Introduction
The Fast Track Land Reform Programme (FTLRP) begun in 2000 resulted in a dramatic change in the agrarian structure of rural Zimbabwe. Moyo (2011a) revealed the emergence of a trimodal agrarian structure in which the peasantry now dominates. The trimodal classification criteria suggest that land endowment is the foremost factor of import for social differentiation among rural farmers, yet land utilisation varies across households within and across settlement models, as Patnaik (1976) and Akram-Lodhi (1993) observed elsewhere. Specifically, Moyo (2011a, 944) argues that the trimodal agrarian structure is ‘based on differences in land size, forms of land tenure, social status of landholders and capacity to hire labour’. In this sense, while the trimodal agrarian structure somewhat explains land ownership patterns after the FTLRP, it falls short in revealing the important drivers for the changing agrarian relations. Moreover, Zimbabwe’s agricultural land sizes are differentiated along agro-ecological regions, with the drier regions hosting larger sizes compared to wet regions (Government of Zimbabwe 2001). However, farmers in the wet regions are more diversified and productive in both crop and livestock production while those in the drier regions mainly specialise in livestock production, if not supported by irrigation facilities.
Scoones et al. (2011, 1) observed the need for ‘solid, on the ground research aimed at finding out what happened to whom and where and with what consequences’ after the FTLRP. To this end, following surveys in Masvingo Province, Scoones et al.’s (2010) study revealed that radical transformation in land and livelihoods had resulted in a new composition of people in Zimbabwe’s countryside. Their study in Masvingo showed the emergence of four categories of farmers, where 10% are ‘dropping out’, 33.6% are ‘hanging in’, 21.4% are ‘stepping out’ and 35% are ‘stepping up’. Their study revealed that land holdings, access to remittances, access to farming assets, farm residence, off-farm activities and agricultural production account for social differentiation in post-FTLRP Zimbabwe. Zamchiya (2012) uses the tri-modal model as an analytic schema for the A11 farms in a case study in Chipinge district; however, the study eschews peasant classification.
The context for Zimbabwe’s reconfigured agrarian relations is important. First, an evaluative study of rural capital accumulation and social differentiation is timely, 19 years after the FTLRP (Shonhe 2017). Second, the FTLRP was implemented amidst political contestations, capital flight and a poorly performing national economy resulting in both disinvestment and reduced investment in critical productive sectors (Kamidza 2013), including agriculture (Nyakazeya 2009). Third, the introduction of heterodox policies culminated in the implementation of neoliberal policies, epitomised by the dollarisation of the economy and refinancialisation from 2009, with a direct and immediate impact on agricultural funding, production and marketing (Shonhe 2017). Fourth, widespread allegations of patronage, politicisation of land allocation and the distribution of agricultural inputs (Alexander 2006; Marongwe 2008; Zamchiya 2012) imply an effect on agricultural output, thereby limiting the scope for capital accumulation and social differentiation at the household level.
Fifth, repeated droughts over the period between 2001 and 2008 also affected agricultural productivity and capital accumulation, stunting class formation. Moreover, de-industrialisation experienced from 1999, when the gross domestic product (GDP) dropped by 50% (Kamidza 2013) and unemployment rose to over 90%, while industrial capacity utilisation dropped from 35.8% to 4–10% by 2008 (Nyakazeya 2009), all combined to trigger an urban–rural migration, which contributed to the reconfiguration of the countryside today. Notwithstanding, the highly variegated settings in which land reform was carried out make drawing conclusions difficult. Moreover, the implications of the refinancialisation of the agrarian economy through domestic capital and international capital under neoliberalism remain sketchy. This article seeks to close this gap in knowledge and to explore its implications for the agrarian economy in Zimbabwe.
The rest of this article is structured thus: first addressed is the changing land ownership after the FTLRP. This is followed by a conceptualisation of peasant classification and agrarian transition. The next section then proffers a historical review of social differentiation, before presenting the methodology and background setting of the study in Hwedza district. This is followed by a section on land utilisation in Zimbabwe in general and in Hwedza district in particular, and an analysis of the Emerging Farmer Classification model, comprising four categories of farmers, in the reconfigured agrarian economy. The final section reveals the significance of differentiated access to capital in social differentiation, followed by a discussion situated in the agrarian transition debate, before rounding off with the conclusion.
Changing land ownership after the FTLRP
Zimbabwe’s FTLRP resulted in a significant structural transformation of agrarian relations, which Moyo (2011a) observed to have brought about a trimodal agrarian structure with diverse land tenure systems. Moyo (2011a) identified the following land tenure systems: large-scale commercial farms (LSCFs), small-scale commercial farms (SSCFs), old resettlement areas (ORAs), communal areas (CAs) and newly resettled areas (A1 – villagised small-scale farms and A2 – mid-sized farms). The FTLRP reversed the iniquitous racially based land ownership pattern which was born out of settler accumulation by dispossession from 1890 (Amin 1976) and the Land Apportionment Act of 1930 (Arrighi 1966). Land alienation resulted in discriminatory land tenure in which 6000 white farmers and colonial capital dominated. Most of the alienated black Africans were turned into wage workers on commercial farms, where the industry was highly dependent on 700,000 peasant families and 8000 yeoman small-scale black commercial farms for social reproduction (Moyo 2011a).
