Agrarian transition to capitalism and the fate of the African peasantry¶
How, when and why did South African agriculture become capitalist, and what caused the rise and decline of the black peasantry, sharecropping and labour tenancy?
Works engaging the debate: 31
Positions¶
- Bundy's rise-and-fall of the peasantry
- Prussian path and structural transition (Morris, Murray)
- Complex, uneven, locally contested transition (Keegan, Bradford)
- 1913 Land Act: segregation ideology vs class restructuring vs political pressure
Evidence base¶
The debate has been argued mostly from qualitative state and commission records read regionally. Few series have been built and tested. The core primary corpus is shared across the social-history contributions. It comprises Native Affairs Department files in the Central and Transvaal Archives, such as the NTS labour and Trust files (Bradford (1990)) (Schirmer (1995)) (Clynick (1990)) (Morrell (1990)). It also includes Department of Justice magistrates' annual reports for 1910-13 and 1918 (Keegan (1985)) (Morrell (1990)) and Bloemfontein Colonial Secretary and Agriculture files (Keegan (1979)) (Keegan (1986)). Commission evidence is equally central: SANAC 1903-05, the Natives Land (Beaumont) Commission Minutes of Evidence UG 22-1916 and the 1918 local committees, the Native Economic Commission 1930-32 and the Native Farm Labour Committee 1937-39 (Keegan (1979)) (Keegan (1985)) (Morrell (1988)) (Brookes (1924)) (Schirmer (1995)). The 1913 Act literature adds Hansard, Governor-General, Prime Minister and private papers, and a microfilmed Chamber of Mines file (Feinberg (1993)). Pondoland adds trader ledgers (O'Donnell), WNLA and Mine Labour Organisation files, Blue Book labour-pass tables and dipping returns (Beinart (1982)). The evidence for most claims is coded as illustrative or partial: named farmers, districts and episodes rather than samples.
Only a handful of works build systematic quantitative evidence. Simkins (1981) reconstructs reserve output from the Union Agricultural Censuses 1918-1969, valued at constant 1946 prices against population census denominators for 1936-70. He tests it with Cobb-Douglas regressions across magisterial districts for 1927, 1937, 1950 and 1960. Lewis (1984a) retabulates household-level Cape census rolls, among them the 1848 Ngqika census, Peddie 1854, Ox-kraal/Kamastone 1861, Healdtown 1869 and the 1875 Tamacha enumeration papers. From these he draws distributions of acres, ploughs, cattle and wives per homestead, alongside weekly market returns from 1857-58. Feinberg (1993) counts speakers in the 1913 Hansard (45 speakers, at least 17 on the bill's principle) and uses corrected Lands Department returns on Transvaal African farm purchases by period. Scattered district figures appear elsewhere. Examples are Bradford's NTS 2233 recruit schedules (over 18,000 northerners, about 75% channelled to farms by 1945), Schirmer's Agricultural Census numbers for Lydenburg farm size, income and tractors, Clyhick's 70,000 bags of African deliveries to the Lichtenburg Co-op in 1944, and Morrell's agricultural export series. Bundy's own work is not in this set. His Healdtown, Victoria East and Thembuland figures enter only second-hand, through (Lewis (1984a)) and Binswanger and Deininger (1993). A large share of the 31 works are syntheses or polemics with no new data: the three Lemon versions, both Mabin versions, both Crush versions, (Murray (1987)), (Murray (1989)), (Keegan (1989)), (Tatham (1990)), (Iliffe (1999)), (Jeeves and Crush (1990)) and Crush and Jeeves (1993).
