1France has always fascinated comparative scholars throughout the world. Its rather original political space, particularly its polarized party system, and sometimes unstable political life make for an especially rich case study. Because of these unique characteristics, numerous specialists in France are quick to cite a “French exception”, justifying their reluctance to adopt a comparative perspective. And yet several recent studies suggest that, on the contrary, French politics and institutions are a great deal less exceptional than we would like to think.  These observations are particularly supported by the study of party competition. In the mid 1990s, French party politics specialists largely adhered to the idea that the French political space was tripartite, organized around three clusters situated on the left, right, and extreme right of the political spectrum. This interpretation has since been challenged, and some have proposed that France has returned to a more classical, essentially bipolar, form of electoral competition.
2This article does not seek to fuel this controversy, but rather sets out to explore a dimension that, though essential in the theoretical literature on party competition, has to date remained largely neglected: the policy issues highlighted in the context of the party game, and the relation between political supply and the structure of political competition. This realignment is promising, not only from a comparative perspective, given the predominance of issue competition theories in the international literature. It also allows for an analysis of the more subtle changes, which the cleavage lens developed by Seymour Martin Lipset and Stein Rokkan would not permit, as its focus is extremely general. In other words, the fact that party competition is structured by major historical cleavages, either by a left-right axis or two approximately orthogonal dimensions, provides insight into the operative “format” and political grammar, but sheds light on neither the “substance” of the electoral agenda nor the dynamics of its evolution.  Similarly, electoral competition remains commonly – albeit more or less explicitly – understood to be largely exogenously determined by the identity, affiliation, and preferences of voters. Viewing this competition as a systemic agenda that each party contributes to shaping but that none controls completely rehabilitates the weight of interactions among parties.
3Based on the analysis of original data collected by the Agendas France project on the attention given to different political issues within electoral platforms, this article seeks to contribute to a renewal of the theories and methodologies employed in explaining French political competition. We will present and briefly discuss the debates around the nature of French party competition, before explaining our methodological approach and presenting our findings.
Understanding the dynamics of French electoral competition: from cleavages to issue competition
4Most research on French party competition is characterized by the adoption of a very broad perspective – the format of the competition  – and by an ascending vision centered on political demand rather than supply. This work is primarily structured by a controversy around the transformations of the French political space, which ties into the broader discussion on the evolution of values in Western democracies since the late 1970s. Like other countries, France experienced “a violent materialist reaction” to post-materialism, which translated, in particular, into the rise of the Front National (FN, National Front). According to Gerard Grunberg and Étienne Schweisguth, these changes resulted in the transformation of the French political space from a bipartite to tripartite structure during the period between the 1988 and 1995 presidential elections.  From that moment on, the attitudes of French voters have been better explained by the juxtaposition of two transversal cleavages: on the one hand, a classical left-right cleavage relative to the desirable degree of interventionism, taxation, and redistribution; on the other, a “cultural” dimension thought to have emerged perpendicularly to the former, associated with issues such as immigration, Europe, or globalization. 
5This hypothesis provoked the publication of numerous works aiming to identify and describe this second dimension. In the French context, their analysis was contested by Robert Andersen and Jocelyn Evans, who challenged its empirical foundations and viewed the persistence of the left-right axis as the most structuring dimension of competition, the cultural dimension having no significant impact.  In an article published in the previous issue of this journal, Vincent Tiberj develops a new approach modeling the impact of both dimensions, and to a certain extent confirms their existence in the French political space. 
6Regardless of the definition of “cleavage” selected, the works referred to above adopt a very broad perspective, developed in order to reconstruct the long-term evolutions of the format and social anchors of competition. While the alignment, disalignment, and realignment phenomena are interesting, they alone do not account for the dynamics of electoral competition, to the extent that they neglect the question of shorter-term changes and the issues that are concretely set forth in different campaigns. Furthermore, the different contributions to the debate seem to share a number of premises that are not necessarily self-evident. In particular, they share a common conception of political life that is essentially “bottom-up”, privileging political demand. For this reason, the evolution of voters’ values and identities, for instance, constitutes a crucial parameter in the explanation of the National Front’s high scores. And yet, although the elections sanction the parties running for election, parties are likely, in the context of uncertainty and electoral volatility, to be primarily concerned with the strategies of their rivals. This article strives to test this intuition by adopting an approach centered on political supply.
7In addition, the debate is dominated by a spatial conception of French party competition, seen primarily as the confrontation of antagonistic points of view. The rich literature developed around the matter of electoral competition and strategies nevertheless suggests that parties demarcate themselves from one another not only by adopting distinct positions, but also by attempting to impose their signature issues on the electoral agenda. From this perspective, it is indispensable to explore the content of this agenda and the dynamics of “issue competition”, a concept first defined by David Robertson,  who views party competition as “selective emphasis rather than direct confrontation”.  This idea gave rise to numerous works that, by mobilizing various notions such as “issue emphasis”,  “salience”,  or “issue ownership”,  conclude that there exist different thematic profiles of parties, and that parties have an interest in attracting attention to the issues that they “own”.  A brief overview of the works published on the idea of issue ownership shows that this concept is both polysemous and multidimensional. It is increasingly mobilized in a “perceptionist”  sense, in order to explain voting patterns by voters who are assumed to make their choices by turning to the party considered the most competent in the areas which they judge to be the most important.  The existence of issue voting implies that party competition is largely constrained by the systematic association of a party with one or more specific political issues. Moreover, numerous studies on issue competition and issue ownership specifically aim to shed light not only on electoral behavior, but also on party strategies.  From this perspective, noting the different levels of attention given to issues and interpreting the observed differences is key to understanding the party game.
