Labour Process and Workers' Bargaining Power in Export Grape Production, North East Brazil PDF

Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. Labour Process and Workers’ Bargaining Power in Export Grape Production, North East Brazil Oxford, Journal JOAC © 1471-0358 Original XXX Labour Ben The Selwyn Author. Process UK of Articles Agrarian Journal andChange Workers’ com...

Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. Labour Process and Workers’ Bargaining Power in Export Grape Production, North East Brazil Oxford, Journal JOAC © 1471-0358 Original XXX Labour Ben The Selwyn Author. Process UK of Articles Agrarian Journal andChange Workers’ compilation Bargaining © Blackwell PowerPublishing in Export Ltd, Grape Henry Production Bernstein and Terence J. Byres. Blackwell Publishing Ltd BEN SELWYN This article uses the Global Commodity Chain (GCC) framework to investigate labour regimes in export grape production in the São Francisco (SF) valley, North East Brazil. A combination of strict northern retailer requirements and producers’ ability to target export windows leads to an increasingly complex labour process. Whilst much GCC literature focusing on export agriculture concludes that labour is relatively powerless, this article presents a rather different case. The need to upgrade production continually in response to retailers’ demands gives workers strategic leverage which, together with a strong and continuing tradition of rural trade union organization, means that they have been able to extract significant concessions from exporting farms. Keywords: agricultural labour, global commodity chains, grape production, North East Brazil INTRODUCTION This article investigates how changes in the organization of northern horticultural markets impact upon producers, local level development and capital-labour relations. It does so by charting the formation and functioning of the São Francisco (SF) grape branch, in particular its export segment, located in North East Brazil, and by examining its evolving labour process.1 It draws on the global commodity (or value) chain framework (Kaplinsky and Morris 2001) and attempts to contribute to it by investigating and theorizing the position, role and power of labour within the chain. Ben Selwyn, Department of Politics and International Relations, University of Southampton, Southampton, SO17 1BJ. e-mail: [email protected] The article is based on fieldwork carried out in 2002 and 2003 for a PhD at SOAS under the supervision of Henry Bernstein, whose advice also helped in the writing of this paper. During this period I conducted research on 26 farms in the SF valley using the participant observation methodology as formulated by Jorgensen (1989). I carried out semi-structured interviews with farm managers, owners and workers. I also carried out off-farm interviews and collected data from development agency staff, producer cooperatives and marketing boards, rural trade unions and the local office of the Ministry of Labour. I presented an earlier version of this paper to the Historical Materialism World Development Research Seminar in London, December 2006. 1 I use the term branch in accordance with the convention of conceptually dividing the economy into sectors (for example, industry and agriculture) and branches (specific activities within each sector, for example grape vs grain production within the agricultural sector). © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Over the last three decades or so, the trading of fresh fruits and vegetables, often referred to as Non-Traditional Agricultural Exports (NTAXs) has become one of the fastest growing branches of world trade (McMichael 1997). Barham et al. (1992, 43) propose that the term ‘non-traditional’ refers to a product (commodity) that has not been produced in a particular country (or region) before, to a product that was previously produced for domestic consumption but is now exported, or to traditional products produced to meet the needs of new markets. The expansion of global NTAX production has gone hand in hand with the rise of global retail (supermarket) corporations which play a central role in transforming food production and sourcing. In addition to falling transport costs, technological innovations in the cool chain such as new transportation and refrigeration techniques have made possible the production of highly perishable NTAXs in regions that were previously too far from northern markets. New agricultural zones have been established analogous in many ways to the opening up of frontiers for grains and cattle during the nineteenth century (Friedmann 1993; Schwartz 2000). The rise of NTAX production and the integrating role of global supermarkets in commodity chain ‘governance’ is a central feature of contemporary globalization. The formation and integration of the São Francisco grape branch into the circuits of global NTAX production is one case of this broader phenomenon. The Global Commodity Chain (GCC) approach initiated by Gereffi and Korzeniewicz (1994) has become well established.2 It enables the researcher conceptually to connect actors involved in the production, distribution and marketing of the same commodity even though they may be geographically dispersed over great distances (Kaplinsky and Morris 2001). It requires a detailed focus on the commodity’s specific characteristics and production conditions. The GCC approach facilitates an analysis of power relations along the chain, by showing which actors make strategic decisions and which actors have to respond to them. By identifying such connections the approach enables tracing how transformations in one part of the world impact on transformations and developments in other parts of the world. Until recently the study of labour was largely excluded from GCC analysis. Henderson et al. note that ‘hardly any work has been done, on . . . the reproduction of labour power within the GCC’s perspective’ (2002, 459). In addition, ‘In so far as “workers” are present in this literature they appear as passive victims . . . as capital