This altered the land ownership pattern, in which 10 million hectares in the hands of white large-scale farmers was transferred to a broad mass of indigenous Zimbabweans, thereby reversing land access inequalities, constraining agricultural production and social differentiation associated with colonial rule (Arrighi 1966). The FTLRP resettled over 146,000 families on A1 plots, allocated an average of 6 ha. A further 23,000 families received A2 commercial farms (Moyo 2011a) averaging 140 ha each. The agro-industrial estates now constitute 240 establishments held on freehold title and include eco-tourism ventures and conservancies (Moyo 2011b). The enlarged peasantry of 1,319,000 households rely on self-employment of family labour, producing food crops, mainly for auto-consumption, selling some surpluses and often engaging in non-farm activities, including selling labour on a casual basis, albeit in a differentiated fashion, on A1 plots and communal land (Moyo 2011a).
Farmers in this medium-scale category have greater access to capital finance given the nature of the tenure system, in which title deeds and the 99-year lease arrangements potentially enable access to credit. The ongoing robust debate on the acceptability of the 99-year leases as bankable security coincides with an increased orientation towards neoliberalism that may lead to new land tenure amendments promoting bankability. Similarly, land rentals and leases under joint ventures and state-led partnerships in the agro-fuel and livestock sectors are on the increase, involving former commercial farmers, some Chinese capitalists, United Kingdom companies and some business executives, some A2 farm beneficiaries and state land. On the A1 farms, some Zimbabweans in the medium-scale category, involved in variegated income-earning activities, are also involved in land rentals and leases, participating in cash and export crop production.
The medium-scale and large-scale farmers in the resettled areas have begun to benefit under Command Agriculture. Command Agriculture, also known as the Special Maize Programme for Import Substitution, encompasses the mobilisation of domestic and international capital towards complex state-mediated contract farming arrangements involving provision of support in input supply and commodity markets for grain foods and some cash crops. Maize and other small grains are sold through the Grain Marketing Board (GMB) via a reintroduced stop order system.2 Command Agriculture is mistakenly narrowly viewed as merely a military-assisted party patronage extension by the ruling Zimbabwe African National Unity – Patriotic Front (ZANU–PF) without unpacking its full architecture and effect on agricultural development (see, for instance, Moyo 2017). Arguably, this complex state-led contract farming system is reconfiguring agrarian relations alongside private-driven international and domestic capital contract farming accessed through tobacco merchants, all combining to restructure Zimbabwe’s countryside. In all, this points to the increasingly changing workings of capital in the agrarian economy.
This reconfiguration of the agrarian structure and associated changes in the tenure system has combined with the negative effects of capital flight triggered by the FTLRP to accentuate changes in the financing, production, marketing, accumulation and social differentiation trajectories in the countryside. In this sense, the trimodal classification criteria do not reveal the full implications of the FTLRP, not least because some variables with significant political economy and statistical significance were missed in the analysis. Importantly, the impact of the changing workings of capital far beyond the role of politics compels reflection.
Methodology and survey data
The setting for this article is Hwedza district in Mashonaland East Province in Zimbabwe, which is located approximately 127 km south of Harare, where a case study was carried out in 2016. The case study used a mixed-methods approach in which 230 respondents were selected for a survey using structured questionnaires, and 20 in-depth interviews were carried out, collecting information on production, marketing, capital accumulation and social differentiation. The quantitative data were processed using SPSS statistics software for data analysis. Hwedza district uniquely straddles four agro-ecological regions (Natural Regions IIa, IIb, III and IV3) that receive different levels of rainfall and therefore are suitable for different cropping and livestock programmes.
Hwedza district encompasses six settlement models. Respondents were selected from five farming sectors, as shown in Table 1: there is one remaining white-owned LSCF and two owned by indigenous black farmers, some medium-scale farms composed of A2 farms and SSCFs, CA plots, ORAs and the A1 farms, as shown in Figure 1. The LSCFs are in located in the natural agro-ecological region (NRIIa) along the Watershed West area and the A1 and ORA are in NRIIb, while the CAs are situated in NRIII and NRIV and the SSCF area is in NRIV in the Zviyambe area.
| Settlement type | Ward no. | |||||
|---|---|---|---|---|---|---|
| 13 | 4 | 3 | 1 | 5 | ||
| Farming sectors | Communal cluster | |||||
| Targeted sector | SSCFs | ORAs | A1 | A2 | CAs | Total |
| Population size | 2423 | 3198 | 2983 | 4813 | 6664 | 20,081 |
| Ward households | 533 | 722 | 627 | 1136 | 1625 | 4643 |
| Household sample size | 53 | 46 | 46 | 32 | 53 | 230 |
Note: the targeted sectors include small-scale commercial farms (SSCFs), old resettlement areas (ORAs), A1 and A2 plots (see note 1), and communal areas (CAs). Source: compiled from CSO census data, Ministry of Agriculture (2016).
Cluster sampling was relied upon and carried out in three stages: first, purposely selecting Hwedza district as the largest cluster; multiple stages of purposely selecting five clusters so that agro-ecological and settlement types are included; and randomly selecting respondents in each of the areas identified to avoid bias. Respondents for 20 in-depth interviews were purposively selected based on their knowledge and community positioning. The study area has granite stones with some intrusions of dolerite and greenstones, which generate richer, redder soils compared to the dominant sandy soils, suitable for maize and tobacco production as well as cattle ranging. The wide range of settlement models emerging from the extensive FTLRP and agro-ecological regions provide the differentiated settings appropriate for a review of the agrarian structure in rural Zimbabwe. While there are three remaining LSCFs, these were not included in this study as they are outliers, by dint of their unique capitalist operations.
