Skip to content

Understanding the Democratic Transition in South Africa

Literature › Works

Robert P. Inman; Daniel L. Rubinfeld (2013). Understanding the Democratic Transition in South Africa. American Law and Economics Review. Oxford University Press.
Type journal article
Scope Economics and economic history of apartheid
Method mixed
Period studied 1977–2010
Regions South Africa, Western Cape, Northern Cape, homelands
Themes transition post apartheid, macro fiscal finance debt, politics ideology elections, inequality poverty unemployment, unions strikes, sanctions divestment, comparative theory, business and state
Citation key inman2013

Argument

The authors argue that South Africa's peaceful transition succeeded because federal governance gave the white economic elite a credible check on maximal redistributive taxation. The majority central government sets taxes, while an elite-run province (the Western Cape) controls low-cost redistributive services, creating a 'hostage game'. They calibrate a formal model (developed in Inman and Rubinfeld 2012) to estimate when federalism is feasible (Border and Assignment Constraints) and sustainable (patient players). They then compute the present value of the transition to the majority and to the elite against continued apartheid, using growth regressions and fiscal data. They conclude both groups gain, but the equilibrium is fragile if the ANC's demand for redistributive services rises.

Evidence and method

Provincial fiscal transfer data (Financial and Fiscal Commission, Division of Revenue Bills), Development Bank of South Africa homeland budgets, Penn World Tables 7.0 growth rates 1950-2008, election results, calibration from Gruber-Saez, Reinikka-Svensson and Karlin-Zinman, Waldmeir's account of the negotiations, and government development indicators.

Place in the debates

Findings

  • Redistributive services to lower-income households rose sharply under democracy compared with late apartheid. (p. 15)

    the level of redistributive services provided to lower income households has been significantly higher under democracy than under apartheid, initially a 400% increase and growing.

  • Late-apartheid redistributive spending was very low per capita, supporting a low level of service inputs per majority adult. (p. 17)

    for an average year during post-Soweto apartheid regime (1977–93), spending for comparable redistributive services is estimated to be about 525 Rand/capita, supporting a level of redistributive inputs per majority adult of only q = 0.17.

  • Constitutional design gave the central government the taxes and the provinces the redistributive services. (p. 5)

    the FFC gave control over all important taxes, including income and profits taxation and the VAT, to the central government (Chapter 13, Section 214), but then assigned control for the provision of redistributive services

  • The Western Cape's border and service assignments fall within the feasible set, but current spending is close to the tipping point. (p. 24)

    Our most recent estimate for FY 2009/10 places q = 0.85 and very close to the maximum value consistent with feasible democratic federalism.

  • Early post-apartheid budgets paid less than the majority's minimum acceptable transfer, a deliberately 'anti-majoritarian' moderation. (p. 26)

    Actual transfers paid were only 75% of the majority’s “required” minimal transfers.

  • Average growth was far higher under democracy than under apartheid in the simple regression. (p. 28)

    The average rate of income growth under apartheid (1950–93) is 1.24% per annum, while that under democracy (DEMOCRACY = 1) is 2.65% per annum.

  • Majority unions had a significant negative growth effect in late apartheid, sanctions a smaller imprecisely estimated one. (p. 29)

    Majority unions and their threat of industrial action had a larger and statistically significantly negative effect, reducing the annual rate of growth by about 1.88%.

  • Democratic federalism cut elite redistributive taxation only slightly below the maximal rate, yet elite taxes are far above apartheid levels. (p. 32)

    democratic federalism lowers elite redistributive taxation by 2400 Rand per year from the maximal rate, or by about 7.5%. Compared to apartheid, redistributive elite taxation is more than 10-fold larger

  • The majority gains immediately and the elite gains after a few years, with federalism slightly better than unitary democracy. (p. 32)

    The aggregate present value of the transition’s economic benefits for a typical majority resident is 154,329 Rand under unitary governance and 161,393 Rand under democratic federalism

  • Elite returns to the transition are large; only the most impatient elite would have resisted on economic grounds. (p. 33)

    The results show a significant internal rate of return from the democratic transition, close to 30% (Table 4, Col. 3). As an economic calculation alone, only the most impatient elite residents should have resisted democracy.