seeks cheap labour’ (Smith et al. 2002, 47). However, recently, several useful studies have incorporated labour into the GCC framework (Collins 2000; Barrientos and Kritzinger 2004; Dolan and Sutherland 2002), and this article follows their lead in order to illuminate the conditions and practices of labour in one branch of export horticulture. GCC-inspired research has mostly arrived at pessimistic conclusions regarding labour’s conditions in NTAX sectors. Kritzinger et al. (2004) find that horticultural 2 See the useful collection of articles at http://www.globalvaluechains.org/ © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 527 Ben Selwyn farms in South Africa supplying large northern retailers are subject to pressures of increased domestic liberalization, competition and increased quality-related buyer requirements, and are subject to the insecurities of price fluctuations for their produce. In these circumstances, they shed permanent labour, replacing it with contract labour, with significant adverse effects for conditions of employment and pay. Similarly, Korovkin (1992) and Gwynne (1999) show that the expansion of export grape production in Chile resulted overwhelmingly in the displacement of permanent workers by especially women temporary workers. Writing about the Chilean and South African deciduous fruit sectors, Barrientos argued that ‘the only group remaining at the base of the value chain, onto whom the growers themselves can offload some of the risk, is the fruit labour force, and particularly the more flexible seasonal workers, where the highest levels of female employment are found’ (2001, 88). More recently however, Barrientos and Kritzinger (2004) have found that increasing buyer quality requirements have led to some fruit exporting farms seeking to improve the skill levels of some of their workforce, with cases of employers preferring permanently employed over temporarily contracted labour. This latter observation is important, showing that GCC research is able to illuminate how changes in global markets, commodity chains and local conditions can, under certain conditions, lead to gains for workers. The following investigation of labour-capital relations is informed by Erik Olin Wright’s (2000) distinction between workers’ structural and associational power. Whilst the former refers to workers’ position within the production process, the latter refers to workers’ collective organization. The distinction is useful because it facilitates a focus on the changing context within which workers find themselves, via, for example, new production techniques, and their ability to respond and to improve their position in processes of production and accumulation. The following sections sketch the emergence of the SF grape branch and its composition. I then highlight northern retailer strategies and SF producer responses, and their effects for gains by labour. I locate those gains in a broader theoretical discussion of the determinants of labour’s bargaining power, and then draw some conclusions. THE EMERGENCE OF EXPORT GRAPE PRODUCTION IN NORTH EAST BRAZIL The focus of this paper is the grape branch located in the São Francisco river valley in the interior of North East Brazil (see map). This region, commonly referred to as the Sertão, is perhaps best known for its regular and crushing droughts as its name suggests (Hall 1978; Davis 2001). However, from the 1960s onwards the then military dictatorship and subsequent civilian governments in Brazil began to invest heavily in irrigation in the region in order to stimulate economic development there. In 1960 the North East as a whole had approximately 52,000 hectares of publicly irrigated land, 11.4 per cent of the national © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 528 Map 1 The São Francisco Valley in Brazil Source: Nilo Coelho Irrigation District (2002). © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 529 Ben Selwyn total. This increased to 261,000 hectares (17.6 per cent) in 1980 and 619,000 hectares (22 per cent) in 1988 (Ramos 1990, 483). By the late 1990s in the São Francisco valley alone there were over 40,000 hectares of publicly irrigated land and an additional 60,000 hectares of privately irrigated land (Selwyn 2007, 86). The valley experienced the emergence and expansion of a major fruiticulture sector and within that a thriving grape branch. In 1991, of the total area of all crops planted on public irrigation projects, fruiticulture accounted for 5,293 hectares (14 per cent) and by 1997 for 17,336 hectares (47 per cent) (Correia and Marinozzi 1999). By 2001, the major fruiticulture crops grown in the region were, approximately, in hectares: mango 16,500, coconut 10,000, banana 5,500, and grape 4,500 (VALEXPORT 2002). Between 1997 and 2001, foreign exchange earnings from grape sales increased from US$4.7 million to US$20.4 million (VALEXPORT 2002, 16). Between 1991 and 2000 grape export volumes increased from approximately 1,050 tons to 13,300 tons (VALEXPORT 2002). The principal grape-producing regions of Brazil are in the South and South East whose growing seasons are mostly confined to the period between December and March due to excessive rain during the rest of the year. The SF valley accounts for over 90 per cent of Brazilian grape exports (VALEXPORT 2002) because it is able to organize production to take advantage of periods of low supply in Europe, so-called export windows (see Figure 1). COMPOSITION OF THE GRAPE BRANCH In 2001 there were 1,407 grape farms in the SF valley comprising approximately 4,500 hectares of vineyards.3 The branch is divided between large, medium and small farms as follows. There were 81 large farms, above 10 hectares, comprising 5.7 per cent of all grape farms with around 46 per cent of all vines cultivated. There were 201 medium farms, of between 3 and 9.9 hectares, 14.4 per cent of the total with 24 per cent of vines. Small farms, below 3 hectares, numbered 1,125, 79.9 per cent with less than 30 per cent of all vines. The valley’s farms sell their grapes to four main markets. These