The second-order character of past European elections is a well-established hypothesis with respect to voter turnout and voting behaviour. This paper presents a conceptual framework for testing this hypothesis on the supply side of European elections. It includes three groups of indicators allowing for the comparison of national and European election manifestos to determine the latter's second-order nature: (1) resource allocation, measured by the number of actors involved in manifesto adoption and by the manifestos' length; (2) the manifestos' content, comparing issue congruence and framing; and (3) political competition, measured by the ideological distance between manifestos. Building on this, we analyse the manifestos of all relevant German parties in the 2009 European and national elections using Comparative Manifesto Project data as well as original, self-created data on election manifestos. We find major variations between parties which can neither be explained by government participation nor satisfaction with the European Union. INTRODUCTION European elections are commonly ascribed as second-order elections. Initially, this thesis was developed for the demand side of elections, or voter behaviour, 1 and has been tested as such on several occasions. 2 Going beyond this perspective, questions about parties and their relationship to European elections arise. More precisely, one can ask whether parties also regard European elections as second-order elections and what impact this type of election has on parties' behaviour. In order to answer these questions, an analysis of the allocation of parties' supply is needed to determine whether parties show different allocation patterns in European as opposed to national elections. In other words, by comparing parties' supply of three commoditiescandidates, campaigns and election manifestos -we can evaluate whether parties treat first-and second-order elections differently. Due to the mix of methods and range of data necessary for these analyses, existing empirical studies concentrate on the analysis of each of these three 'commodities' separately. Thus far, only studies on campaigning provide a clear analytical framework for testing the second-order nature of European elections as well as empirical results for single countries, in particular Germany. More precisely, comparative analyses of election campaigns reveal that German parties invest far fewer resources in European than Downloaded by [Monash University Library] at 15:12 03 February 2015 national election campaigns. 3 Additional studies of German election campaigns show, in general, a lower degree of the campaigning professionalism often found in European elections but, furthermore, highlight an increasing professionalism between the campaigns in 2004 to 2009. 4 Whether this finding however is influenced by the subsequent national election four months later is not clear yet. 5 Moreover, a second-order nature can be partially confirmed for individual candidates' campaigning behaviour as they spend less time...