Studies of electoral campaigning normally focus on national, parliamentary elections. Contrary to this, this article looks at campaign efforts related to the local council elections in Denmark on 21 November 1989. The exploratory of the paper - primarily due to the lack of a coherent body of previous research on local electoral campaigning - is emphasized. Two questions are central:

(1) Is it possible to point to significant differences between national and
sub-national election campaigns?

(2) Is it possible to demonstrate that electoral campaigns at the sub-national
level matter, i.e. that differences in the conduct of electoral campaigns
can explain significant differences as regards election results?

One often meets politicians who argue that it is necessary to think of the
next election all the time between one election and the next and to see all

Resumé

Electoral campaigning is studied almost without exception at the national level. This article has chosen another road, claiming that electoral campaigning can also be studied at the local election level. Campaigning before the Danish local elections of 21 November 1989 is studied. The design permits comparisons between the two levels (national/local) as well as between different units at the local level. It furthermore provides an opportunity for studying the influence of local party systems as well as local mass media on election campaigning. A substantial part of the article discusses the institutional frameworks surrounding electoral campaigning in the municipalities studied and in general. It is maintained that the electoral system, the mass media structure, and the (local) party system are important contextual factors or frameworks. Given this, it is argued that organization, past performance, and campaign focus as a mix of policy proposals and leader image are paramount in affecting the local election vote. The main conclusions are: local election campaigning differs from national election campaigning; local election campaigning matters, i.e. it has a direct effect on the vote; and the functions of local party organizations in connection with iocal elections and local performance make them less vulnerable to organizational decline, which most mass membership political parties are experiencing at the national level.

---

This is a revised version of a paper to the ECPR workshop on Electoral Ciimpiugning in the Western World. Bochum. Germany. 2-8 April 1990.

one's efforts as directed towards doing well when that election eventually comes. One local politician claimed that the duration of his electoral campaign was two years.' The importance of such extended electoral campaigns for what is achieved during the formal campaign has been stressed repeatedly - and for good reason (Pennimann 1981, 114f).

Danish local elections are 'fixed date'-elections, with the (identical) date of local and regional elections being set in the electoral law.² The obvious advantage of 'fixed dates' for political parties and candidates is that they can plan their future campaign activities in greater detail. Interviews with local party organization personnel support the suitability of the proposal by Farrell & Bowler of separating the period of campaign planning and preparation from the subsequent period of campaign implementation (Farrell & Bowler 1990, 7f).

The period of planning and preparation has an introvert character: candidates are to be found, the local party organization has to develop an electoral programme or manifesto, and plans for the period of campaign implementation have to be considered and agreed upon. The period of implementation, by comparison has an extrovert character. The direction of efforts of the party members and campaign workers might be used to separate the two phases: introvert campaign behaviour belongs to the preparatory phase, extrovert behaviour to the formal, implementing phase of the campaign.

It might be worthwhile to consider some of the factors - or 'contextual circumstances' (Farrell & Bowler 1990, 2) - which constitute frameworks around election campaigning activities. I shall concentrate here on the electoral system, the party system, the media structure, and the national polling prospects, respectively. My first assertion is that these factors function as important frameworks for local election campaigning - in Denmark as well as elsewhere.

Four Frameworks for Local Election Campaigning

1. The Electoral System as a Framework

The Danish electoral system at the national level is a complex PR system (Johansen 1979; Tonsgaard 1986). The complexities ensure proportionality between votes cast and seats won for parties who poll at least 2.0 percent of the vote, and reflection of the regional distribution of the party vote in the allocation of seats within each party. The low threshold and the high degree of proportionality are important as partial explanations of the format of the Danish national party system, which since 1973 has consisted of 8-11 parties having representation in parliament (Pedersen 1987; Elklit 1990a).

Sub-national electoral systems are simpler than the national system. A general account of local elections and local politicians - based on data from the 19705, but still useful - is given by Bentzon (1981b). An administrative reform in 1970 established 14 counties (amter) at the regional level and 275 municipalities (kommuner) at the local level. The reform was carried out by amalgamation and some redistricting of the smaller and older administrative units. The two metropolitan municipalities, Copenhagen and Frederiksberg, function both as local as well as regional administrative units.

Elections at the local level can be considered as 275 individual elections in multi-member constituencies, and at the regional level as 14 other individual elections, also in multi-member constituencies. They all take place according to the same electoral law, on the same day and with the same electoral system and procedures.

The simplicity of the electoral system at the sub-national level is due to the electoral independency of the individual units and the use of the d'Hondt PR system. This system is slightly favourable to the largest party (Taagepera & Shugart 1989, 32). This quality has been known for a long time, and electoral alliances between any number of political parties and non-partisan lists (in order to redress this överrepresentation) are therefore allowed for.

Before 1985 each party's candidates in sub-national elections were placed in a specific rank-order on the ballot paper. This also applied to the candidates of the non-partisan lists. The selection of candidates within a list took place in a way very favourable to the candidates at the top of the list. If a list won four seats, it was difficult - although not totally impossible - for candidates placed from number five to the bottom of the list to get elected. This system is also used in Belgium for parliamentary elections (van den Berghe 1979, 239).