are, in order of most stringent buyer requirements and highest prices received: UK, mainland European, top-end domestic and bottom-end domestic markets. The UK market consists primarily of giant (corporate) retail chains such as Tesco’s and Sainsbury’s; the mainland European market of a mix of wholesale markets and demanding corporate retailers such as Carrefour; and the top-end Brazilian market primarily of large supermarket chains such as Bom Preço (Dutch owned) and Pão de Açúcar (part French owned) and Carrefour (French). The bottom-end Brazilian market is made up of large wholesale markets, small local open-air markets and small fruit and vegetable outlets catering to lower-income consumers. Up to 10 per cent of farms sell onto the UK market, up to 36 per cent sell onto the mainland EU and top-end Brazilian markets, and up to 69 per cent sell onto the 3 This and the following data are derived from the Censo Fruticultura 2001 do Vale do São Francisco produced in CD format by CODEVASF, the latest available. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 530 Figure 1 Major grape sales in Europe: source of origin and month of supply Source: Information provided by UK importers and Kolla Hamburg. bottom-end domestic market (author’s calculation; Selwyn 2007). These figures exceed 100 per cent as some farms sell onto more than one market. For example, some farms targeting export markets sell to both the UK and mainland EU, and also sell produce that does not reach export standard on the top-end domestic market. Most large farms export. Medium farms are divided between those that aim to export exclusively, those that produce for both export and top-end domestic markets, and those producing for the bottom-end domestic market. The majority of small farms sell onto the bottom-end domestic market. Farms vary in their ownership structure, social origins and organization as well as size, and scale of grape production. The majority of large farms originate from international and national industrial and commercial capital. For example, the French retail giant Carrefour owns three large export grape farms in the region. Brazilian capital in large-scale export grape production include chemical firms from the South of Brazil and engineering firms from the North East. Some large farms originated in earlier agricultural cooperatives, in particular the Juazeiro Agricultural Cooperative (Cooperativa Agricola do Juazeiro – CAJ). CAJ is the biggest grape cooperative in the SF valley. In 2003 it had 45 members with approximately 300 hectares in grape production (just over 6 per cent of the total area). Medium farms are divided mainly between private investors who © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 531 Ben Selwyn have entered grape production quite recently, including those from agricultural backgrounds such as agronomists, as well as those coming from non-agricultural backgrounds and members of the CAJ. The majority of small farms in the valley are owned and managed by colonos: family farmers who were able to access land and settle on public irrigation schemes. Many exporting farms, including those belonging to the CAJ, are members of the Brazilian Grape Marketing Board (BGMB). BGMB has a quality control team that communicates and assists its members to produce and package export standard grapes. It provides packaging material and facilities, and most crucially, coordinates marketing in the UK and mainland Europe. A second significant export channel for producers is to sell their grapes to large valley-wide marketing firms that purchase grapes from mostly medium-sized farms for export to the mainland EU market. They assist their suppliers in order to guarantee the production of export quality grapes. The majority of farms producing for the domestic market are not organized in such ways for marketing their produce. Estimates of the number of workers in fruiticulture branches in the SF valley vary. One suggests that approximately 40,000 wage workers across the valley were involved in irrigated agricultural production in the mid to late 1990s (Damiani 1999, 14). According to Daimiani, 29,000 worked in the grape and mango branches, these being the principal export crops. In the mid 1990s, Bloch (1996, 49) estimated approximately 20,000 workers employed in the grape branch alone (but did not explain the basis of his estimate). Rural trade unions and the local Ministry of Labour provide similar estimates, although accuracy of their data is difficult to assess. Data for individual farms suggest significant variations in labour density according to farm size and market orientation. Large exporting farms employ more workers per hectare per production cycle than other farms. For example, Special Fruit farm cultivated 140 hectares of grapes in 2003 and employed 680 workers in grape production (an average of 4.8 workers per hectare); Brasil Uvas cultivated 100 hectares with 516 workers (5.1 per hectare); and Vale das Uvas 153 hectares with 700 workers (4.6 per hectare). Of the farms I surveyed, the average on large specialized export farms was 4.7 workers per hectare per cultivation cycle. These figures contrast to labour densities for medium farms selling onto the bottom end domestic market which average 2.5 workers per hectare. Small, colono farms selling onto the bottom end domestic market rely predominantly on family labour and employ labour on short-term (often daily) temporary bases. NORTHERN SUPERMARKETS AND SÃO FRANCISCO SUPPLIERS The sourcing and retail sale of fruit and vegetables in mainland Europe and particularly the UK has undergone a major transformation over the last three decades. One indication of this is the concentration of retailer market share. In the UK the total grocery market increased from a value of £45.5 billion in 1987 to £93.3 billion in 1998. Supermarkets increased their market share of fresh fruit © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 532 and vegetable sales from around 22 per cent in 1975 to 48 per cent in 1985, 78 per cent in 1995 and approximately 80 per cent