation, Privatisation and Representativeness in Local Government" in P Bogason (ed) New Modes of Local Government Organizing: Local Government Fragmentation in Scandinavia Commack NY: NOVA Science Publishers

Bäck Henry 2000 Kommunpolitiker i den stora nyordningens tid Malmö: Liber

Bäck Henry & F Johansson 2000 "Reinventing the wheel: Institutional reforms in the city of Stockholm" EURA Workshop, University College Dublin, April 13-15

Bäck Henry & F Johansson 1997 Medborgarna, politikerna och stadsdelsnämndsreformen. Redovisning av enkätundersökningar hösten 1996 Stockholm: IKE

Clarke M & J Stewart 1989 "The Future of Local Government: Issues for Discussion" Birmingham: Institute of Local Government Studies

Eulau H, J C Whalke, W Buchanan & L C Ferguson 1959 "The Role of the Representative: Some Empirical Observations on the Theory of Edmund Burke" American Political Science Review No 3

Hague R, M Harris & S Breslin 1992 Comparative Government and Politics: An Introduction London: Macmillan

Jönsson S, S Rubenowitz & J Westerståhl (eds) 1995 Decentraliserad kommun - exemplet Göteborg Stockholm: SNS Förlag

Jönsson S, L Nilsson, S Rubenowitz & J Westerståhl 1997 Decentraliserad välfärdsstad. Demokrati, effektivitet och service Stockholm: SNS Förlag

Larsen H O & A Offerdal 1992 Demokrati uten deltakere? Arbeidsvilkår og lederroller i kommunepolitikken Oslo: Kommuneforlaget

Lewin L 1998 "Bråka inte!": Om vår tids demokratisyn Stockholm: SNS Förlag

Milbrath, Lester W. 1972. Political Participation. Chicago: Rand Mc.Nally & Company.

Montin S 1989 "Från demokrati till management. Decentralisering inom kommunerna" Statsvetenskaplig tidskrift Vol 92 Nr 2

Nilsson L 1993 Stadsdelspolitikerna i Göteborg 1989-1991 Göteborg: Statsvetenskapliga institutionen

Parry G, G Moyser och N Day 1992 Political Participation and Democracy in Britain Cambridge: Cambridge University Press

Pateman, Carole, 1980. Participation and democratic theory. Cambridge: Cambridge University Press.

Premfors R, M Sanne & J Sandqvist 1994 Demokrati i storstad. Stadsdelsnämnder i Stockholm Stockholm: Carlssons

Rothstein B, P Esaiasson, J Hermansson, M Micheletti & O Pettersson 1995 Demokrati som dialog. Demokratirådets rapport 1995 Stockholm: SNS Förlag

Rubenowitz S 1995 "Medborgarna och stadsdelsreformen" in S Jönsson, S Rubenowitz & J Westerståhl (eds) 1995 Decentraliserad kommun - exemplet Göteborg Stockholm: SNS Förlag

Sandqvist J 2000 Hur Stockholm fick stadsdelsnämnder. Den historiska bakgrunden till SDN-reformen och beslutet om indirekta val Stockholm: IKE

Solli R 1988 Decentralisering i kommuner. Om politik och administration på lokal nivå Lund: Studentlitteratur

SOU 2000:1 En uthållig demokrati

Westerståhl J 1995 "Rollen som stadsdelspolitiker" in S Jönsson, S Rubenowitz & J Westerståhl (eds) Decentraliserad kommun - exemplet Göteborg Stockholm: SNS Förlag

Whalke J C , H Eulau, W Buchanan & L C Ferguson 1962. The Legislative System New York: Wiley

Young K 1990 "The Changing Context of the Member’s Rôle" in L Greyson, M Hobson & K Young Inlogov informs on The Members Role Birmingham: Institute of Local Government Studies