Conceptualising peasant classification and agrarian transition
In Bernstein’s (2009, 249) view, ‘much is obscured by characterizing social formations in the South today as peasant societies, or contemporary classes of petty-commodity producing small farmers as peasants’. In Marx’s ([1852/1853] 1977) writings, the peasantry is a pre-capitalist remnant to be dragged into modernity and through the full development of a capitalist mode of production. Marx ([1852] 1971, 230) therefore maintained the view that smallholder production did not promote division of labour, introduction of science, diversity of development, talent or the development of social relations. Moyo and Yeros (2005), however, observed that a process of re-peasantisation is ongoing following the FTLRP, discounting the possibility of repeating the European pre-deterministic experience propounded by Marx (1967). In this sense, for Zimbabwe, following the urban–rural migration consequent to de-industrialisation in the early 2000s (Shonhe 2018) and the redistribution of former white commercial farm land to indigenous black farmers, the countryside is now dominated by small-scale family farmers (Moyo 2011a).
The use of peasantry in depicting farming families in the countryside has therefore been subject to much debate. Arguably, the changing dynamics of small-scale farming and the ‘intensifying systemic crisis’ in the countryside due to globalisation (Akram-Lodhi and Cristóbal 2010, 178), and their increased incorporation into domestic and global commodity chains, render depiction problematic. Moyo (2016, 1) defines peasants as ‘small-scale family farmers who mainly depend on family labour and produce a significant share of their own food and may use their labour for off-farm activities and wage labour’. This article applies this conceptual view in peasant classification and uses ‘peasants’ simply to mean small-scale family farmers, to whom we now turn.
Patnaik (1988) discusses a simple ‘Labour Use Index’, suggesting that the complex farm economics data collected from cross-sectional samples of holdings make their use quite meaningless. As a result, the index does not consider different scales of capital investment (green revolution inputs, technology such as irrigation and equipment, cropping patterns, size of family labour) per unit area. This results in a cross-sectional comparison of diametrically opposite types of households placed in the same group based on an analysis of acreage size data, obscuring their capital investment and social differentiation. Whereas Patnaik’s work is conceptually grounded in political economy tradition, the avoidance of analytical complexity that accommodates and achieves the fullness of significant variables in preference to a simple Labour Use Index, which relies on labour hiring tendencies, is self-defeating. Notwithstanding, Patnaik (1988) agrees with Kautsky (1988) and Lenin (1964) that land size is just one of the factors that affect farm-scale production, to which must be added, as Akram-Lodhi (1993) observed, the cropped area, rented-in land, cattle holdings, access to water, quality of soil, green revolution seed and fertilisers, level of mechanisation and access to labour.
Patnaik’s (1976) labour-exploitation index, while congruent with the conceptualisation of the small-scale farmer mode of farming, emphasises the use of hired-in labour, reliance on family labour self-employment and family labour hire-out for empirical classification of households, therefore rendering it narrow and simplistic. It eschews the impact of a green revolution linked to many variables, mainly access to finance resources. Thus, as Akram-Lodhi (1993) concedes, this index has not been widely used and has faced unrelenting criticism from many scholars, particularly Athreya et al. (1987). Nonetheless, Patnaik’s (1988) argument that differences in resource endowment within the heterogeneous small-scale farming sector present real chances for qualitatively differentiated trajectories for agricultural investment and social differentiation unobscured by land size is on point. The scale and intensity of investment and advancement in technology are critical variables that drive surplus generation per unit area and ought to be included in contemporary political economy analysis of rural agrarian economies. Zimbabwe’s peasantry is now significantly incorporated into global commodity chains through the input and output markets, including contract farming, joint ventures and partnerships now fashioning new agrarian change trajectories for Zimbabwe’s agrarian economy.
Social differentiation in Zimbabwe: a historical perspective
Studies on peasant social differentiation have informed dynamics on the rural agrarian economy since the Russian Revolution. Arrighi (1966) undertook a comprehensive review of how social differentiation developed during the colonial period. In the main, the workings of capital in the agrarian sector, mining and later the manufacturing sector shaped social differentiation. As Arrighi asserts, an increase in the European population from 1902 triggered a demand in agricultural commodities, supplying mines and emerging urban centres. A white rural bourgeoisie then developed, ‘consisting of owner-workers of small and medium-sized mines and farmers’ (Arrighi 1966, 36). During this period, Africans were mainly ‘a class of self-employed rural cultivators’, while ‘African wage-workers and African middle-classes and petty bourgeoisie’ were an extension of the rural peasant cultivators (37). Later, the promulgation of the Land Apportionment Act of 1930 alienated the productive land from the Africans (Ibid.) and created reserves in less productive areas that became the virtual dormitories of labour reserves (Hendricks and Lushaba 2005). The act pushed the male population to work in the mines and emerging industries owned by the white bourgeoisie. Common interests between the white classes and the British South African Company resulted in combined programmes in investments and legislative initiatives aimed at proletarianising the African male population.
Therefore, two fundamental events combined to drive racialised social differentiation in the 1930s. First, the land dispossessions undermined Africans’ accumulation and as such converted them into wage workers on farms, in mines and in a nascent manufacturing industry, in response to unrelenting demands for taxes and fees under colonial laws. Second, the First World War created the need to produce locally, as external supply could not be guaranteed. This resulted in the development of the white bourgeoisie, running a nascent industry, who employed white wage workers and African wage workers. Later, the influx of Europeans after the Second World War increased the importance and role of agriculture, particularly tobacco, if not the replacement of food with cash crops during the 1950s. However, through semi-proletarianisation, on the back of white capital, wage workers became a source of social reproduction to their families, who remained on the land under the traditional land tenure system in the countryside. An African bourgeoisie was therefore formed on this premise, and so was the nationalist class.