8Attention to issues in the media agendas during electoral campaigns in France has regularly been the object of research.  These studies have sometimes studied the interactions between media agendas and public agendas in terms of agenda, framing, or priming effects.  However, save for the work of Jean-Louis Missika and Dorine Bregman  dedicated to the contents of parties’ official televised campaigns, party agendas have not sparked much interest in the French academic community, despite the data available on the thematic content of media coverage of different candidates.  In particular, in the French case, the hypothesis of issue ownership has not been tested in a systematic and longitudinal manner using data centered on the parties themselves. However, it has been contested by several authors working on other cases, who empirically ascertain a strong thematic overlap between different parties’ platforms.  Beyond these observations, there exist sound theoretical reasons to doubt the existence of stable issue ownership. Indeed, if a party lays claim to an issue and is able to garner an electoral benefit from this “property”, rival parties will be strongly driven to contest the legitimacy of the “owner” in order to avoid granting the party a monopoly over the framing and orientation of the issue, and to deflect criticism for having ignored the subject. Tracy Sulkin  describes the process of appropriation of rivals’ political priorities (issue uptake) in her study of the political agenda of outgoing members of the US congress. Outgoing elected representatives, including in non-competitive constituencies, are active in the main areas emphasized by their opponents. The fact that the idea of issue ownership seems to suffer from an overly static vision of politics and struggles to obtain empirical support does not necessarily indicate that issue competition does not exist. Thus, a dynamic analysis of the content of French parties’ electoral agenda has promise, both to better understand how these parties compete with one another and to progress towards deeper knowledge of the articulation between positional competition and issue competition.
An original methodological approach
French electoral competition: an interesting case
9As mentioned previously, French party competition has never been the object of a systematic analysis of issue competition. However, it proves to be an interesting case to the extent that the French party system experienced profound changes between 1981 and 1993, and subsequently relative stability. During the first decades of the Fifth Republic, the executive functions were exercised exclusively by Gaullist or centrist formations, with the Socialist Party (PS) and Communist Party (PCF) remaining confined to the opposition. The early 1980s marked a true turning point. On the one hand, the PCF electoral results entered a long decline, falling from 20.7% in the first round of the 1978 legislative elections, to 4.3% in the first round of the 2007 legislative elections. On the other hand, 1981 marked the first Socialist victory in both the presidential and legislative elections, and initiated a period of regular alternation between the PS and the RPR-UDF alliance.  France experienced its first cohabitation between 1986 and 1988, then a succession of minority left-wing governments until 1993.
10Even more importantly for the issues of interest to us, this period witnessed the confirmation of two minor parties. The National Front (FN) joined the National Assembly in 1986, following a spectacular breakthrough in the 1984 European elections (during which it reached nearly 11% of votes cast). A few years later, the Ecologist Party (les Verts) found its place on the national stage by obtaining a score of 10.5% in the 1989 European elections and almost 14% in the 1992 regional elections.  Given the priority accorded by the FN to immigration and security themes and by the Verts to environmental protection, these two parties can be considered examples of niche parties (see below). Their emergence enables us to analyze the impact of the transformation of the party system on the content of the electoral agenda, by observing when and how the attention dedicated to these issues by mainstream parties evolved, using the period of relative stability that began after 1993 as a point of comparison. In line with this perspective, this article concentrates on the priorities of the two primary political formations, the PS and the RPR-UDF alliance.
The Agendas Project data: new perspectives
11Empirically analyzing the dynamics of issue competition implies access to data on the attention given to different issues by French parties over time. The classic sources for this are expert survey data (in the tradition of the poll described by Peter Mair and Manuel Castells in 1984) and the data collected by the Comparative Manifesto Project (CMP).
12While certain expert surveys are to be commended for including salience measurements (notably in the case of Chapel Hill data collected by Leonard Ray),  this type of poll suffers from important reliability issues when producing series a posteriori, to the extent that experts are not retrospectively capable of determining the weight of different issues in each party’s argument with sufficient precision.  Additionally, observations are limited by the pre-selection of a limited number of issues on which experts comment.
13The CMP has the advantage of being based upon a relatively unified empirical foundation, composed primarily of political platforms, but using its database does entail several difficulties. On the one hand, the categories used to characterize the content of each quasi-sentence of a political program are not sufficiently precise to reconstruct the modalities of issue competition in their complexity: categories such as “social justice”, “democracy”, or “welfare state” encompass numerous issues that should be distinguished. Furthermore, the CMP grid combines true public policy issues such as “the welfare state”, and conceptual categories, such as “Marxist analysis”, “multi-culturalism” or “degrowth”, which have more to do with the parties’ positioning than their thematic profiles. Finally and most importantly, the French CMP data is characterized by specific weaknesses. Given the difficulty of obtaining certain French programs as well as their characteristic length, the CMP has sometimes coded other types of texts, or even included data from previous elections,  and France is the only country in which the coding unit is not the quasi-sentence but the paragraph.
14Conversely, the data collected by the Comparative Agendas Project (CAP) opens up promising possibilities to better understand the scope and modalities of issue competition. The CAP’s approach is certainly less ambitious than that of the projects cited previously, since it focuses exclusively on the level of attention granted to different issues and does not take into account the positions adopted by the parties. However, this lens is precisely what is needed for a more rigorous (the directional and non-directional approaches are not combined), fine, and precise operationalization of issue competition. The sources used are the platforms collected by the CMP, supplemented in cases where it was both necessary and possible. With the exception of the platform presented by the FN in 1988, which could not be found, this corpus covers all of the legislative programs presented by parties having obtained more than 3% of votes in the legislative elections since 1981: the Parti Communiste Français (PCF, French Communist Party), Parti Socialiste (PS, Socialist Party), Union pour la Démocratie Française (UDF, Union for French Democracy), Rassemblement pour la République (RPR, Rally for the Republic) (and the Union pour un Mouvement Populaire, UMP [Union for a Popular Movement] from 2002),  as well as the two main rivals we alluded to earlier, the FN and the Verts.