Contested evidence¶
The rise-and-fall thesis is contested on Bundy's own numbers. Lewis (1984a) re-reads the Healdtown 1858 marketed surplus against subsistence needs and finds it inadequate. He reads the 1857-58 market returns as many small sales of 50-100 lb made to pay hut tax. He reads the 1882 Thembuland Commission stock and plough data as showing most households were no better off than their grandparents. On this reading, apparent prosperity reflects stratified homesteads and distress sales rather than a market-responsive peasantry. Binswanger and Deininger (1993) uses Bundy's Victoria East data in the opposite direction, as evidence that the Land Act destroyed viable farming. For the twentieth-century reserves, (Simkins (1981)) finds total output flat from 1918 to 1965, with per-capita collapse only after 1948/1955. Yet (Mabin (1991a)) and Mabin (1991b) cite Simkins for the claim that the reserves were already overcrowded and deteriorating by 1948. Clynick (1990) meanwhile finds African maize deliveries substantial in Lichtenburg until about 1950. On the 1913 Act, the same corpus is read three ways. Keegan (1985) and Keegan (1979) read the Beaumont evidence, magistrates' reports and Plaatje as showing that the Act's ban on sharecropping was ineffective but catalysed evictions of wealthier stock-owning tenants, with Act 42 of 1908 as forerunner. Feinberg (1993) reads Hansard, Cabinet-crisis correspondence and Free State caucus papers as showing segregation politics driven by Hertzog's Free State bloc. He argues from silence in a three-item Chamber of Mines file against labour-supply explanations. Brookes (1924) treats the Act as the execution of SANAC recommendations. On timing and character of the transition, the Morris/Murray versus Keegan/Bradford exchange is conducted almost entirely through rival readings of Keegan's Rural Transformations and Bradford's ICU study. Murray (1989) uses quotation, and Keegan (1989) uses theory and analogy; neither presents census, tenancy or credit series. Keegan dates the decisive break to post-1945 tractor mechanisation without tractor statistics. Iliffe (1999) and Schirmer (1995) cite 1946-60 tractor counts, and Jeeves and Crush (1990) doubts mechanisation explains the 1950s shift from labour shortage to surplus, though he too gives no data. The 1948 watershed is contested on incompatible evidence. Bradford (1990) uses pre-1948 NTS and GNLB files to show state-brokered forced labour. Mabin (1991b) rests the scale of post-1948 tenant removals on a single Platzky and Walker estimate of probably at least two million. Morrell (1990) reads Natal Midlands justice and NAD files against Keegan's picture of a distant rural state.
Underused and missing sources¶
The evidence ledgers repeatedly identify the same unexploited sources. The first is district-level agricultural and population census series on labour tenants, squatters, farm employees, tractors and African output on white-owned land. These would date the shift from sharecropping and labour tenancy to wage labour. They would also test Morris's 1920s thesis against Keegan's post-1945 thesis, and the late-1960s farm employment peak (Iliffe (1999)). A second gap is Land Bank mortgage data and deeds-registry bond and transfer records. These would test whether farm capital came from outside agriculture (Keegan (1986)) and whether credit went to debt or investment (Keegan (1985)). Further gaps are Chamber of Mines maize contracts and mine purchasing records for the gold-maize alliance (Morrell (1988)), and labour bureau registers and influx-control prosecution data (Lemon (1984a)) (Bradford (1990)). Co-op delivery ledgers by race, and Trust tenant registers, would quantify African commercial production and stratification (Clynick (1990)). Master and Servants case registers and eviction orders would test farmer capture of magistrates (Morrell (1990)). Simkins's series on farm employment and incomes outside the reserves for 1923-69 is cited by Jeeves and Crush (1990) and Schirmer (1995) but not analysed. The Nel Commission and the national Department of Agriculture files on labour tenancy abolition remain unused.
Strength of key claims¶
| Basis | Share of key claims |
|---|---|
| Systematic data | 2% |
| Partial data | 57% |
| Illustrative or anecdotal | 24% |
| Asserted | 17% |
Comparative figures for all debates are on the Literature overview.
Best-evidenced works¶
Simkins (1981), Lewis (1984a), Feinberg (1993), Beinart (1982), Keegan (1985), Schirmer (1995)
Most-used primary sources¶
Holdings are matched by machine against the archive's file manifest; a dash means no match was found, not that the project certainly lacks the source.
Works and their positions¶
| Work | Position |
|---|---|
| Beinart (1982) | Critiques market-centred peasant and underdevelopment models, and the Mfengu-based model of the 'progressive peasantry'. Argues that pre-colonial social relations and local political struggle shaped Pondoland's trajectory, and drops 'peasant' as too blunt a category. |
| Binswanger and Deininger (1993) | Follows Bundy, Wilson and Davenport. The land laws served mining labour needs and white farmers, and African farmers' competitiveness was deliberately destroyed to produce wage labour. |
| Bradford (1990) | Argues for separating monopoly from small-scale capital and for attending to links between merchant, financial and productive capital, rather than treating 'agriculture' as a single class interest. |
| Brookes (1924) | Treats the Act as hasty and impolitic but not iniquitous. It was based on the 1903-5 Commission's recommendations. Brookes favours partial, gradual and non-compulsory segregation, with added neutral areas where Africans may buy land. |
