Journal of Comparative Politics 52 DIMENSIONS OF PARTY ELECTORAL PROGRAMMES: SLOVENIAN EXPERIENCE Simona KUSTEC LIPICER and Samo KROPIVNIK 1 Party programmes, frequently also called manifestos or platforms, are written documents in which a political party sets out the programme either for its general inter-party making or for broader pre- and/or post-election purposes. As such, the party programme is regarded as an important constitutive part of each democratic party ‘body’ or, even wider, the element of democratic election processes or even the democratic political system in general. Analysing party programmes also represents one of most visible theoretical and methodological research topics inside various political science disciplines. Based on such understandings of the roles of party programmes, the main aim of this article is to disclose the prevailing characteristics of those party programmes prepared or used for the purposes of Slovenian national parliamentary elections (e.g. electoral programmes). With this in mind, we expect to be able to give mainly a general descriptive assessment of the status and importance of party electoral programmes in the case of a young post-socialist democratic state, actually the first of that kind in the Slovenian case. Accordingly, a comparative approach of analysing the outlook and content dimensions of party programmes is applied. 1 Simona Kustec Lipicer is Assistant Professor at the Faculty of Social Sciences, University of Ljubljana, Slovenia, Chair for Policy Analysis and Public Administration, and researcher at the Centre for Political Sciences Research at the Institute of Social Sciences, University of Ljubljana. She is a head of an ongoing basic research project entitled “Pre-election campaign and democratic evolution of state and society, ” conducted by the research team at the Centre for Political Science Research, financed by the Slovenian Research Agency; a member of the Management Committee and Slovenian researcher of the EU COST Action IS0806 “The True European Voter: A Strategy for Analysing the Prospects of European Electoral Democracy That Includes the West, the South and the East of the Continent, ”; Slovenian Researcher of the ongoing international projects of coding electoral programmes MARPOR and Euromanifestos. Samo Kropivnik is Associate Professor of Social Sciences Research Methodology at the Faculty of Social Sciences, University of Ljubljana, Slovenia where he teaches courses on methodology and research methods. His research focus is on a choice of a research design in applied studies and on creative use of multivariate methods in social sciences. Since 2008 he is local academic coordinator of ECPR Summer School in Methods and Techniques. Journal of Comparative Politics 53 The comparison is made from three different perspectives: a) party perspective that relates to the comparison of the programme content of the competing political parties, where their pre- and post-electoral parliamentary status is especially considered; b) time perspective relating to the comparison of the last two national parliamentary elections (2004 and 2008); c) coalition agreement perspective, according to which party electoral programmes are compared to the contents of the crucial post-electoral document. The conclusions, on one hand, expose many similarities among the compared elements regardless of party status, as well as some major differences on the other hand. The latter relate mostly to the contextual dimensions of the programmes and enable us to classify electoral party programmes into four different types according to their prevailing characteristics. 1 On party electoral programmes Party programmes, and in this regard especially electoral programmes, have for quite a long time – for a certain group of political scientists who claim that party programmes count or matter 2 – been an important subject of political science research for many reasons and from many different perspectives. 3 There is not one unique definition of what party programmes are. Mostly we find definitions stating that party programmes are hierarchically the most important written and publically available party documents, in which broad party statements, claims, interpretations, analyses, proposals, recommendations etc. are stated for the 2 See David Robertson, A Theory of Party Competition (London: Wiley, 1976); also see Ian Budge and Richard I. Hofferbert, “Mandates and Policy Outputs: U.S. Party Platforms and Federal Expenditures, ” American Political Science Review, 84 (1990), 111–131; also see Ian Budge and Michael Laver, “Policy, Ideology and Party Distance: Analysis of Election Programmes in 19 Democracies, ” Legislative Studies Quarterly, 11, 4 (1986), 607–617; also see Ian Budge et al, Mapping Policy Preferences (Oxford: Oxford University Press, 2001); also see Hans-Dieter Klingemann et al, Mapping Policy Preferences II: Parties, Electorates and Governments in Eastern Europe and the OECD 1990–2003 (Oxford: Oxford University Press, 2006); also see Hans-Dieter Klingemann, Richard I. Hofferbert and Ian Budge, “ A Theory of Democratic Policymaking, ” in Parties Policies and Democracy, ed. Hans-Dieter Klingemann et al (Boulder, CO: Westview, 1994), 240– 270. 3 We would especially like to emphasise at this point that as big as the mentioned group of party programme ‚advocates’ is, there is an equally large group of the programme adversaries who claim that party programmes are only words on the paper with no real message, power or impact, being more or less a cosmetic or symbolic premium of party-making. Although aware of the latter standpoint, we consciously take the relevance of party programmes for granted. See Gerald Pomper, “If Elected, I Promise: American Party Platforms, ” Midwest Journal of Political Science, 11 (1967), 318–352; also see Louis Sandy Maisel, “The Platform-Writing Process: Candidate-Centered Platforms in 1992, ” Political Science Quarterly, 108, 4 (1994), 671–698. Journal of Comparative Politics 54 purposes of party making. 