Can Civil Society Control Politicians?

Transcription

1 European Consortium for Political Research Joint Sessions of Workshops Uppsala, Sweden, April 13-18, 2004 (Workshop 8: Changing Structure of Civil Society) Can Civil Society Control Politicians? A Case Study From Local Politics in Hungary by G á b o r S o ó s Tocqueville Research Center 7 Bathory, 1054 Budapest, Hungary gabor.soos@t-rc.org

2 Societal Accountability Liberal democracies are built on representation and, as a consequence, the relative autonomy of ruling elites. Therefore, they need accountability as a mechanism to link representatives to those who are represented. Representation is a basic characteristic of modern government. Dahl (1970) lists three reasons why citizens accept the delegation of decisions to their representatives: (1) Criterion of Personal Choice: decisions correspond with personal choice. (2) Criterion of Competence: decisions require a level of competence ordinary people do not have. (3) Criterion of Economy: delegation economizes on the amount of time, attention, and energy, which otherwise citizens should invest. Strøm (2002:3-4.), in the social choice tradition, identifies three motivations for political delegation: (1) Capacity: Citizens' capacity is limited by the time and transaction costs they are able and willing to spend on collective decision-making. (2) Competence: Policy-making often require an amount of specific knowledge and skills. (3) Collective action problems: Citizens may have different interests as individuals and as a group. Agency is needed to coordinate the activities of group members. The point is that, for good reasons, citizens accept the transfer of their powers. In a democratic polity, the principle of the popular rule entails political accountability. Delegation in general, and representation in particular, implies accountability as remedy for the abuse of political power. A basic distinction is made between horizontal and vertical accountability (see e.g. O'Donnell 1999). In a bit simplified way, the former refers to the intrastate checks and balances, while the latter to the relationship between representatives and the electorate. My interest here is the vertical accountability. 'Political accountability' refers to a relationship in which A is accountable to B when A is obliged to to inform B about A's (past and future) actions and decisions, to justify them, and to suffer punishment in the case of eventual misconduct. (Schedler 1999: 17) The definition highlights two duties of representatives: transparency (information giving) and justification (explaining position). The disciplining mechanisms built in accountability is the enforcement of responsive behavior by sanctions, i.e. a system of reward and punishment. The most often discussed of accountability is electoral accountability. Electoral accountability is invariably regarded as a crucial function of competitive elections. For Powell (2000: 10-12), for instance, the two fundamental role of elections are the evaluation of the incumbent government (i. e. accountability) and the choice of future government (i. e. electoral mandate). Citizens have the opportunity to throw the rascals out. Removal from office is the central concern of the definition of Przeworski et al. (1999): Governments are 'accountable' if citizens can discern representative from unrepresentative governments and can sanction them appropriately, retaining in office those incumbents who perform well and ousting from office those who do not. An 'accountability mechanism' is thus a map from the outcomes of actions (including messages that explain these actions) of public officials to sanctions by citizens. Elections are a 'contingent renewal' accountability mechanism, where the Page 2 of 21 pages

3 sanctions are to extend the governments tenure. (Przeworski et al. 1999:10) Electoral accountability has an effect on the policy-making between elections too. The anticipation of future evaluation makes policies more responsive to citizens' needs and wants. Thus, electoral accountability links decision-makers' behavior to citizens' demands. Electoral accountability has some well-known weaknesses. First, citizens often select based on promises (prospective voting) and not the evaluation of past performance (retrospective voting). Second, the average citizen does not always have an adequate level of information. Thus, their decisions are often superficial. Third, elections are rare events and voters' memory is short. Consequently, politicians can act quite freely between elections. (See Manin et al ) Civil society may play an important role to overcome these problems. Peruzzotti and Smulowitz (forthcoming; 2000) conceptualize this civic role as 'societal accountability'. They acknowledge that competitive election as authorization mechanism have no practical alternative. Nevertheless, there are (and must be) other vertical tools for holding representatives accountable. Societal accountability is a different form of vertical control of political players, based on institutional as well as non-institutional mechanisms: How do these civil society-actors accomplish such goals? They do so by a) monitoring the behaviour of public officials and agencies to make sure that they abide by the law, b) uncovering and exposing cases of corruption or governmental wrongdoing and pushing to bring sanctions against the lawbreakers, c) activating, in many instances, the operation of intrastate agencies of control, such as the judiciary or parliamentary investigative commissions, that otherwise would not be activated or would be activated in a biased and politicised way and, d) pressing for institutional reforms that could improve the effectiveness of mechanisms to make authorities accountable. (Peruzzotti 2003a) Societal accountability relies on the initiatives of civil society organizations and media outlets. As this type of accountability does not require the support of the majority, it can effectively work between elections and with the participation of a few committed citizens. It does not claim to be an authorization mechanism, it preserves the separation between representatives and voters (Peruzzotti 2003a). Without authorization, civic initiatives for societal accountability cannot directly sanction unresponsive or illegal political practices. Nonetheless, they have indirect tools by mobilizing other kinds of accountability mechanisms. First, civil society organizations may resort to the mechanisms of horizontal accountability. They can force higher administrative bodies or the judicial system to step in and punish illegal activities. Second, civic groups through the media can call voters' attention to the gap between promises and actual policy-making. By naming and shaming, civic organizations can influence the electoral chances of unresponsive representatives. As professional politicians want to get reelected, the anticipation of electoral defeat prevent the abuse of power. The introduction of this concept has a considerable theoretical relevance, as not only does it give a responsibility to civil society actors in controlling politics, but it also shows a mechanisms through which this control is (or may be) exercised. The Case This paper sets out to assess the degree to which civic organizations in Hungary hold local government representatives accountable. Specifically, it investigates the effect of local Page 3 of 21 pages

