This article tests Duverger's law through analysis of the Polish Senate elections in 2011 and 2015. These two elections were held under the new first-past-the-post (or single-member plurality) system, which replaced formerly used unlimited vote. The main aim of the article is to test, whether we can confirm the expectations of strategic voting in the context of the so-called second-order elections, as the Polish Parliament is a classical example of the asymmetrical bicameralism, with the secondary role of the upper chamber, the Senate. The results show that the strategic voting was not universal phenomenon under the plurality rule, as indicated by many violations of Duverger's law. Our research confirmed that the effect of electoral institutions (institutional structure) is contingent and (at the district level) inhibited by country-specific conditions, with potentially strong influence of the second-order character of the Polish Senate elections.
KEY WORDS:Poland; second-order elections; party system; strategic voting; Duverger's law
INTRODUCTIONRecent empirical research on voting in single-member districts (SMDs), based on extensive datasets of election results, demonstrated the general (although not perfect) validity of Duverger's law (1954), i.e., that the average outcome under plurality rule is generally consistent with two-party competition (Singer and Stephenson, 2009;Singer, 2013). However, most authors dealing with Duverger's original assumptions have been usually given considerable attention to the effects of electoral rules in the so-called first-order elections, i.e. elections to the lower chambers of the national parliaments, just because the lower chamber elections are crucial in the process of national government formation.3 1 This article is one of results of the project "Spatial analysis of voter behaviour and strategic voting", which was financially supported from funds for institutional research of the Faculty of Arts, Jan Evangelista Purkyně University, for 2016.