Since 1985 it has been possible for parties and non-partisan lists to declare that their candidates are running on a parallel basis.? When a party (or a list) chooses to run by this system, the selection of individual candidates to fill the seats won takes places exclusively on the basis of the number of preferential votes cast for each of the candidates. The traditional view that it is advantageous to have a position at the top (or close to the top) of the ballot paper has recently been challenged by Darcy & McAllister (1990). It is nevertheless evident that candidates in Danish local elections with such positions - and especially those placed first - receive considerably more votes than candidates with more humble positions. And with the traditional rank-ordering system, it is definitely an advantage to be placed as close to the top of the list as possible, since it is possible to get elected by help of a supplement of the votes cast for the list as such. Both systems allow the

voter to cast his vote either for the party as such or as a preferential vote
for an individual candidate.

The new system has several consequences for the parties and the candidates, it might also have consequences for local election campaigning. Local party organizations hold different opinions on what to do: some want to keep their influence in matters of candidacy, and they therefore stick to the old rank-ordering system; others want to demonstrate democratic attitudes by granting the voters the final say on the selection of the individual council members - hoping that this will attract more votes than they would otherwise have been able to.

One finds no clear picture as regards party preferences vis-å-vis the two systems. Some local party organizations even change their opinion from one election to the next, according to actual experiences. The availability of two different systems of putting up candidates for local and regional office⁴ gives the local party organizations a free choice. Most parties show no clear pattern as to their preferences for either of the two systems, even though left-wing parties tend to prefer the traditional rank-ordering system, i.e. letting the party and its active members have as much influence as possible.

Another point is the very low threshold for entering the elections. In most municipalities only 25 signatures are required to enter an election, be it for a party or a non-partisan list of candidates. Larger towns can require the presentation of 50 signatures, Copenhagen 150! It is therefore easy for parties - some of which have little organizational strength locally - and for local non-partisan lists to enter local elections. Groups of concerned or discontented citizens can easily present themselves to the voters.

There are at least two other consequences: since sub-national elections are by PR with the administrative unit as a multi-member constituency, all votes count; and the two different systems of putting up candidates have different effects, as regards both the relative influences of party and voters, and the campaign style, which in the latter case tends to be more focused on individual candidates. Thus, there is no doubt that the electoral system per se is an important framework for the conduct of local elections and local election campaigning. This is in accordance with the conclusions reached by Sundberg in his comparative study of local elections in four Nordic countries (Sundberg 1989a). It also contributes to a substantiation of the claim by Farrell & Bowler (1990) that the electoral system as such is an important context for any electoral campaign.

2. The Party System as a Framework

This study concentrates on election campaigns in municipalities in the
county of Aarhus on the east coast of peninsular Jylland. Nineteen of the

26 individual municipalities in the county are quite small, with electorates varying from 3500 to 8000 voters. Four have between 10000 and 16 000 voters, and the three largest 37 000, 49 000 and 207 000 voters, respectively. The latter is the municipality of Aarhus; 45 percent of all voters in the county live in this municipality.

The 26 individual elections for municipality councils differ in regard to the number and the spectrum of parties and non-partisan lists running for office. It is not easy - even when all 275 municipalities are looked at - to find cases where the voters have a choice which includes all 12 parties that ran for parliament in the 1988 election (Elklit & Tonsgaard, 1989, 17; Elklit 1990a). The national party system consists of the eight parties represented in parliament since 1988. Only in the two largest municipalities (Randers and Aarhus) did the voters find these eight parties on the ballot paper. Table 1 gives the number of municipalities in this county in which the eight parties with actual parliamentary representation ran.

Even though the national party spectrum was only available to the voters in two municipalities, the five largest parties were a possible choice for the voters in 20 municipalities, and all five did present themselves in 15 municipalities. So even though no municipal election was exactly identical to any other in the county, most major parties were on the ballot in most municipalities - especially in the larger ones.

Table 2 combines the number of national parties running in the 26 municipalities with the number of local, non-partisan lists. Most of these lists claim to be non-political, even though some have a definite bourgeois character, others are declared socialist, and some defend the interests of some smaller part of the municipality in question. Table 2 supports the conclusion that the 26 individual elections differed considerably in their overall combination of ordinary parties and non-partisan local lists.

The conclusion is that the local party system should not be disregarded

as a framework within which local election campaigning takes place, and
which might condition the character of the campaign.

3. The Media System as a Framework

Television has also become the most important mass media in national election campaigning in Denmark. One reason for this is that the national electoral system is a PR system: all votes count, no matter where they are cast. This means that the personal achievements of the national leadership of a political party and the party's electoral bureau are beneficial to all the party's candidates. The importance of television for national electoral campaigning has been demonstrated by Karen Siune (1984, 1987, 1989; see also Bille 1991).