15These programs were divided into quasi-sentences, or in other words, into units of meaning taking the form of a grammatical sentence or phrase. Each of the 33,400 quasi-sentences thus defined was then coded according to its content, following a code grid translated and adapted from the original American codebook, developed by Frank R. Baumgartner and Bryan Jones.  The French codebook, like its counterparts used in the United States and elsewhere, includes 24 general themes (cf. Table 1) that are, on average, divided into ten subcategories, for a total of 250 exhaustive and mutually exclusive categories.  One of the strengths of the CAP data is that it is well-suited to the extremely precise and differentiated analysis of issue competition. But above all, this data opens up new prospects for the analysis of party competition. The inclusion of the original text of the programs in the database helps to make the analysis comprehensive by studying the framing of issues and the subsequent positions of parties, and thereby reveal the consensual, proprietary, or positional nature of each issue.  Finally, the use of the same schema to code party platforms and diverse institutional agendas – governmental, parliamentary, presidential, etc. – enables us to explore the links between party priorities and the themes put on the agenda by elected representatives.
Towards a new indicator of parties’ thematic priorities
16French electoral competition has been analyzed through the lens of issue competition, starting from the two hypotheses outlined above: issue ownership  and issue uptake. Exploring the dynamics of issue competition implies observing the level of attention dedicated to each issue by the various parties. We suggest approaching this question from the angle of the natural thematic hierarchy operated by parties. At a given moment, each of them strives to push certain topics to the forefront, while letting others fade into the background. In doing so, parties can choose to focus on or ignore the hierarchy of issues that has prevailed in the past.
17Thus, this study focuses on how the priority assigned to different issues has evolved, taking the manner in which governmental parties have ordered the 24 general themes listed in Table 1 above as a starting point. Our interest in this aspect of ordering drove us to use an indicator that has rarely been used: rather than comparing the level of relative attention dedicated to each issue, we based this study on the classification of issues in order of salience. From our point of view, privileging an ordinal measure over a cardinal measure boasts numerous theoretical and methodological advantages. 
18First of all, the classification indicator offers a more precise measure of the distribution of attention among issues, considered from the angle of their prioritization. Indeed, studying the share of attention can be misleading, since the share can increase or decrease without altering the issue’s place in the hierarchy of issues, and conversely, the classification of an issue can be affected regardless of change in the share of attention it receives. Furthermore, the variations in attention share are only rarely associated with changes in the ordering of issues.  Beyond the advantages of our indicator in giving value to the idea that issue competition is first and foremost a matter of priority assigned to different themes, this option is better adapted to the study of the intrusion of new issues, a still rarely studied aspect.
Issue ownership or convergence of thematic profiles?
19The hypothesis of issue ownership implies that certain parties are generally associated with certain themes. From this perspective, it is probable that they seek to impose their signature issues as electoral campaign themes, judged important by both their members and voters. This type of situation should translate into distinct thematic profiles, with each party focusing on different bodies of issues and achieving relative stability, to the extent that ownership must be sustainable over time in order to structure the behavior of voters. The issue ownership hypothesis can be tested by comparing the priority accorded, on average, to each theme in the PS and RPR-UDF alliance platforms between 1981 and 2007, using the share of attention (Figure 1) and classification (Figure 2) indicators.
Average share of attention given to different issue areas in the platforms of the two principle French parties (by percentage of quasi-sentences)
Average share of attention given to different issue areas in the platforms of the two principle French parties (by percentage of quasi-sentences)
20The difference between the two averages is usually slight, and significant for only two issues. As we might expect, socialist platforms pay more attention to issues related to work (p=.02 for both measurements) and civil rights (p=.00 for the percentage comparison and p=.03 for the classification comparison). On the other hand, contrary to what the theory of issue ownership might suggest, the RPR-UDF alliance does not assign significantly higher priority to issues pertaining to justice, crime, and immigration but, surprisingly, sets itself apart with its greater attention to education and social policy. This counter-intuitive finding reveals the absence of contrasting and stable thematic profiles. While it goes against the hypothesis of issue ownership, it is not incompatible with the ad hoc definition of distinct electoral priorities by the two parties, as illustrated by the campaign led by the UMP in 2002, largely centered on the theme of insecurity. Furthermore, it should be noted that our data do not imply that the two groups hold electoral discourses of similar substance, since they do not tell us anything about the stances adopted. The appropriation of rivals’ issues is liable to give rise to a confrontation of antagonistic analyses and proposals inspired by opposing value judgments. It is therefore very possible that the priority assigned by the RPR-UDF alliance to matters of education and social policy indicates an attempt to appropriate the PS’s signature issues.
21We explored this idea in greater depth by observing the degree of similarity among the priorities stated by the two groups during each election. This was done on the basis of an indicator measuring the share of issues common to both parties, by adding the absolute attention differences, issue by issue, between the two parties for a given election (Table 2).  These results line up with the previous findings and indicate a relative convergence of the PS and RPR-UDF alliance priorities, which share approximately half of their issues and operate a similar ordering (on average, a difference of three classification points).
Thematic convergence between PS and RPR-UDF alliance
Thematic convergence between PS and RPR-UDF alliance
22A second manner of testing this hypothesis consists of looking at the degree of stability of parties’ electoral priorities. Christopher Green-Pedersen and Peter Mortensen  developed an indicator of intra-party coherence allowing the exploration of this aspect, by adding the absolute attention difference issue by issue given by each party from one election to another. We completed this indicator of the proportion of attention dedicated to each issue using a second indicator of ranking (Table 3).
Intra-party coherence on issue attention
Intra-party coherence on issue attention
23These two measures reflect a certain level of stability in the priorities of the PS and RPR-UDF, both of whom, over time, have maintained stable attention to at least half of the themes. The highest level of stability was attained recently by the PS, which transferred barely 15% of the attention stated in its 2002 platform to other themes in 2007. Nevertheless, our analysis confirms that the priority issues are not fixed and that, as Stefaan Walgrave and Michiel Nuytemans suggest,  parties adjust their priorities to different extents depending on the election. For instance, between 1988 and 1993, the parliamentary right shifted more than a third of its attention towards new issues. The ranking of issues by the two groups varies on average by almost four places between 1981 and 1986, by 5.5 places between 1986 and 1988, but only by 2.5 places between 2002 and 2007. Therefore, the two indicators highlight the existence of considerable fluctuations in parties’ priorities over time. The level of intra-party coherence (over time) in the prioritization of issues is similar to the level of inter-party convergence (in a given election). The hypothesis of issue ownership is therefore not confirmed by the observation of a stable thematic profile for the two party groups.