in 1999 (Rabobank 2002, 14). Across the rest of Europe large retailers’ food market share is highly concentrated – in Sweden, Norway, the Netherlands, Switzerland, France and Belgium, the top three retail chains have a market share of over 60 per cent, and in Germany, Ireland and Portugal over 50 per cent (Rabobank 2002, 25). Retailers have tightened their control over the procurement of their supply in a number of ways. Whilst in the early 1980s UK retailers procured most fruit and vegetables from wholesale markets, by the late 1990s they collaborated with importers to procure directly from producers (Dolan and Humphrey 2000, 2004; Blythman 2004). Part of the reason for this was increasing public awareness about the risks of food production and marketing, leading to the UK 1990 Food Act which requires retailers to demonstrate ‘due diligence’ in their sourcing and sales of produce (Marsden and Wrigley 1995). Retailers’ greater market share and their sourcing directly from producers enable them to impose strict requirements on the latter. As Dolan and Humphrey (2004) note, producers are not only expected to supply a physical commodity, but also numerous additional services. In these ways, what was once a solidly price-driven commodity chain has evolved into an increasingly quality-driven commodity chain. The large UK retailers just mentioned source grapes on the basis that they meet well-defined quality standards such as bunch and berry size, berry colour, sugar levels; that they are consistent (including that grapes of the same variety sourced from different growing regions/countries are indistinguishable); and that supply is reliable, enabling them to sell grapes 52 weeks a year. Large retailers increasingly source produce from a list of approved suppliers that conform to their company standards. Typically these standards cover uses of pesticides, fertilizers and manures; pollution prevention and protection of human health. In response to these rising market entry requirements, producers in the SF valley have transformed the production process, making it increasingly scientific, and have formulated and implemented a product and process upgrading strategy along a number of fronts. From the mid 1990s the main exporting association (and employers organization VALEXPORT (Associacão dos Produtores de Hortigranjerios e Derivados do Vale do São Francisco, São Francisco Horticulture Exporters Association) – in collaboration with the state development agency CODEVASF (Commisão do Vale do São Francisco, São Francisco River Basin Commission) and some large farms – implemented a project to introduce seedless grapes into the SF valley in line with northern retail market demands. This project entailed research in other grape-producing regions around the world and attempts to adapt its findings to SF conditions. Whilst the processes of research, learning and adoption were initially difficult, by the early 2000s all large farms and increasing numbers of medium farms were producing export quality seedless grapes. A second important area of innovation and improvement in product quality is evidenced by the ability of SF farms to produce larger and higher quality © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 533 Ben Selwyn grapes. According to agronomists in the region, in the early 1990s farms were producing grapes with a 17 mm diameter. By the early 2000s UK retailers were requiring grapes with a diameter of up to 25 mm. SF producers have been able to enlarge grapes through rigorously scientific production methods. For example, producer organizations have quality control teams run by experienced agronomists that instruct members on the preferable number of bunches per vine, and berries per bunch, as well as training them in shoot, soil and leaf analysis which contribute to raising plant productivity. Other areas of upgrading pursued by SF producers include market diversification (selling onto different export markets, and using different brand names to sell to different buyers within the same markets), sourcing of cheaper inputs, and the establishment of, and training in, production methods required by corporate supermarkets such as meeting Eurepgap standards,4 and export standard packing houses. THE ORGANIZATION OF PRODUCTION The above shows that the production of export quality grapes is increasingly demanding. How do farm owners and managers in the SF valley respond effectively to these demands, and, more specifically, how do they organize labour and production processes in order to produce the grapes required by northern retailers? This section outlines how farms design and implement production calendars, and how they attempt to raise productivity in various ways. By investigating how production is organized, it is possible to illuminate the sources of workers’ structural power in the export grape branch. Production Calendars and Export Windows The SF valley benefits from a climate that provides conditions for year-round production. An annual average temperature of 28°C, over a narrow range of 25–30°C; low rates of average annual rainfall of about 540 mm; high sunshine intensity of about 3,000 hours per year; relatively dry air, discouraging the proliferation of pests and diseases; and a ready and controlled water supply through irrigation, enable farmers to produce all year round (EMBRAPA 2000). Consequently, they are able to produce and export to Europe when they choose, as opposed to other world market grape regions constrained by less advantageous climatic conditions. The two principal exporting periods (windows) for SF farms are May–June and November to mid January (Figure 1 above). To target these export windows, SF farms use detailed production calendars and require a managerial cadre and workforce capable of organizing work and working according to strict timetables. The following provides a stylized account of production and delivery-to-market timetables. 