In post-colonial Zimbabwe, the study by Cousins, Weiner, and Amin (1992), based on communal area surveys in different agro-ecological settings in the 1980s, is informative. First, reviewing Amin’s (1991) survey, Cousins et al. observe that the distribution of means of production for agriculture (cattle, land and implements), income from crop sales and other off-farm activities was highly skewed, with a high impact on inequality indices. Households attaining lower agricultural production were more likely to have family members locally employed as wage labourers and less likely to have migrant workers in skilled jobs, with the former group highly dependent on wage labour for subsistence (Cousins, Weiner, and Amin 1992). Amin’s (1991) study, conducted in the 1987–88 farming season in Chirau and Magondi districts in agro-ecological regions II and III, respectively, revealed the existence of a three-strata society – poor (22%), middle (60%), and rich (18%) – in the district.
Similarly, based on a 200-household survey in Mtoko communal areas in regions III and IV carried out in 1985, Coudere and Marijsse (1988) found that 20% of the population earned 50% of the total income, while wage earners were the richest, and widows who had no access to remittances were the poorest, even though such access was the most important variable in explaining social differentiation. The provision of agricultural financial and material support by migrant workers is a colonial design where capitalism in the form of industrial development was in turn subsidised by agricultural production under the traditional land tenure system (Arrighi 1966). A similar study by Jackson, Collier, and Conti (1987), during the same period, based on 600 households in five agro-ecological regions, revealed that 10% of the households earned 43% of the income and controlled 40–60% of the marketed commodities, while 50% of the poorest households controlled as little as 10% of the income.
It was observed that during drought years, most income came from wage income. For instance, Bonnevie (1987) discovered that, due to drought, 57% of the income came from wage remittances, while 17% came from farming. This was based on a survey of 72 households in Mangwende communal area in 1983. Wage earners were better placed to produce and earn farming income. For instance, in 1984, Weiner and Harris (1991) interviewed 417 households in five agro-ecological regions and found that households with access to wage income were more integrated into the market through selling of crops and hiring of labour, but that this had a strong association with access to cattle for draught power and access to land. With regard to concentration of livestock, Jackson, Collier, and Conti (1987) found that 44–54% of the cattle were held by the top 10% of the households in the surveyed areas.
Previous writings on peasant classification in Zimbabwe observed that class formation was stirred by state interventions in communal areas (see Cousins, Weiner, and Amin 1992) and post-FTLRP reconfigurations (Scoones et al. 2010), mainly the broadened land ownership, capital flight and heterodox economic policies adopted by the government in response to them (Moyo and Yeros 2005; Moyo et al. 2009; Moyo 2011a, 2011b; Moyo and Nyoni 2013). Some authors also observed the pre-eminence of the semi-proletarianisation thesis (see Moyo and Yeros 2005), where remittances from formally employed family members, who straddle town and rural homes, finance agricultural production. However, the role of remittances declined following de-industrialisation beginning in 1999; instead, massive re-peasantisation from then onwards shows increased access to land among peasant farmers following the FTLRP (Moyo and Yeros 2005) and changing access to working capital (Moyo and Nyoni 2013), in particular, tobacco contract farming (Scoones et al. 2017; Shonhe 2017). The trimodal agrarian structure suggests that 78.6% of agricultural land is held by 1.3 million small-scale farmers, while 13.4% is held by medium-scale farmers and 8% is held by black large-scale farmers, remaining white large-scale farmers, agro-estates, corporates and state institutions.
Land utilisation
Given that Moyo (2011a) based the trimodal classification on land sizes, as well as on forms of land tenure, social status of landholders and their capacity to hire labour, upon which crop production and capital accumulation tendencies were to be predicated, it is apt to examine its impact on social differentiation. First, smallholder farmers under the CA, ORA and A1 are automatically classified as peasantry; smallholder farmers are in SSCF; A2 farmers are then identified as medium-scale farmers; while the LSCF and the corporates, conservancies, parastatals and other institutional land holdings are agro-estates. However, empirical evidence reveals that land sizes contribute less significantly to agricultural commodity production patterns, marketing trends and capital accumulation processes in rural Hwedza district, compared to other factors. Second, changing access to and workings of capital affect the total cropped area, which influences crop production trends and asset endowment. While there are huge differences in land sizes across settlement models, the average cropped area for most commodities is unconnected to sizes of land owned.
Third, moreover, the cropped area per crop per farming sector reveals that the sizes of the average cropped area for the major crops, maize and tobacco, do not correlate with the accessible land sizes, as Table 2 reveals. The land sizes fall short in aligning with the respective capital accumulation tendencies among farmers. To illustrate this point, while A1 farmers surveyed held an average of 7.6 ha, their average cropped area stood at 1.8 ha (23.7%), compared to 9.5 ha held by the ORA group who had an average cropped area of 1.3 ha (13.7%), and A2 farmers who hold an average of 56.3 ha of arable land but till an average of 7.6 ha (13.5%). Plainly, the cropped area is lopsided, with the CA peasant farmers (who own an average of 3.7 ha) utilising 45.9%, representing 1.7 ha of accessed land, compared to SSCF who own 96.5 ha and utilise a mere 3.6% (3.5 ha) of it.