24Finally, the evolution of attention to different areas can be analyzed by using kurtosis (k) and L-kurtosis (L-k) measures, which are common in the literature on agenda-setting to measure the distribution of variations in issue attention.  A high value for these indicators (k > 3 and L-k >. 123) indicates a leptokurtic distribution of these variations; in other words, a distribution in which very small and very large variations are over-represented relative to what is observed in a normal distribution. “The General Punctuation Hypothesis” formulated by Bryan Jones and Frank Baumgartner posits that the changes in the ordering of political priorities are distributed leptokurtically, both for cognitive and institutional reasons. The issue ownership hypothesis also predicts an over-representation of small variations, but offers an alternative explanation: parties are attached to the issues that belong to them, to the extent that their politicization would enable them to obtain votes. Yet the kurtosis and L-kurtosis values characterizing the distribution of variations in attention within the PS and RPR-UDF (measured in terms of classification) indicate a nearly normal distribution (k=3.4 and L-k=.183),  which confirms the graphical estimation of the distribution and several complementary statistical tests.  This last finding also runs counter to our expectation of a leptokurtic distribution that would go hand in hand with issue ownership.
25In sum, our analyses of the changes in the thematic priorities of the PS and the RPR-UDF do not enable us to corroborate the existence of “issue ownership”. None of the issues studied seem to belong to one party rather than another, and the two primary political groups share a majority of themes. Studied in terms of hierarchy, the changes in attention were slighter and more regular than what is generally expected: parties do not cling to a few signature issues and only exceptionally modify their priorities. Fierce electoral competition, in the context of which “constant adaptation is simply a matter of survival for parties”, seems to counterbalance the effects of cognitive limits, institutional friction, and the potential perception of electoral gains related to the politicization of certain issues. The parties appear constrained to dynamically adapt to their environment and to respond to challenges posed by their rivals’ strategies.
26Thus, party ideology does not appear to reduce the attention spectrum of the parties studied. While this situation can be viewed in conjunction with the thesis of the decline of ideologies in contemporary party competition, we can also suppose that adhesion to a particular ideology is first and foremost determinant in defining the party’s orientation and positioning. Because an ideology is a weltanschauung, it is never short of answers and offers a lens allowing parties to produce proposals for each new problem placed on the political agenda.
Niche-party priorities and their influence on the party agenda
27The absence of distinct thematic profiles could be due to the specificities of the two largest French parties, which can be characterized as “catch-all” parties. This type of organization builds its success on the coverage of a broad spectrum of areas and claims to represent the largest possible electoral clientele. If the primary parties share numerous concerns, it may be because they are competing for the same electoral sectors. Niche parties, on the other hand, base their legitimacy and their electoral appeal on the politicization of a limited number of issues.
In the context of their studies of the French party system, Bonnie Meguid and other niche-party specialists have mainly focused on two parties: the FN and the Verts. Table 4 examines the question of whether there exists a specificity of thematic profiles for these two parties, by providing a glimpse of the degree of variety in the areas addressed by each party in its programs, using the entropy H indicator developed by Shannon.  If, as Bonnie Meguid asserts, niche parties target a restricted number of issues, this should translate into a significantly lower level of entropy relative to that of other parties.“Instead of prioritizing economic demands, these parties politicize sets of issues that were previously outside the dimensions of party competition […]. The issues are not only novel, but they often do not coincide with the existing, ‘left-right’ lines of political divisions […]. Niche parties differentiate themselves by limiting their issue appeals. They eschew comprehensive policy programs common to their mainstream party peers, instead adopting positions on and prioritizing only a restricted set of issues” 
Entropic values for political programs in each election 
Entropic values for political programs in each election 
28Our data do not confirm the existence of an opposition between “monothematic” niche parties and “catch-all” parties, with more diverse preoccupations. The Verts’ and FN’s scores fall into the average, or even slightly above. We should note nevertheless that the PCF scores seem more revealing of a “monothematic” way: the platforms presented by this party in the 1981 and 1988 elections concentrated on a small number of issues, and the questions pertaining to economic policy and labor represented over 30% of the program, whereas other issues such as liberties, health, or energy policy were practically absent. Despite this, the PCF remains an “old” party that played a central role in the politics of the after-war years and cannot quite be designated as a niche party.
29In order to more accurately determine the weight of niche parties in issue competition, another strategy consists of concentrating on their ability to place their signature issues on the party agenda. According to Meguid, the combined strategies of mainstream parties – particularly those ideologically situated the farthest from the niche party – are decisive for the success or failure of such a party. A party situated to the center-right could, for example, deliberately politicize an issue dear to the radical left in order to weaken its center-left rival. The issue uptake hypothesis suggests a different scenario: if we view party strategies as intrinsically connected and developed within the limits of a highly competitive context, it seems unlikely that a mainstream party, that is to say one of those liable to form a government, can afford to ignore the discourse of a more radical rival if it is appealing to similar segments of voters. From this perspective, it is probable that the affirmation of niche parties is reflected first and foremost in the closest mainstream party.
30These different hypotheses have been explored by observing the changes in the PS and the RPR-UDF alliance’s attention to themes of justice and police, immigration – areas claimed by the FN – and the environment – a theme classically associated with the Verts – while taking into account the electoral performance of these two niche parties (Figure 3).