4 Some authors see their importance mainly in the context of effective, efficient and transparent party politics. 5 Besides the described candidate-centred analytical focus, numerous analyses, as well as theories, suggest that party programmes should be considered more broadly than only as phenomena of party politics. This relates to the role of party electoral programmes within the election process (for example, as part of an electoral campaign, 6 or even to the much broader role of understanding them as an integrant constitutive part of a democratic political system. 7 In general party programmes can be analysed through various contextual insights: 1) Some would analyse their outlook, which includes the scope of the programme, like the length (number of pages, words, sentences, [sub]chapters), type of programme document, etc. 2) Others would analyse the processes of its making: where and how the party programme is prepared and who prepares it (centred on the party leader or institutionalized in the domain of various public-policy experts); 3) Some would analyse the programme’s content, like whether it consists of 4 For example see Annika Werner and Andrea Volkens. Manifesto coding instructions, 2009. Available at http:// www.wzb.eu/zkd/dsl/daten/marpor/MDCodingManual.pdf (3 November 2010); also see Simona Kustec Lipicer and Nikolina Bilavčić, “Volilni programi in volilne vsebine skozi volilno izkušnjo volitev v Evropski parlament 2009 (Electoral programs and contents through the lens of European Parliament 2009 electoral experience), ” in Politične vsebine in volilna kampanja: slovenska izkušnja z volitev v evropski parlament 2009 (Political Contents in Electoral Campaig: Slovenian Experience with the 2009 European Parliament Elections), ed. Simona Kustec Lipicer (Ljubljana: Fakulteta za družbene vede, 2010), 71–97 . 5 Peter Mair, “Myths of Electoral Change and the Survival of Traditional Parties, ” European Journal of Political Research, 24 (1993), 121–133. Craig Allen Smith and Kathy B. Smith, “ A Rhetorical Perspective on the 1997 British Party Manifestos, ” Political Communication, 17 (2000), 457–473. 6 See Hermann Schmitt. Determinants of Dyadic Correspondence in European Parliament Elections, 2008. Available at http://www.jhubc.it/ecpr-riga/virtualpaperroom/122.pdf (15 September 2010); see also Michael Laver and Ben W. Hunt, Policy and Party Competition (New York: Routledge, 1992); William Keefe, Parties, Politics, and Public Policy in America (Washington DC: CQ Press, 1994); Alan Ware, Political Parties and Party Systems (Oxford: Oxford University Press, 1996); Gary W. Cox, Making Votes Count: Strategic Coordination in the World’s Electoral Systems (Cambridge: Cambridge University Press, 1997); Pippa Norris, The Battle for the Campaign Agenda in Anthony King and others, New Labour Triumphs: Britain at the Polls (Chatham, NJ: Chatham House, 1998); David Farrell and Rudiger Schmitt-Beck (ed.), Do Political Campaigns Matter? (New York: Routledge, 2002). 7 Richard S. Katz, Democracy and Elections (Oxford: Oxford University Press, 1997); S.C. Stokes, “Political Parties and Democracy, ” Annual Review of Political Science, 2 (1999), 243–267; Ingrid van Biezen, How Political Parties Shape Democracy (Irvine: Center for the Study of Democracy: 2004); Simona Kustec Lipicer (ed.), Politične vsebine in volilna kampanja: slovenska izkušnja z volitev v evropski parlament 2009 (Political Contents in Electoral Campaigns: Slovenian Experience with the 2009 European Parliament Elections) (Ljubljana: Fakulteta za družbene vede, 2010). See also Hermann Schmitt. Determinants of Dyadic Correspondence in European Parliament Elections, 2008. Available at http://www.jhubc.it/ecpr-riga/ virtualpaperroom/122.pdf (15 September 2010). Journal of Comparative Politics 55 ideological positions and/or policy statements, what kind and type of them, etc. 4) Others would analyse their future impacts, like what happens after the election (i.e., how the content is ‘translated’ to further party-, political- and policy-making and processes in general from, for example, the coalition agreement point of view, state strategic documents, budget perspectives, etc.). All this leads us further to the time dimension of analyzing electoral party programmes, which is according to the already existing studies oriented either in 1) The pre-election period; or 2) The post-election period. The described complex understandings of the positions and roles of electoral programmes can be summarized in Figure 1 below. It is also true that this kind of understanding has so far been presented prevailingly in Western countries with long democratic traditions and much less in former undemocratic ones, including those from the so-called post-communist territory. The main aim of this article thus lies primarily in the ambition to narrow this gap, and to disclose the prevailing characteristics and types of party electoral programmes in the case of Slovenia, as one of the younger post-communist ‘success stories’ of democracy. Doing so, the above analysis focuses on the prevailing characteristics (similarities and differences) and hierarchy of electoral party programmes according to their outlook (e.g. type of programme document and length) and content dimensions, being compared through time from pre- and post-electoral perspectives (in the last two national parliamentary electoral processes); through the up-to-date status of the political party (parliamentary vs. non-parliamentary; coalition vs. opposition); and through the further impacts of the coalition agreement content. Journal of Comparative Politics 56 Figure 1: Possible dimensions of party electoral programmes 2 On applied methodology As can already be seen from this short theoretical introduction, the central units of analysis in this article are party electoral programmes, to be specific those prepared by the Slovenian political parties for the purposes of the 2004 and 2008 national parliamentary elections. In total we are dealing with 21 programmes, 8 equally split between elections (11 for the 2004 elections and 10 for the 2008 elections). 