  • Security spending was not reduced after apartheid, so the elite's 'peace dividend' is limited. (p. 33)

    We estimate that from 1977 to 1993, the apartheid regime spent 33.7 billion (2000) Rand annually on military, police, and prison services.

Evidence ledger

Sources used

Source Kind Period How used
RSA, President's Office, Development Indicators 2008 commission or official report 1993-2007 Establishes that incomes, poverty, housing, schooling and other welfare indicators improved after apartheid, while HIV and crime worsened.
Financial and Fiscal Commission, The Allocation of Financial Resources Between the National and Provincial Governments: FY 1997/98 (Tables 2, 3, 6b) commission or official report FY 1995/96-1997/98 Supplies per-capita intergovernmental transfers and service grants by province for the first three post-apartheid budgets in Table 1.
Minister of Finance, Division of Revenue Bill, Various Years, Part 4: Provincial Allocations official statistical series FY 1998/99-2009/10 Supplies provincial grant and service-grant data for FY 1998/99 to 2009/10 used to construct Table 1 and Table 2.
Financial and Fiscal Commission, Recommendations for Fiscal Year 1996/97 (September 1995); FFC, The Allocation of Financial Resources, FY 1996/97 commission or official report 1995-96 Documents the supplemental Western Cape medical grant and the FFC's recommended service standards (teacher ratios, clinic visits, transfer levels) behind q = 0.53.
Development Bank of South Africa, Annual Report, Various Years company records or reports 1977-93 Source for homeland budget/transfer payments (about 26.3bn Rand a year, 1500 Rand/capita incl. leaders' payments) used to estimate apartheid-era redistributive spending and elite taxation.
South African Department of Information, Perskor, South Africa, Various Years official statistical series 1977-93 Source for apartheid-era annual military, police and prison spending (33.7bn Rand) in the 'peace dividend' calculation.
South African Budget, Department of Finance, Various Years official statistical series post-1994 Source for post-apartheid police and prison expenditures (39.1bn Rand average) in the peace dividend comparison.
Department of Finance, 1998 Budget Review (as quoted in FFC: A Ten Year Review) commission or official report 1998 Shows that Mandela-era budgets followed FFC recommendations.
Penn World Tables 7.0 (GRGDPCH, South Africa 1950-2008) international trade or finance data 1950-2008 Dependent variable (annual real per capita GDP growth) in regressions estimating growth under apartheid and democracy, used to project incomes.
Karlin and Zinman (2008) estimates of time preference for credit-constrained South African households secondary literature compilation c. 2008 Provides low majority discount factor (delta=0.33) used for gmin sensitivity and welfare robustness.
Gruber and Saez (2002) elasticity of taxable income estimates secondary literature compilation n/a Used to calibrate the tax-avoidance parameter beta and thus the revenue-maximizing tax rate (32,000 Rand).
Reinikka and Svensson (2004) evidence on local capture of transfers in Uganda secondary literature compilation n/a Used to set bounds on the capture parameter phi in the South African application.
Election results, www.elections.org.za election returns 1994-2009 Shows ANC never exceeded 45% in the Western Cape and opposition coalitions held majority, establishing the Western Cape as the elite-controlled province.
Constitution of the Republic of South Africa, 1996 (Statute No. 108 of 1996), incl. Interim Constitution provisions legislation gazette 1993-1996 Establishes the assignment of taxes to the centre, services to the provinces, revenue sharing, and provincial boundaries.
Press reports (New York Times 2009 and 2012, Financial Times 2007 and 2011, The Economist 2011, New Yorker 2010) newspapers periodicals 2007-2012 Used to document service-delivery protests, ANC internal pressure for redistribution, and nationalization debates.
Waldmeir (1997) and Muthien and Khosa (1998) accounts of transition negotiations and provincial demarcation secondary literature compilation 1990-1996 Provide the narrative of negotiations and province boundaries supporting the institutional background.