| Model | No. | Total area per sector (ha) | Average area available (ha) | Arable land (ha) | Average arable land (ha) | Total cropped area (ha) | Average cropped area (ha) | Irrigated area (ha) | Average irrigated area (ha) |
|---|---|---|---|---|---|---|---|---|---|
| A1 | 46 | 348.2 | 7.6 | 263.5 | 5.7 | 84.7 | 1.8 | 0 | 0 |
| % of arable land | 4.70 | 75.67 | 32.1 | 23.7 | 0.00 | ||||
| SSCF | 51 | 4924 | 96.5 | 4747 | 93.1 | 177 | 3.5 | 0 | 0 |
| % of arable land | 67.17 | 96.41 | 3.73 | 3.6 | - | ||||
| A2 | 26 | 1462.5 | 56.3 | 1264 | 48.6 | 198.5 | 7.6 | 4.5 | 0.17 |
| % of arable land | 19.90 | 86.40 | 15.7 | 13.5 | 2.30 | ||||
| ORA | 45 | 428.2 | 9.5 | 371.4 | 8.3 | 56.8 | 1.3 | 0 | 0 |
| % of arable land | 5.80 | 86.70 | 15.3 | 13.7 | 0.0 | ||||
| CA | 46 | 168.25 | 3.7 | 85.85 | 1.9 | 80.3 | 1.7 | 6.5 | 0.14 |
| % of arable land | 2.30 | 51.03 | 93.54 | 45.9 | 8.56 | ||||
| Total | 230 | 7331.15 | 34.26 | 7077.0 | 33 | 686.38 | 3.2 | 13.31 | |
| % of arable land | 100 | 96.53 | 9.69 | 9.3 | 1.94 |
Note: SSCFs - small-scale commercial farms; ORAs - old resettlement areas; CAs - communal areas.
Source: Shonhe survey, 2016.
The implications of this revelation are far-reaching for the African agrarian economy. First, differentiated crop production within and across settlement models implies that the types of crops and marketing channels are equally varied. The peasantry is no longer confined to food crops, and as such, the medium- and large-scale capitalist farmers no longer have a monopoly over cash crops. On the contrary, cash crops such as tobacco are now more predominant among the peasantry compared to the capitalist classes identified in the trimodal agrarian structure (see Moyo 2011a). Second, beyond land sizes, production and accumulation processes are being shaped by access to the workings of capital, where a variety of variables have gained importance, warranting a more nuanced analysis of the ongoing reconfiguration of agrarian relations. Third, the peasant classification model needs revisiting. A simplistic, qualitative classification approach will not capture the complex dynamics of agrarian change in variegated settings where a large-scale land redistribution programme has been implemented and the macro-economic environment and political settings are undergoing a crisis, as has been the case for Zimbabwe from 1999.
The Emerging Farmer Classification model
This article uses component factor and cluster analysis and political economy lenses to decrypt the Emerging Farmer Classification in rural Hwedza district. Component analysis is a useful procedure to synthesise quantitative data with beneficial applications to peasant classification. Using the clustering method aided by an in-depth political economy analysis captures the variegated and differentiated settings in which the reconfiguration is emerging, as it provides scope for the consideration of the many variables driving agrarian change.
Using the reduction dimension factor (RDF), the first part of the component factor analysis (CFA), variables pertinent to the classification of farmers in the sample are derived. This process, with the aid of a literature review and knowledge gained from in-depth interviews, produces predictors for social differentiation (see Moyo et al. 2009; Shonhe 2017). Besides identifying the predictors, the RDF produces the total variance explained, the scree plot, the communalities and the Kaiser-Meyer Olkin measure of sampling adequacy, and Bartlett's test, which are all valuable for the classification process. Through an elimination process, total variance is explained, and the scree plot depicts the number of variables with an eigenvalue above 1 and therefore significant for the classification of the farmers. Using standardised variables, the communalities derived identify the extraction values through ranking of the predictors based on their levels of significance. The CFA tools (total variance explained, and the scree plot and communalities) produced five predictors, namely ‘cattle owned’, ‘family labour hired out’, ‘food security’, ‘maize output’ and ‘tobacco output’. Overall, the nominal regression analysis gave a p value of 0, which implies that the hypothesis should be rejected and that there is a relation among the factors being considered. The SPSS-based regression significance testing therefore gives comfort that the clustering process carried out in this study is both reliable and valid.
These variables are linked to access to capital, cropped land and subsequent changes in asset endowment. The variable regression significance testing indicates the factors that are useful for clustering of the farmers. Variables with significance below 0.1 are considered relevant for the farmer clustering process, while those equal to or above this figure are not significant and are to be excluded from the analysis. Table 3 shows that some variables were revealed as not being significant (age, income earned, year settled, non-farm activities, credit and gender) and they were therefore not emphasised in this analysis. While this contradicts factors identified by Moyo (2011a), some of the factors used in the classification process coincide with those identified by Akram-Lodhi (1993). The latter identified ‘cropped area’, ‘amount of rented land’, ‘number of animals’, ‘water availability’, ‘quality of soils’, ‘quality of seed and fertilisers’, ‘degree of mechanisation’ and ‘availability and use of labour’ as important variables in the classification process. The African Institute of Agrarian Studies (AIAS) 2005–06 survey also relied on the CFA to derive predictors for peasant classification and identified ‘livestock holdings’, ‘food security’ and ‘access to capital’, among others (AIAS 2007).