9 Some analysed documents (7 in each sample) were composed especially for the elections (electoral programmes); others are long- term party programmes that were used in unchanged format to compete for votes (party programmes). 10 Included in the analysis, there are also two post- election documents, namely the Coalition Agreements of 2004 and 2008. All 23 8 See Table A in the Attachment. 9 All the analysed documents were voluntarily submitted by political parties in 2010 with the main aim of being analysed as a part of an ongoing basic research project entitled “Pre-election campaign and democratic evolution of state and society, ” conducted by the research team at the Centre for Political Science Research at the Institute of Social Sciences, University of Ljubljana, financed by the Slovenian Research Agency (1/5/2009–30/4/2012). 10 Most of the time we will not distinguish between these two formats and use the same terms (programme, document) for both although the parties that officially competed based on their general programs claimed that they used them as the electoral programs at the same time. Journal of Comparative Politics 57 documents (units of analysis) are presented in Table A of the Appendix in total with their basic characteristics and the characteristics of the parties’ electoral outcomes. Due to incomplete responses, the sample of documents is a non-random convenient sample and cannot be used to draw conclusions about the population. On the other hand, the sample consists of all parliamentary 11 parties’ programmes and a good part of the non-parliamentary 12 parties’ programmes, which makes it a supreme purposive sample. Additionally, all the documents are considered as the best presentation of party positions by the parties themselves which resolves the data validity question. There is a reasonable symmetry in other sample characteristics, as well. There are seven parliamentary party programmes in the 2004 sample and eight in the 2008 sample. There are two strong parties (high share of votes and seats won) and three or four parties that won no seats in each sample (the rest won modest shares). There are four coalition members in the 2004 and three in the 2008 sample, meaning that the electoral programmes for all coalition members for both periods are included in the sample. The methodology of content analysis of party programmes has been used for the set purposes whereas for the outlook part of the analyses, a simple descriptive content analysis and word count have been used, while for the content part of it, MARPOR coding methodology and schema have been applied. 13 As a result, each party programme in the sample (including coalition agreements) is characterised according to seven policy domains. The more sentences in the document that resemble the domain, the more important the domain is in the document. Analytical conclusions are drawn mostly from graphical presentations, supported by multivariate analytical methods results. 11 Parties that won seats in parliament either at previous elections or at the elections in question (or both, as most parliamentary parties did). 12 Parties that didn’t hold or win any seats in the period under investigation. 13 See Annika Werner and Andrea Volkens. Manifesto coding instructions, 2009. Available at http://www. wzb.eu/zkd/dsl/daten/marpor/MDCodingManual.pdf (3 November 2010); also see Ian Budge et al, Mapping Policy Preferences (Oxford: Oxford University Press, 2001); also see Hans-Dieter Klingemann et al, Mapping Policy Preferences II: Parties, Electorates and Governments in Eastern Europe and the OECD 1990–2003 (Oxford: Oxford University Press, 2006). Journal of Comparative Politics 58 3 Electoral programmes outlook The variation in document length, measured by the number of nuclear sentences, varies quite obviously between the programmes as well as between both periods (the variation in 2008 is higher due to extremes, but differences are still modest and therefore acceptable). Due to a relatively low number of units of comparison between document length and type against electoral results, these can be best illustrated in graph format. Graph 1 below demonstrates that all but two of the parties that won no seats presented party programmes and not electoral programmes and that all their documents but one do not exceed a thousand sentences in length (see a relatively short column of mostly grey circles rising from zero on the vertical axis). Additionally, it appears that as a trend the number of seats is rising with the length of the programme and that the grey circles disappear as the number of seats goes up. Most successful parties (according to number of seats and coalition membership) presented electoral programmes, close to 2000 sentences in length (see a cloud of black circles in the far right-hand part of graph). 14 The trend seems to be independent of election year (no part of the graph hosts particular year documents). There is one obvious outlier to these trends, namely LDS in 2008. The party has gone through a lot of internal changes and re-organisations 15 just before the 2008 elections, ending in its drastic lost of power and votes. Apart from parliamentary LDS and other, less obvious outliers-like NSi in both analysed years or parliamentary SLS in 2008 (all won a lower number of seats as expected according to the trend) - there is a lot of empty space in the graph (between 10 and 23 seats and above 2000 sentences where no party programme could be found). So, from methodological point of view, speculations about the trend have weak fundaments and could only be applied on a limited scale (e.g. 5 to 10 seats and up to 700 sentences or 23 to 29 seats and 1500 to 2000 sentences). 14 The exception is DeSUS, being the only party that has parliamentary and even coalition member status even with a low number of seats and only a relatively short party program. 