Contract farming and the reorganisation of agricultural production within the Chókwè Irrigation System, Mozambique

Abstract

This paper contributes to the discussion about ‘inclusive business models’ as alternatives to large-scale land acquisitions by analysing a case in which a foreign agribusiness investor, within an impact investment paradigm, acquired and rehabilitated a rice processing plant in Chókwè, Mozambique. A contract farming programme drawn up to source raw produce for the factory led to radical shifts in control over land and water resources. The case is set against the background of a large-scale irrigation system that has played an important role in national agricultural policies since colonial times. The private sector-led development approach is strongly supported by the state which portrays it as an opportunity to reduce its dependency on rice imports. The investor redirected the benefits of land and water use through taking a role of coordination and control within the irrigated production system.

The ensuing rural transformation reminds one of earlier periods in the development of the irrigation system: the central control over production, and the very high level of agricultural inputs and mechanisation, remind one of the state socialist period, while the removal of smallholders from their land and its concentration in the hands of a few bears a strong resemblance to the scheme's establishment under colonial rule.

Keywords:

• contract farming
• irrigation
• Mozambique
• land grabbing
• water grabbing
• rice

Acknowledgements

This contribution builds on the work conducted by the CP66 team (Munguambe, Chilundo, and Julaia 2009). I am particularly grateful to Nynke Post Uiterweer, Alex Bolding and Pieter van der Zaag for sharing their experiences and insights. Discussions in the context of the supervision of the MSc thesis by Felipe van der Struijk (2013) provided additional insights, while an assignment for The Royal Netherlands Embassy in Maputo provided the opportunity to conduct additional fieldwork (see Veldwisch, Beekman, and Bolding 2013).

I thank all interviewees for their time and openness. MIA has been particularly open and helpful in providing me insights into their ways of working. Its working culture of openness and approachability came as a welcome surprise.

Two anonymous reviewers went to considerable lengths to improve the paper, amongst other things leading to a stronger historical contextualisation and a more pronounced argumentation.

Notes

¹An earlier version of this paper was published in Spanish as Veldwisch (2012).

²The system is located about 200 km northeast of Maputo. The scheme takes its water from the Limpopo River at the Macarretane diversion weir, where water enters the canal system by gravity.

³The terminology is part of the government's discourse and rhetoric to show its continued commitment to make the irrigation system ‘work’ and dates back to the Third FRELIMO (Frente de Libertação de Moçambique) Congress in 1977 (Pellizzoli 2010a, 2010b); see section 2 of this paper.

⁴See Amilai (2008, 67–73) for details on the design and construction period.

⁵A second-class Portuguese citizenship for Mozambicans, who received special education and adopted Portuguese language and customs.

⁶Also called chibalo.

⁷The designation ‘complexo agro-industrial’ (CAI) is a specific category of state farms with an organisation based on that of Soviet sovkozes and with similar organisational structure to the agro-industrial complexes in certain socialist countries in Europe. Together with the state enterprises producing for export, the CAIs were generally considered more important than other state farms and were under the direct control of the Board of Ministers (Mosca 1996, 25–26).

⁸For Mozambique as a whole, there was an indexed decline for marketed rice from 100 in 1975 to 79.9, 46.4 and 30.7 for 1976, 1980 and 1981 respectively (Mosca 1996).

⁹Eventually, all 10 state farms went bankrupt (West and Myers 1996).

¹⁰Many households cultivated much more than 1 ha, as these areas were allocated to individuals rather than households.

¹¹Rice production grew ‘to 12,000 tons in 1982–3, 16,000 tons in 1983–4, 28,000 tons in 1984–5, 42,000 tons in 1985–6, and an estimated 50,000 tons in 1986–7’ (Bowen 1989, 367).

¹²Hermele (1986, cited in Amilai 2008) reports that in the 1984–1985 season, there was a cultivated rice area of 8150 ha with an average yield of 3.4 t/ha and a total production of 28,015 tons.

¹³This was the worst flooding in 150 years (Silva et al. 2010).

¹⁴In 2005, this was called GPSA, Gabinete de Promoçao do Sector Comercial Agrario (Office for the Promotion of the Commercial Agricultural Sector), which was later transformed into CEPAGRI, Centro de Promoção da Agricultura (Centre for the Promotion of Agriculture).

¹⁵The increase was mainly a result of area expansion after the peace treaty of 1992; yields per hectare hardly increased at all during this period.

¹⁶‘Among major food products, rice is the only product facing an import tariff’ (ACI 2005, 46).

¹⁷In 2010, MIA had the technical capacity to process 20,000 tons of paddy annually, and a matching storage capacity of 20,000 tons. In April 2010, MIA was in the final stages of installing a hi-tech mechanical dryer with a capacity of 350 tons/day.

¹⁸Previously, this company was known as Moçfer. The name change was announced on Moçfer's website on 4 June 2010 (Mozfoods 2010).

¹⁹Interview, 30 March 2010.

²⁰Personal observation during my first field visit in November 2008, backed up by Munguambe et al. (2009, 25, Fig. 13) which also clearly illustrates this; the area that was used by ARENDONZE was fragmented by intersecting waterlogged and saline areas, while MIA's 500 ha were in one piece and located on the other side of D11, where no problems of water logging are reported.

²¹Munguambe et al. (2009, 23), for instance, note that ‘it is “impossible” to plough heavy clay soil with hoe or animal traction in most of the areas within the CIS’.

²²Interview, 30 March 2010.

²³Source: sample contract provided by MIA during a field visit.

²⁴In practice, this works out to 4.5–4.8% effective interest over the production period.

²⁵Field notes, 30 March 2010.

²⁶Field notes, 30 March 2010.

²⁷Field notes, 30 March 2010.

²⁸CIS is divided into blocks of 16 ha that are referred to as ramal. Each ramal is served by a caleira (also called a regadeira), which, in irrigation design terms, would be called a tertiary canal.

²⁹Field notes, 30 March 2010.

³⁰The total land area reported as being used in D11 during the 2009/2010 season was 680 ha: 500 ha by APs, 50 ha by experimental APs and 130 ha by small producers. This is considerably less than the 1000 ha reported for the 2006/2007 and 2007/2008 seasons. The reasons for these differences are not clear but are possibly due to larger parts of D11 not being very suitable for crop production anymore, due to water logging and salanization.

³¹Field notes, 29 March 2010. Also see Van der Struijk's MSc thesis (2013) which illustrates the role of the WUA board members in land concentration within the WUA boundaries.

³²Field notes, 31 March 2010.

³³Ideally, land preparation starts in July, while transplanting and seeding take place in October and November (Chilundo et al. 2012).

³⁴MIA had aimed for 5000 ha of land cultivated by APs, but did not manage to finish the selection process in time.

³⁵See also the discussion by Pellizzoli (2010a) about this issue.

³⁶It was not possible to cross-check the numbers presented by MIA, but they are within the ranges presented in Tables 6 and 7.

Additional information

Gert Jan Veldwisch is an assistant professor of Irrigation and Development at the Water Resources Management group of Wageningen University, The Netherlands. Gert Jan coordinates the Water Equity Network for Southern and Eastern Africa, a regional collaboration under the Water Justice Alliance (www.justiciahidrica.org), a comparative action-research and capacity-building programme which focuses on processes of accumulation of water rights, conflicts and civil society responses.