| Parameter estimates | ||||||||
|---|---|---|---|---|---|---|---|---|
| Average maize cluster | B | Std. error | Wald | df | Sig. | Exp(B) | 95% confidence interval for Exp(B) | |
| Lower bound | Upper bound | |||||||
| Intercept | −2.591 | 24.121 | .012 | 1 | .914 | |||
| Respondent’s age | .009 | .018 | .270 | 1 | .603 | 1.009 | .974 | 1.045 |
| Household size 2015 | .138 | .114 | 1.474 | 1 | .225 | 1.148 | .919 | 1.435 |
| Arable | −.060 | .214 | .080 | 1 | .778 | .941 | .618 | 1.433 |
| Scotch carts owned 2014 | −.255 | .536 | .226 | 1 | .634 | .775 | .271 | 2.216 |
| Income average | .000 | .000 | 2.044 | 1 | .153 | 1.000 | .999 | 1.000 |
| Average family labour involved in agriculture | −.232 | .273 | .723 | 1 | .395 | .793 | .465 | 1.353 |
| Average family labour hired out | −.607 | .357 | 2.895 | 1 | .089 | .545 | .271 | 1.097 |
| Average male permanent workers | −.631 | .386 | 2.673 | 1 | .102 | .532 | .250 | 1.134 |
| Maize average value | .489 | .520 | .886 | 1 | .346 | 1.631 | .589 | 4.515 |
| Year settled | .000 | .000 | .046 | 1 | .831 | 1.000 | .999 | 1.001 |
| Number of months maize to last | .064 | .064 | .992 | 1 | .319 | 1.066 | .940 | 1.208 |
| Average tobacco output | .000 | .001 | .315 | 1 | .574 | 1.000 | .998 | 1.001 |
| Average maize output | .000 | .001 | .472 | 1 | .492 | 1.000 | .999 | 1.002 |
| Average cattle owned | .162 | .069 | 5.449 | 1 | .020 | 1.176 | 1.026 | 1.347 |
| Average dryland tobacco sold | −.142 | .295 | .232 | 1 | .630 | .868 | .487 | 1.545 |
| Involvement in non-farm activities | .099 | .525 | .036 | 1 | .850 | 1.104 | .394 | 3.091 |
| Access to credit | −3.746 | 8.851 | .179 | 1 | .672 | .024 | 6.908E-10 | 806640.642 |
| Access to contract farming | .479 | 1.056 | .205 | 1 | .650 | 1.614 | .204 | 12.798 |
| Access to irrigation | 2.891 | 4.461 | .420 | 1 | .517 | 18.009 | .003 | 112968.480 |
| Gender | 2.298 | 24.087 | .009 | 1 | .924 | 9.958 | 3.128E-20 | 3.170E21 |
| Availability of electricity | 1.634 | 7.194 | .052 | 1 | .820 | 5.124 | 3.852E-6 | 6816780.745 |
| Availability of documentation | −.792 | .631 | 1.577 | 1 | .209 | .453 | .131 | 1.560 |
Source: Shonhe survey, 2016.
Using pre-set clusters in SPSS, the two-step cluster analysis developed four categories based on five factors (‘average maize output’, ‘average tobacco output’, ‘average cattle holdings’, ‘number of months maize harvest will last’ and ‘family labour hired out’), hereinafter referred to as the Emerging Farmer Classification model. The four categories of farmers in the Emerging Farmer Classification model are poor emerging farmers, medium-scale emerging farmers, medium-scale to rich emerging farmers; and rich emerging farmers, in the proportions shown in Figure 2. The distribution of farmers along the Emerging Farmer Classification model shows that poor peasants constitute 31.7%; medium-scale emerging farmers 51.7%; the medium-scale to rich emerging farmers 14.8%; and the rich emerging farmers 1.7%. The characteristics of the different categories of farmers emerging can be explained based on how the different variables in Table 4 played out during the studied period (March 2012–May 2014). The five variables identified through the CFA are used to assess the emerging clusters.
| Criteria | Poor emerging farmers | Medium-scale emerging farmers | Medium-scale to rich emerging farmers | Rich emerging farmers |
|---|---|---|---|---|
| 1. Maize output (kg) | 821 | 1487 | 2680 | 6917 |
| 2. Tobacco output (kg) | 268 | 448 | 768 | 19,433 |
| 3. Cattle owned | 4 | 5 | 25 | 25 |
| 4. Months of harvested maize supply | 0.7 | 9 | 10 | 11 |
| 5. Labour hired out | 0.3 | 0.1 | 0.3 | 8.7 |
Source: Shonhe survey, 2016.
Poor emerging farmers
The poor peasants have a mean average maize output of 821 kg and a mean average tobacco output of 268 kg, own on average four head of cattle and have barely a month of food supply (see Table 4), and are differentially located in different settlements, mainly within the small-scale family farming models: 28.3% of the A1 settlement households, 33% of the ORA households and 49% of the CA households (see Table 5). Agro-ecologies do not impact on agricultural productivity and poverty levels given that most of the poor located in the CA, ORA and A1 models are in NRII, which receives high rainfall adequate to support rain-fed agriculture. Food security is the most important variable for the poor households. The poor peasants located in the CA are observed to have low receipts in local and diaspora remittances and have limited proceeds from agricultural commodity sales. With less than 18% of the male population in the poor emerging farmers category, the female farmers who have no access to remittances and savings dominate this category of emerging farmers.
| Cluster | N | % | A1 | % | SSCF | % | A2 | % | ORA | % | CA | % |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Poor emerging farmers | 73 | 31.70 | 13 | 28.26 | 10 | 18.87 | 9 | 28.13 | 15 | 33 | 26 | 49 |
| Medium-scale emerging farmers | 119 | 51.70 | 32 | 69.57 | 32 | 60.38 | 13 | 40.625 | 20 | 43 | 22 | 42 |
| Medium-scale to rich emerging farmers | 34 | 14.80 | 1 | 2 | 8 | 15.09 | 9 | 28.13 | 11 | 24 | 5 | 9 |
| Rich farmers | 4 | 1.70 | 0 | 0 | 3 | 5.66 | 1 | 3.13 | 0 | 0 | 0 | 0 |
| Combined | 230 | 100 | 46 | 100 | 53 | 100 | 32 | 100 | 46 | 100 | 53 | 100 |
Note: SSCFs - small-scale commercial farms; ORAs - old resettlement areas; CAs - communal areas.
