Local Districts, National Contexts, and the Number of Parties

Electoral Institutions, Social Diversity and Fragmentation of Party Systems: A Reassessment

“…Minimal linguistic diversity at district level is possibly also the reason why, in the interaction with the exclusion magnitude, model performance does not improve in our study. This is in contrast to other more recent studies based on international comparative parliamentary elections (for example Clark and Golder ; Geys ; Rashkova ; Singer and Stephenson ; Tiemann ), but in line with the findings of Li and Shugart (: 25f.) who qualify findings on the interactive nature of permissive electoral systems and high social heterogeneity as weak and inconsistent in parliamentary systems.…”

“…Moreover, multilevel models permit the analysis of contextual effects while still allowing for heterogeneity between contextual units (Steenbergen and Jones : 127). In recent years, the estimation of multilevel models has become the standard approach in similar studies on party system fragmentation (Riera ; Rashkova ; Tiemann ; Singer ). Additionally, null models for the abovementioned dependent variables demonstrate that the variance at cantonal level is 71 percent for the raw number of parties standing for election and 60 percent for ENPV.…”

“…In this case, neither psychological nor mechanical effects play a measureable role. However, low fragmentation can exist where there is a high demand for different parties but the electoral system is restrictive (Tiemann : 130). The interaction between sociological heterogeneity and the corresponding demand for parties on the one hand, and party system fragmentation on the other hand, should therefore be more pronounced in open electoral systems than in restrictive systems.…”

“…Subsequent studies no longer restricted their attention to the electoral formula, but also included other factors of the electoral system. According to large parts of the research, district magnitude represents one of the most important institutional determinants of party system fragmentation (Amorim Neto and Cox ; Geys ; Ordeshook and Shvetsova ; Rae ; Singer and Stephenson ; Taagepera and Shugart ; Tiemann ). The larger the districts, the better societal heterogeneity in the districts can be translated into the party system.…”

“…Most of these studies, however, use only mean values at the national level to estimate the effect of sociological variables (for example Riera : 129; Singer : 211f) it they account for sociological factors at all. As districts within a country may vary considerably in their socio‐structural design, there is a danger that distorted estimates will result (Rashkova : 250; Tiemann : 140). Consequently and following the recommendation of Amorim Neto and Cox (: 168), we utilise district level data for sociological variables as far as possible…”