15 Twelve (out of 23) deputies left the party in the outbreak of 2007 . The core of the deputies who left and other past party members formed a new party, Zares. Journal of Comparative Politics 59 Graph 1: Document length and type against electoral results 16 Even more important, from a theoretical point of view the relation is likely to be spurious; it is not the type or the length of the programme that creates electoral success but wider dimensions- including the party’s political competence, organisation, selection process, tradition, strength, expertise- that produces more relevant programmes and attracts voters at the same time. But although it might be true that a strong party simply can’t afford not to produce a well- prepared electoral programme, at the same time this might not guarantee electoral success (see the example of LDS or NSi in 2008). Therefore, without applying any causal relations, it is safe to conclude that most successful parties use electoral programmes of considerable length while unsuccessful parties mostly use short party programmes. In between, there are units that behave according to this pattern as well as units that deviate, and most of the variation can be explained only on an individual basis. 16 The vertical axis depicts document length and the horizontal axis the number of seats won (parliamentary threshold marked by a dashed line is 4%). Each electoral programme is presented as a black circle and grey circles represent party programs. All circles are labelled and the election year appears right after party name. For the details about the voting system see Jurij Toplak, “The parliamentary elections in Slovenia 2004, ” Electoral Studies, 25, 4 (2006), 825–831; also see Danica Fink Hafner, “Slovenia, ” European Journal of Political Research, 48, 7–8 (2009), 1106–1113. Journal of Comparative Politics 60 4 Electoral programme content So far, the characteristics of analysed documents are merely superficial. In political science the party programme content is more valued than the outlook. In Table B in the Appendix, the content of the analysed programmes is presented according to the MARPOR scheme, as a share of sentences devoted to the domain. As can be seen from the table, there are many values, numerical in character, and they vary considerably. Apart from the simple fact that the rank order of domains remains basically the same between both points in time, it is difficult to draw any conclusions solely on visual inspection of the table. Therefore, in the next analytical steps, we first focus on the general characteristics of each period and then on more detailed comparisons between all analysed programmes. In line with previous conclusions, we will draw mostly on graphical presentations, supported by multivariate analytical methods results. In the first step, both compared periods -2004 and 2008 -are characterised by the documents' “general structure” 17 and by the post-election Coalition Agreement as the most important single document. The domain structure of these four documents is presented in Graph 2 by the height of the bars. 18 Graph 2: General structure (all party documents total as a single document) and Coalition Agreement structure in 2004 and 2008 17 See Table B in the Appendix. 18 Each document’s presentation is composed of seven bars, shaded in from white to black, representing seven domains. The higher the bar, the more the domain is significant (higher share). The vertical axis is by percentage, from 0% to 40% and is fixed for all following graphs. Journal of Comparative Politics 61 There are similarities and differences between the four documents, but apparently the similarities prevail. From a reasonable distance all four shapes appear to be roughly the same. All four are clearly characterised by the largest share of economy (around 30%), followed by political system (around 25%) in second place and welfare and quality of life in third place (around 15%). The first three domains overcome the next four, i.e. fabric of society, social groups, external relations and freedom and democracy, with shares up to 10% (around 5% to 9%). Such a structure seems universal in party documents (represented by the merged document called general structure) and coalition agreements, is therefore mainstream in both periods. Differences between time points seem to be smaller than differences between various types of documents. Both general programmes are more similar to each other than to any of coalition agreements and vice versa. All differences regarding less represented domains are almost negligible and all differences regarding the top three domains are more noticeable. Both coalition agreements contain more about the political system and less about welfare and quality of life as compared to general structure, while economy varies unsystematically between the years (but remains the most represented domain). Small differences in coalition agreements regarding reductions in economy and political system shares from 2004 to 2008 lead to more attention to welfare, quality of life, and social fabric. Modest differences in general structure can be characterised as a shift towards the economy on behalf of political system share from 2004 to 2008 as the financial crisis approaches. There is also a tiny increase in social group share and a tiny decrease in external relations share from 2004 to 2008. Further comparisons between individual documents, created in two time periods lead to higher complexity that can’t be directly depicted in a single graph. To compare all possible pairs of 23 units according to the seven numerical characteristics and to pay additional attention to pairs of programmes from the same party, as well as all pairs including coalition agreements, along with bringing the characteristics of all documents and parties into account requires the application of multivariate clustering methods followed by graphical presentation of the supplemented results. 