Distortions to Agricultural Incentives in South Africa¶
Johann Kirsten; Lawrence Edwards; Nick Vink (2007). Distortions to Agricultural Incentives in South Africa. Agricultural Distortions Working Paper 38 (World Bank Development Research Group project led by Kym Anderson). World Bank, Washington DC.
| Type | working paper |
| Scope | Economics and economic history of apartheid |
| Method | mixed |
| Period studied | 1910–2005 |
| Regions | South Africa |
| Themes | agriculture land, business and state, macro fiscal finance debt, sanctions divestment, transition post apartheid, methods data sources, origins segregation |
| Citation key | kirsten2007 |
Argument¶
The paper traces how the state supported white commercial farming. It used the Land Acts, the 1937 Marketing Act and its Control Boards, subsidies, tax concessions and credit. After 1955 this support grew until about 1980, in a climate of growing international isolation. Deregulation began in the 1980s, driven by fiscal and macroeconomic pressures, within the old institutions. It was completed after 1994 through trade liberalisation and the 1996 Marketing of Agricultural Products Act. The authors quantify the changing distortions with Nominal and Relative Rates of Assistance (NRA, RRA) for about 80 percent of field crops and animal products, plus consumer tax equivalents, for 1961-2005. They find assistance peaking in 1980-84 and near zero by 2000-05. They conclude that remaining distortions are sectoral (sugar, processing) and that support for emerging black farmers needs rethinking.
Evidence and method¶
Secondary literature and policy documents (White Papers 1946 and 1984, the Kassier and Brand reports, the De Kock Commission, OECD 2006). Official statistics (Abstract of Agricultural Statistics, Marketing Board annual reports, Customs and Excise, SAGIS, Reserve Bank, Statistics SA). International price series (FAOSTAT, IMF IFS, World Bank). The authors' own spreadsheet of NRAs, with annual price tables and comparisons with OECD estimates. De Klerk (1983), Vink (1993, 1999) and others are drawn on for the narrative.
Place in the debates¶
- Causes of farm mechanization and labour shedding (Habakkuk-type thesis) (see Farm labour: mechanisation, labour shedding, wages and labour control). Links combine-harvester adoption in maize to preferential tax treatment and says it coincided with the peak of forced removals of permanent farm labour. Wage-driven or labour-supply explanations are not tested.
- Sanctions effectiveness and the cost of isolation (see Sanctions, disinvestment and the anti-apartheid movement). Treats sanctions and isolation as a secondary influence. Sanction-busting costs and targeted export campaigns (Outspan oranges) hit fruit-exporter returns, but domestic support policy and fiscal and macroeconomic pressure drove the timing of reform.
- State intervention versus market in agriculture, and the case for supporting emerging farmers (see Agricultural policy, productivity, farm size and land reform). Finds apartheid-era controls costly, fiscally burdensome and redistributive towards large producers. Argues that some 1950s-60s Afrikaner-farmer support programmes deserve re-examination for black empowerment, in WTO-compatible 'green box' form.
- Measuring agricultural protection (NRA versus OECD PSE/NPC) (see Data reliability and econometric method). Allows negative NRAs and splits traded and non-traded fruit, where the OECD imposes zero. This yields greater volatility, and the authors caution that the estimates are volatile and indicative.
Findings¶
-
From 1955 policy meant extensive support and regulation alongside growing isolation, followed by rapid deregulation in the 1990s. (p. 3)
After 1955 the story of agricultural policy towards commercial farmers involved widespread support, regulation and control in a climate of increasing isolation from the rest of the world, especially in the 1980s
-
Marketing Boards covered most agricultural output by the 1970s. (p. 7)
By the 1970s more than 20 Boards were in operation, covering some 80 percent of total agricultural production.
-
Combine-harvester adoption in maize was rapid and tax-induced, and it marked the turning point in farm employment. (p. 7)
The turning point around 1970 coincides with the introduction in the late 1960s of combine harvesters, stimulated by preferential tax treatment.
-
Mechanization in maize harvesting was fast, 1968-77. (p. 7)
the share of the maize crop that was harvested with combine harvesters grew from 16 percent in 1968 to 81 percent by 1977.
-
The red meat scheme redistributed rents to large producers and away from smaller ones. (p. 9)
the largest effect of the intervention lies in the redistribution of wealth towards larger producers (and speculators) and away from smaller producers.
-
Rising fiscal costs of apartheid and deficits shaped macro policy and farm interest costs. (p. 12)
One important consequence was that the budget deficit reached a peak of 7.3 percent of GDP in 1993 (Strydom 2002), necessitating high real interest rates.
-
Farmers received large implicit tax concessions in the early 1980s, later reduced. (p. 13)
income tax concessions to farmers amounting in 1981-84 to 70 percent of their theoretical tax bill (Lamont 1990).
