Revisiting the Culprits of Party System Fragmentation in Brazil

• In Event: New Data New Approaches: Political Science in Brazil Today

Abstract

The very high fragmentation of Brazil's legislative branch is often assumed as the outcome of the combination between extremely proportional electoral rules and high social heterogeneity. Legislative elections are often associated to an overly permissive Open List PR system, with seat allocation taking place in very large electoral districts. Undoubtedly, a combination of the very same institutions that comparative literature usually points as being the main villains of countries' party system fragmentation.

However, while being partially guilty as charged, these institutions are neither the sole nor necessarily the main culprits. A simple but compelling evidence of that lays in the frequently forgotten fact that only the Brazilian Lower Chamber is elected under those rules. The Brazilian Senate, although also presenting a high party system fragmentation, is actually elected by an one-turn majoritarian system of 1-to-2 district magnitudes. Consequently, the famously accused electoral institutions of Brazil are not to be blamed alone. In this article, my intent is to indict other suspects.

First, I use electoral data on more than 100 countries since 1945 to demonstrate that even among other countries with similar macro electoral institutions, Brazil is an extreme outlier when it has to do with legislative party system fragmentation. Its strikingly high effective number of parties – be it at large or within districts – deviate from any pattern of permissive electoral rules and also from the comparative expected effect of social heterogeneity.

Next, I resort to computational simulations and to brand new electoral data on the Brazilian national elections and on partisan pre-electoral coalitions. I start showing that the low level of party nationalization plays a crucial role in increasing the final national fragmentation in Brazil, in comparison to the average within-district fragmentation. This is specially true regarding the Senate, but also true for the Lower Chamber. In order to show that, I use Monte Carlo simulation to algorithmically estimate a distribution of what would have been the shares of seats of each party in the two national chambers in each general election at many different levels of nationalization of the parties’ electoral support – where nationalization is understood as the territorial homogeneity of the spread of votes.

Then, I investigate to which extent adopting a minimum threshold of votes for distributing seats would have changed the seats distribution in all general elections since 1994. In order to do that, I simulate the seat shares at varying levels of national and regional thresholds. By doing that, it becomes possible to not only estimate for the first time the effect of the lack of minimum threshold on the Brazilian legislative party system fragmentation, but to also identify the optimum thresholds. I also simulate seat shares with and without the actual allowance of legislative electoral coalitions per district, showing that those have a strong effect in boosting within-district fragmentation - actually similar to the lack of formal electoral threshold.

Even more importantly, I will argue that the key to understand the high fragmentation of the Brazilian party system lays in the multi-level electoral coalition strategies that are coordinated by parties across the different offices that are at stake in the multiple federal levels of the country. The reason for it is that the electoral rules of Brazil tie the electoral coalitions that a party can enact for each office: in each of the Brazilian states, a party has to have the exact same coalition for state governor and for the Senate, from which the coalition for the Lower Chamber has to be a subset of.

Finally, I will also show that recently, the soft rules for creation of new parties, associated with the low entrance-barrier for the Lower Chamber elections, started to act as a new strong factor of party system fragmentation. I will also show that there has been several increases in the public resources that are directed to the parties (public funds and publicly paid TV and radio time for propaganda), thus creating additional incentives for such creation of new parties.

As I will show, that way a more accurate story can be told about the highest party system fragmentation in the world. I propose that it is the result of a very specific combination of institutional and political choices, not to be reduced to macro institutions or social characteristics. With more precise and fine-grained diagnostics of the problem, it becomes easier (and safer) to design eventual solutions than to resort to the always debated macro reforms of electoral rules.

Author

• Fabricio Vasselai, University of Michigan