The situation regarding local election campaigning is different, because the national television channels have difficulty in covering simultaneously 289 individual local or regional elections. In 1989 - and at previous local elections - the national channels encountered problems covering these elections, both when sub-national elections were considered as news items in their own right and when the use of television as a channel for election campaigning was on the agenda.

One solution was to select one or two typical municipalities and then elucidate the problems facing them. It remains an open question if, or to what degree, the ordinary televiewer succeeded in translating the problems in such sample municipalities to problems in his or her own municipality. The conclusion suggests itself that television plays a much lesser role in sub-national than in national elections, since local and regional elections by standard news criteria are only of minor importance - and they are covered accordingly.

The second Danish television network has regional programmes - includingregional
programmes - in some parts of the country, but it is not

very probable that the problems referred to were solved within programmes competing with national programmes during a period characterized by national political conflict over the budget and economic policies, and dramatic developments in Eastern Europe. The county studied did not have regional television programmes at the time of these elections.

It thus appears that the role of television is very different in national and sub-national election campaigning. Television coverage is of prime importance in national electoral campaigns, but it is - as far as the scarce evidence permits - only of limited importance in Danish low level elections. In national elections, television has been found to be 'the most used source of information and most influential on those less interested in politics and having less knowledge about politics, but who nevertheless vote on election day' (Siune 1987, 2). The different coverage of the two kinds of elections might be an important factor if we are to explain why local elections - generally speaking - are low interest elections with a turnout of 20 percent below the national level!⁵

National newspapers face the same problems with local elections as other national media. Local newspapers are therefore the obvious candidate as the most important medium for local election campaigning. Developments over time in the structure of Danish newspapers, however, have brought about a situation where most local and regional newspapers are the only newspaper in their area; the remaining newspapers have given up their previous commitment to one of the four old parties (the Social Democratic Party, the Radical Liberals, the Conservatives, and the Liberals), making them less susceptible to political propaganda than they used to be.⁶ They also tend to cover larger areas than previously.

The ensuing commitment to a broader and more neutral coverage of the different political views, building primarily on traditional news criteria, has not made the local newspapers better suited as media for local election campaigning. It can even be argued that this enhances the tendency to emphasise the coverage of campaign events instead of issues or policy proposals contained in election manifestos, which Ware stresses as being important for the impact of television on party competition (Ware 1987, 81). Ware also mentions reported differences between television and newspapers, the development in newspaper structure, nationally as well as locally, seems to entail a further development in this direction, as is the case with television. Smaller communities without a local newspaper often have a small, commercial, local advertisement weekly, distributed free of charge to all households in the area. These weeklies often use a few columns to cover local news items. They can of course also be used for electoral campaigning.

The media structure in every single municipality is thus a unique combinationof
media (national television and radio networks and

national newspapers) and a number of specific media, including regional television programmes on the second television channel, local television channels, local radio channels, local and regional newspaper coverage, and local advertisement weeklies. Our hypothesis, therefore, is that the media structure explains a considerable part of the differences between campaign strategies in different municipalities, and, furthermore, that the general unavailability of national media explains the differences between national and local election campaigns.

4. The Polling Prospects of National Parties as a Framework

The electoral swing phenomenon is well documented, also at local levels, and so is the parallel development of party strength at the national level and the outcome of local and regional level elections. Danish studies covering these effects are Bentzon (1981a) and Risbjerg Thomsen (1986).

The Progress Party can be used as an example. The party did very badly in the January 1984 general election with an all-time low of 3.6 percent and opinion polling figures during 1984 and 1985 were even worse. In the local and regional elections of November, 1985 this was reflected in considerable losses of votes as well as seats across the country. The party ran in only 17 municipalities in the county of Aarhus, and the average poll in these municipalities was a mere 3.0 percent. The Progress Party nevertheless recovered, contrary to the prophecies of political and academic observers and commentators. In May 1988, the party was supported by 9.0 percent of the voters. Opinion polling figures were rising after the election, and in November 1989 they were at the 10-12 percent level nationally. In 1989 the Progress Party put up candidates in 21 of the county's 26 municipalities, and on average the party polled 6.2 percent in these municipalities. But the electoral campaign of the Progress Party in all municipalities was poor and feeble, as regards both quality and quantity. The election result was impressive, however, and the only conclusion possible is that parties sometimes succeed electorally without really trying!

This example illustrates an important point: the rise or fall in a party's national Strength might contribute - independently of other factors - to the outcome of elections on other electoral levels. This kind of influence works in all municipalities, thereby creating a swing phenomenon. This is not to say that all electoral swings are of equal size: local circumstances - including the local party system and the local media structure - moderate the influence of the national level.

The conclusion is that sub-nationa! elections take place under considerableinfluence the national level. The voters see the local political parties as local representatives of national parties, and they punish or reward them correspondingly. Therefore, it is difficult to study local electionson

tionsontheir own: they are to some degree but reflections of general attitudes towards national parties, leaders and policies. The national politicalsituation terms of actual party strength) is thus the fourth frameworkwhich be integrated in the study of local election campaigning.