4 representatives' formal commitment to civil society. The focus is on two specific aspects of the political involvement of civil society: membership of local representatives in civil society organizations and the electoral support (nomination and endorsement) given by associations to local representatives. Do these commitments result in more control of civil society over local politicians? Do civil society organizations hold local representatives formally committed to them accountable and are those local politicians more responsive to societal needs? These are the questions to be answered here. First, I present the level and determinants of civic membership and electoral support. Then, I analyze the degree of societal accountability by means of survey data. Finally, I attempt to show to what extent civic commitments and societal accountability matter. I use the term 'societal accountability' in a broader sense than Peruzzotti and Smulowitz (forthcoming). I refer to all kinds of accountability claims made by civic organizations, even if those claims do not receive much publicity. I focus here on only one possible source of societal accountability, the formal commitments of representatives to civil society organizations. 'Formal commitment' here includes membership in civil society organization and the electoral support (nomination and endorsement) of those organizations. By 'civil society organization' or 'civic organization' I mean formal voluntary organizations such as associations, trade unions and foundations. Although informal groups are certainly important parts of civil society (especially in the Post-communist context), my focus is precisely on the influence of the formal commitments of local politicians. The term 'local representatives' includes both mayors and councilors. Mayors are directly elected in Hungary. Mayors are professional politicians in municipalities of more than 3,000 inhabitants (though full-time mayors are not uncommon in smaller villages). Their election takes places at the same time and for the same period (four years) as that of the council. Depending on the population of local government, the size of council varies from 3 to 65. Only a few councilors work full-time, usually vice-mayors in larger cities. Municipalities of less than 10,000 inhabitants directly vote on candidates on a common list ( short ticket ). Voters can select as many candidates as there are places in the council. This number ranges from three (below 100 inhabitants) to thirteen (between 5,000 and10,000 inhabitants). Municipalities of 10,000 or more inhabitants have a mixed electoral system. Approximately 60 percent of representatives are elected in individual electoral districts. The representatives in constituencies are elected by simple majority. To make representation more proportional, the votes of the candidates who are on party lists and were unsuccessful in their individual district are summed up and about 40 percent of the places are distributed among the lists. Representatives elected from the list represent the whole city, not a smaller territory. As each settlement has the right to form an own local government, the Hungarian local government system is highly fragmented. In a country of ten million, there are more than 3100 local governments. Most of them have 1,000 or less inhabitants. This has a consequence for the contribution of civic organizations to local democracy, since there is no, or only a limited number of, organizations in many small places. The main source of data is a survey of almost 1,000 Hungarian local representatives which has been conducted in the beginning of 2003 in the framework of the Indicators of Local Democratic Governance project, initiated and funded by the Local Government and Public Service Reform Initiative ( and administered by the Tocqueville Research Center ( The paper also draws on the surveys of the Hungarian Central Page 4 of 21 pages

