Decentralisation in South Africa

Literature › Works

Martin Wittenberg (2003). Decentralisation in South Africa. School of Economic and Business Sciences and ERSA, University of the Witwatersrand (first draft, May 2003).

Type	working paper
Scope	Wider apartheid literature
Method	synthesis review
Period studied	1910–2003
Regions	South Africa, Pietermaritzburg/Natal-KwaZulu, Bantustans (TBVC)
Themes	bantustans removals influx, business and state, politics ideology elections, transition post apartheid, macro fiscal finance debt, resistance social movements, demography urbanisation, comparative theory, origins segregation
Citation key	wittenberg2003

Argument

Wittenberg argues that apartheid was itself a form of decentralisation, designed to maintain central control and reduce accountability to the African majority. Its fragmented racial administrative tiers (provinces, white municipalities, Administration Boards, homelands, TBVC states) produced fiscal crises, corruption and a spatial and bureaucratic mess. Later coordinating layers (planning regions, Regional Services Councils, the NSMS) were added to manage this. A Pietermaritzburg case study shows how micro-differences in administrative regulation shaped local interests and fed civil war. Post-1994 decentralisation, shaped by the balance of forces at the negotiations, inherited these regions and tiers. The paper gives a preliminary, impressionistic assessment of the new system, noting a centralising drift under ANC dominance.

Evidence and method

Secondary literature (Hindson, Christopher's Atlas of Apartheid, McCarthy, Niehaus, Kentridge), SAIRR Race Relations Surveys, Weekly Mail & Guardian and Natal Witness press reports, statutes, National Treasury budget and fiscal review documents (Division of Revenue table, equitable share formulae), Bahl's working papers, and the author's own earlier Local Government Policy Project report. The author discloses he was Secretary of the UDF.

Place in the debates

• Theory of decentralisation (accountability, efficiency, development vs capture and macro-stability) (see Post-apartheid economic policy, governance and corruption). Rejects a purely abstract case for decentralisation; costs and benefits depend on institutional design and the social context, following Bardhan.
• Purpose of apartheid decentralisation and homeland 'independence' (see Homelands, separate development, decentralisation and removals). Decentralisation aimed to preserve central control by fragmenting opposition and coercing Africans; partition failed because land and economic resources were niggardly. Independence was 'not a complete sham' because it complicated direct intervention.
• Causes of the 1980s-90s Natal Midlands violence (see Resistance and liberation movements). Differences in administrative regimes (freehold, tribal authority, township) shaped interests and conflict; Gwala's 'bureaucratic entry points' is incomplete, and state security-force involvement is stressed.
• Is post-apartheid South Africa centralising or decentralising? (see Post-apartheid economic policy, governance and corruption). Fiscally very decentralised (per Bahl) but with a centralising drift via ANC party discipline and Treasury dominance; evidence of efficiency gains is as yet inconclusive.

Findings

• Apartheid is framed as decentralisation intended to reduce government accountability to the majority. (p. 4)

  Indeed apartheid could be seen as form of decentralisation intended to reduce government’s accountability to the majority.

• Black housing in the cities was financed through the municipal beer monopoly, and Soweto through the Native Services Levy, which also discouraged labour-intensive urban production. (p. 13)

  By making the employment of African labour more expensive, it discouraged labour intensive forms of production in the urban areas.

• In 1971 the state removed control of the townships from municipalities and placed it with Administration Boards, which controlled housing, employment and urban access. (p. 13)

  So the Administation Boards controlled accommodation, employment and access to the cities.

• The lack of an economic base for the bantustans was the fatal weakness of partition, because National Party supporters would not give up economic resources. (p. 20)

  The problem of the lack of an appropriate economic base for the bantustans was the Achilles Heel of the entire partition project.

• The primary function of decentralisation was coercive control, which made corruption endemic and bred unaccountable bureaucracies and administrative chaos. (p. 25)

  Since an equitable partition was never going to be entertained, the primary function of decentralisation was always going to be a more coercive one, viz. to maintain control over the African population.