The Study of Local Election Campaigning

It appears that two of the four frameworks are of a more general character: (1) the electoral system is the same for all parties in all municipalities and counties; and (2) the national polling prospect is an - almost - identical factor for all the local branches of a party. The two remaining factors are specific, i.e. bound to individual municipalities: (3) the local party system is specific to the individual municipality, even though it might be possible to construct a workable categorization of local party systems, if all 275 municipalities were taken into consideration; (4) the mass media factor is also specific, since individual municipalities tend to have their own mass media structure, which to some degree determines the election campaign strategies of the parties. Neighbouring municipalities might have media structures which are close to one another, and which therefore will tend to influence local election campaigning in identical ways.

A thorough study of a major political party's national election campaign is a demanding effort, since data-collecting efforts and participant observation must take place during the campaign. With campaign preparation in many committees and staff meetings - and all the extrovert activities - it cannot avoid being a full-time occupation during the period in question. If such analysis is to cover several major parties a team of researchers is required.

The same problem arises when empirical studies of local election campaigning parties and municipalities are on the research agenda. If a considerable amount of local data-collection and participant observation is foreseen, a team of researchers is a necessity. If such a team is not available, a choice has to be made between concentrating on a selection of parties in one municipality (or a very limited number of municipalities) and conducting ex post facto study, i.e. collecting the data after the election. One way of collecting data is by interviewing relevant persons, but with this research design subjects will know the election result and therefore may be biased in their recollection and reporting of the campaign.

The study reported here was designed as an ex post facto study where a small selection of local campaign officers or local party organization chairmenwere after 21 November 1989. Twelve interviews were carried out, covering five different parties in seven municipalities (and the county). Most interviews were with Social Democrats, since the Social Democratic party is normally the largest party. The number of parties

interviewed in each municipality varied from one to three. The length of the interviews, which also covered organizational issues, varied from U to 2\ hours. Interviews with some of the national parties' general secretaries were conducted during the initial phase of the study.

Studies of election campaigns have normally been individual case studies. In order to improve the prospects for comparative research within this field, Farrell & Bowler in a recent paper (1990) have advocated the use of a common set of research questions and conceptualizations. They argue that election campaigns can be broken down into three constituent parts: (1) campaign organization and control, (2) preparations and strategic planning, and (3) implementation and campaign tactics (ibid., 8). The first two of these three elements belong to the period of campaign preparations referred to in the introductory sections, while the third obviously is the period of campaign implementation.

Farrell & Bowler are primarily concerned with national election campaigns, while the emphasis in this article is on local election campaigning. This allows testing of the suitability of their proposals on a different level of elections, thereby adding an intra-system type of test to the traditional inter-system comparisons.

1. Campaign Organization and Control

Since local campaigns are mainly fought locally, it follows that the question of staffing, equipment, and so forth is primarily a question of equipment in local organizations. These are usually small, and often the office of the organization is the desk (or less than that) of the chairman or the secretary in their private home. The staffing and the equipment at hand depends directly (and positively) on the size of the municipality and the party in question. The larger the party - in terms of members and voters - and the larger the municipality, the better equipped and staffed the local organizations. Most of these organizations are poor, with no fax, personal computers, or printing and photocopying facilities. It follows that there is no reason to believe that this variable by itself can explain differences between campaigns at the local level.

As to the question of who is in charge of the campaigning efforts and coordination, the answer is that it is normally a separate election committee. The election committee is composed of some combination of ordinary committee members, candidates, and specialists (on this level maybe someone in his ordinary job has access to a word processor or a photocopier). If parties with more than one branch in the municipality are considered, the electoral committee will also comprise representatives of the coordinating committee of the party 'umbrella' organization.

The smoothness of the operation of the organization is a matter of

judgment, even though answers to this question tend to be more positive, the better the election result - at least as compared to expectations. Traces of conflicts and grievances between members of the electoral committees were, however, also heard during some of the interviews. They were often the result of conflicts over the composition of the list of candidates and the numbering on the ballot paper - conflicts, in short, caused by unfulfilled ambitions.

2. Campaign Preparations and Strategic Planning

Only very rarely can local party organizations afford to use various types of consultants or agencies in their campaigning efforts or to get 'objective information' from opinion polls, etc. (Rose 1967, 152-154). Only in a few cases was a professional lay-out specialist asked to help in designing campaign material.

More interesting are the preparations for distributing the various messages concerning candidates and policy proposals) to potential voters. It was expected that the parties would consider it appropriate to use door-to-door canvassing as the most important means of communication in local elections - also because most candidates are well-known figures locally. But it turned out during the interviews - and from participant observation - that door-to-door canvassing is used only rarely in Denmark (both in local and national elections). Actually, it was only used by one party of those interviewed.

This form of campaigning is known to be relatively effective (Weir 1985), but the answers to questions were almost invariably that it was too difficult to find candidates and campaign workers who would do door-to-door canvassing, even in their nearest neighbourhood. A certain (somewhat surprising) shyness on the side of the candidates and the campaign workers was also referred to. It is interesting that the only party organization using this campaign method (the Social Democrats in Hadsten, an average size municipality) was very satisfied with the result, and they did remarkably well in the election.⁷ There are, however, other possible explanations for the Social Democrats doing well in Hadsten.