| Clynick (1990) | Homelands are read 'not so much as the creatures of a Verwoerdian nightmare, but as the products of a complex historical process of rural class struggle'. Apartheid is held not to have been crucial in setting the Trust's class character. |
| Clynick (1990) | Argues a prosperous, differentiated 'progressive' peasantry persisted in Lichtenburg into the 1940s and 50s and continued to grow on Scheduled land even as opportunities for African farmers declined elsewhere. It compares this to the Marico peasantry studied by Manson and Drummond. |
| Crush (1982a) | Disputes that the sequence was universal. Some societies (Swazi, Zulu) had no 19th-century commodity production, while others (Pedi) had migrant labour before it. |
| Crush (1982b) | Holds that the peasant-to-proletarian sequence was not universal. Migrancy and commodity production followed varied sequences, and internal division of labour (e.g. BaKoena chiefs) was a key determinant before 1886. |
| Crush and Jeeves (1993) | Notes the recent taxonomic debate (Morris, Keegan, Murray) and, following Bradford, treats it as a vital prerequisite for understanding and changing the present. |
| Crush and Jeeves (1993) | Says a growing body of research (Simkins, Beinart, Levin and Neocosmos, and the Vaughan–McIntosh and Drummond–Manson articles) reassesses and nuances Bundy's account of complete destruction. The peasant boom was uneven and benefited a small group. |
| Feinberg (1993) | Rejects the capital, farm-labour and mining-labour explanations as unsubstantiated. Holds that the stated aim was segregation, and that the Act was passed because of Free State pressure in a political crisis within the South African Party. |
| Feinberg (1993) | Finds no evidence that the Chamber of Mines lobbied for the Act. Farm labour was mentioned by few MPs and opinion was divided on whether the Act would increase labour supply. |
| Feinberg (1993) | Agrees with Wickins's caution about simple explanations but argues that his own wider evidence points to political pressure as the decisive cause of passage in 1913. |
| Iliffe (1999) | Holds that proletarianization seeped rather than swept, owed little to legislation and more to local struggles between farmers and tenants, and that a minority of African smallholders survived even apartheid. |
| Jeeves and Crush (1990) | Sides with the critics. It cites Keegan's view of a 'complex, ambiguous and multi-faceted process' and calls for geographically and temporally disaggregated, regional study. |
| Keegan (1979) | Treats the 1913 anti-squatting provisions as the culmination of Free State landed interests' drive to curb peasant independence and as rooted in agrarian class restructuring, not only racial ideology, with poor-white anxieties tied in. |
| Keegan (1979) | Rejects 'proletarianization' as a label, preferring 'enserfment' to describe the transition to labour-tenants with restricted movement and rights over their produce. |
| Keegan (1979) | Holds that imperial intervention (capital transfers and coercion) was the essential precondition, though administrative fiat alone could not restructure rural relations. |
| Keegan (1985) | Rejects the liberal image of a stagnant, backward countryside. Accepts the revisionist stress on early black peasant success and white agriculture's vigour, but criticises it as unilinear and functionalist, ignoring local specificity and the gap between law and practice. |
| Keegan (1985) | Only partly. The prohibition of sharecropping was largely unenforced and the main effect was tighter landlord control and extraction. It was neither the decisive blow nor mere consolidation of earlier law, as Wickens implies. |
| Keegan (1985) | Narrow economic calculation often favoured sharecropping, so populist ideology, poor-white anxiety and organised white mobilisation must be added to explain the assaults. |
| Keegan (1986) | Rejects it. He denies continuity between the old Boer landowners and the new capitalist farmers, and denies that proletarianisation followed automatically from industrialisation. He stresses discontinuity and contingency. |
| Keegan (1986) | Critiques both for assuming a single universal dynamic. He argues peasant persistence reflects peasant resilience and a weak colonial state, and may constrain accumulation rather than serve capital. |
| Keegan (1986) | Argues that the revisionist literature ascribes the decline solely to state action. He adds white farmers' day-to-day struggle, racial ideology, mechanisation and subsidy; the state was important but not decisive. |
| Keegan (1986) | Sharecropping was a response to white landholders' lack of capital and black tenants' resources, not a feudal survival. Labour tenancy was a rational solution to supervision and capital constraints. |
| Keegan (1989) | Rejects Morris. Keegan accepts Lenin's account of the capitalist character of some tenancy forms but argues the 1920s countryside was still dominated by primitive accumulation and share-cropping. Capitalism became dominant only with mechanisation after the 1940s. |
| Keegan (1989) | Accepts the Prussian analogy, but reinterprets east-Elbian and South African systems as products of external capitalist expansion mediated by merchant capital rather than as feudal. |
| Lemon (1984a) | Treats both as partly valid: concern over the blackening of the platteland coexisted with restructuring that permitted capitalist agricultural expansion, with Morris's account given more weight. |
| Lemon (1984b) | Treats both as valid: the 'verswarting' concern was real, but state intervention restructured relations of production and permitted capitalist farm expansion. |
