South Africa's Party System¶
Susan Botha (1996). South Africa's Party System. Journal of Theoretical Politics 8(2). Sage Publications.
| Type | journal article |
| Scope | Wider apartheid literature |
| Method | descriptive quantitative |
| Period studied | 1910–1994 |
| Regions | South Africa, Homelands (Transkei, Bophuthatswana, Venda, Ciskei) |
| Themes | politics ideology elections, transition post apartheid, resistance social movements, origins segregation, bantustans removals influx |
| Citation key | botha1996 |
Argument¶
Botha argues that South Africa's party system before 1994 was not a single system but three interacting ones: a dominant white party system, legally prescribed subsystems for the 'quasi-enfranchised', and a banned or excluded 'quasi-party system' of extra-parliamentary movements. Legislation on the franchise and on party membership and legality largely shaped this structure, so western indicators (Sartori's counting rules, Rae's fractionalization) fit poorly. She applies indicators (participation, number of parties, competition, ideological distance, party characteristics, volatility, change and stability) to the pre-1994 system and to the 1994 election. Her conclusion is that the post-1994 system is predominant-party, with a risk of ANC hegemony, and is still being normalized, so its future direction is uncertain.
Evidence and method¶
Election statistics for white elections 1915-1989, compiled from Schoeman (1977), Faure et al. (1988), SAIRR Race Relations Surveys, Van Vuuren et al. (1987), Weide and Weide, and newspapers; the 1994 results and turnout estimates by race and province from Reynolds (1994); secondary literature on party systems (Sartori, Rae, Lijphart).
Place in the debates¶
- Applicability of western-derived party-system typologies and indicators to South Africa (see White electoral politics, delimitation and party systems). Holds that Sartori's counting rules and Rae's fractionalization index were unsuitable before 1994 because legislation barred organizations and voters, so the system needs to be analysed as a set of legally constructed subsystems.
- Dominant/predominant versus hegemonic party in post-apartheid South Africa (see White electoral politics, delimitation and party systems). Classes the post-1994 system as predominant-party, with warning signs that the ANC could become hegemonic in Sartori's sense.
- Interpretation of white voter swings to the right in the 1980s (see White electoral politics, delimitation and party systems). Argues that apparent swings to the right were often voter fall-out caused by the NP's own ideological shift to the left, rather than a rightward movement of voters.
Findings¶
-
The pre-1994 party system was three-fold, and each part was shaped by legislation. (p. 3)
a dominant party-system that organized and represented the white electorate for parliamentary purposes and that had established its hegemony over the other 'party systems'
-
Exclusion from the franchise at Union drove non-whites into extra-parliamentary organizations such as the ANC, which later became anti-system. (p. 3)
Initially, the aforementioned organizations campaigned for inclusion in the latter system, but later on aimed at replacing it with their own system and thus became anti-system.
-
White turnout declined in the later years of apartheid, and the NP's vote share peaked in 1977 before falling. (p. 4)
Overall, voter participation in the dominant party-system (i.e. among whites) was relatively high but with indications of a decline in the later years of the old system.
-
Political participation by the unenfranchised took the form of stay-aways and boycotts, with sharp regional variation in their uptake. (p. 5)
On election day (for the tricameral Parliament) in 1987, 96 percent of workers in the Port Elizabeth/Uitenhage area observed a call for
-
The NP became a predominant party from 1958, helped by the plurality electoral system. (p. 7)
-
Legislation banned multiracial parties, which limited party membership to a single race group. (p. 9)
The Prohibition of Political Interference Act (Act no. 51 of 1968), adopted by Parliament, prohibited membership of 'mixed' political parties.
-
Turnout in the 1994 election was high across all race groups. (p. 12)
Voter participation was therefore high among all race groups.