The most prevalent means of distributing campaign material in local elections was sending election pamphlets etc. as unaddressed mail to all households in the municipality (or to have campaign workers - including the candidates themselves - distribute the folders in order to save money). This brings the campaign message to all potential voters, but in a way which very much parallels the distribution (and consequently maybe also the reception) of other sorts of advertising folders (i.e. throwaways!). Public meetings were also used - mainly in the countryside - to establish contact

with potential voters. Such meetings were rare and poorly attended in the
towns.

It is essential that not only distributional, but also strategic preparations (Farrell & Bowler 1990, 15) are analysed carefully if one is to understand the campaign and its possible effects. Local party organizations consider two tasks most important for the general strategic preparatory phase of the campaign: (1) the selection of the candidates, who in their mix of various qualities (e.g. personality, political attitudes, gender, occupation, age group, hobbies, maybe functions as shop steward, living area within the municipality, etc.) will hopefully appeal to a considerable proportion of the electorate; and (2) the working out of the electoral manifesto.

With few media available, issues such as timing, pacing and so forth become less important, since it is extremely difficult to force local issues on the political agenda by using the mass media, especially with no daily newspaper in the area, and with many parties from a number of municipalities to get access to television, radio, and regional newspapers for the very same purpose. Most Danish newspapers have special columns (on special pages) for local news, edited by a 'local news editor', who will try to avoid being (too) biased in his selection of what to publish. Hence, the two main dimensions in strategic campaign preparation - after candidacies the electoral manifesto - appear to be: (1) the campaign mode (or style); and (2) the campaign focus.

Re (1): Farrell & Bowler (1990,15) propose that in the case of incumbents a defensive campaign mode is to be expected, while the previous opposition will probably behave more aggressively. In the local cases under scrutiny, however, it is difficult to see that this should be the case. A major reason for this is that the appointment of members to the different council commissions committees is also carried out by PR. This calls for alliances between parties, the result being that a majority coalition divides the majority of positions among themselves, leaving the smaller part to a minority coalition or to individual minority parties. The result is a strange and confusing mixture of incumbency and opposition, which cannot avoid blurring the question of responsibility, making it difficult to campaign on the basis of some parties being incumbents and others opposition. If an alliance gives the mayoralty to the Social Democrats and the presidency over important commissions - such as the Commission for School and Culture or the Commission for Public Works and Environment - to the Conservatives and something else to the Liberals, then most of the local parties share responsibilities for what has been carried out - even for the local taxation level. Who, then, are to campaign aggressively and who defensively on the basis of past performance?

If a municipality happens to have a party with a majority of seats, then
the situation is different. But it is a rare phenomenon in a PR system.

When it is found or approached,⁸ campaigning tends to be rather selfconfidentand ('See all the good things we have been able to do, and we will do even better during our next term in office'), not primarily defensive or aggressive.

It appears that the Danish mixed mayor-committee system in local administration entails an orientation in campaigning more directed towards policy proposals and future activities than would otherwise be the case. A future-orientated campaign style also entails discussions of past performance the present situation - and the responsibilities for these - at least to some degree. It nevertheless appears that the mixed mayor-committee system in the municipalities has the abovementioned for local election campaigning.

Re (2): If we turn to campaign focus, it follows from the preceding paragraphs that an emphasis on policy proposals and issues from the election manifesto is to be expected. To some degree this is also found to be the case, but the tendency in national elections to focus on leader images is certainly also found in municipalities, especially when a mayor has served well for a number of years (Elklit 1990b).

As the electoral system gives a certain weight to preferential voting (especially when the system with 'parallel' candidates mentioned above is used), local party organizations sometimes decide that equal emphasis is to be given to all candidates, to give all an equal chance for getting elected, e.g. in the form of securing equal appearances on folders, placards, at meetings, etc. Such a practice is not only functional internally, but also externally (i.e. as regards the voters) since it will attract more neighbourhood and at least in principle demonstrate that the voters have the final say in choosing among all the candidates, who stand equal chances. Deliberate use of all the candidates also gives the party the possibility of stressing the party image insofar as this image suggests that the party represents various social groups (classes/categories, etc.). The number of young people, women, workers, shop stewards, farmers, or people in private urban business signals this part of the party image to the voters.

The campaign focus, speaking generally and on the basis of the parties studied so far, has been on policy proposals and future needs of the local community, building in most cases on a broad and imprecise assessment of the actual situation, particularly as far as the economic and financial situation is concerned. 'Responsibility in economic affairs" was a key electoral slogan for most parties. If, however, a popular person is heading the list, most parties exploit his or her image, parallel to what is seen in national electoral campaigning. Parties with such candidates are neverthelessanxious the leader's image does not totally overshadow the policy proposals in the election manifesto. Other considerations - internal

and/or external - might also entail the stressing of the availability of many
more candidates to choose among.

