Strategic voting in two-round elections: confirming the learning hypothesis in the Czech presidential elections of 2013-2023?

This paper explores whether strategic voting can also be identified in two-round electoral systems. We use aggregate data at the level of Czechia's 206 microregions to assess the incidence of strategic voting in the Czech presidential elections using two-round elections between 2013 and 2023. We find the following evidence of a considerable rise of strategic voting in the 2023 presidential election: (1) rise of support for the two strongest candidates nationally, as well as at the level of individual microregions; (2) decline of strategic failure and sincere voting indicated by shifting values of the second-first loser ratio; (3) increased electoral competition approaching strong competitive two-party dominance as shown by Nagayama diagrams. These results point to the importance of the 'learning hypothesis' that strategic voting in new democracies tends to increase as voters (but also political elites) become more experienced with the electoral process. Furthermore, while we present a case study of a Czech presidential election, we also make an important contribution to the literature on strategic voting. Therefore, a theoretical contribution of this study also lies in its effort to apply the Duvergerian logic to a very different context of presidential elections in which electoral competition is highly personalised.