4 For Eurepgap see http://www.eurepgap.org/Languages/English/index_html and Rabobank (2001). © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 534 Table 1a. Production calendar for exporting farms: May–June window Start date 16 Dec 23 Dec 30 Dec 7 Jan 14 Jan 21 Jan 28 Jan 4 Feb Harvest date 15 22 1 8 15 22 1 8 April April May May May May June June Arrival in Europe 1 May 8 May 15 May 22 May 1 June 7 June 15 June 22 June Source: Interviews with farm managers. Table 1b. Production calendar for exporting farms: November–December window Start date 5 July 12 Jul 19 Jul 26 Jul 2 Aug 9 Aug 16 Aug 23 Aug Harvest date 1 Nov 8 Nov 15 Nov 22 Nov 1 Dec 8 Dec 15 Dec 22 Dec Arrival in Europe 15 22 1 8 15 22 1 8 Nov Nov Dec Dec Dec Dec Jan Jan Source: Interviews with farm managers. First Window: May–June: During this period global grape supply is sporadic. South Africa and Chile (and, for smaller quantities, India) supply the European market between January and April and by May exports from these countries are waning. The USA begins to export grapes to Europe from mid-June onwards, and Italy only begins supplying the European market in June before reaching its supply peak between July and October. Second Window: November–December: Italian grape supply declines in November. Spain supplies northern Europe in October and November, and the USA and India export some grapes to Europe during November and December respectively. South Africa and Chile begin exporting to Europe from late December through to May. SF grape farmers have a competitive advantage during the two seasonal windows in the annual calendar of global production and supply to European markets. Tables 1a and 1b show the two ‘typical’ exporting farm production © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 535 Ben Selwyn calendars.5 Production takes approximately 17 weeks from initial fertilizer application to harvest. The average time between harvest and arrival in European supermarkets is approximately two weeks. On farms that export during all eight weeks of the window, vineyards are usually divided into eight sections with one harvested each week. Cultivation of the first section begins in the first week of the calendar, and a further section is brought into production each subsequent week. Production Operations and Timing of the Cultivation Cycle Grape production has become more complex over the last two decades because of the rising buyer requirements detailed above. An indication of the complexity is the relatively high number of operations involved in the cultivation cycle. Table 2 shows the operations necessary to meet the standards for the four main market destinations. In the table, dark horizontal stripes denote operations always carried out, dark grey denote operations sometimes but not always performed, and those not shaded denote operations not performed. Column two refers to the timing of operations with numbers signifying weeks before, during or after the major cutting back of branches. Table 2 shows that the number of operations per cultivation cycle that are performed is determined primarily by export destination. Farms producing for the top-end UK market usually require up to 34 operations per cultivation cycle, grapes destined for mainland Europe require up to 31, top-end domestic require up to 25 and bottom end domestic up to 13 operations. Not carrying out operations on time, or without sufficient attention, results in lower quality fruit. A farm owner explained that ‘a bad pruning delays the whole cycle, leaves bunches too tight and can spoil the whole bunch. Also, with bunch thinning the berries may be touched too much by the workers, which leaves them with stains later on, and you have lower quality berries’.6 A high and increasing level of scientific plant management is thus required for export production. Whilst farms have detailed production calendars, variations in each plant cycle, depending on climatic conditions, mean that agronomists and managers must continually monitor plant, berry and bunch growth in order to ensure the operations are carried out at the optimal moment of the cycle. In contrast to exporting farms, colonos selling on to the bottom end domestic market do not have access to agronomists or strict production calendars, often find it difficult to recruit more skilled labour at the right time, and regularly find themselves lacking inputs such as fertilizers and agro-toxins at the necessary moment. As one explained ‘to produce quality fruit you need the right conditions and those require money. We don’t have enough to invest properly and so 5 I write ‘typical’ because not all farms export throughout the full eight-week window. Some concentrate their production and exports earlier or later in the cycle in partnership with their importers. 6 Interview with Kashaki Kabaka, CAJ, 2003. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 536 Table 2. Operations and timing of the cultivation cycle according to market destination Operation Timing of operations Market destination UK 1. Soil analysis 2. Soil aeration 3. Chemical fertilization 4. Organic fertilization 5. Major cutting back of branches 6. Painting shoot stubs 7. Collecting the branches 8. Breaking branches 9. Agro-toxin application 10. Dry-tying 11. Dry-tying 12. Cutting back shoots 13. Application of Giberelic acid 14. Tying back shoots 15. Application of Giberelic acid 16. Cutting back shoots 17. Leaf analysis 18. De-leafing/freeing bunches 19. Application of Giberelic acid 20. Cutting back of shoots 21. Re-tying of shoots 22. Application of Giberelic acid 23. Thinning young berries 24. Application of Giberelic acid 25. Freeing the bunches + selection 26. Cutting back shoots 27. Application of Giberelic acid 28. Pruning 29. Chinese hat 30. Cutting back shoots 31. Repeat major bunch pruning 32. Pre-harvest cleaning 33. Brix (sugar level) analysis 34. Harvest Mainland Europe Top-end Brazilian Bottom-end Brazilian −1 −1 −1 −1 1 2 2 2 2 2 3 4 4 5 5 5 5 5 6 6 7 7 7 8 8 9 9 9 11 11 12 15 16 16 Key: always performed sometimes performed never performed Source: Observations and interviews from different farms. