Learning the Incentives of Differing Mandates Under a Mixed-Member System: Party Loyalty and Legislative Voting in Lithuania

According to the differing mandates hypothesis, under a mixed-member electoral system re-election seeking legislators from single-member districts (nominal tier) should defect from party line more often than legislators that are elected trough the proportional system (list tier). On paper, Lithuania looks like a very convenient case to find support for this hypothesis, as electoral system here is a pure mix of majoritarian (71 MPs; two-round system) and proportional formulas (70 MPs; list preferential). However, all the major studies (Ishiyama 2000; Clark et al. 2008; Preece 2014) on Lithuania up to date found a contrary result: list legislators defect significantly more often than the nominal tier legislators. Weak social cleavages, low re-election rate, candidate-selection procedures and geographical dispersion of preference votes were raised as explanations for this peculiar finding. On the other hand, it is also possible that previous studies were inconclusive due to methodological problems, as they analyzed quite short time frames (four or six years) and their data was collected from a period (2000-2006) when party system exhibited highest-ever levels of volatility. In our paper, we analyze all the available roll call votes in the Lithuanian parliament from 1997 to 2018. We look for a learning effect: with the increasing institutionalization of party system and rising re-election rate of legislators, incentives of mixed-electoral system should become clearer. We also add two theoretical contributions to the discussion on the differing mandates hypothesis and test them empirically. First, party discipline in a post-communist country could be related to the year of election cycle and government/opposition status. Due to the hyper-accountability of post-communist voters, almost all governments lose their popularity eventually. Therefore, incentives for the nominal tier legislators to defect from party line may be strongest in the year preceding the parliamentary elections and among the government parties. Second, we also take into account the phases of lawmaking. Not all the roll call votes are equal. There are three legislative votes in case of every Lithuanian law: submission of draft, consideration and adoption. As the last phase and respective roll call vote is the most important, differing incentives should manifest most clearly precisely in this group of legislative votes.