Nationwide Vote Research Articles

Empirical threshold of representation

An operational method using data from previous elections is proposed for determining the vote share a small party needs to have a fifty-fifty chance of winning its first seat. The resulting median value for 23 electoral systems is 1.0 per cent of the nation-wide vote, with a range from 0.1 to 8 per cent. This empirical threshold of representation is affected by assembly size, legal representation threshold (if any exists), and geographical concentration of small party votes. In turn, this threshold affects the number of seat-winning parties and the effective number of parties in the system. Empirical thresholds can also be calculated on the district level. They can then be compared with theoretical thresholds of representation, and unanticipated discrepancies occur, because apparently minor aspects of electoral rules can alter the outcome.

Proportional representation in variable-size legislatures

This paper examines a method to achieve proportional representation (PR) for parties in a legislature that allows the size of the legislature to vary as a function of voting results. The legislature's base consists of one elected candidate from each voting district. Seats are then added to underrepresented parties to approximate their nationwide vote proportions. Additions are governed by rules designed to honor the single-member district concept, to keep the increase manageable, and to satisfy other desirable criteria.