5 Statistical Office in 1990, 1994, 1998, and Electoral data and general data on civil society are also used. Local Representatives in Civic Organizations As Table 1 demonstrates, belonging to civic organizations is a fairly important element of local representatives' life. Altogether approximately half of the Hungarian local representatives have membership or position in civil society organization. The first sign of political importance is the fact that the ratio of mayors holding membership is higher (61% in 2002) than the proportion of councilors (48% in 2002). The more professional, full-time politicians seem to need more formal contacts than the more amateur ones. The second important observation is that more and more representatives are involved in civil society activities. Table 1 shows a clear upward trend. The increase is especially large among mayors whose total membership ratio went up by almost 12 percentage. Thirdly, membership organizations (mainly associations) are more important for local representatives than foundations. Two explanations, which are not mutually exclusive, offer themselves. A supply-side interpretation would refer to the fact that there are much more membership-based organization than foundations; therefore, a local representative (as any other citizen) have more chance to join an association than a foundation. The demand-side, political explanation would focus on the political ambition of politicians. In most cases, associations or trade unions can enhance the political capital of local politicians more effectively than foundations. Table 1. The proportion of local representatives holding membership in civil society organizations Local Representatives Mayors Councilors Foundation Association Total 11,80% 17,90% 17,60% 21,40% 27,10% 24,10% 10,60% 16,60% 17,20% 37,90% 41,90% 43,10% 39,80% 48,90% 54,80% 37,60% 41,00% 42,50% 42,70% 48,70% 49,40% 48,90% 58,20% 61,30% 41,90% 47,40% 48,00% Source: Central Statistical Office Surveys of local representatives in 1994, 1998, and The logistic regression analysis of civil society membership (see Table 2) makes the profile of the socially active local representative more accurate. Municipality size has a great effect and consistent effect on membership ratio: the larger the municipality, the more likely the membership of local representatives. This can be put down to two factors. First, the density of associational life goes parallel with the population of municipalities. More urbanized life offers more opportunities to join. On the demand side, population well correlates with societal complexity, which requires more organization from politicians. In complex local societies, successful politicians must extend their network through organizational means. The most puzzling phenomenon coming out of the analysis is the gender gap concerning civil society membership. Women councilors and mayors are consistently less involved in civil society organizations than men, even if lots of other variables are taken into account. Page 5 of 21 pages

6 The influence of education seems clear: the more educated a representative, the more likely he or she committed itself to a civil society organization. The interpretation of this finding is far from being self-evident. One may argue that education increases the chances of formal networking, while less educated local politicians rely more on informal contacts. The mayoral effect, observed above, may also be found in all regression analyses. One aspect of the increasing trend also appears in the results. Reelected councilors and mayors are more likely to have membership in civic organization than newcomers. Accumulated political capital includes civil society membership, and, often, board memberships in associations and foundations. An even more detailed picture can be gained from the analysis of the Local Representative Survey, conducted by the Tocqueville Research Center. Respondents in this survey indicated their function (not simply membership) in foundations, associations, and trade unions. Thus, the results are somewhat different from the above-presented ones. Altogether 43 percent of local representatives are formal leaders in civil society organizations. A small percentage (3.6%) of these local representatives have an office in trade union. Membership in associations is the most frequent form of civil society activity (61%), while approximately one third (35%) of councilors and mayors are board members in foundations. The analysis (see Table 3) confirms the importance of population size, but shows a declining effect in larger municipalities (above 50,000 inhabitants). This is a point I will return to in the next section. The gender difference is just the opposite as in the previous analysis, adding even more to the puzzling nature of the problem. More women (47%) have function than men (43%). Education becomes insignificant when cultural capital (measured by and internet usage and knowledge of foreign languages) and position in workplace (measured by the number of subordinates) are included in the analysis. The influence of cultural capital demonstrates that it is not formal education that matters, but the most mobile and specific forms of knowledge. The importance of workplace position still supports the hypothesis of formal/informal networking. Mayoral position and reelection once more proved to be important and party membership unimportant. Two variables indicating political ambition reinforces the supposition that membership is partially of political nature. Those who aspire a higher position (mayor, MPs, etc.) and those who participated in some kind of political training are more often involved in civil society activities. Interestingly, ideological self-placement also correlates with position in civil society organization. The more rightist the respondents regard themselves, the more likely they have an office in civil society organization. Socialization also seems to be important. Those whose parents had an office in social organization are much more likely (59% of them) to have a similar office than those who did not have such function (40%). To conclude, it is quite clear that personal (sociability) and political (coalition building) motivations are mixed. While certain indicators of social background (education, cultural capital, managerial position, socialization) shows social motivation, the importance of political variables (mayoral position, reelection, ambition) clearly points to a political incentive to be socially active. Page 6 of 21 pages