4.1 Type of electoral programmes content Focusing on document content, to recognize various levels of data structure without serious information loss, multivariate hierarchical agglomerative cluster analysis 19 has been applied. Applied methods are capable of producing 19 For details about agglomerative hierarchical cluster analysis and applied methods and distances see Richard A. Johnson and Dean W. Wichern, Applied Multivariate Statistical Analysis: IV/12 Clustering (London: Prentice Hall, 1992). Journal of Comparative Politics 62 classifications in such a way that units inside a class are as similar as possible according to all characteristics, and classes are as different as possible according to the same characteristics. Methods rely on all variables and do not reduce the number of dimensions as various scaling of factor analysis methods do. Results vary with (dis)similarity measures and linking methods, but can be evaluated as reliable if a similar structure is uncovered by comparable measures and methods. In our case, while classifying 23 documents according to seven domains similar results have been provided with all combinations of Euclidean and squared Euclidean distance with centroid, within-group, complete and Ward method but one. A hierarchical agglomeration tree produced with the most commonly employed combination of Euclidean distance and Ward linkage, depicted in Figure 2, has been recognized as a typical representative of all created agglomeration trees. Journal of Comparative Politics 63 Figure 2: Dendrogram depicting data structure regarding 23 documents characterised by seven domains (Euclidean distance and Ward method) Drawing on the dendrogram (agglomeration tree), there are four types of documents, three of them more similar and one quite distinct. Internally (within groups), the level of similarity remains high and equal across groups. On the following pages, group composition, as depicted by the dendrogram, will be analysed with additional focus on interparty stability between elections and resemblance of coalition agreements. Simultaneously group structure according to all domains, depicted in Graph 3, will be studied to realize group content and Journal of Comparative Politics 64 meaning. 20 Graph 3: Group structure and total structure according to the seven domains. The first group, with the most similar documents, contains both coalition agreements and SDS programmes. Almost perfect stability in coalition agreements and highly similar SDS programmes can be traced between elections. In 2004, SDS won elections, formed a coalition, and led the coalition until the next election in 2008 that brought a completely different coalition to power. Yet, the coalition agreement structure was not changed much from 2004 to 2008 and the SDS programme structure remains closest to it, although the party was in opposition at the time of the 2008 coalition agreement. The first group differs from the general structure 21 in paying far more attention to the topic of the political system (the highest share in all groups) and less to almost all other domains, especially welfare and quality of life (the lowest share in all groups), while roughly resembling the total share of economy. Political system minus welfare and quality of life seems to be the formula of coalition agreements and SDS programmes that compose the first group of party programmes. On the other end of the scale, the fourth group of party programmes is composed of four small parties, which appeared only at one election, either 20 Domain structure of groups is depicted as the height of bars. Each group’s presentation is composed of seven bars, shaded in from white to black, representing seven domains in exactly the same manner as in Graph 2. The higher the bar, the more the domain is significant (higher share). 21 In sum, the total structure is almost identical to the previously in-depth described general structure in 2004 and 2008: approximately 30% of economy, 20% of political system as well as of welfare and quality of life, 10% of fabric of society and social groups, 6% of external relations and freedom and democracy. Journal of Comparative Politics 65 2004 or 2008, and won no seats. That group is the most different from the others (including the coalition agreements) according to all domains. The last group is clearly characterised by the prevailing domain of welfare and quality of life (the highest share in all groups; the second highest share of all domains in all groups), highest shares of fabric of society, social groups and freedom and democracy as well as the lowest shares of economy and political system. Almost exclusive interest in welfare, society and democracy domains seems to be hallmark of the small but not lasting SEG and GZS in 2004 along with Lipa and LPR in 2008. In between, there are two similarly structured groups of documents (according to their medium size and high internal similarity), but only modestly similar according to domains. In the first of the two, i.e. in the second group, there are three parties that so far have had an opportunity to lead the government, at least indirectly, and that have all undertaken important inter-party changes. In this group, the following programmes can be found: programmes of the LDS party that was for 12 years (1992–2004) the dominant coalition party 22 and was significantly reorganised before the 2008 elections; the electoral programme of the newly established (from a part of LDS) Zares party; the programme of a current coalition power, SD (re-named from ZLSD in 2005 with the aim of distancing itself from the ex-communist legacy);and the programme of NSi, which was, in a way, indirectly in power for a few months in 2000 when the coalition was led by SLS+SKD, a party which afterwards split apart in to an independent SLS party and the newly established NSi. The interparty stability of their programmes is very high, despite tectonic changes in their political strength over the last two elections. Similarity of their documents with coalition agreements (and matching SDS programmes) is less than modest, although higher than in the case of the fourth group. The second group is concerned mostly about the economy (the highest share in all groups; the highest share of all domains in all groups) and proportionally less about all other domains except perhaps external relations. All domains but political system are more present than