31The three examples presented above indicate that, in accordance with our hypothesis, the rise of niche parties is accompanied by an appropriation of their signature issues by mainstream parties, and particularly by the one that is the closest ideologically. The breakthrough of the Ecologists in the 1989 European and municipal elections – as well as in the 1992 regional elections, whose results are not represented here – led to a spectacular increase in the PS’s attention to the issue of environmental protection during the 1993 legislative elections: with an increase of around 5%, the environment moved up 5 places in the classification of socialist priorities between 1988 and 1993. Similarly, following the good scores recorded by the FN from the 1984 European elections, immigration appeared in the RPR-UDF alliance’s program – where it had been completely absent in 1981 – and immediately reached 5%. “Justice and police”, which includes issues such as “social order” and “homeland security”, follows a different dynamic. In contrast with topics like environmental protection or immigration, it is an older theme, historically claimed by the parliamentary right. The FN seems to have succeeded in contesting its ownership, which was translated by a peak in attention by the RPR in 1986, then by the increased weight of the security issue in the RPR’s programs in the 1990s, with a historic peak in 2002.
32Note that the appropriation of a new issue by one of the two mainstream parties can lead to adaptation by the other mainstream party. While the PS was the first to integrate environmental protection into its priorities, the RPR and UDF’s programs displayed increasing interest in this topic from the second half of the 1990s. Similarly, the politicization of security questions by the FN has stronger repercussions on the themes highlighted by the RPR, but the PS also seems to take up this issue, albeit later and to a lesser extent. Although this observation can be generalized, the variations in attention to different issues from one election to another should be positively correlated among competing parties, and particularly so among mainstream parties. Our data confirm that the changes in the classification of the 24 themes operated by the PS and the RPR-UDF alliance between 1981 and 2007 are very strongly correlated (.83**).
33Therefore, the attention given to issues follows a rationale neither of opposition nor differentiation. Whatever the reasons leading to changes in attention to issues – exogenous or endogenous – in political platforms, there seems to be a systemic rationale that drives parties to modify their attention towards the same issues. As expected with the issue uptake rationale, when a given party devotes more attention to an issue, its rival tends to imitate it and place this issue on a higher level in its own prioritization of issues.
34In sum, this analysis of the characteristics of the thematic profile of niche parties and their impact on “catch-all” parties’ agendas confirms the existence of strong interactions, even if the spectrum of niche parties’ priorities is less focused on a few themes than we might have expected. The ordering of electoral themes therefore does not seem to follow a rationale of opposition or differentiation. Regardless of the reasons behind a change in attention within a party, a systemic rationale appears to drive other parties to turn their attention towards the same issues. Niche parties could thus be one of the primary forces within the evolution of the electoral agenda: mainstream parties are attentive to these new competitors and strive to integrate their priorities, in particular when they are relatively close in ideology. In other words, the transformations of the party system are likely to translate into important modifications of the electoral agenda.
Party system transformations and the electoral agenda
35While the emergence of new party organizations and the dynamics of interaction among parties are essential to understand the composition of the electoral agenda, we should observe the different fluctuations in the ordering of issues according to the degree of change of the party system. Over the period studied, these fluctuations are normally distributed, but this does not mean that they continue to be over time. On the contrary, the issue-uptake hypothesis leads us to posit that the ordering of issues experiences more marked variations when the party system is transformed, which we test by comparing the period from 1981-1993, characterized by important changes, to the period from 1993-2007, during which the French party system experienced greater stability.
36The average variation in attention measured in terms of ranking is indeed considerably higher between 1981 and 1993 (4.4 ranks on average) than between 1993 and 2007 (2.8 ranks on average). Similarly, the proportion of attention dedicated to each issue changed 2.9 points on average before 1993, and only 1.7 points afterwards.  On average once again, the change in this proportion of attention is 159% before 1993 and 74% after. In accordance with our expectations and regardless of the indicator selected, the average stability of the electoral agenda seems to be significantly higher since 1993.
37Nevertheless, understanding the rationale of issue competition in France necessitates going beyond these average values, and grasping the timing of the changes. In order to do so, we looked at the punctuations in the distribution of variations in attention by the PS and RPR-UDF alliance.  Since 95% of the values in a normal distribution are included in an interval of approximately two standard deviations from the average, the values outside of this interval are treated here as punctuations (Table 5). Following this definition, we have established, in particular, the list of variations in attention greater than 9 ranks. The issues listed experienced either a strong increase or a strong decline in attention. 
Punctuation in the change in classification of issues between 1981 and 2007 
Punctuation in the change in classification of issues between 1981 and 2007 
38Very large-scale variations in attention – punctuations – were identified in 26 cases over the entire period. Half of the topics were affected by these variations, which are nearly uniformly distributed. The RPR-UDF alliance only modified its priorities slightly more frequently (14 times) compared to the Socialist party (12 times). On the other hand, the vast majority of variations did occur between 1982 and 1993 – between 80% and 88%, depending on the indicator used to measure the level of attention. The changes identified as punctuations are not always identical depending on the indicator  used but the relation between party system and electoral agenda is robust in all three cases. We are not affirming the existence of a status quo since 1993, but the priorities stated by parties in their platforms experienced slighter fluctuations over this period.
39Thus, the study of the relation between the transformations of the party system and the changes in the electoral agenda has provided further confirmation of the issue-uptake hypothesis. Indeed, the discussion has demonstrated that the partisan context is more important in determining the characteristics of issue competition than the supposed identity of the party, based on hypothetical issue ownership.
42Approaching French electoral competition through the lens of issue ownership enhances the understanding of the dynamics of competition among French parties. The data from electoral platforms coded by the Agendas France project, on which these first analyses are based, offer a solid foundation to progress in this direction. From a methodological point of view, measuring the attention to different issues in terms of ranking helps to overcome the biases from which the most common indicators suffer, and seems robust and appropriate to address issue competition.