7 Supporting Candidates The Hungarian electoral allows civic organizations to nominate candidates in both the short ticket and mixed system and have 'compensation lists' in the mixed system (municipalities of 10,000 or more inhabitants). One may draw some conclusions from the electoral data (Table 4). First, the proportion of party-nominated councilors and mayors has been decreasing since This may not be related to the ever stronger anti-party feeling in Hungary (and elsewhere in the contemporary democracies). The explanation may lie in the increasing trend of Hungarian politics toward a two-bloc system, in which the two big parties (ex- Communists and rightists) and their allies run only two candidates. As a consequence, less candidates are nominated by parties, but they are nominated by more than one party (often three or four). Second, civic organizations, alone or in coalition with political parties, nominate more and more candidates. This process is not parallel with the decline of party nominations; it is not former party candidates change their political color (though this is not uncommon either). The rest of the section will portray the profile of these representatives. Third, it is clear that mayors are less dependent on the electoral support of organizations. This may partly be explained by their higher name recognition (especially incumbents enjoy this advantage). Participation in local political life has accumulated the political capital of some local politicians since 1990 and, consequently, they stand in less need of the orientating labels of political parties or civic organizations. Another reason for this phenomenon is the different character of mayoral elections. The political 'trade marks' of organizations often suggest the representation of particular interests and a 'quarrelsome' political culture. The organizational life is usually weak in smaller local governments and the issues are rather nonpartisan, so such negative associations do not help would-be mayors, who must represent the whole municipality Table 4. Nominations at local elections Mayors Councilors Nominating organization Independent % Party Civic organization Party and civic organization Total Party nominated Civic nominated Source: Central Statistical Office (2003). Bőhm and Szoboszlai (1992). The breakdown of the electoral results by electoral systems (in Table 5) provides another view. The importance of the electoral support of parties and civic organizations is minimal in local governments of 10,000 or less inhabitants. Neither the level of organization of local society nor the electoral system encourage the seeking of organizational support. The situation is completely different in larger municipalities. The ratio of independent representatives is very low in towns and cities, while nine out of ten representatives were Page 7 of 21 pages

8 nominated by at least one political party and half of them by at least one civic organization in In this category, the significance of both parties and civic organizations has been growing since the establishment of the local government system. More and more civic organizations run compensation list on the local elections. Table 5. Nominations at local elections by electoral systems Nominating organization 'Short ticket' (less than 10,000 inhabitants) Electoral districts Compensation list Independent 81.8% 86.8% 86.8% 8.7% 9.3% 5.9% Party 10.1% 6.1% 4.7% 48.2% 43.0% 42.8% 64.8% 52.8% 38.2% Civic organization 4.5% 4.9% 6.5% 7.8% 9.7% 9.1% 13.8% 22.7% 29.1% Party and civic organization 3.6% 2.2% 2.0% 35.3% 38.0% 42.2% 21.4% 24.5% 32.7% Total 100.0% 100.0% 100% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Party nominated 13.7% 8.3% 6.7% 83.5% 81.0% 89.0% 86.2% 77.3% 70.9% Civic nominated 8.1% 8.3% 8.5% 43.1% 47.7% 51.3% 35.2% 67.2% 61.8% Source: Central Statistical Office (2003). Unfortunately, there are no data collection of the political nature of civic organizations. Thus, nobody knows how many of the civic organizations nominating candidates at the elections belong to the organizational background of political parties, how many function as local party and how many are truly civic groupings of citizens. Who are nominated by civic organizations? The most important factor in the analysis (see the logistic regression in Table 6) is size. Civic nomination is the most frequent in town local governments. In all probability, this can be explained by the combined effect of the electoral system and the characteristics of the local society. The importance of organizational support is much higher in the mixed system where only organizations can have compensation lists and the duels in electoral districts require a great deal of effort of coalition building. On the other hand, parties in this municipality category are not so well organized as in cities, providing much political room for local civic organizations. Membership in civic organization obviously play a big role in getting the nomination of civic organizations, which apparently like supporting their own members. As in the case of membership, women are less likely to get nominated by civic organizations. Party members are less likely to have civic nomination in addition to their party's. There is a big difference among right and left in this respect, as candidates of rightist (noncommunist) parties are more frequently receive a civic nomination than leftist candidates. Newly elected representatives were significantly more often nominated by civic groups. Nomination by civil society organizations seems a means of recruitment into local political elites. Civic organizations may politically support candidates. The determinants of public endorsement (shown in Table 6 in the Appendix) are similar to those of nomination. One important difference is that mayors, who often prefer running as independent candidates, received the support of civic organizations much more often than any other category of local representatives. There is two apparent explanations for that. The electoral support of civil society groups divides the electorate much less than that of parties. On the other Page 8 of 21 pages