The party image as such appears to be of lesser importance when the campaign focus is being decided. This is even more so when national polling prospects are negative. In November 1989, this was the case for the Conservatives. One local Conservative organization consciously tried to dissociate itself from the image of the PM's national party by stressing in their campaigning that they were 'The Conservatives of Odder'. They did well in the local elections,⁹ but this was primarily due to a new top candidate, a popular and respected figure locally. The main point here, however, is not possible reasons for the election result, but the fact that a local organization tried to include in their campaign message that they were not identical with the national party.

3. Implementation and Campaign Tactics

Three examples of various campaign tactics and of implementation problems suffice. First, some parties did develop tactical plans which said, for example, that four main issues were to be emphasized during the last four weeks of the campaign period. Advertisements, folders, themes for meetings, and news coverage were to go hand-in-hand and demonstrate the cohesiveness of the party's planning and considerations about future activities. Both the Social Democrats and the Socialist People's Party in Aarhus nevertheless experienced that it is no easy task to set the communication agenda for the mass media, i.e. to make the local radio station and the regional/local newspaper act according to the wishes of the parties. Hence, 'the themes of the week' were not transmitted to the electorate by the media and therefore failed to catch the attention of the voters.

Second, the Centre-Democrats, a minor and organizationally weak party in the municipality of Aarhus, had certain success due to rather crafty campaign planning. Locally the party is more or less a one man-party, and the top candidate - and main campaign manager and activist - decided to launch his campaign very early, when access to the media was easy. This made sure that he was noticed in the early phase of the campaign period, making him less vulnerable to the crowding of parties and candidates in the media during later phases of the campaign, where he, representing a minor party, would have difficulties in getting access.

Third, in a minor municipality it was decided that all parties who wanted to, and who would pay a certain fee, could have videos shown in the central shopping centre. One party paid the fee, but did not - for various reasons - have its video produced. So the party organization lost both their money and demonstrated to the public that they were not able to deliver the

party's own advertising video tape. The party suffered heavy electoral
losses, but for a variety of reasons.

Some Results, or: Measuring Direct Campaign Effects

Did the local election campaigns matter? Discussion of this question concentrates the effect on the voters, i.e. on the election outcome, leaving aside those aspects of the question which touch upon campaign effects on the politicians, the parties themselves, etc. It might be appropriate to use the concept 'direct effects' - or 'intended effects' - to cover campaign effects on the voters as they are expressed very tangibly in election results. The term 'indirect' effects (sometimes 'unintended') could then be used to cover the various effects of the campaign on politicians, be it 'third-person effects' (Davison 1983) or more direct influences on the politicians, or even the kind of prolonged influence on interpersonal or interparty attitudes which, for example, also affects post-election negotiations on alliance formation in the council.

This section concentrates on 'direct' campaign effects. Since it is not possible to include surveys of voter reactions to different campaign modes, mixes of campaign foci, etc., the argument will necessarily have to rely on inherent assumptions, tendencies in aggregate statistics, and the author's perceptions, and less on traditional hard evidence.

The influence of the contextual frameworks is not easy to assess, since so many factors vary from one municipality to another. One can. however, argue that the expected effects of the two different ways of ordering candidates on the ballot paper have not been demonstrated to be more than marginal. It might even be that the differences between national and local election systems are less important than expected. Differences between various PR systems and the various majority electoral systems are still believed to be important.

The national polling prospects of the different parties have not been shown to have important consequences for local election campaigning, apart from supplying some optimism (or pessimism) to the local party organization. Yet the effect on voter behaviour per se is still believed to be important. Recent development in electoral volatility supports this suggestion. National polling prospects, of course, are outside the control of the individual party branches. That is why it is a "contextual" factor (or 'framework'). Since the term 'framework' could pass on the impression that it is something common to all parties and lists in a municipality, we may exclude it from the category of frameworks; it is primarily an external

factor like other such factors. This point of view is reflected in the drafting
of Figure 1.

The importance of differences in mass media structure between local communities on local election campaigning has not yet been demonstrated convincingly. Differences between local and national media systems for electoral campaigning, however, are still considered paramount.

Neither has the influence of various local party systems been assessed

rigorously. It is, however, difficult to point to important differences in campaigning between municipalities in the various cells in Table 2 above. Might it be more rewarding to control for the presence of locally dominant parties?

It thus appears that at least three of the four general frameworks might be useful - and necessary - in describing the specific political and electoral situations within which election campaigning takes place - locally or nationally - and that these frameworks can be used to establish if (to a reasonable degree) 'all things are equal' or not. But they do not encompass all relevant factors: they affect and determine the campaigns to a considerable but they remain frameworks surrounding the central topic of our study! The campaigns themselves are still the key element.

The influence of specific campaign modes, campaign foci, various strategic tactical planning elements, or different distributional considerations are extremely difficult to assess and to study in a systematic and gratifying way. Claims to the opposite remain unproven. It is not difficult to find specific situations which appear to demonstrate the importance of the various factors. The research problem, however, is still that it is difficult to rule out the influence of other causal factors also present.