• Post-1994 South Africa is fiscally more decentralised than expected, with subnational expenditure share never below 40% of the total. (p. 38)

  Even if the national debt is added in to the budget and conditional grants are viewed as really central government expenses, the share of subnational government never falls below 40% of the total.

• Formula-driven transfers broke apartheid-era patronage and targeted poor, formerly unserved rural areas. (p. 46)

  The move to formula driven allocations was an enormous step forward in breaking some of those patronage networks.

• Municipalities were consolidated from 843 to 284 in the 2000 demarcation. (p. 34)

  This led to a rationalisation of the 843 municipalities down to 284 municipalities: six type A, 231 type B and 47 type C municipalities.

Evidence ledger

Sources used

Source	Kind	Period	How used
National Treasury, Division of Revenue table (Intergovernmental Fiscal Review 2003, Table 1.1)	official statistical series	1999/00 to 2005/06	Table 1 shows the vertical split of nationally raised revenue between national departments, provinces and local government, to assess whether provinces are losing share.
Budget Review 2003 (provincial equitable share formula description)	commission or official report	2003	Supplies the components and weights of the provincial equitable share formula and the statement that allocations are indicative rather than earmarked.
National Treasury 1998, The introduction of an equitable share of nationally raised revenue for local government	commission or official report	1998, with 2003/04 allocations	Describes the local government equitable share formula (S-grant, I-grant, free basic services, nodal, R293, guarantee) and the percentage weights of each component in 2003.
Intergovernmental Fiscal Review 1999 and Intergovernmental Fiscal Relations Act of 1997	legislation gazette	1997-1999	Source for the legal and consultative framework of fiscal decentralisation, the PFMA 1999 and personnel expenditure trends.
Intergovernmental Fiscal Review 2003 (personnel expenditure, municipal debt, audit outcomes)	official statistical series	2000/01-2002/03	Provides figures on municipal debtor balances, Auditor General audit status, property rates, and personnel expenditure shares used to assess the post-1994 system.
Bahl (2001) calculations of expected subnational expenditure share	secondary literature compilation	c. 2001	Used to claim South Africa is fiscally more decentralised than expected (15.7% expected subnational share).
Bahl and Solomon (2000), Regional Services Council Levy evaluation	secondary literature compilation	c. 2000	Supplies the data on distribution of RSC turnover levy revenues and compliance used to criticise the levy.
Alderman et al. (2000) poverty mapping; 1996 Census; October Household Surveys; 1995 Income and Expenditure Survey	household or labour survey microdata	1995-1996	Describes the imputation method combining the 1995 IES and 1996 census to produce poverty counts for the equitable share formula, and its provincial biases.
Christopher (1994), The Atlas of Apartheid	secondary literature compilation	1913-1990s	Source for trust land, resettlement, border industry, and Regional Services Council counts (42 RSCs) and homeland consolidation.
Hindson (1987), Pass controls and the urban African proletariat	secondary literature compilation	1950s	Source for the description of pass laws, labour bureaux and Section Ten rights.
Statutes: Native Laws Amendment Act 1952, Urban Areas Act, Bantu Authorities Act 1951, Promotion of Bantu Self-Government Act 1959, Status of Transkei Act 1976, Regional Services Council Act 1985, etc.	legislation gazette	1913-1998	Legislative framework forms the basis for describing apartheid's decentralised structures and the post-1994 local government system.
SAIRR Race Relations Survey 1976, 1988/89 and South Africa Survey 2001/02	newspapers periodicals	1976-2001	Cited for Transkei repression in 1976, the SADF intervention in Bophuthatswana in 1988, and a municipal official's 2001 complaint on districts.
RSA (1990) Handleiding Nasionale Koordineringsmeganisme (secret government memo)	state archives	1990	Quoted to show security element persisted in the National Coordination Mechanism.
Newspapers: Weekly Mail and Guardian and Natal Witness (1997-2002)	newspapers periodicals	1997-2002	Used to document corruption, misadministration, political appointments and provincial divergence on AIDS policy after 1994.
Pietermaritzburg conflict literature: Kentridge (1990), Aitchison (1988) mimeo, Cross et al. (1988), Gwala (1988), Wills (1988) census figures, Lawrance (1988)	secondary literature compilation	1985-1990	Supplies death counts, displacement figures, population by race (1985 census) and the narrative of the Midlands conflict.
Wittenberg and McIntosh (1992) Local Government Policy Project report; author's participation as UDF Secretary	author survey or fieldwork	1992	Earlier unpublished material underpinning the sections on pre-1994 decentralisation, RDACs, NCM, and the Pietermaritzburg case study.