Source: Shonhe survey, 2016.
This category also comprises the female-headed households that have limited access to cattle holdings and therefore rely on hiring out casual labour, as indicated by informant MN during an interview (personal interview, MN, 13 July 2016, Hwedza district).
The poor emerging farmers are also characterised by limited access to inputs, which impacts negatively on agricultural production, leading to food insecurity and limited integration into the global food circuits. The poor emerging farmers category relies on hiring labour out for social reproduction. Importantly, the study revealed that there are some poor families in all the farming models, of which 18.87% are SSCF households and 28.13% are in the A2 model even though they hold average land sizes of 96.5 ha and 56.3 ha, respectively.
Medium-scale emerging farmers
Hwedza district has 51.3% medium-scale peasants, consisting of 69.6% of the A1 households, 60.4% of the SSCFs, 40.6% of the A2 households, 43% of the ORAs, and 42% of the CAs. The main variables influencing placement in this category are cattle ownership, food security and maize production. On average, medium-scale farmers produce 1500 kg of maize and 450 kg of tobacco annually, keep an average of five cattle, have food stored that would last up to 9 months and hire out family labour to supplement family income. A small proportion of the male population is involved in migrant labour and remits money to finance agricultural production. As Moyo (2016) observed, these are semi-subsistence producers focused mainly on producing food crops for auto-consumption. The involvement of family labour, mainly women, is higher in this category. There are more male-headed families, including youths, involved in semi-subsistence farming in this category.
Medium-scale to rich emerging farmers
A medium-scale to rich emerging farmer category is also identified, including 14.8% of the peasantry in Hwedza district and consisting of 15.1% of SSCF households, 28.1% of the A2 farmers, 24% of the ORA households, 9% of the CA households and 2% of the A1 households. The medium-scale to rich emerging farmers have an average of 2680 kg in maize output and 770 kg in tobacco output, maintain an average of 25 head of cattle and 10 months' food supply, and rarely hire out labour. Food security and cattle ownership are the most important predictors for this category of farmers. In particular, cattle provide draught power and guarantee wealth, while cattle sales proceeds are important for the purchase of agricultural inputs and social reproduction of the household. The majority of these farmers are located in SSCF and A2 land types and have access to tenural systems that promote access to credit, even though this has been limited due to economy-wide crises in Zimbabwe. Yet some CA households (9%) and 2% of the A1 households also fall in this category.
Due to high participation in contract farming and the production of cash crops, the medium-scale to rich emerging farmers are highly intergrated into the international and domestic financial and commodity markets. Based on the Hwedza district case study, a combination of medium-scale peasants and medium-scale to rich emerging farmers constitutes 66.1% of the farming households. Faced with limited state support in agricultural production over the period of the survey (2012–2015), medium-scale emerging farmers and many other farmers across the farming models are now resorting to proceeds from sale of agricultural produce. This notwithstanding, the ‘independent’ medium-scale farmers have begun to accumulate more, becoming incorporated into global commodity markets. This category of farmers has high cattle holdings which serve as a source of wealth, draught power and sales proceeds and as an important indicator of one’s social standing in the community (personal interview, JP, 23 July 2016, Hwedza district). The category comprises state bureaucrats, industrialists and members of the security sector (personal observation, July 2016). A few divorcees and widows continue to run farm enterprises in the A2 and SSCFs sectors, respectively.
Rich farmers
The rich farmers constitute a mere 1.7% of the smallholder farmers; they are situated in SSCF and A2 farms in equal proportions and produce an average of 6920 kg of maize and 1950 kg of tobacco per annum, maintain an average of 25 head of cattle and have 11 months of food supply. The rich farmers have higher resource endowment compared to other categories, and as such contribute most in tobacco and maize sales and are therefore more integrated into the commodity and financial markets. In the main, these farmers have greater access to scotch carts, green revolution inputs, tractors and technology (including irrigation) and hire in more labour. This category comprises family members who are gainfully employed and are able to finance farming operations. They hold key farming equipment, employ farm labour and are male dominated compared to the other categories. Some of them are connected to party hierarchy or have military links and therefore tend to benefit from patronage networks, mainly in the form of farming inputs. Moreover, both the state bureaucracy with strong party links and some military personnel oversee the agricultural distribution channels.
Social differentiation and access to capital finance
The study revealed that access to finance has a significant bearing on social differentiation in Hwedza district. Farmers have a wide range of other sources of income and produce other agricultural commodities (Scoones et al. 2017), mostly from non-traditional sources. For instance, the Hwedza district survey revealed that farmers rely more on proceeds from agricultural sales (52.6%) than tobacco contract farming (117%). However, more proceeds from the sale of contracted tobacco are reinvested into the next season than those from any other sources of finance. In any event, tobacco farmers engaged in contract farming have higher agricultural sales proceeds compared to independent farmers. This is influenced by the fact that contract farming offers better prices compared to auction floors. Most tobacco contract farmers move to the medium-scale and medium-scale to rich emerging farmer category, depending on their scale of production. Similarly, semi-proletarian farmers who remit some financial resources applied towards agricultural production are in the medium-scale category. These farmers hire in labour as opposed to hiring out, due to the intensity of labour required for tobacco production and their increased capacity to do so by virtue of increased revenue.