in the first group. Economy with a fair share of everything else seems to be the motto of LDS, SD, NSi and Zares. In the third group, there are five parties, with two of them experienced only in the 2004 elections, namely AS and SEG. For the rest, high interparty stability can be confirmed since the SNS, SLS and DeSUS programmes from both elections are included. The structure of all documents is quite similar to the structure of the second group and not very close to the coalition agreements (and matching SDS programmes). For this group, the economy is less significant (but not least) while welfare and quality of life, and to a lesser extent fabric of society and 22 With a slight interruption for a few months in 2000 (see also Table A in the Appendix). Journal of Comparative Politics 66 social groups, are more significant (but not most). A balanced interest in welfare and society domains seems to be typical of DeSUS, SLS and SNS as well as of AS and SSN in 2004. 5 Electoral Programme inter-relation complexity In the final step, to present relations between 23 documents forming four groups, together with initially introduced programme and party characteristics in a single picture, the network analysis graphical potential has been exploited. 23 In Figure 3, documents are nodes, depicted as geometric shapes. Circles represent parties in opposition, squares represent parties in power, and triangles represent coalition agreements. The size of nodes is proportional to document length (the longer the document, the bigger the circle, square or triangle). The name of the party appears on the right-hand side of each geometric shape. If the name is underlined, the document is a party programme and, if not, an electoral programme. The number of seats won by each party is printed inside the geometric shape and font size is proportional to the value. If no seats were won, there is a large X to the left of the node. The shading represents election year; black for 2008 and grey for 2004. Relations between the nodes are Euclidean distances, taking into account all seven domains. The spatial distribution of nodes (documents) is parallel with dissimilarity measures that have been used in hierarchical agglomerative cluster analysis. The visual distance between documents corresponds to their differences in all seven domains: the further apart the nodes are, the more different the documents are; the closer the nodes are the more similar the documents are in structure. 24 Added curves separate groups obtained with cluster analysis and added text is used to annotate the personality of each group. 23 A freeware program, NetDraw, has been used; see http://www.analytictech.com. 24 Distortion due to two-dimensional presentation is rather small (Stress coefficient equals 0.17). Journal of Comparative Politics 67 Figure 3: The big picture Drawing on Figure 3 it is easy to recognize that election year has absolutely no impact on document content; grey and black shapes are mixed all over the diagram. 25 The logic behind such a mixture is in the stability of document content through time, meaning that the 2008 version of a party’s programme is always close to the 2004 programme of the same party. Other characteristics seem to be more significant although there are no strict rules and exceptions are always present. Starting from coalition agreements (very similar although serving very different coalitions), one can conclude that a focus on the economy or political system pays off since there are impressive election results enclosed within an imaginary circle around these coalition agreements (SDS4, SDS8, SD8, LDS4, NSi4 and also DeSUS8, SLS4 and SNS4 with less impressive results as well as a more balanced programme structure). Most of these documents are extensive electoral programmes and all but one of their creators is associated with the actual or previous coalition. The further from the core group we move, the less of its just described characteristics are present; quite balanced coverage of domains with stress on 25 Although it appears that in 2004 the documents were more different than in 2008 since there is more grey than black on the margins. Journal of Comparative Politics 68 economy or political system is gradually replaced by a narrow focus on welfare and various society domains. At the same time electoral results are diminishing toward zero on the margins, leading to a clear prevalence of circles indicating opposition parties (most of them also without parliamentary status), using mostly party programmes of modest length. All exceptions to the above are to be found in the economy stressing group: NSi8 and LDS8 with poor electoral results and SD4 and Zares with better-than-expected results. But there are obviously other more conceptual reasons for that than documents structure and outlook. 6 Conclusion Based on the theoretical predisposition that party electoral programmes count and should be given special research consideration, especially when they relate to youngish, post-communist democracies, we first proposed a three- dimensional perception of their possible analysis (see Figure 1) and then presented the case study of all available Slovenian party electoral programmes prepared for the last two national parliamentary elections, analysing the first, inner-circle party programme outlook and content dimensions. Research has shown that focusing on each party separately reveals a great deal of similarities between programmes that can be traced through time, across the party’s parliamentary/non-parliamentary status, and even across the party’s coalition membership status. Parties do keep their programmes rather stable despite the passage of time and changes in the policy arena and the same is true even for post-election coalition agreements. Consequently, differences in outlook between initially different groups of parties- e.g. between parliamentary and non-parliamentary parties- remains quite stable despite modest changes in group composition. Similar conclusions could also be revealed for the content dimensions of the analysed