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 537 Ben Selwyn Table 3. Incentive schemes according to farm size and market orientation Farm size and type Type of incentive scheme Number of operations covered by scheme Impact on worker productivity Large/medium, exporting Daily task target and piece rate (DTT+PR) plus bonus and gifts Daily task target (DTT) Most High Few Low Medium/small, domestically orientated Source: Interviews with farm managers and owners. cannot do things at the right time. Sometimes we know what we need to do but are not able to do so at the time, for example, a soil analysis which then demands fertilizer’.7 Hence, even when they do know how to produce high quality grapes, colonos often lack the resources to do so. Raising Productivity Farms in increasingly quality-driven chains supplying higher value markets thus require considerable investment in upgrading while also subject to the effects of competition. The latter include some decreases in farm gate prices since the early 2000s. Hence, managers need to reduce costs, including unit labour costs, whilst raising the quality, productivity and effectiveness of labour in order both to meet retailer requirements and to maintain profitability. They seek to do so through a number of productivity incentive schemes (Table 3). Daily task targets: The daily task target (DTT) productivity scheme where farms specify a daily number of tasks required of workers is most widely used on colonos farms. Once workers complete these they can stop work. On farms distant from their homes, workers stop and usually rest until their transport arrives. Where farms are close to their homes, workers often leave once they have completed the operations. The DTT scheme operates better on farms where the workers live nearby, for reasons discussed below, and is used most by colonos, albeit less precisely than on other farms. Daily task targets plus piece rates: The daily task target scheme is declining in popularity on exporting farms. Instead they are using a task target plus piece rate system (DTT+PR) because it enables them to utilize workers’ labour for a full eight hours per day. Workers are required to complete a given number of task targets daily, and after achieving this target, continue working on a piece rate. 7 Interview with Antonio Jose de Almeida, August 2003. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 538 At the time of this research, export farms required their workers to perform the following numbers of operations per day, before qualifying for bonus pay: cutting back shoots 100 plants, branch pruning 80 plants, tying back branches 80 plants, bunch pruning 500 bunches. Many farm managers regard this system as preferable to the daily task target system because having workers stop early often causes problems: We found out that this system [DTT] was no good since it left many workers with nothing to do for part of the day, and they would hang around on the farm, sometimes interfering with the work of others. With the new system the worker can achieve the number of operations that are required per day, and then continues working, and earns extra on top.8 Choosing Productivity and Incentive Schemes How do farm managers create, decide upon, and implement productivity and incentive schemes? Many have close relationships with each other. There are plenty of opportunities for interaction between them through inter-farm visits, and through meeting at producer organization lectures. As a manager explained, ‘we speak with technicos (agricultural technicians) and agronomists on other farms and so we all know what the best productivity rates are in the valley . . . then we try to reach these rates on our farm . . . there are not many secrets here.’9 In these ways common standards for productivity schemes spread across the valley: We know that a good podadore [pruner] can work on at least 80 plants per day, so we pay him the basic salary and then for every extra plant he works on he gets an extra so-many centavos . . . we know that this is the number that a good worker can achieve in the São Francisco valley, so we make sure that our workers are doing the same.10 Improving the Quality of Work Managers are not only concerned with increasing the quantity of operations performed by workers, but also the quality of their labour. Improving quality means training workers to avoid waste and inefficiency, and to perform new tasks that contribute to the production of higher quality fruit. For example, pruning vines at the start of the cultivation cycle is necessary in order to produce healthy vine shoots and high quality fruit. The operation itself is physically demanding and designated as a job for male workers, as opposed to ‘more dexterous’ female workers (Collins 1993). However, the job increasingly requires a close knowledge of vines. On exporting farms it is increasingly 8 9 10 Manager, Brasil Uvas, Juazeiro, July, 2003. Interview with manager, JMM, Petrolina, May 2002. Interview with manager, Frutivit, Petrolina, May 2002. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 539 Ben Selwyn common for managers to carry out a shoot analysis to assess whether they are fertile or not. They then tell workers which shoots to cut off, and which to leave. A good vine pruning leaves more fertile branches from which better quality grapes can be grown. Prior to this emphasis on raising the quality of labour it was common practice during pruning for workers to leave the same number of branches per plant. The new approach facilitates higher vine productivity, as fertile stems are identified through laboratory-based shoot analysis before pruning. More sophisticated technology goes hand in hand with raising worker skill levels as they are trained to understand how vines react to different activities. It is also increasingly the case that workers are taught about the consequences of the tasks they perform, rather than being told to simply carry out any particular operation. A manager explained: Today what is more important is their heads, rather than their strength . . . The workers need to understand which branches they need to cut back, which they