9 hand, mayors organize their followers often in the form of associations. Such organizations may not qualify as ideal civic groups. Another difference between nominations and supports is that incumbent representatives were more likely to get civic endorsement than newcomers. While the nomination of civil society organizations is an effective way to enter into the club, club membership can be preserved by a coalition of supportive organizations. The results of the Local Representative Survey of the Indicators Project are similar, so they are not presented here. Summing up, a significant minority of mayors and councilors were nominated or publicly endorsed by civil society organizations. In the category of 10,000 or more inhabitants (where almost sixty percent of the population live), more than half of the local representatives were nominated or at least supported by civil society organizations. The question now is to what extent civic organizations can translate political support into influence. Holding Accountable Does membership in civic organizations and their electoral support matter? Do civic organizations contribute to societal accountability? A tentative answer can be gained by the analysis of a question of the Local Representative Survey: How frequently are you requested by the following people to explain your position on a local public issue? The following table shows the level of societal accountability as the answerability of local representatives. Table 7. Frequency of accountability claims Actors (1) Not at all (2) Only a few times a year (3) Once or twice a month (4) Every week Mean Friends, family members 3% 13% 35% 50% 3,3 910 Colleagues at work 12% 20% 36% 33% 2,9 835 Neighbors 7% 34% 40% 19% 2,7 892 Constituency 4% 24% 42% 30% 3,0 875 Members of civic organizations 40% 31% 22% 8% 2,0 983 Fellow party members 76% 7% 10% 6% 1,5 983 Businesspersons 31% 45% 20% 4% 2,0 769 Journalists 68% 26% 5% 2% 1,4 983 Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project Who holds councilors and mayors accountable? First of all, their family, friends and colleagues! While local representatives claim they represent the whole or a part of the municipality, or a political group, or a social group, they actually have more political communication with their inner circle. As Table 8 shows, this is not a consequence of the fragmented nature of the Hungarian local government system, i.e. the large number of small municipalities. Representatives in small and large municipalities alike explain their political record to people who are very close to them. Table 9 demonstrates that there is no difference between mayors and councilors: full-time politicians are as often (and very often) are asked by friends and family as the usually amateur councilors. A survey of urban mayors also confirm this: powerful leaders of cities are primarily held accountable Page 9 of 21 pages N

10 by their closer friends and family members. This is a form of interpersonal accountability we know little. More than 70 percent of local representatives must respond to political questions of their constituency. However, the relatively high correlation (see Table 10 in the Appendix) between the frequency of accountability claims by constituency, on the one hand, and family, friends and colleagues, on the other hand, points to a partial overlapping of these groups. What local representatives perceive as 'constituency' often means the people who they more frequently meet at home or workplace. What makes the difference between small and large municipalities (see Table 11 in the Appendix) is the frequency of accountability claims made by organizations: parties, media, and civic organizations. While local representatives in smaller local governments receive accountability claims from 3.1 types of actors, this number is 4.4 in municipalities of 10,000 or more inhabitants. This increase is due to the higher accountability to the representatives of political parties, media outlets and civic organizations. Civic organizations relatively often ask local politicians to explain their political position (see Table 12 below). Even in local governments of less than 10,000 inhabitants, where only a few organizations exist, every fourth representatives are raised questions at least once a month. This proportion is much higher (46%) in larger municipalities. Belonging to civil society organization induces more frequent accountability claims. The difference between municipality categories is not that large, though civic organizations in towns and cities seem somewhat more active. More than half of the representatives in large municipalities and a little less than half of them in small municipalities are regularly asked by civic organizations. Every sixth of those who have a position in civic organization receives accountability claims every week. Table 12. Accountability claims by civic organizations Frequency of accountability claims Representatives with position Small municipalities Large municipalities Total Small municipalities All representatives Large municipalities (1) Not at all 22% 10% 20% 43% 17% 40% (2) Only a few times a year (3) Once or twice a month Total 33% 37% 34% 30% 37% 31% 30% 37% 31% 20% 34% 22% (4) Every week 15% 16% 15% 7% 12% 8% N Small municipalities: less than 10,000 inhabitants. Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project The multivariate analysis (logistic regression Table 13 in the Appendix) does not reveal surprising associations. The frequency of accountability claims increases by municipality size. Unlike membership and electoral support, the correlation is linear. The accountability relationship is between two parties;therefore, the willingness of politicians also matters. This is the reason why mayoral position, ambition and sophistication (cultural capital) have a significant effect. Both position in civic organization and the electoral support of such organizations increases Page 10 of 21 pages

11 the likelihood of accountability claims by civic organizations. What is even more important, the combined effect of position and support is the most significant factor in the multivariate analysis. Civic organizations that supported one of their leaders at local election seem to impose the heaviest demands for accountability. To sum up, local politicians are most often asked by their inner circle, but accountability claims are often made by civil society organizations too. Almost half of all local representatives (and even more in towns and cities) are asked at least once a month to explain their political position. The extent of 'civic accountability' is clearly influenced by electoral support and position in civic organization. Shaping Opinion Do civic commitments and accountability have a political relevance? Do local representatives having a membership and/or supported by associations at elections listen more to civil society organizations? Are local politicians more often forced to give answer to uneasy question more responsive to the needs of civil society? I attempt to respond to these questions by analyzing the variance of a multi-item question of the Local Representative Survey. The question was as follows: How influential does the opinion of the following people seem on the formation of your own opinion on local government issues? Respondents might indicate the level of influence on a seven-point scale. The next table shows the perceived influence of various actors listed in the questionnaire. Representatives of civic organizations have a medium influence (3.6) on councilors and mayors. (The low influence of media can be attributed to the simple fact that there is usually no media outlet in villages, i. e. in the majority of Hungarian local governments.) Table 14. Opinion formation of local representatives Political actors Mean influence Local citizens 5.51 Members of own faction 5.09 Mayor 4.14 Representatives of own party 3.97 Head of the municipality administration 3.92 Representatives of NGOs 3.55 Members of other factions 3.39 Civil servants of local administration 3.24 Representatives of local business 3.39 Local church leaders 2.94 Representatives of other parties 1.97 Journalists 1.93 Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project What determines the level of perceived influence of civic organizations? The first linear regression analysis (shown in the Appendix as Table 15) clearly demonstrates that, even if many other potentially relevant variables are controlled for, both civil society membership and the electoral support of civil society organizations matter. It is very important that having an office in civic organization is of more importance than electoral mandate. What seems to count is rather the frequency of contacts than the one-time support at election Page 11 of 21 pages

