Party Switching Over a Century: The French Case (1910-2024)

Legislative party switching (LPS) – when elected officials change their parliamentary party group (PPG) affiliation during their term – poses a fundamental challenge to representative democracy by disrupting the electoral mandate and weakening party labels. While extensively studied in new democracies and presidential systems, the phenomenon remains underexplored in consolidated parliamentary democracies, particularly over extended historical periods. This paper addresses this gap by providing the first systematic analysis of LPS in France across more than a century (1910-2024), encompassing three distinct regimes and 26 legislatures. Drawing on an original database of 19,262 parliamentary affiliations systematically collected from parliamentary archives (Official Journal 1910-2024), we document 2,886 switches involving 2,087 deputies. This unprecedented temporal coverage allows us to observe LPS dynamics through major institutional transformations: from the absolute parliamentarism (1910-1940), through an unstable regime (1946-1958), to a rationalized parliament (1958-2024). Our analysis proceeds in two stages. First, descriptive analyses map temporal evolution and qualitative modalities of switching across regimes (directions, collective/indivdual). Second, we employ multivariate regression models to identify causal determinants of switching, testing the effects (office-, vote- and ideology-seeking). Preliminary descriptive findings reveal exceptional instability under the Fourth Republic (rate of 37.99%) and nearly identical switching rates between the Third and Fifth Republics (18.72% and 18.50%). However, qualitative differences emerge: the Third Republic shows more individual and structural switching, while the Fifth Republic experiences more collective and episodic switches. The multivariate stage will test competing explanations for these patterns through multilevel models accounting for temporal dependencies and institutional contexts. We will examine whether the counter-intuitive similarity between Third and Fifth Republic rates persists after controlling for party system characteristics, and whether the positive association between polarization and switching reflects genuine causal mechanisms or spurious correlation driven by electoral realignments. Additional analyses will explore interaction effects between institutional centralization and party system fragmentation, hypothesizing that centralized institutions may simultaneously reduce individual switching while incentivizing collective exits to capture institutional prerogatives. This research contributes to comparative legislative studies by demonstrating the limits of institutional determinism in explaining parliamentary behavior. The persistence of relatively high switching rates despite the Fifth Republic's rationalization suggests an "irreducible share" of switching shaped by partisan dynamics rather than formal rules alone. By integrating long-term institutional variation with party system evolution, this paper advances our understanding of how formal and informal constraints interact to shape legislative defection in consolidated democracies.