-
Post-1994 land reform transferred little land. (p. 15)
After 12 years of state-sponsored land reform, less than 4 percent of the land had been transferred.
-
The NRA for primary agriculture peaked in the early 1980s and fell to near zero. (p. 19)
The five-year average NRA for primary agriculture rose to a peak of 31 percent between 1980 and 1984, but then fell to less than 10 percent in the 1990s and close to zero since then.
-
The Maize Board's export losses in the late 1980s required a state bailout. (p. 21)
The government bailed out the Maize Board with a payment of R400 million to cover the shortfall, but then indicated that this will not be repeated.
-
Sanction-busting costs reduced fruit-farmer returns in the early 1980s. (p. 28)
During some years, especially the early 1980s, the costs of sanction busting impacted negatively on farmer returns.
-
Orange exporters were targeted during apartheid, with a measurable negative effect after Soweto 1976. (p. 28)
Outspan oranges were targeted world-wide during the apartheid years, and the resulting losses can be seen in the negative spike in 1977 (after the 1976 Soweto uprisings).
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Abstract of Agricultural Statistics (various years, incl. 2006), Department of Agriculture | official statistical series | 1960-2005 | Source of production volumes, farm-gate and wholesale prices, export realization prices and sugar fob prices used to compute NRAs, and of livestock/meat composition trends. |
| Annual reports of the various Agricultural Marketing Boards (pre-1995) and producer organizations | commission or official report | pre-1995 (boards); post-1994 (producer organizations) | Provide farm-gate prices, export realization figures (fruit) and adjusted production data for the NRA computation. |
| BFAP (2006) commodity market database (Bureau for Food and Agricultural Policy, University of Pretoria) | secondary literature compilation | 1961-2005 | Unified quantity database and industry processing-cost margins for maize meal, sunflower oil, wheat flour and sugar that feed the NRA estimates. |
| South African Grain Information Service (SAGIS) data | official statistical series | various, esp. 2000-05 | Used to adjust production data, derive wheat export parity quality adjustment, and supply freight and discharging costs. |
| Customs and Excise trade data | international trade or finance data | 1988-2004 for sugar; various | Export and import volumes and the processed-to-raw sugar export ratio used to build the sugar reference price. |
| IMF International Financial Statistics world commodity prices | international trade or finance data | 1960-2005 | World prices for grains, mutton, sugar, US HRW wheat, yellow maize (US No.2 Gulf), Argentine beef and UK sunflower oil used as border prices. |
| OECD (2006) Review of Agricultural Policies: South Africa, PSE/NPC database | international trade or finance data | 1994-2003 | Used to complete the poultry border price series and as comparator for the authors' NRA estimates (Appendix Table 8). |
| Statistics South Africa retail price data (Statistical release P3051.1) | official statistical series | various, esp. 1961-2005 | Retail prices of cake flour, maize meal and sunflower oil converted to wholesale prices using margins for processed-product NRAs. |
| Food Pricing Monitoring Committee Final Report (NAMC 2003) | commission or official report | 2000-03 margins | Supplies wheat, maize and sugar chain margins used to derive wholesale prices and processing margins. |
| South African Reserve Bank official exchange rates and Quarterly Bulletin of Statistics | official statistical series | 1960-2005 | Official exchange rates convert border prices; parallel rates deemed inapplicable. |
| Government policy documents and budget reports of the Department of Agriculture | commission or official report | 1950s-1990s | Source of production, consumption, input and trade taxes and subsidies, and the direct subsidies tabulated in Appendix Table 2. |
| Edwards (2005) tariff and collection-rate data | secondary literature compilation | 1961-2005 | Tariff data and manufacturing collection rates used for non-agricultural NRAs and agricultural tariffs. |
| Input-Output tables | official statistical series | various | Production values at distorted prices used to weight NRAs for processed products and non-agricultural sectors. |
| World Bank (2007) World Development Indicators and Quantech (2005) South African Standardised Industry Database | national accounts macro | 1965-2004 | Basic economic indicators (GDP shares, employment, trade) in Appendix Table 1. |
| Union of South Africa White Paper on Agricultural Policy (1946); RSA White Paper (1984) and (1997) land policy | commission or official report | 1946, 1984, 1997 | Quoted to document the policy rationale for farm protection and the 1980s policy pillars. |
| Marketing Act 1937/1968 and Marketing of Agricultural Products Act No. 47 of 1996; Land Acts; Water Act No. 36 of 1996; Agricultural Labour Act 147 of 1993 | legislation gazette | 1912-1998 | Describes the institutional framework of control boards, land segregation and later deregulation that defines the policy periods. |
| De Kock Commission (1985), Brand Report (1988), Kassier Committee Report (1992) | commission or official report | 1984-1992 | Cited as milestones for monetary reform and for deregulation of marketing and price setting in the 1980s and 1990s. |