12 times. This hypothesis is also supported by the fact that the electoral support of civic organizations in which the respondent is a member has a higher coefficient than that of organizations which nominated or endorsed the respondent as an outsider. The population size of local governments also has a large influence, but this effect is not linear. Similarly to what was presented earlier, it reaches its peak (4.04) in the town category (between 10 and 50 thousands inhabitants) and is somewhat lower in the cities (3.50). Mayors are more exposed to external influence, so their perception of stronger influence is not surprising. What is startling is the fact that political ambition has no significant effect on the degree to which local representatives listen to civil voices. When, however, age is taken out of the analysis and a multicollinearity effect is eliminated, ambition becomes significant. (Age correlates with ambition very well, as you have more limited ambitions after a certain age.) The effect of socialization is also significant: those whose parents had function in some kind of social organization tend to have a leading position. The inclusion of accountability (see Model 2 in Table 15) largely increases the explained variance (by 11 percentage) and the coefficients radically change. All the variables mentioned above lose a large part of their explanatory power. The electoral support of civic organizations becomes insignificant, which clearly shows that a one-time support without subsequent monitoring or the support of organizations that are formed around, and for, an influential politicians have less effect on the opinion-formation of local representatives. To summarize, accountability has a highly significant influence on the opinion formation of councilors and mayors and the effect of membership and electoral support are to a large extent mediated through accountability claims. Conclusions Civil society is often claimed to have democratic functions. For instance, Diamond (1995) discusses ten different democratic effects of civil society, while Mark Warren (2001) identifies twelve democratic outcomes of civil society in three groups (developmental effects, public sphere effects, and institutional effects). All such theoretical efforts includes the idea that civil society has a role to prevent abuse of office by controlling politicians. The goal of this paper was to investigate to what degree this control actually exist. The case study focused on mayors and councilors in Hungarian local governments. Do membership in civic organizations and their electoral support have a political relevance? The answer is rather affirmative. About half of the local representatives in Hungary have a membership in civil society organizations and this proportion is increasing. Half of the candidates in the mixed electoral system (i.e. in the municipalities of 10,000 or more inhabitants) were supported by civic organizations. These formal linkages often lead to accountability claims by civic organizations, which attempt to control their members who hold a political position and those who ran under their flag at the last local election. What is more, these accountability claims actually affect the opinion formation of local representatives by making politicians more attentive to the voice of civil society organizations. The quantitative analyses also showed that the frequent monitoring of representatives' activities between elections is more important than the simple electoral support. Page 12 of 21 pages

13 This paper is not more than a first step. There are many pending questions here. I am presenting only some of them here. As many point out (e. g. Edwards et al. 2001), some civil society organizations are conducive to democracy, but others are not. The simple fact that the civil society commitments of local representatives matter may not be a good news to democracy. A similar problem is that certain accountability claims may not serve the common good, but particular interests. The data used here cannot distinguish between civic organizations that have public concerns and 'local parties', i.e. organizations that fully devote themselves to (municipal) politics and function as a political party. A functional, not legal distinction between political and civil society may result in different findings. To continue with theoretical problems, Schedler's (1999) definition quoted in the beginning is not fully covered in the paper. The evidence presented here has told much about answerability. Nonetheless, the survey questions used here do not adequately distinguish between transparency (a simple provision of information) and justification (defense of political position). More importantly, the available quantitative data did not shed much light on the sanctioning mechanisms through which representatives are disciplined. An additional theoretical problem that did not receive enough attention here is the deliberative nature of everyday accountability discourses. A measurement problem is the validity of the accountability question analyzed in the paper. One cannot know to what extent the questions were actually uneasy. I may be more appropriate to measure the strength of accountability relationships not only by the frequency of accountability claims, but also by the depth and publicity of questions. It would not be difficult to raise even more theoretical and methodological questions. What seems sure to me is that the problem of societal accountability deserves more attention and require more quantitative and qualitative research. Page 13 of 21 pages