43From a more substantive point of view, our findings show the advantages of the issue competition perspective by confirming that French political parties do not confront one another only by adopting opposing stances, but also by modulating their attention to the different issues. However, while certain scholars observe a systematic association of certain themes to a party – among voters or in the media – and note the existence of issue voting, our analyses do not verify the hypothesis of issue ownership with regard to the existence of stable and distinct thematic profiles of French political parties. In accordance with the prior observation of the great similarity of right and left-wing governments’ concerns in France, the PS and RPR-UDF alliance seem to operate similar hierarchical ordering from one election to another.
44Political parties have adjusted their priorities on a regular basis since 1981. The evolution in political supply can be observed in conjunction with “real world” events and voters’ evolving preferences. However, in agreement with the thesis of issue uptake, our analyses suggest rather that these variations are largely determined by processes endogenous to the party system, such as the strategies developed by partisan actors in order to avoid leaving the monopoly over certain themes to their rivals. Effectively, these variations follow the dynamics shared by different parties – endorsing the idea of the existence of a relatively constraining “systemic” party agenda – and their amplitude varies with the degree of stability of the party system. In particular, we have observed the direct impact of the affirmation of the FN and the Verts on electoral agenda-setting for questions related to immigration, security, and environmental protection. Thus, the relative stabilization of the French party system and the consecutive decline of the endogenous sources of change in the 1990s may explain why electoral priorities have become more stable, despite the increasing volatility of media coverage.
45The debates around cleavages and alignments cannot give a full and accurate picture of agenda-setting that is primarily endogenous to the party system. On the other hand, they remain necessary and complementary to understand the manner in which parties appropriate salient electoral issues. Are the majority of issues shared by all parties simply the object of a competition to appear the most active in that area, or do they give rise to the confrontation of antagonistic positions, potentially structured by major historic cleavages?
46Finally, a deeper understanding of how electoral issues are positioned within a hierarchy from one electoral campaign to another presupposes a better understanding of the impact of the “real” world, and particularly of social mobilizations and media attention. Their role constitutes a promising but complex new direction for research. 
The ‘party’ data from the CAP
47The CAP was initially launched by Frank Baumgartner and Bryan Jones to compare American political agendas: legislative, governmental, presidential and media agendas, and so forth. On the basis of the American project an international network was established, involving many European researchers, and the project has recently been extended to include the party agenda via the coding of electoral programs.
Electoral programs used for the PS and RPR-UDF alliance
48We have usually been able to obtain the text of each party’s electoral program for each legislative election. The only difficulties we encountered concerned legislative elections organized after the Assembly had been dissolved by the newly elected president: in this case the elections of 1981 (for which we do not have the PS’s legislative project) and 1988 (for which we were unable to locate either the agreement between the RPR and UDF, or the RPR program). It is conceivable that the fact that these elections were called early did not leave all parties enough time to draw up and circulate a proper program. For the PS in 1981 we have coded François Mitterrand’s 110 Propositions for France from his presidential campaign. For the RPR-UDF alliance the best source we found was Jacques Chirac’s campaign speech – which, moreover, has also been used by the Comparative Manifesto project.
Overview of the PS and RPR-UDF programs
Overview of the PS and RPR-UDF programs
49Following the approach developed by the CAP, the electoral programs were divided into quasi-sentences – arguments which might consist of a whole grammatical sentence, or part thereof. The rules governing what constitutes a quasi-sentence are precisely defined in the party handbook which is used by several of the CAP teams. Once the quasi-phrases have been isolated, coding consists of attributing one or several thematic codes (up to three) to each unit, referring to the thematic grid used to code the other agendas, which includes 28 general categories (macroeconomic policy, health, energy, etc.) and more than 250 subcategories (fiscal policy, organization of the healthcare system, management of nuclear waste, etc.).  Moreover, a series of additional dichotomous variables enable us to establish whether the quasi-sentence is declarative or contains a concrete declaration; whether it refers to the EU; whether it’s a title or subtitle; or whether it’s a passage criticizing the policies of other parties. The data thus obtained allow us in particular to retrace how the share of attention awarded by each program to each issue has evolved.
50The coding of the French platforms took place between October 2009 and June 2010. The five coders were trained on the basis of a test file and a series of exercises which enabled us to verify their reliability and clarify the rules of coding. During the coding the seminars run by the Agendas France project provided an opportunity for coders to discuss any issues, for the codebook to be refined, and, above all, for ensuring inter-coder agreement between those involved in the processing of the different agendas.
51Beyond these meetings, co-ordinating the work was facilitated by the use of a virtual coding interface. The text of the platforms was uploaded to a server in Amsterdam which was shared by a number of the CAP teams. The AmCat program, developed for our project by Wouter van Atteveldt, allows the team leader to create coding tasks in relation to these texts and allocate them to different coders; to have an overview of coding progress; and to compare the results from several coders working on the same platform. AmCat is connected to a coding interface, Inet, which enables each coder to access his or her texts easily, isolate the quasi-sentences, and allocate codes to them. The data are recorded on the central server and can be downloaded in a table format which contains not just the codes associated with each quasi-sentence, but also the original text. The inclusion of the text in the data guarantees maximum transparency in coders’ choices, at the same time as facilitating the creation of corpora of texts relating to particular areas with a view to qualitative analysis (on the framing of problems, for example, or the value judgments expressed by the parties).
Sylvain Brouard, Andrew M. Appleton, Amy Mazur (eds), The French Fifth Republic at Fifty? Beyond Stereotypes (Basingstoke: Palgrave, 2008); Emiliano Grossman, Nicolas Sauger (eds), “France’s political institutions at fifty”, West European Politics, 32(2), 2009.
Christoffer Green-Pedersen, “The growing importance of issue competition: the changing nature of party competition in Western Europe”, Political Studies, 55(3), 2007, 607-28.Online
C. Green-Pedersen, “The growing importance of issue competition…”.
Gérard Grunberg, Étienne Schweisguth, “Vers une tripartition de l’espace politique”, in Daniel Boy, Nonna Meyer (eds), L’électeur a ses raisons (Paris: Presses de Sciences Po, 1997), 179-218.