The trimodal structure inadequately captures the ongoing class formation in rural Zimbabwe. First, the classes are based on state categorisations of farm sizes (Government of Zimbabwe 2001) and agro-ecological regions and are therefore devoid of class consciousness. Notably, farmers occupying land on the same tenure system may be in different agro-ecological regions and may end up in different categories due to differentiated access to capital and agricultural production patterns. Second, a review of the farm-size categorisation process reveals that not all the farms in the A1 and ORA are smallholders, and neither are all A2 farms found in the medium-scale class. Moreover, the categorisation adopted by Moyo (2011a) assumes homogeneous farm sizes per settlement model. Even though A1 farms are generally 6 ha in size for arable land and 9 ha for grazing land, the trimodal model identifies them as smallholders. Yet studies in Africa generally define smallholders as those holding under 5 ha of land, and medium-scale farms as those holding 5–100 ha (Jayne et al. 2016) consisting of A1, ORA, A2 and SSCFs models. In this sense, the categorisation offered by the trimodal structure is limiting and does not sufficiently reveal ongoing social differentiation in Zimbabwe. Third, even if these misgivings are discounted at the national and subnational levels, the trimodal structure remains trapped by policy rather than technical considerations. As the survey revealed, the cropped area is far smaller than the total land endowment.
Consequently, peasant classification must consider the important role of capital in influencing agricultural productivity, defining marketing options and influencing capital accumulation processes, and ultimately in social differentiation. Differentiated variables associated with the workings of capital influence agricultural productivity, capital accumulation and social differentiation across settlement models, far beyond the social status and total land held. By identifying the five critical variables for the classification process, namely ‘average maize output’, ‘average tobacco output’, ‘average cattle holdings’, ‘number of months maize harvest will last’ and ‘family labour hired out’, the Emerging Farmer Classification model illustrates the central importance of access to agricultural finance in social differentiation.
This indicates an increased significance of capital in shaping agricultural production patterns and capital accumulation, indicating that land size, social status and land tenure and capacity to hire labour, while important, do not fully explain social differentiation among the farming households. As the Emerging Farmer Classification model shows, households in the four categories of farmers are located across the settlement models, distinguished mainly by their differentiated access to capital finance. Essentially, the Emerging Farmer Classification model reveals how the increasing role of international and national/local capital has begun to shape social differentiation at micro and macro levels within the agrarian economy.
The preponderance of medium-scale farmers affirms tendencies established in other African countries such as Ghana, Kenya, Tanzania and Zambia (Jayne et al. 2016). The medium-scale farms control about 20% of total farmland in Kenya, 32% in Ghana, 39% in Tanzania and above 50% in Zambia (Ibid). The role of off-farm income invested by urban-based professional and influential people in the rural areas was noticeable. The medium-scale farmers were seen to be gaining influence in policy and prioritisation of expenditure. In Zimbabwe, excluding the LSCFs and CAs and therefore focusing on 72.3% of the farmland, 90.9% of agricultural land is held by medium-scale farms. The Emerging Farmer Classification model for Hwedza district indicates that the medium-scale farms hold 66.1% of the farmland, quite in tandem with trends in other countries in Africa.
What drives this agrarian reconfiguration in Africa, and possibly in Zimbabwe? In the face of high levels of de-industrialisation after 2000 and urban-to-rural migration which triggered re-peasantisation through the FTLRP, it has not been possible for the agrarian economy to rely on investment from urban-based professionals. Equally, capital flight associated with sanctions imposed in response to political, governance and land reform issues limited access to international and domestic finance capital prior to 2009. The reinsertion of capital thereafter has not been adequate to shape agrarian relations in Africa. Moreover, notwithstanding the reintroduction of the Maguta – Command Agriculture – from 2016, the government’s ability to dispense patronage-based support has been limited by its own capacity challenges. Notwithstanding the differences in sources of capital driving the agrarian economy, there is an ongoing structural transformation (Jayne et al. 2015, 2016) occurring in many African countries. However, for Zimbabwe, an extensive land reform programme and subsequent de-industrialisation after 2000 have meant that reinvestment of agrarian capital from commodity sales predominates. This is generating new farmer classes, knowledge of which is critical for policy, practice and theoretical reflections on agrarian change in post-FTLRP Zimbabwe.
Conclusions
In this article, the increased importance of capital in the structural transformation of Zimbabwe’s agrarian structures was revealed, and an Emerging Farmer Classification model dominated by medium-scale emerging farmers was observed. The accumulation trajectory shows an upward mobility of rural emerging farmers from poor to medium-scale emerging farmers across farming models and agro-ecological regions. This movement is fuelled by reinvestment of agricultural sales proceeds, a phenomenon observed across the settlement models. Whereas monopoly capital, both international and domestic, was reinserted through contract farming targeting cash crops following an earlier flight from 2000 to 2009, local (self-financing) agrarian capital has begun to trigger the scope for sovereign agrarian transformation. Rather than land endowment, access to capital is driving accumulation in rural settings. The introduction of state-driven Command Agriculture (a form of contract farming), from the 2016/17 agricultural season, is revitalising this tendency, focusing on food crops.