| Authors' compilation of financial aid to farmers from official documents (Appendix Table 3); Agricultural Credit Act programmes | commission or official report | 1981-1992 | Summarizes Rand amounts of financial assistance programs to farmers in the 1980s. |
| De Klerk (1983) thesis on maize harvesting mechanisation, Western Transvaal | secondary literature compilation | 1968-1981 | Provides the share of the maize crop harvested by combine harvesters (16% in 1968 to 81% in 1977) to document labour displacement. |
| Vink (1993, 1999) and Vink and Schirmer (2002) narrative summaries; Vink and Kirsten (1999) employment data | secondary literature compilation | 1955-2000 | Underpins the policy-history narrative and farm employment trends; sections explicitly borrow from these works. |
| Lamont (1990) study of agricultural tax concessions | secondary literature compilation | 1981-84 | Source for income tax concessions to farmers amounting to 70% of their theoretical tax bill. |
| Authors' spreadsheet of NRA, RRA and CTE estimates (Anderson et al. 2007/2008 methodology) | author survey or fieldwork | 1961-2005 | Basis for all NRA, RRA and CTE figures, tables and figures presented. |
New data¶
Consistent database of annual nominal rates of assistance, relative rates of assistance and consumer tax equivalents for about 11 primary agricultural products (maize yellow and white, wheat, sunflower, beef, mutton, poultry, apples, oranges, grapes, sugar) and four lightly processed foods in South Africa, 1961-2005, plus non-agricultural tradable NRAs, built from domestic prices, border prices and policy data.
Inference¶
Quantification: Statistical analysis.
Comparison of domestic prices with border reference prices (adjusted for margins, freight and quality) to compute NRAs and RRAs over time and across commodities, supported by narrative policy history, commodity-scheme case studies (maize, red meat, deciduous fruit) and comparison with OECD estimates; no regression or causal identification.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Agricultural support peaked in the early 1980s (five-year average NRA of 31% in 1980-84) and then fell to under 10% in the 1990s and near zero after 2000, consistent with Control Board abolition and trade liberalization. (p. 19) | Authors' NRA spreadsheet from domestic and border price series (Table 1, Figure 1). | Systematic data |
| Agricultural tradables were highly assisted relative to non-agriculture in the 1960s, late 1970s and 1980s, with the incentive tilting slightly against agriculture by 2000-04. (p. 22) | RRA computed from agricultural NRAs and non-agricultural NRAs based on tariff collection rates (Table 2, Figure 2). | Systematic data |
| Marketing schemes (maize, red meat, deciduous fruit) redistributed wealth to large producers and cooperatives and probably inhibited export growth, e.g. the maize stabilization fund was perpetually in arrears. (p. 8) | Descriptive account drawing on Vink (1999), Karaan and Myburgh (1993) and Chile comparison; no quantitative test. | Illustrative or anecdotal |
| The peak in exportable NRAs in 1985-89 reflects large maize exports at a loss, as world prices fell 33% and domestic prices rose 44% from 1985-87, leading to a R400 million state bailout. (p. 20) | Maize price series and NRA calculations (Appendix Table 4); Maize Board stabilization fund account. | Partial data |
| Mechanisation via combine harvesters, stimulated by preferential tax treatment, caused farm employment to turn after 1970 and coincided with forced removals of permanent labour. (p. 7) | De Klerk (1983) on combine harvester share; Marcus (1989); Appendix Table 1 employment series. | Partial data |
| Direct state support to farmers (interest subsidies, distress relief, price stabilization) rose sharply through the 1980s and was suspended around the 1994/95 transition, with state farm spending and research funding falling in real terms afterwards. (p. 19) | Department of Agriculture budget reports compiled in Appendix Table 2 and 3; budget and parliamentary grant figures. | Partial data |
| The authors' NRAs are more volatile than OECD estimates, largely because negative NRAs are kept rather than set to zero and because reference prices and coverage differ. (p. 23) | Side-by-side comparison with OECD (2006) NPC-based estimates in Appendix Table 8. | Systematic data |
The author's caveats¶
Authors note that distortion estimates are very volatile because of exchange rate volatility and imperfect pass-through; identifying international prices and margins was difficult; sudden sign switches were not smoothed; a consistent database could be constructed only from 1965; collection rates understate non-agricultural protection; dairy and highly processed products are excluded, so processed NRAs may be biased; white maize reference price is imperfect; wheat flour wholesale prices were estimated for lack of official data.
Gaps in the evidence¶
The policy-history and scheme-level claims (maize, red meat, fruit) rest on secondary literature and assertion rather than board records or firm-level data, so distributional effects on small versus large farmers and black farmers are not measured. Non-agricultural protection uses tariff collection rates, which understate protection; effective-protection or quota-equivalent data would test the RRA better and exist for part of the period. Labour, land-price and farm-profitability data would be needed to link subsidies to mechanisation and employment claims.
Replicability: Data described but not public.