need to leave, which vara [the shoot from which the grape bunches will emerge during the current cycle] and which osporao [the shoot that the grape bunches will emerge from during the following harvest cycle] they need to leave for the following harvest cycle. They are told before pruning which branches to cut back and leave based on the shoot analysis.11 Improving work quality involves greater expenditure by the farms in training and management costs. However, through increasing efficiency, this improves both the quality and the yields of the crop. It also reduces the necessity for the further application of labour later in the production process. Dolan and Sutherland write with reference to the Kenyan horticultural chain ‘While upgrading to more sophisticated processing has increased the demand for labour in developing countries (particularly in pack houses) this growth does not appear to be associated with an attendant demand for skilled labour’ (2002, 20). Whilst it is difficult to draw a precise distinction between skilled and unskilled labour, the present research suggests that in export grape production integration into more qualitydriven chains leads to a greater demand for relatively higher skilled workers. Displacing Labour through New Technology As noted above, more highly capitalized exporting farms employ more workers per hectare than less capitalized farms. However, over the last decade or so grape farms have been able to reduce the numbers of workers they hire per hectare per annum. While comparative data from ten years ago for the farms studied in this research is not available, a number of studies from the mid 1990s estimated that average numbers of workers per hectare on exporting farms was higher than it 11 Interview with Edwardo Shagawa, Curaca, March 2002. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 540 is today. Research carried out by the Brazilian Ministry of Agriculture in the mid 1990s estimated that one hectare of grape production required 5.4 workers per year (Ministerio da Agricultura e Abastecimento 1997, cited in EMBRAPA 2000). Damiani estimates that ‘growers of table grapes of [high quality] employ an average of . . . six workers per hectare’ (1999, 178). There is also local anecdotal evidence to suggest that there has been a general tendency on exporting farms for a reduction in the number of workers employed per hectare. Whilst new technology is often used to displace labour across different industries there are a number of limits to this option in the SF grape branch. Operations that comprise the cycle of production are based mainly on the application of direct human labour: pruning branches or bunches, cutting away flowers and young berries from the bunch or tying back and directing branches. These operations cannot be mechanized due to their complexity. As one manager said, ‘Grapes are a very old culture, so everyone knows how to produce them. There are not many new ways of mechanising the process.’12 Nevertheless there are two areas where farms are introducing labour-saving/displacing methods. Table 4 shows how a number of tasks are being mechanized or rationalized. For example, the practice of workers walking up and down vineyards spraying vines in order to eliminate pests is less common than previously. Instead fly-nets are placed between vines to trap flies, thereby reducing labour requirements. Reducing levels of pesticide application also helps meet production process standards now increasingly required by northern retailers. Displacing Labour through New Methods of Vine Cultivation A second method of displacing labour is to alter the methods of vine cultivation. The most common trellis system used is where vines are directed along the trellis about 2 metres above the ground. At the time of research a new ‘Y’ trellis system was being introduced on exporting farms. Increased productivity is derived from eliminating various difficulties workers experienced in the previous overhead system. Previously they had to look upwards continually, often through the vines and towards the sun, reducing their quality of vision, often leading to eye damage, and resulting in neck and back muscle injuries. Because bunches were above their heads workers only saw them from the bottom up, necessitating more work in the packing house to get them into the shape required by retailers. In tasks such as bunch pruning, shorter workers often stood on metal stools to reach bunches and had to move them from plant to plant, in the process wasting time. The stools were not very stable because of uneven ground and often caused accidents when workers fell off them, or led to reduced levels of concentration when standing on them due to fear of falling. The new trellis is about 180 cm high. Grape bunches grow down the side of the trellis rather than overhead. Bunches are never above shoulder height, making the use of stools unnecessary. 12 Interview with manager, Special Fruit, June 2003. © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 541 Ben Selwyn Table 4. Labour saving technologies Operation Old system New system How labour is saved Agro-toxin application General application Application according to integrated fruit production methods Eliminating pests Agro-toxin application Fly nets Applying fertilizers Manual chemical fertilizer application Ferti-irrigation Cutting grass Manually Mechanically Soil preparation Manual digging of 1.5 m holes to apply fertilizers (deep fertilizing) Mechanically Monitoring of pests/ diseases allows a targeted application of agro-toxins: reduces quantity of agro-toxins and labour necessary Fly nets trap pests, reduce their numbers and allow managers to identify which pests are present to target them Fertilizers applied via drip irrigation: reduces labour, increases specification of quantities and timing of application Grass cut by mower attached to tractor The 1.5 m holes are dug by machines saving time and labour Source: Interviews with farm managers and author’s observations. Workers have a better sight of grape bunches, can concentrate on pruning