Past performance seems to be an important factor - maybe the central factor. Satisfaction with performance in the council during the latest term influences politicians (as candidates) as well as campaign mode and focus, and party organization behaviour. Campaigning under such circumstances is expected to be different, and more influential, than a campaign starting from dissatisfaction.

We still have the problem of distinguishing between the direct campaign effect on the vote and the direct, or immediate, reaction of the voters to an administration which has been doing well during its latest term in office. Only in some situations will it be possible to prove which of these effects are the more important (cf. Figure 1). The relevance of the second factor follows from electoral research that claims that voters react to their perceptions matter how they are formed) of party behaviour during the last electoral period, i.e. that voters (re-)act rationally, at least to some degree (see Nannestad 1989).

Positive evaluation of past performance might be connected with a longtime, and popular mayor, who is heading his party's list of candidates. The election campaign focus in this case is probably going to be a mix of leader image, past performance, and policy proposals (which the voters believe will be implemented, because of the party's past performance). relative importance of these different factors can only be determined on the basis of detailed interviews with a sample of the electorate, from intensive interviews with party personnel, and even less from electoral statistics.

Organizational strength (not only a question of membership size), often entailing good and effective organizational planning, is also considered a key factor. It is not enough by itself, however, since it requires a specific situation, e.g. a specific media structure, which entails receptiveness for the messages from the organization, in order to function optimally. Effectiveness putting across the communication to the electorate turns out to be the central issue!

On the national level, communication with the electorate can be done most effectively in a centralized fashion by taking advantage of modern technology, but for economic reasons and due to the local media structure this is not possible in individual municipalities. Organizational effectiveness in planning and implementing local election campaigning thus depends on the capacity to use the communication means available. It appears that those few party branches which were consciously engaged in direct personto-person did better than those who were hesitant in arranging canvassing or who did not send their candidates out to meet the voters on all possible occasions.

Another important factor - which is outside the influence of the party itself - is the campaigns of competing parties. If the opponents' campaign efforts are poor, then the party's own chances of success are obviously better, and vice versa.

These various factors are integrated in Figure 1, which models the explication of the local election campaign and its direct effect. The provisional of the model, as well as the question of its applicability to other levels of electoral campaigning, must be stressed.

Conclusion

Two questions were raised in the introduction. The first asked if it was possible to point to significant differences between national and subnational campaigns. The answer to this question depends on the level of generalization. Since local election campaigns almost exclusively take place locally, they cannot exploit the national mass media. To a considerable degree they are therefore contingent upon conditions of the local media system, and are likely to involve more direct personal communication and the utilization of other means of local knowledge and personal networks. For this reason such campaigns must be conducted on a decentralized basis. This is contrary to the main tendency in national election campaigning. So there certainly is a difference between national and sub-national electoral campaigning.

But if the municipalities are considered individual and independent
political systems (which to some degree they are, especially in a local

election connection), then the parallels between local and national elections become more obvious: election campaigning depends unavoidably on the mass media system available and has to accept this system as it is, and make the best of it. In this respect and on this level of generality it is difficult to claim that campaigning on one of these levels differs from campaigning on the other.

The other introductory question asked was 'Can differences in the conduct electoral campaigns explain differences in election results?', or - to phrase it otherwise - 'Does local election campaigning matter?' The answer to this question is definitely in the affirmative. Also, on this level how election campaigns are organized, how well organizations operate, and how qualified and careful the planning is are important. Strategic and tactical planning, and the implementation of plans account for more than a difference in degree.

But it is still difficult to measure how much local election campaigning matters. Other factors, some of which are of considerable importance, also matter when the voters decide to cast their vote. And even though some voters claim that they have made their decision well before election campaigning the reinforcing effects which election campaigning has on exactly this category of voters should not be forgotten.

Local election campaigns are the responsibility of local party organizations branches. Their specific content and contexts make a continuation this decentralization necessary, even when national elections - at least in PR systems - tend to become increasingly centralized.

It has, moreover, been claimed that this local election function could be a safeguard against the continuous erosion of the mass membership in political parties (Ware 1987, 133-135; Sundberg 1989b, 127). If parties are able to present themselves and are accepted as influential local organizations, integrate local interests, who select candidates, and who have a say on policy proposals, if parties perform well in the local community administration, and if parties have local leaders of some quality, then there is no reason to fear that voters, or members, should disregard the local party organizations. It would still be worthwhile to support them, at least on polling day. This perspective is perhaps particularly relevant for the larger and more 'visible' parties, locally as well as nationally, as Selle &. Svåsand have claimed (1983, 228). Thus, it can be argued that local party organizations will not disintegrate, since they are more needed in the local communities and in local election campaigning than the traditional mass membership parties on the national level.

---

NOTES

1. Erik Svensson of the Centre-Democrats. He succeeded in winning a seat both in the

---

2. The third Tuesday of November every fourth year.

3. The Danish term is 'sideordnet opstilling'.

4. The same possibilities have for a long time been available in general elections.

5. 65-67 percent compared to 85-88 percent.