New data

The authors assembled a 1995/96–2009/10 per-capita provincial transfer and service-input dataset (Table 1) from FFC and Treasury documents, plus calibrated model parameters and projected 70-year welfare streams (Tables 2 and 4), and ran a small time-series growth regression on 1950–2008 PWT data (Table 3). Much of the calibration is in an unpublished Data Appendix.

Inference

Quantification: Econometric (causal design).

Formal game-theoretic model (repeated hostage game) calibrated to South African fiscal data; feasibility and sustainability tested by comparing observed budgets to model-derived bounds (qmin, qmax, gmin, gmax). Welfare gains from transition are computed by simulation of discounted 70-year income streams using growth rates from a simple time-series regression with regime dummies.

Key claims

Claim Basis Strength
Redistributive service levels per capita under democracy are far higher than under apartheid (about 2189 Rand/capita in 1995/96 versus about 525 Rand/capita in 1977-93). (p. 17) FFC and Division of Revenue tables (Table 1) compared with an apartheid-era estimate based on Development Bank homeland budgets. Partial data
The Western Cape borders and service assignments meet the Border and Assignment Constraints; q rose to 0.85 in FY 2009/10, near the upper bound of 0.86, so federalism is close to a tipping point. (p. 23) Calibrated model (mu = 0.184, mu_max = 0.192, qmin = 0.58, qmax = 0.86) against Table 1 service inputs, with elite and majority population figures and a Data Appendix. Partial data
Democratic federalism has been sustainable under all presidencies since 1996: actual transfers lie within or near the model's bargaining range [gmin, gmax], although early Mandela and Mbeki transfers were only about 75% of gmin. (p. 25) Model-computed gmax and gmin (Table 2) compared with actual per-capita transfers from Table 1. Partial data
Democracy raised annual growth from about 1.24% under apartheid to 2.65%, while sanctions and COSATU depressed late-apartheid growth to about -0.8%. (p. 29) Time-series regression on PWT GRGDPCH 1950-2008 with regime dummies (Table 3), with only 59 observations and a low adjusted R2. Partial data
The majority gains immediately from the transition (PV 154,329 Rand unitary, 161,393 Rand federal), and the elite gains in net present value (about 395,909-412,922 Rand) with a roughly 30% internal rate of return, falling about 20% after counting police costs. (p. 33) Simulation of 70-year income streams using regression growth rates, Table 1 transfers, homeland budgets from the Development Bank, and calibrated tax rates. Partial data
Democratic federalism lowers elite redistributive taxation by about 7.5% relative to the maximal tax rate in a unitary democracy. (p. 32) Calibrated revenue-hill model (N0 = 9.6m, beta = 0.00015) and Table 1 transfers, assuming phi = 0.15 capture. Asserted
Mandela-era budgets followed the bipartisan FFC recommendations and equalized effective public employees per majority resident across provinces; Mbeki's later budgets raised redistribution under ANC rank-and-file pressure. (p. 17) Table 1 time path of grants and q by province and presidential regime, plus FFC and Finance Department documents. Systematic data

The author's caveats

The authors note a break in the data series from FY 2002/03 due to Treasury accounting changes, that FY 1997/98 uses projected grants, that homeland payments likely overstate redistribution because of leaders' personal use, that reliable apartheid-era crime rates are unavailable, and that rights and peace dividends are omitted. Many calibrations are deferred to a Data Appendix available on request.

Gaps in the evidence

The elite capture parameter phi and the protest penalty rho are calibrated from Uganda evidence rather than measured in South Africa, so the central mechanism is untested empirically. Provincial expenditure and audit data, public-employee training and quality records, and household income surveys could test actual capture and service quality in the Western Cape, and they exist. The counterfactual apartheid growth path rests on a short dummy-variable regression of 59 annual observations and assumes identical growth for unitary and federal democracy. Negotiation archives or interviews with FFC members could test whether the hostage logic motivated actors.

Replicability: Data described but not public.