Our understanding of how and why federal systems exhibit diverging trajectories is still limited. Scholars have advocated different explanatory factors so as to account for federal system dynamics, most notably territorial cleavages~for example, Erk!. This paper argues that historical institutionalism offers a valuable meta-theoretical approach capable of overcoming the shortcomings resulting from the isolation of one explanatory factor at the expense of others. In particular, as historical institutionalism puts emphasis on the historical construction of both institutions and ideas when accounting for political dynamics, it is suggested that it bears considerable potential to address important research questions raised in the literature on comparative federalism.In order to assess the value of this approach to the analysis of federal system dynamics, the paper proceeds in three steps. First, it demonstrates how federalism can be conceptually reformulated as a multilayered political order, comprising an institutional and an ideational layer. Second, it introduces two models of change, the model of path dependence and the process sequencing model, both providing different types of historical explanation. Finally, the paper tests how each Acknowledgments: An earlier version of this paper was presented at the 2007 CPSA annual conference. I would like to thank my discussant André Lecours for his excellent comments. I am also grateful to Karl Froschauer, Nina Gerstenkorn, Bettina Helbig, Rainer-Olaf Schultze and the CJPS reviewers for their valuable comments and suggestions.