Methodology
South Africa allocates seats in the National Assembly, provincial legislatures and half of the seats at the municipal level using the Droop method of proportional representation. Under this system a quota of the number of votes a party must obtain to win a seat is calculated by dividing the total number of votes cast by the number of seats to be allocated plus 1. Each party’s vote is then divided by the resulting quota plus 1 to determine the number of seats the party receives. If the total number of seats awarded by this calculation is less than the total number of seats to be filled, the remaining votes for each party, i.e., the number of votes each party receives above the multiple of the quota for a single seat, are then rank ordered to determine which parties, if any, are awarded additional seats. South Africans also cast two ballots—one for the national party list of their choice, and one for the provincial party list of their choice. To allocate the 46 seats that Gauteng Province elects to the National Assembly, the total vote in the province for the provincial lists (3,707,595) was divided by 46+1, i.e. 47 to establish a quota for each seat of 78,885. Each party’s vote was then divided by the quota + 1, i.e. 78,886. Thus, the ANC which obtained 2,488,780 votes in Gauteng for its provincial list, was awarded, 32 seats. The ANC received an additional 31 seats from its vote in Gauteng for its national list for a total of 63 seats from the province.
Because Gauteng currently elects an estimated 92 members of the National Assembly, all simulations of alternative systems are based on a total of 92 seats. To simulate the results of alternative electoral systems that are partially based on PR, we used the Droop formula. Thus, for the parallel system of 46 seats elected from single member districts and 46 elected by PR, we established the quota for the 46 PR seats by dividing the total provincial vote by 46+1, i.e. 47, as described above. For the multi-member systems of 12, 18, 23 districts, we first established the number of seats for each district and then allocated these seats to the various parties, one district at a time, by computing the quota for the district and then by dividing the each party’s vote by the quota. In the example of 12 multi-member districts, 7 or 8 seats were assigned to each district computed by the SDSS depending on the number of votes cast in each district. In the example of 18 multi-member districts, 5 or 6 seats were assigned to each district computed by the SDSS. In the example of 23 multi-member districts, 3 to 5 seats were assigned. For the last example of the “multi-mixed member proportional” system, 69 seats were allocated in groups of 2 or 3 seats to each of the same 23 districts. The remaining 23 seats were allocated to achieve the same overall result as a system based solely on PR.
The data used for these simulations consist of population and electoral data from the 1999 elections for each of 1,837 voting districts in Gauteng Province and the shape files (i.e. digitized maps) for each voting district to facilitate the application of the SDSS. The data was provided by the Independent Electoral Commission (IEC) of South Africa. Responsibility for the choice and construction of the alternative systems simulated on this website is solely that of the authors.
Centre for Policy Studies (CPS), Johannesburg, www.cps.org.za
Electoral Institute for Southern Africa (EISA), Johannesburg, www.eisa.org.za
Institute for Democracy in South Africa (IDASA), Pretoria and Cape Town, www.idasa.org.za
Independent Electoral Commission (IEC), Pretoria, www.elections.org.za
Copyright © 2001 by Joel D. Barkan, Paul Densham and Gerard Rushton