14 Appendix Table 2. Determinants of civil society membership (CSO data) Variables Population (reference group: less than 2,000 inhabitants 2,000-10,000 10,000-50,000 more than 50, *** 2.20*** 2.60*** 1.59*** 2.26*** 2.20*** 1.57*** 1.95*** 2.20*** Gender (male) 1.33*** 1.43*** 1.32*** Education (reference group: primary) trade school secondary school degree 1.78*** 2.30*** 3.80*** 1.50*** 2.17*** 3.60*** 1.67*** 2.35*** 4.12*** Mayor 1.20** 1.39*** 1.67*** Not independent candidate 1.52*** 1.45*** 1.98*** Party membership (reference group: rightist) leftist no party membership Councilor in local administration before *** ** * 1.12* 1.06 Reelected representative 1.33*** 1.43*** 1.56*** Nagelke R % 15.2% 16.8% Source: Central Statistical Office Surveys of local representative in 1994, 1998, and Page 14 of 21 pages

15 Table 3. Determinants of civil society membership (Indicators data) Variables Odds ratios (significance) Population (reference group = less than 1,000 inhabitants 1, ,000-4,999 5,000-9, ,000-49,999 more than 50, *** 2.89*** 2.99*** 4.51** 4.12** Gender (female) 1.61** Age not sign. Education not sign. Cultural capital 1.26* Manager 1.53** Party membership not sign. Ambition 1.70** Political training 1.86*** Mayoral position 2.29*** Ideological self-placement (left to right) 1.18*** Reelected representative 1.66*** Parents' civil society position 1.69** Nagelke R % N 781 Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project Page 15 of 21 pages

16 Table 6. Determinants of nomination and endorsement by civil society organization (CSO data) Variables Nomination Endorsement Population (reference group: less than 2,000 inhabitants 2,000-10,000 10,000-50,000 more than 50, *** 26.10*** 9.33*** 3.80*** 5.80*** 5.20*** Gender (male) 1.53*** 1.15 Education (reference group: primary) trade school secondary school degree Way of election (ref. group: short ticket) Mayor Electoral district Compensation list *** 2.27*** Membership in civil society organization 6.10*** 2.80*** Party membership (reference group: no party membership) leftist rightist 0.06** 0.18** 0.44*** 1.58* Councilor in local administration before Reelected representative 0.77*** 1,43*** Nagelke R % 18.0% Source: Central Statistical Office Surveys of local representative in 1994, 1998, and Page 16 of 21 pages

17 Table 8. Frequency of accountability claims Mayors of municipalities of more than 10,000 inhabitants Actors (1) Not at all (2) Only a few times a year (3) Once or twice a month (4) Every week Mean Friends, family members 3% 13% 35% 50% People with whom you work 12% 20% 36% 33% Your neighbors 7% 34% 40% 19% Your constituency 4% 24% 42% 30% Members of civic organizations 40% 31% 22% 8% Fellow party members 76% 7% 10% 6% Businesspersons 31% 45% 20% 4% Journalists 68% 26% 5% 2% Source: European Mayor Survey 2003 conducted by the Tocqueville Research Center in Hungary. Table 9. Frequency of frequent accountability claims by population categories Mayors of municipalities of more than 10,000 inhabitants Actors Smaller municipalities (less than 30,000 inhabitants) Larger municipalities (more than 30,000 inhabitants) Total Friends, family members 92.80% 91.30% 92.40% People with whom you work 91.20% % 93.40% Your neighbors 56.50% 47.80% 54.30% Members in civic organization 75.40% 91.30% 79.30% Fellow party members 57.10% 90.50% 65.50% Buisnesspersons 66.70% 91.30% 72.80% Journalists 78.30% % 83.70% Your constituency 75.40% 90.90% 79.10% Frequent accountability relationships: accountability requested at least once a month, i.e. categories 3 and 4 in the above table are combined. Source: European Mayor Survey in Hungary, conducted by the Tocqueville Research Center in N Page 17 of 21 pages

18 Table 10. Correlations of accountability types (Spearman's rho) Friends, family Collea gues Neigh bors Constit uency Businessp ersons Civic organizations Party members Journa lists Friends, family Colleagues Neighbors not sig..084 Constituency Businesspersons Members of civic organization Fellow party members Journalists Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project Table 11. Frequent accountability relationships by municipality size Actors Smaller municipalities (less than 10,0000 inhabitants) Larger municipalities (more than 10,0000 inhabitants) Total Friends, family members 85% 84% 84% Colleagues at work 68% 78% 69% Neighbors 60% 54% 59% Constituency 72% 74% 72% Businesspersons 22% 39% 17% Members of civic organizations 27% 46% 30% Fellow party members 9% 62% 16% Journalists 3% 27% 7% Mean number of strong accountability relationships ,3 Frequent accountability relationships: accountability requested at least once a month, i.e. categories 3 and 4 in the above table are combined. Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project Page 18 of 21 pages