documents, although in this case more significant differences can be detected and patterns can be recognized. Accordingly, four different types of electoral party programmes appear: 1. Those emphasising the issues relating to political-system characteristics (both SDS programmes and coalition agreements); 2. Those emphasising economy related issues (LDS, SD, NSi and Zares); 3. Those oriented primarily to welfare and society issues in connection with the economy (parliamentary parties DeSUS- the only permanent coalition party, SLS and SNS- the only permanent opposition party; and non-parliamentary parties of AS and SSN); and 4. Those programmes, typical of the non-parliamentary SEG, GZS, Lipa and LPR, that almost neglect economic issues and focus instead on welfare and quality of life, social groups, democracy and freedom in their Journal of Comparative Politics 69 programmes. Last, but not least, we would like to emphasise again that the article is deliberatively descriptive in its nature since no corresponding kind of research data and analysis have existed so far in the case of Slovenia. We are aware that such pioneer role can not completely justify the fact that no initial argument per se is developed in the article or that a comparison with at least one other country that would enable for a more complex assessment of Slovenian party programs is missing. But, assuming that this first preliminary analysis represents a useful ground for further research while opening up many important and broader research questions and challenges we are convinced that this contribution could be seen as a trigger for a range of many other that could be written on its initiative basis. In general the subject of the article addresses one of the currently important phenomena related to the characteristics of democratic political system making and the deficits connected with them by asking about the ‘real’ nature and characteristics of electoral programme as an important component of democratic electoral process. In this regard analytical results of the article reveal the necessity for further research to focus on more in-depth explanations and discussions about programme characteristics, the reasons for them, and possible post-election impacts - which all move us forward to more general questions of whether and what kind of potential party programmes can nowadays in a specific democratic political system framework become important capital of the broader electoral scene and wider political processes. References Biezen, van Ingrid. How Political Parties Shape Democracy. Irvine: Center for the Study of Democracy: 2004. Budge, Ian, and Richard I. Hofferbert. “Mandates and Policy Outputs: U.S. Party Platforms and Federal Expenditures. ” American Political Science Review, 84 (1990): 111–131 . Budge, Ian, and Michael Laver. “Policy, Ideology and Party Distance: Analysis of Election Programmes in 19 Democracies. ” Legislative Studies Quarterly, 11, 4 (1986): 607–617 . Budge, Ian, Hans-Dieter Klingemann, Andrea Volkens, et al. Mapping Policy Preferences. Oxford: Oxford University Press, 2001 . Cox, W. Gary. Making Votes Count: Strategic Coordination in the World’s Electoral Systems. Cambridge: Cambridge University Press, 1997 . Farrell, David, and Rudiger Schmitt-Beck (ed.). Do Political Campaigns Matter? New York: Routledge, 2002. Journal of Comparative Politics 70 Fink-Hafner, Danica. “Slovenia. ” European Journal of Political Research, 48, 7–8 (2009): 1106–1113. Johnson, A. Richard, and Dean W. Wichern. Applied Multivariate Statistical Analysis: IV/12 Clustering. London: Prentice Hall, 1992. Katz, Richard S. Democracy and Elections. Oxford: Oxford University Press, 1997 . Keefe, William. Parties, Politics, and Public Policy in America, Washington DC: CQ Press, 1994. Klingemann, Hans-Dieter, Andrea Volkens, Judith Bara, Ian Budge, and Michael D. McDonald. Mapping Policy Preferences II: Parties, Electorates and Governments in Eastern Europe and the OECD 1990–2003. Oxford: Oxford University Press, 2006. Klingemann, Hans-Dieter, Richard I. Hofferbert and Ian Budge. “ A Theory of Democratic Policymaking. ” In Parties Policies and Democracy, ed. Hans-Dieter Klingemann, Richard I. Hofferbert and Ian Budge with Torbjorn Bergman, Hans Keman, Francois Pétry and Kaare Strom. Boulder, CO: Westview, 1994. Kustec Lipicer, Simona (ed.). Politične vsebine in volilna kampanja: slovenska izkušnja z volitev v evropski parlament 2009 (Political Contents in Electoral Campaigns: Slovenian Experience with the 2009 European Parliament Elections). Ljubljana: Fakulteta za družbene vede, 2010. Kustec Lipicer, Simona and Nikolina Bilavčić. “Volilni programi in volilne vsebine skozi volilno izkušnjo volitev v Evropski parlament 2009 (Electoral programs and contents through the lens of European Parliament 2009 electoral experience). ” In Politične vsebine in volilna kampanja: slovenska izkušnja z volitev v evropski parlament 2009 (Political Contents in Electoral Campaig: Slovenian Experience with the 2009 European Parliament Elections), ed. Simona Kustec Lipicer. Ljubljana: Fakulteta za družbene vede, 2010. Laver, Michael and Ben W. Hunt. Policy and Party Competition, New York: Routledge, 1992.. Mair, Peter. “Myths of Electoral Change and the Survival of Traditional Parties. ” European Journal of Political Research, 24 (1993): 121–133. Maisel, Louis Sandy. “The Platform-Writing Process: Candidate-Centered Platforms in 1992. ” Political Science Quarterly, 108, 4 (1994): 671–698. Norris, Pippa. The Battle for the Campaign Agenda in Anthony King and others, New Labour Triumphs: Britain at the Polls. Chatham, NJ: Chatham House, 1998. Pomper, Gerald. “If Elected, I Promise: American Party Platforms. ” Midwest Journal of Political Science, 11 (1967): 318–352. Robertson, David. A Theory of Party Competition. London: Wiley, 1976. Schmitt, Hermann. Determinants of Dyadic Correspondence in European Parliament Elections, 2008. Available at http://www.jhubc.it/ecpr-riga/ virtualpaperroom/122.pdf (15 September 2010). Smith, Craig Allen, and Kathy B. Smith. “ A Rhetorical Perspective on the 