WikiSummary, the Social Science Summary Database

Desposato: The impact of electoral rules on legislative parties

From WikiSummary, the Free Social Science Summary Database


Desposato. 2004. The impact of electoral rules on legislative parties: Lessons from the Brazilian Senate and Chamber. Unpublished.

Research Question

Do electoral institutions affect (legislative) party organization?

Place in the Literature

Many argue that electoral institutions determine whether politicians have personal (pork) or party (public goods) incentives. In the Brazilian case, they cite open-list PR as the cause of Brazil's weak party system. But the fact that electoral institutions create incentives for personalistic politics in the electorate does not imply that party discipline will be weak in the congress. For a supporting argument, see Figueiredo and Limongi (2000).


Though many study the Brazilian Chamber (lower house), few study the Senate, which has similar powers but very different electoral rules. Senators are elected in plurality districts (more or less). Thus, if the literature is correct, we should observe less developed parties in the Chamber than in the Senate. Desposato looks specifically at a few indicators (Y) of this:

  • Y1: Roll call voting (party cohesion)
  • Y2: Calculate each legislator's ideal point and compare parties' dispersion in a spatial analysis
  • Y3: Use a spatial model of party-switching to measure party influence over legislators

How the Senate is elected

Each state gets three Senators, serving staggered eight-year terms. Thus, one Senator seat was open in each state in 1990 and 1998, and two seats were open in each state in 1994 and 2002. So 1990 and 1998 were plurality elections, and 1994 and 2002 were two-member district plurality elections (each voter could vote for two candidates).


  1. There are no significant differences between legislative parties in the two institutions.
  2. The sign and significance of estimates flips when he limits the analysis to only those bills that were voted on by both chambers (which test is the basis of his conclusions). Thus, we must be careful trying to test similar hypotheses by comparing legislatures from different countries--since they will be voting on dramatically different bills.