| Lemon (1984c) | Treats the two as compatible. Anxiety about 'verswarting' of the platteland was real, but the removals restructured production relations and permitted capitalist agricultural expansion in white areas. |
| Lewis (1984a) | Rejects the periodisation. Argues most households never experienced a general rise, and that impoverishment began soon after conquest. |
| Lewis (1984a) | Stratified. Homestead heads controlled cattle, wives and clients, and chiefly redistribution reinforced inequality. Bundy and Wolpe are criticised for assuming egalitarianism. |
| Lewis (1984a) | The state did wish to preserve the reserves, but their persistence also reflects the internal dynamics of household growth and decline. |
| Mabin (1991a) | Follows Bundy: peasantry was progressively squeezed over a century, concentrating people into migrant labour and farm wage work, though it was not the sole foundation of the migrant system. |
| Mabin (1991a) | Acknowledges the issue is hotly contested (cites Keegan); sees the state accelerating a long, hesitant shift from labour tenancy to wage labour, without unanimous farmer support. |
| Mabin (1991b) | Accepts the argument that a peasantry was largely destroyed over a long period, so that labour tenancy gave way to wage labour. He notes the matter is contested (citing Keegan 1989) and stresses regional variation in agrarian history. |
| Mabin (1991b) | Treats the Acts as fundamental but warns against attributing too much to them, given regional variation. |
| McCann (2001) | Basotho and black sharecroppers responded vigorously to mining markets after 1867. They were later undercut by dispossession (1913 Land Act), labour loss and state favouring of white farms, which turned them into subsistence labour-reserve farms. |
| Morrell (1988) | Notes that farmers held tenants by legislation but also by concession (citing Keegan). Argues that for progressive farmers neither route suffices once year-round wage labour is required. |
| Morrell (1990) | Disputes Keegan's view of the state as a distant, unobtrusive actor; argues that local officials, especially magistrates, had real impact and were intertwined with farmers. |
| Morris (1987) | Presents the two farms as one of the few places where a black peasantry has survived, with substantial self-built infrastructure and largely subsistence production. The claim is based on local sources and is preliminary. |
| Murray (1987) | Murray endorses the books' view that the transition was complex, uneven and indeterminate, shaped by local class struggle and regional variation rather than a uniform schema. |
| Murray (1987) | Endorses Keegan's view that state support was a political necessity for undercapitalised white farmers, while outcomes were still shaped by local struggle and not simply determined by legislation such as the 1913 Land Act. |
| Murray (1987) | Sides with Keegan and Robertson that the seigneurial stereotype does not fit South Africa or Africa generally, and that share-cropping is inimical to class formation. |
| Murray (1987) | Argues that before mechanisation capitalist farming had no clear scale advantages over peasant farming, so white agriculture depended on state support. |
| Murray (1989) | Murray finds Keegan's rejection of the Prussian path neither definitive nor devastating, because it conflates Lenin and Barrington Moore. He sides with Morris's location of the transition in labour tenancy and the 1920s. |
| Murray (1989) | Murray argues that explaining white farmers' drive for control by racial ideology reproduces the liberal paradigm and leaves the explanation incomplete. |
| Murray (1989) | He rejects defining class by consciousness. In his view labour tenants should be analysed by how their labour was organised and surplus extracted, not by their 'peasant consciousness'. |
| Murray (1989) | Murray says Keegan denies the decisiveness of coercive state policy yet repeatedly relies on state intervention (Natives' Land Act, credit, marketing) to explain outcomes. |
| Schirmer (1995) | Rejects the view that rural South Africa moved towards a developed capitalist mode of production, in which resistance only drives change between predetermined stages. Argues that African resistance shaped the distinctive path of change. |
| Schirmer (1995) | Class analysis is 'helpful but inadequate'. Racial domination, patriarchal family structures and ideas of autonomy and slavery explain tenant behaviour, and racial solidarity often overrode class divisions. |
| Schirmer (1995) | Abolition came mainly from the failure of state regulation in the face of African resistance, together with massive racial coercion by the apartheid state, rather than a fully developed capitalist transformation of the countryside. |
| Simkins (1981) | Challenges the view of long-run decline in total output from early in the century. Total output held steady to the mid-1960s, and the dramatic per-capita decline is located after 1948/1955. |
| Simkins (1981) | Emigration sustained reserve subsistence capacity until the 1950s. Tighter influx control and relocation ended this, producing rapid decline and rising dependence on remittances. |
| Simkins (1981) | Cross-section regressions find no region or year with zero marginal productivity of labour, although the Ciskei and Natal showed diminishing returns to density. |
| Tatham (1990) | Reports that Morris's 'Prussian path' thesis of an entirely capitalist agriculture by 1920 was challenged and 'proved wrong on empirical grounds'. She sides with Keegan's view that fixing a point of transition does violence to real historical processes. |
| Vink (2012) | Treats the Land Acts and marketing legislation as complementary instruments that cut out black competition and secured cheap labour for mines and farms (following Bundy). |