1997 Journal of Comparative Politics 71 British Party Manifestos. ” Political Communication, 17 (2000): 457–473. Stokes, S. C. “Political Parties and Democracy. ” Annual Review of Political Science, 2 (1999): 243–267 . Toplak, Jurij. “The parliamentary elections in Slovenia 2004. ” Electoral Studies, 25, 4 (2006): 825–831 . Ware, Alan. Political Parties and Party Systems. Oxford: Oxford University Press, 1996. Werner, Annika, and Andrea Volkens. Manifesto coding instructions, 2009. Available at http://www.wzb.eu/zkd/dsl/daten/marpor/MDCodingManual.pdf (3 November 2010). Journal of Comparative Politics 72 8 Appendix Legend of party abbreviations used: AS – Active Slovenia DESUS – Democratic Party of Pensioners of Slovenia DS – Democratic Party of Slovenia GŽS – Women’s Voice of Slovenia LDS – Liberal Democracy of Slovenia LPR – List for Justice and Development NSi – New Slovenia – Christian Peoples Party SDS – Slovenian Democratic Party SDSS – Social Democratic Party of Slovenia SDZ – Slovenian Democratic Union SDZS – Slovenian Democratic Union SEG – Ecological Movement Party of Slovenia SKD – Slovenian Christian Democrats SLS – Slovenian People’s Party SLS+SKD – Slovenian People’s Party + Slovenian Christian Democrats SNS – Slovenian National Party SSN – Party of the Slovenian Nation SOS – Slovenian Craftsman’s Party SSS – Socialist Union of Slovenia SKZ – Slovenian Farmers Union ZLSD (now SD) – United List of Social Democrats (now Social Democrats) ZS – Green Party of Slovenia ZARES – Zares– New Politics Journal of Comparative Politics 73 Table A: Analysed documents in the electoral context (*Non- parliamentary political party; + marks internal documents) 26 Party and Election Y ear No. of Sentences Programme T ype Share of Votes No. of Seats (total 90) Coalition 24 Member before/ after the elections SDS 2004 1972 Electoral 29.08% 29 No/Yes LDS 2004 1782 Electoral 22.80% 23 Yes/No NSi 2004 1530 Electoral 9.09% 9 No/Yes Coal.Agr. 2004 1382 Non applicable SEG 2004* 882 Party 0.41% 0 No/No SD (ZLSD) 2004 643 Electoral 10.17% 10 Yes/No SSN 2004* 634 Party 0.21% 0 No/No AS 2004* 608 Electoral 2.97% 0 No/No DeSUS 2004 254 Party 4.04% 4 Yes/Yes SLS 2004 195 Electoral 6.82% 7 Yes/Yes SNS 2004 148 Electoral + 6.27% 6 No/No GŽS 2004* 83 Party 0.54% 0 No/No LDS 2008 3649 Electoral 5.21% 5 Yes/No SD (ZLSD) 2008 1802 Electoral 30.45% 29 No/Yes SDS 2008 1505 Electoral 29.26% 28 Yes/No Coal. Agr. 2008 1340 Non applicable NSi 2008 1197 Electoral 3.40% 0 Yes/No SLS 2008 748 Electoral 5.21% 5 Yes/No DeSUS 2008 363 Party 7 .45% 7 Yes/Yes 26 So far we have had nine different governments in independent Slovenia: 1) in the period 16/5/1990– 14/5/1992 coalition DEMOS constituted of the following political parties SDZ, SDZS, SKD, SKZ, SOS and ZS (lead by Lojze Peterle, SKD); 2) in the period 14/5/1992–25/1/1993 coalition of SDS, DS, ZS, LDS, ZLSD and SSS (lead by Janez Drnovšek, LDS); 3) in the period 25/1/1993–27/2/1997 originally the coaltion of LDS, SKD, ZLSD and SDSS (lead by Janez Drnovšek, LDS); SDSS exist from coalition in 1994, and ZLSD in 1996; 4) in the period 27/2/1997–7/6/2000 coalition of LDS, SLS and DESUS (lead by Janez Drnovšek, LDS); 5) in the period 7/6/2000–30/11/2000 coalition of SDS and SLS+SKD (lead by Andrej Bajuk, SLS+SKD); 6 and 7) in the period 30/11/2000–3/12/2004 coalition of LDS, ZLSD, SLS and DESUS (lead by Janez Drnovšek, LDS (in the period 30/11/2000–19/12/2002) who was elected for the president of the republic in 2002; and by Anton Rop, LDS in the period 19/12/2002–3/12/2004); 8) in the period 9/11/2004–21/11/2008 coalition of SDS, NSI, SLS and DESUS (lead by Janez Janša, SDS); 9) in the period beginning 21/11/2008 with a coalition of SD (re-named from ZLSD), Zares, DESUS, LDS (lead by Borut Pahor, SD). Journal of Comparative Politics 74 ZARES 2008 329 Electoral 9.37% 9 Not existed/Yes Lipa 2008* 255 Party 1 .81% 0 No/No SNS 2008 131 Electoral + 5.40% 5 No/No LPR 2008* 38 Party 0.56% 0 No/No Journal of Comparative Politics 75 Table B: Programme relative structure according to MARPOR domains 27 Document Economy PoliticalG System W elfareG andG QualityG ofGLife FabricGofG Society SocialG Groups ExternalG rela- tions FreedomG andG demo- cracy AS 2004 24.34 13.82 35.53 8.55 11 .18 5.26 1 .32 DeSUS 2004 29.92 14.57 22.05 7 .09 7 .09 8.27 11 .02 GZS 2004 2.41 2.41 43.37 19.28 9.64 3.61 19.28 LDS 2004 38.83 22.95 14.81 2.97 5.16 7 .13 8.14 NSi 2004 35.62 18.76 12.55 10.59 10.72 6.60 5.16 SDS 2004 23.93 37 .02 12.52 6.08 8.32 7 .51 4.61 SLS 2004 27 .69 20.51 14.87 8.21 15.90 8.21 4.62 SNS 2004 22.30 20.27 18.92 10.14 12.16 11 .49 4.73 SD (ZLSD) 2004 33.44 12.13 20.06 7 .78 15.09 4.35 7 .15 SEG 2004 11 .11 7 .71 40.93 20.86 7 .37 7 .03 4.99 SSN 2004 18.14 19.72 19.40 26.34 8.99 4.73 2.68 Coalition Agr. 2004 35.17 31 .55 9.55 4.63 7 .45 7 .96 3.69 TOTAL 2004 29.03 23.00 17.93 9.07 8.75 6.87 5.35 DeSUS 2008 21 .76 17 .63 22.59 5.51 13.22 9.64 9.64 LDS 2008 42.94 9.67 16.03 8.08 10.11 5.51 7 .65 NSi 2008 36.09 13.78 18.05 11 .28 11 .19 4.93 4.68 SDS 2008 29.90 30.56 15.15 6.38 9.70 4.98 3.32 SLS 2008 25.13 12.43 28.34 10.43 12.30 7 .62 3.74 SNS 2008 24.43 10.69 26.72 19.08 9.92 8.40 0.76 SD 2008 33.35 21 .59 16.37 5.83 11 .21 4.27 7 .38 Lipa 2008 14.12 6.27 32.16 16.08 17 .25 3.53 10.59 LPR 2008 13.16 7 .89 26.32 15.79 7 .89 13.16 15.79 ZARES 2008 41 .03 9.12 25.53 2.74 8.21 2.43 10.94 Coalition Agr. 2008 30.75 28.06 12.99 7 .84 10.15 4.10 6.12 TOTAL 2008 34.67 17 .28 17 .64 8.06 10.69 5.21 6.45 27 Since each sentence has been linked to a single domain only, the sum of shares is always 100. According to the election year, the table is split into two parts. At the end of each part of the table the general structure of all documents in total is presented, i.e. number of sentences in each domain summed and recalculated as a share applicable if all individual documents in a period were only one large document. Naturally, longer documents have a stronger effect on the general structure, but that is reasonable. Parties are arranged in alphabetical order; domains are arranged in descending order according to their frequency.