See Simon Bornschier, Romain Lachat, “The evolution of the French political space and party system”, West European Politics, 32(2), 2009, 360-83. This second dimension was later named an “open-closed” or “universalist-authoritarian” cleavage. It is clearly related to the opposition between materialists and post-materialists identified by Ronald Inglehart and to the hostile reactions to post-materialism highlighted by Piero Ignazi. See Piero Ignazi, “The silent counter-revolution: hypotheses on the emergence of extreme right-wing parties in Europe”, European Journal of Political Research, 22(1), 1992, 3-34; Ronald Inglehart, The Silent Revolution. Changing Values and Political Styles among Western Publics (Princeton: Princeton University Press, 1977).Online
Robert Andersen, Jocelyn A. J. Evans, “Values, cleavages and party choice in France, 1988-1995”, French Politics, 1(1), 2003, 83-114.Online
Vincent Tiberj, “La politique des deux axes. Variables sociologiques, valeurs et votes en France (1988-2007)”, Revue française de science politique, 62(1), 2012, 71-108.
David Robertson, A Theory of Party Competition (Oxford: Wiley, 1979).
Ian Budge, Dennis J. Farlie, Explaining and Predicting Elections. Issue Effects and Party Strategies in Twenty-Three Democracies (London: G. Allen & Unwin, 1983).
Edward G. Carmines, James A. Stimson, “Issue evolution, population replacement, and normal partisan change”, American Political Science Review, 75(1), 1981, 107-18; Issue Evolution. Race and the Transformation of American Politics (Princeton: Princeton University Press, 1990), and “On the structure and sequence of issue evolution”, American Political Science Review, 80(3), 1986, 901-20.
Ian Budge, Richard I. Hofferbert, “Mandates and policy outputs: US party programs and federal expenditures”, American Political Science Review, 84(1), 1990, 111-31; Richard I. Hofferbert, Ian Budge, “The party mandate and the Westminster model: election programmes and government spending in Britain, 1948-85”, British Journal of Political Science, 22(2), 1992, 151-82; Bonnie M. Meguid, “Competition between unequals: the role of mainstream party strategy in niche party success”, American Political Science Review, 99(3), 2005, 347-59, and Party Competition Between Unequals. Strategies and Electoral Fortunes in Western Europe (Cambridge: Cambridge University Press, 2008).Online
John R. Petrocik, “Issue ownership in presidential elections, with a 1980 case study”, American Political Science Review, 40(3), 1996, 825-50.Online
For a review of the definitions of issue ownership and the work developed from this idea, see Stefaan Walgrave, Jonas Lefevere, Anke Tresch, “The associative dimension of issue ownership”, Public Opinion Quartely, forthcoming.
The authors who adopt this “perceptionist” perspective have not reached a unanimous definition of issue ownership. As demonstrated by S. Walgrave et al. (see previous footnote), we can differentiate between an evaluative definition, tied to the perceived competence of parties in a given area, and a cognitive definition, accentuating the spontaneous association of a party with a theme.
G. Grunberg, É. Schweisguth, “Vers une tripartition de l’espace politique”; Wouter van der Brug, “Issue ownership and party choice”, Electoral Studies, 23(2), 2004, 209-33; Éric Bélanger, Bonnie M. Meguid, “Issue salience, issue ownership, and issue-based vote choice: evidence from Canada”, Electoral Studies, 27(3), 2005, 477-91; Bernt Aardal, Pieter van Wijnen, “Issue voting”, in Jacques Thomassen (ed.), The European Voter (Oxford: Oxford University Press, 2005), 192-212.
I. Budge, D. J. Farlie, Explaining and Predicting Elections…; C. Green-Pedersen, “The growing importance of issue competition…”; Jane Green, Sara B. Hobolt, “Owning the issue agenda: party strategies and vote choices in British elections”, Electoral Studies, 27(3), 2008, 460-76.Online
Jacques Gerstlé, “Les campagnes présidentielles depuis 1965”, in Pierre Bréchon (ed.), Les élections présidentielles en France (Paris: La Documentation française, 2007), 67-104; Christophe Piar, Jacques Gerstlé, “Le cadrage national d’un scrutin local” in Bernard Dolez, Annie Laurent, Claude Patriat (eds), Le vote rebelle (Dijon: Éditions Universitaires de Dijon, 2004), 89-102.
Christophe Piar, “La coproduction de l’actualité télévisée et son impact sur la formation des opinions des citoyens en campagne électorale”, PhD thesis in political science, Paris, Université Paris I-Panthéon Sorbonne, 2009.
Jean-Louis Missika, Dorine Bregman, “La campagne: la sélection des controverses politiques”, in Élisabeth Dupoirier, Gérard Grunberg (eds), Mars 1986: la drôle de défaite de la gauche (Paris: PUF, 1986), 97-116.
Jacques Gerstlé, Christophe Piar, “Les campagnes dans l’information télévisée”, in Pascal Perrineau (ed.), Le vote de rupture (Paris: Presses de Sciences Po, 2008), 21-50.
Lee Sigelman, Emmett H. Buell, “Avoidance or engagement? Issue convergence in U.S. presidential campaigns, 1960-2000”, American Journal of Political Science, 48(4), 2004, 650-61. David F. Damore, “The dynamics of issue ownership in presidential campaigns”, Political Research Quarterly, 57(3), 2004, 391-7.Online
Tracy Sulkin, Issue Politics in Congress (Cambridge: Cambridge University Press, 2005).Online
In 2002, the RPR and the majority of the UDF merged to form the UMP.
Another fleeting green party, Génération Écologie (Generation Ecology), allied itself with the Verts for the 1992 and 1993 elections.
Leonard Ray, “Measuring party orientations towards European integration: results from an expert survey”, European Journal of Political Research, 36(2), 1999, 283-306.