-
The ANC won a large majority in 1994, well ahead of its rivals. (p. 14)
On the national level the ANC polled 62.65 percent of the votes, three times that of its nearest rival, the NP, with 20.39 percent
-
The post-1994 system is classed as predominant-party, with a warning of possible ANC hegemony. (p. 14)
the present party system shows characteristics of a predominant-party system but there are also warning signs of the ANC being a hegemonic party in Sartori's terms
-
Party support after 1994 was less strictly racial than before, though the ANC's votes remained overwhelmingly black. (p. 15)
Though an estimated 94 percent of the ANC's votes came from blacks (i.e. 80 percent of the black vote), it was estimated that it secured 27.8 percent of the Coloured vote
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| White election turnout series 1915-1989 (compiled from Faure et al. 1988; Rand Daily Mail 2 Dec 1977; SAIRR 1987/88, 1989/90; Schoeman 1977; The Friend 2 May 1981; Van Vuuren et al. 1987; Weide and Weide) | secondary literature compilation | 1915-1989 | Establishes the percentage of white voters participating in each general election, showing high turnout with later decline (Table 1). |
| Van Vuuren et al., South African Election 1987 (stay-away figures, 1987 tricameral election day) | secondary literature compilation | 1987 | Used to illustrate 'voter participation' of the unenfranchised as stay-aways and the contrasting Coloured response. |
| Schoeman 1977, Parlementere Verkiesings in Suid-Afrika 1910-1976 (contested seats in 1933 election) | election returns | 1910-1976 (1933 cited) | Supports the point that many seats were uncontested, e.g. only 72 of 150 seats in 1933. |
| National Party vote share series 1915-1989 (SAIRR 1987/88, 1988/89, 1989/90; Schoeman 1977; Sunday Tribune 2 May 1981; Van Vuuren et al. 1987; Weide and Weide) | election returns | 1915-1989 | Used to show a degree of voter volatility in the dominant party-system (Table 3). |
| SAIRR Race Relations Survey 1989/90 (homeland legislature composition) | secondary literature compilation | 1989/90 | Establishes that only Bophuthatswana, Gazankulu and KaNgwane had a majority of elected legislators, supporting reduced competition. |
| Van Vuuren et al. 1987 (estimates of party support by language group, 1987 election) | secondary literature compilation | 1987 | Gives estimated NP and CP support among English-speaking whites and NP support among Afrikaners. |
| Reynolds (ed.) 1994, Election '94 South Africa (electorate size, turnout by race and province, seats, party shares, vote by race) | election returns | 1994 | Source of all 1994 data: eligible voters, turnout by race (Table 4) and province (Table 5), seats won (Table 6), national vote shares, and racial composition of ANC/NP support. |
| Prohibition of Political Interference Act (Act no. 51 of 1968) | legislation gazette | 1968 (repealed 1985) | Cited as legislation prohibiting mixed-race parties, shaping party membership structure. |
| Black Authorities Act of 1951 and homeland/council legislation (Natives Representative Council 1936, Coloured Persons' Representative Council 1968, South African Indian Council 1964, Community Councils 1977, 1983 constitution) | legislation gazette | 1936-1983 | Describes the legislative creation of party subsystems for the 'quasi-enfranchised'. |
| Kleynhans 1987, South African General Election Manifestos 1910-1981 | secondary literature compilation | 1910-1981 | Referenced for the claim about NP ideology reflecting socialist yet anti-communist tendencies. |
Inference¶
Quantification: Descriptive tables.
Descriptive, conceptual analysis of the South African party system using Sartori's and Rae's party-system criteria; three eras (pre-1994, post-1994, future). Claims rest on descriptive tables of election turnout and vote shares compiled from secondary sources, with interpretive narrative and speculation; no statistical testing.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| The party system was shaped by legislation as an external factor, producing a dominant white party-system, quasi-enfranchised subsystems and a parallel quasi-party system. (p. 210) | Asserted through narrative of franchise exclusion, homeland and council legislation and the Prohibition of Political Interference Act; no data. | Asserted |
| White voter participation was relatively high but declined in the later years of the old system. (p. 211) | Table 1, compiled turnout percentages from 1915 to 1989 drawn from Schoeman, SAIRR, Van Vuuren et al. and others. | Systematic data |
| Voter participation in the Coloured/Indian and homeland subsystems was low and statistically unreliable, reflecting illegitimacy and boycotts, with the unenfranchised participating through strikes and stay-aways. (p. 212) | Table 2 turnout of registered voters; SAIRR and Van Vuuren et al. stay-away figures for 1987. | Partial data |
| The NP was a predominant party from 1958 to 1994, helped by plurality voting overrepresentation; earlier periods saw multiparty competition and coalition. (p. 214) | Narrative of election outcomes and Table 3 vote shares; seat-vote disproportionality asserted, not tabulated. | Partial data |
| The 1994 election had high turnout across all race groups and provinces, with low spoilt ballots. (p. 219) | Tables 4 and 5 compiled from Reynolds 1994 (IEC estimates). | Systematic data |
| After 1994 the system is predominant-party, with the ANC winning 62.65 percent against the NP's 20.39 percent, with warning signs of hegemony. (p. 221) | Reynolds 1994 election results and Table 6 seat counts; hegemony claim interpretive via Sartori. | Systematic data |
| Parties are now less racially structured, apart from the left and right: the NP won 66.7 percent of the Coloured vote and the ANC 27.8 percent of the Coloured vote. (p. 222) | Reynolds 1994 estimates of vote by race; author's calculations from chapter 11. | Partial data |
| Future party-system development (multiparty polarized pluralism or one-party dominance) depends on continued participation, volatility and grassroots organization. (p. 223) | Speculation and comparison with European and African trends; no data. | Asserted |
The author's caveats¶
The author notes that statistics on voter participation in the subsystems and homelands are controversial owing to low voter registration and residence outside the homelands. Opinion polls for banned quasi-parties were unreliable. Allegations of serious shortcomings in the 1994 election are set aside by accepting the outcome as a true reflection of how people voted.
Gaps in the evidence¶
The argument lacks systematic data on party membership, organization, and opinion polling or survey data on voter attitudes and ideology; ideological positions are asserted rather than measured. No formal indices (effective number of parties, volatility, disproportionality) are calculated, though election returns existed to do so. Only aggregate turnout is reported; no cross-tabulation of race, class or region would test the claims about party-race alignment. Post-1994 survey and exit-poll data now exist to test them, as do archival records of banned organizations.
Replicability: Data described but not public.