Inference

Quantification: Descriptive tables.

Historical-institutional narrative based largely on secondary literature, plus a single-locality case study (Pietermaritzburg) and descriptive analysis of official fiscal tables and formulae. The post-1994 assessment is explicitly impressionistic, with no econometric work.

Key claims

Claim	Basis	Strength
Apartheid decentralisation (homelands, Black Local Authorities, Administration Boards) aimed at maintaining central control and fragmenting the opposition, and produced fragmented, fiscally weak structures and corruption. (p. 24)	Synthesis of secondary literature (Hindson, Christopher, McIntosh, Niehaus) and legislation; no new data.	Illustrative or anecdotal
Black Local Authorities were financially unviable, lacking a commercial and industrial base, with rising unemployment and removal of the beer monopoly, and their 1983 elections sparked urban unrest. (p. 14)	Asserted narrative without cited data on BLA finances.	Asserted
The state-created administrative boundaries in Pietermaritzburg shaped the Midlands civil war, as different administrative regimes produced different interests and conflict patterns. (p. 31)	Case study drawing on Kentridge, Aitchison, Gwala, Cross et al., Wills and local histories.	Illustrative or anecdotal
Post-1994 South Africa is fiscally highly decentralised, and the provincial share of nationally raised revenue fell then is budgeted to recover, reflecting provincial underspending rather than national dominance. (p. 38)	National Treasury Table 1 (Division of Revenue 1999/00-2005/06) and Bahl (2001) benchmark.	Partial data
The formula-driven equitable share is redistributive and poverty-targeted, benefiting poor rural areas that previously received no transfers, though poverty data are imperfect. (p. 46)	National Treasury descriptions of the provincial and local formulae and the Alderman et al. imputation method.	Partial data
The district council layer, derived from the apartheid Regional Services Councils, is the most problematic part of the system and tiers overlap. (p. 35)	A SAIRR-quoted municipal official, the author's institutional analysis and Bahl and Solomon on RSC levies.	Illustrative or anecdotal
Corruption and misadministration are prevalent at subnational levels, and political centralisation through ANC party structures limits the effective autonomy of provinces and municipalities. (p. 44)	Press reports (Weekly Mail & Guardian, Natal Witness) and the author's observations of premier replacements.	Illustrative or anecdotal
There is as yet little evidence that decentralisation is producing more efficient government; local government financial control is weak. (p. 44)	National Treasury (2003b) audit statistics (5% failed 2000/01 accounts, 77% failed 2001/02) and personnel expenditure ratios.	Partial data

The author's caveats

The author notes that econometric work is yet to be done and the assessment is preliminary and impressionistic. He also discloses non-neutrality as Secretary of the UDF and consultant on the transfer system. He notes that the Income and Expenditure Survey likely overstated poverty in the Free State and understated it in Limpopo.

Gaps in the evidence

The argument on apartheid-era decentralisation relies almost entirely on secondary works and a few mimeos, without systematic archival data on Administration Board, BLA or homeland finances or transfers, which exist in provincial and departmental archives and Auditor-General reports. The post-1994 assessment lacks municipal-level panel data on transfers, service delivery and spending that could test claims on capacity, corruption and poverty targeting; such data are now available from National Treasury and Statistics South Africa. The Pietermaritzburg causal claim that boundaries generated violence would need systematic data on deaths, locations and administrative status, or comparison with other regions.

Replicability: Data poorly documented.