Ian Budge, Hans-Dieter Klingemann, Andrea Volkens, Judith Bara, Eric Tanenbaum (eds), Mapping Policy Preferences, Estimates for Parties, Voters and Governments, 1945-1998 (Oxford: Oxford University Press, 2001).
Thus the data on the UDF are constant for three successive elections in the CMP database.
The RPR and the UDF are analyzed here based on their shared electoral platforms (“shared” in that they are based on pre-electoral coalition agreements). The only exception is the 1988 election, for which we do not have a shared program and had to use that of the RPR alone. These documents, systematically signed and agreed before elections, were preferred over the RPR programs since the two parties generally fielded joint candidates. Moreover, this choice guarantees a greater continuity given the fusion of the RPR and a majority of the UDF in 2002. Note that the thematic priorities stated in each election by the RPR and by the union with the UDF are largely similar.
See Frank R. Baumgartner, Bryan D. Jones (eds), Policy Dynamics (Chicago: University of Chicago Press, 2002). For the original codebook, see <http://www.policyagendas.org>.
The codebook is available from the authors on request.
Isabelle Guinaudeau, Simon Persico, “L’UE dans la compétition électorale en France, en Allemagne et au Royaume-Uni: un enjeu consensuel, ‘propriétal’ ou positionnel?”, in Mathieu Petithomme (ed.), L’européanisation de la compétition politique nationale. Adaptations et résistances en perspective comparée (Grenoble: Presses Universitaires de Grenoble, 2011), 73-95.
Our analysis is related to partisan competition dynamics, rather than vote determinants. When we discuss issue ownership, we are referring to the work that observes or assumes the existence of distinct and stable thematic profiles of political parties, and to the implications of issue voting for party strategies. Our observations thus do not directly contribute to the analysis of issue ownership in the “perceptionist” sense.
The results based on our new indicator (classification) were systematically controlled for using pre-existing indicators (the percentage of percentage and percentage index method).
The percentage index – the difference between the percentage of attention at time t0 and time t1 for each theme – is inconvenient in that it is sensitive only to the most important changes in absolute terms. For instance, a fall in average importance leading to the total disappearance of an issue from a party’s agenda could not be detected using this approach. The inverse is also true: an average increase in the weight of a previously absent issue risks passing by unnoticed. Finally, major negative changes are conditioned by the issue’s prior salience. The percentage of percentage method – the difference between the percentage of attention at time t0 and time t1 divided by the percentage of attention at time t0 for each theme T – constitutes the most common indicator. It is biased in that it is unable to address the lack of attention and overestimates the most minute changes for the most marginal issues. See Stefaan Walgrave, Michiel Nuytemans, “Friction and party manifesto change in 25 countries, 1945-98”, American Journal of Political Science, 53(1), 2009, 190-206.Online
This indicator was originally developed by L. Sigelman and E. H. Buell, “Avoidance or engagement?…”
Christopher Green-Pedersen, Peter B. Mortensen, “Issue competition and election campaigns: avoidance and engagement”, Potsdam, ECPR Conference, 2009, available at <http://www.agendasetting.dk/papers/uploaded/115201044905PM.pdf> (viewed in November 2011).
S. Walgrave, M. Nuytemans, “Friction and party manifesto change in 25 countries, 1945-98”.
F. R. Baumgartner, B. D. Jones, Policy Dynamics.
We conducted analyses including and excluding the changes in classification of issues receiving little or no attention. There is no impact on the result of the analysis of the distribution of changes.
This refers to a series of standard tests: kurtosis, Shapiro-Wilk W test, and Shapiro-Francia W’ test. However, punctuations appear in the distribution of changes in attention given to issues using the percentage of percentage method (k = 6.35 and L-k =. 230) and the percentage index (k = 45.6 and L-k =. 357).
B. M. Meguid, Party Competition Between Unequals…, 3-4.
The H indicator developed by Shannon is a very widespread diversity indicator, which has the advantage of being less sensitive to changes in magnitude than comparable indicators, such as the Hirschmann-Herfindahl index. Shannon’s H is written as: . The version used here is relatively easy to read: the closer it comes to zero, the more the program focuses on a limited number of issues; conversely, a score close to 1 indicates a perfectly even distribution of attention among all themes.
The calculation was done using the 24 primary themes in the codebook.
Figure 3 indistinctly presents the election results for niche parties in the European and legislative elections.
In both cases, the average difference is highly statistically significant (p <. 001). The percentage-percentage method also provides congruent and significant results (p <. 005).
We previously established that the ordering of issues in party programs is not punctuated. This implies that punctuations are not over-represented (such as in a leptokurtic distribution), but not that all the changes in the order of priorities are of the same magnitude.
Given that certain issues fall into the same ranking among the 24 differentiated themes, the classification by each party in each election includes between 18 and 23 ranks depending on the case (20.7 on average). A variation of 9 rankings then means that the relative position of the issue has experienced a change of 45% on the scale of values.
The numbers in bold indicate which indicator(s) identified the changes as a punctuation.
As we assumed, the diverging results between the ranking change and the percentage index have to do with the divergences between the changes in quantity of attention and the changes in priority of attention. With good reason, a variation of attention that only marginally affects the hierarchy of issues is not detected as a punctuation by the approach centered on the rankings of issues. Conversely, and in contrast with the percentage index method, it allows us to grasp slighter variations that upset the ordering of issues. The biases inherent to the percentage-percentage method are illustrated by these results, since this method does not allow detection of decreases in attention and leads to an over-representation of slight variations in attention concerning initially marginal questions. This is notably the case with the energy issue, which gained 2 places and 1.3% of attention in the 2007 socialist program compared to 2002, for which the relative proportion of the increase reached 710%.
This article would not have been possible without the collaboration and invaluable research work of Simon Persico and Caterina Froio, as well as the generous funding granted by the Agence nationale française de la recherche (National French Research Agency) and the Agenda France project (“ANR-Gouverner-55” grant).
The coding table is available at <http://www.agendas-france.fr>.