6. This does not amount to saying that one cannot occasionally detect a certain political bias - and most often in the direction of the party which the paper used to be committed to.

7. The rise from 33.3 percent to 41.9 percent of the total vote was well above the average Social Democratic rise in municipal elections.

8. As with the Social Democrats in Randers or Grenå.

9. The rise from 18.8 percent to 22.9 percent of the total vote was a positive deviation both from the general pattern in the local elections and from the national polling prospects.

REFERENCES

Bentzon, K.-H. 1981a. 'De politiske partier og de kommunale valg'. Økonomi og politik 55,
141-158.

Bentzon, K.-H. 1981b. Kommunalpolitikerne. Copenhagen: Samfundsvidenskabeligt Forlag.

Darcy, R. & McAllister, I. 1990. 'Ballot Position Effects', Electoral Studies 9, 5-18.

Elklit, J. 1990a. 'A Survey of Research on Postwar Party Behaviour in Denmark', paper
prepared for a meeting of the 'Party Behaviour Group', Bochum, Germany, 1-2 April
1990.

Elklit, J. 1990b. 'Byrådsvalget i november 1989', in Vedel-Smith, H., ed., Århus-Årbog 1990.
Aarhus: Aarhus Byhistorisk Udvalg/Erhvervsarkivet.

Elklit, J. & Tonsgaard, 0., eds. 1989. To folketingsvalg. Vælgerholdninger og vælgeradfærd
i 1987 og 1988. Aarhus: Politica.

Farrell, D. M. & Bowler, S. 1990. 'Comparative Election Campaigning: Towards a Framework of Analysis', paper prepared for ECPR Workshop on Electoral Campaigning in Western Europe, Bochum, Germany, 2-8 April 1990. To be included in Farrell, D. M., & Bowler, S., eds., Election Campaigning in a Comparative Perspective (working title), London: Macmillan (in preparation, 1992).

Johansen, L. N. 1979. 'Denmark', in Hand, G., Georgel, J. & Sasse, C, eds., European
Electoral Systems Handbook. London: Butterworths.

Nannestad, P. 1989. Reactive Voting in Danish General Elections 1971-1979. A Revisionist
Interpretation. Aarhus: Aarhus University Press.

Pedersen, M. N. 1987. 'The Danish "Working Multiparty System": Breakdown or Adaption?',
in Daalder, H., ed., Party Systems in Denmark, Austria, Switzerland, The Netherlands and
Belgium. London: Frances Pinter.

Pennimann, H. R. 1981. 'Campaign Styles and Methods', in Butler, D., Pennimann, H. R.
& Ranney, A., eds., Democracy at the Polls. A Comparative Study of Competitive National
Elections. Washington: American Enterprise Institute.

Risbjerg Thomsen, S. 1986. 'Kommunale valg 1970-85: Landspolitisering eller lokalpolitisering?',
Elklit, J. & Tonsgaard, 0., eds., Valg og vælgeradfærd. Studier i dansk
politik. 2. ed., Aarhus: Politica.

Rose, R. 1967. Influencing Voters: A Study of Campaign Rationality. New York: St. Martin's
Press.

Selle, P. & Svåsand, L. 1983. 'The Local Party Organization and Its Members: Between
Randomness and Rationality', Scandinavian Political Studies 6, 211-229.

Siune, K. 1984. 'Bestemmer TV valgresultatet?', in Elklit, J. & Tonsgaard, 0., eds., Valg og
vælgeradfærd. Studier i dansk politik. Aarhus: Politica.

Siune, K. 1987. 'Mass Media between Parties and Voters', paper prepared for ECPR
Workshop on 'Party Strategies and Party-Voter Linkages: Neglected Dimensions in Election
Studies', Amsterdam, The Netherlands, 10-15 April 1987.

Siune, K. 1989. 'Valgkampen og vælgerne', in Elklit, J. & Tonsgaard, 0., eds., To folketingsvalg.
og vælgeradfærd i 1987 og 1988. Aarhus: Politica.

Sundberg, J. 1989a. 'Politiseringen af kommunalvalen i Norden', Statsuetenskaplig Tidsskrift
92, 25-40.

Sundberg, J. 1989b. Lokala partiorganisationer i kommunale och nationale valg. Fdrandringar
i organisatorisk styrke och medlemsaktivitet i politiska lokalfbreningar i Finland åren 1945—
1987. Helsinki: Finska Vetenskaps-societeten.

Taagepera, R. & Shugart, M. S. 1989. Seats and Votes. The Effects and Determinants of
Electoral Systems. New Haven & London: Yale University Press.

Tonsgaard, O. 1986. 'Valglovgivningen - en kritisk gennemgang', in Elklit, J. & Tonsgaard.
0., eds., Valg og vælgeradfærd. Studier i dansk politik, 2. ed., Århus: Politica.

van den Berghe, G. 1979. 'Belgium', in Hånd, G., Georgel J. & Sasse, C, eds., European
Electoral Systems Handbook. London: Butterworths.

Ware, A. 1987. Citizens, Parties and the State. A Reappraisal. Cambridge: Polity Press.