19 Table 13. Logistic regression: Frequent accountability claims Variables Population (ref. category: less than 2,000 inhabitants) 2, ,000-49,999 50,000 or more Age Gender Education Odds ratio 1.8*** 2.8*** 3.2** Cultural capital 1.3* Relative income Managerial position Position in civic organization 2.3*** Electoral support of civic organization in which member 4.4*** Electoral support of civic organization in which not member 1.4* Electoral support of political party Political ambition 1.5* Participation in political training Mayor 3.0*** Ideological self-placement (left to right) Reelected representative Parents' civil society position N=781; Nagelke R 2 29% Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project Page 19 of 21 pages

20 Table 15. Logistic regression on the civic organizations' effect on opinion formation Variables Model 1 Model 2 Population (log).22***.13*** Age Gender Education Cultural capital Relative income Managerial position Position in civic organization.20***.11*** Electoral support of civic organization in which member Electoral support of civic organization in which not member.11***.07** Electoral support of political party -0.08* Political ambition Participation in political training Mayor.14***.07** Ideological self-placement (left to right) Reelected representative Parents' civil society position 0.09**.06* N=688; Nagelke R 2 16% 27% *** Correlation is significant at the.01 level (2-tailed); ** Correlation is significant at the.05 level (2-tailed); * Correlation is significant at the.10 level (2-tailed). Source: Local Representative Survey 2003 Indicators of Local Democratic Governance Project Page 20 of 21 pages

21 References Bőhm, Antal and Szoboszlai György Önkormányzati választások Budapest: MTA Politikai Tudományok Intézete. Central Statistical Office Az önkormányzatok döntéshozói Budapest: KSH. Dahl, Robert After the Revolution? New Haven: Yale University Press. Diamond, Larry Rethinking Civil Society: Toward Democratic Consolidation Journal of Democracy Vol 5. No. 3. Kohn, Margaret Panacea or Privilege? New Approaches to Democracy and Association. Political Theory, Vol. 30, No. 2, Manin et al Elections and Representation in Adam Przeworski, Susan C. Stokes, Bernard Manin (eds.) Democracy, Accountability, Representation. Cambridge Studies in the Theory of Democracy. Cambridge: Cambridge University Press. O'Donnell, "Horizontal Accountability in New Democracies" in Andreas Schedler, Larry Diamond and Marc F. Plattner. eds. The Self-Restraining State: Power and Accountability in New Democracies. Boulder and London: Lynne Rienner Publishers. Peruzzotti, Enrique and Catalina Smulovitz. Forthcoming. Societal Accountability: The Other Side of Control. Forthcoming in Enrique Peruzzotti and Catalina Smulovitz (eds.) Enforcing the Rule of Law. The Politics of Societal Accountability in Latin America. Pittsburgh, Pittsburgh University Press. Peruzzotti, Enrique. 2003a. Improving Governance: Civil Society's Efforts For More Accountable Government in Latin America. IDS Civil Society and Governance Policy Briefs, Institute of Development Studies. Peruzzotti, Enrique. 2003b. Media scandals and societal accountability: Assessing the role of the senate scandal in Argentina Paper prepared for the Conference Estrategias de Accountability Social en América Latina. Acciones Legales, Medios de Comunicación y Movilización, Department of Political Science and International Relations, Torcuato Di Tella University, Buenos Aires, April Powell, G. Bingham Elections as Instruments of Democracy: Majoritarian and Proportional Visions. New Haven: Yale University Press. Przeworski et al Introduction in Adam Przeworski, Susan C. Stokes, Bernard Manin (eds.) Democracy, Accountability, Representation. Cambridge Studies in the Theory of Democracy. Cambridge: Cambridge University Press. Schedler, Andreas Conceptualizing Accountability. in Andreas Schedler, Larry Diamond and Marc F. Plattner. eds. The Self-Restraining State: Power and Accountability in New Democracies. Boulder and London: Lynne Rienner Publishers. Smulovitz, Catalina and Enrique Peruzzotti Societal and Horizontal Controls Presented at Institutions, Accountability and Democratic Governance in Latin America The Helen Kellogg Institute for International Studies, University of Notre Dame, May 8-9, Strøm, Kaare Parliamentary Democracy as Delegation and Accountability. ESRC Research Seminar on Modelling Political Accountability: The Principal-Agent Model. School of Management, University of East Anglia, December 6, Warren, Mark E Civil Society and Good Governance. U.S. Civil Society Project. Georgetown University: Georgetown Public Policy Institute. Warren, Mark Democracy and Association. Princeton, NJ: Princeton University Press. Page 21 of 21 pages