and move from bunch to bunch more quickly. One manager estimated that this system (Table 5) has reduced the farm’s labour requirements from roughly 5.5 to 4 workers per hectare. The previous analysis shows how increasing northern retailer power manifests itself through ever-stricter product and process demands, which in turn push farms to increase the complexity of production methods whilst simultaneously rationalizing production. For farms to respond successfully they require increasingly skilled workers. Managers are keen to stress how workers must not only be able to carry out various tasks well, but must be able to understand how the tasks they perform fit into the overall process of grape production. Greater buyer demands, more complex production processes and farms’ demand for an increasingly skilled workforce enhance workers’ structural power, because even short work stoppages can disrupt the entire harvest calendar and prevent farms producing export standard grapes. However, this structural power of workers in the labour process is realized through associational power, via trade union © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 542 Table 5. Daily (task target) productivity rates according to trellis type per worker Operation Traditional trellis Y-shaped trellis Cutting back shoots Major pruning Tying back stems Pruning 100 –110 (vines) 80 (vines) 80 (vines) 500 (bunches) 150–160 (vines) 150 (vines) 110–120 (vines) 650 (bunches) Source: Interviews with managers on Vale das Uvas and Brasil Uvas. organization and representation which gives them notable bargaining power as is shown next. GAINS MADE BY LABOUR13 The main trade union representing workers in the SF grape branch is the STR or Sindicato de Trabalhadores Rurais, originally formed in 1963 to represent small-scale farmers in the SF region. The 1964 military coup and subsequent 21 years of military rule not only disabled much rural trade union activity until the abertura (political opening) in the early 1980s (Cammack 1991), but also influenced the content of trade union activity. Prior to the coup, the STR and other rural trade unions’ main objectives had been distributive land reform to benefit small-scale family farmers. An important objective of the military government was to remove the question of agrarian reform from the political agenda, thus limiting possibilities of rural-based revolts and even revolution, as the 1959 Cuban revolution loomed large in the thinking of the military’s leadership (Cammack 1991). The military coup, whilst effectively killing off the question of agrarian reform for a generation, contributed to a shift in rural trade union activities towards wage and welfare issues (Perreira 1997, 1999). In the early 1970s the government established limited but significant dental, medical and retirement funds for rural workers and family farmers to be organized and distributed by rural trade unions, thus creating an important material incentive for their expansion. Brazilian rural trade unions increased from 266 to 2,144 between the end of 1963 and 1980 (Perreira 1999, 10). The issue of agrarian reform has been taken up again since the political opening, only this time by the Movimento dos Trabalhadores Rurais Sem-Terra (Landless Workers’ Movement, or MST). Rural trade unions in the Brazilian North East tend to focus on wages, work conditions and welfare issues. From the 1970s onwards, therefore, the SF valley’s STR activities centred on providing welfare to dryland family farmers. By the early 1980s however, and with the expansion of the irrigation districts, the majority of the STR membership 13 See also Damiani (1999). © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Labour Process and Workers’ Bargaining Power 543 Ben Selwyn and leadership comprised colonos based on them. Its main activities thus shifted towards securing land on irrigation districts both for farmers attempting to resettle in the region, who had been provided with small amounts of compensation after their displacement by the state during its construction of the irrigation projects from the late 1960s onwards (see Collins and Krippner 1999), and assisting colonos already settled on the irrigation districts. The STR did this through helping them with their applications for land (many of them were illiterate), with gaining access to credit and with negotiating lower water fees and more flexible payment times with CODEVASF. The STR’s activities then shifted again with the continued expansion of fruiticulture production, the emergence of an export agriculture sector employing significant numbers of workers on permanent contracts, and the political opening. These processes contributed to the STR reorienting its principal activities towards improving the wages and work conditions of workers in the irrigated agriculture sector. The shift was greatly facilitated, and to an extent guided by the Pernambuco Rural Workers Federation, FETAPE (Federação de Trabalhadores Agricolas de Pernambuco). Not only did FETAPE assist in the general strategic shift towards supporting rural workers, it also played an important part in formulating the STR’s tactical approach of focusing specifically on larger export farms (see below). In assisting the STR in this way, FETAPE brought its own experience of organizing workers in struggles for higher wages and improved conditions in Pernambuco’s coastal sugar zone. Damiani notes: FETAPE had become one of the strongest rural unions in Brazil since the late 1970s, having a large membership, a capable leadership, and experience with organising wage workers and negotiating with tough employers . . . FETAPE had organised the first strike in rural Brazil in 1978 (i.e. during the repressive conditions of the military government), [and] it had negotiated with growers agreements that increased wages. (1999, 185–6) At the time of research the STR and FETAPE were both affiliated to the Workers Party PT (Partido dos Trabalhadores). FETAPE sent representatives and lawyers to the SF region to assist the STR in organizing the expanding rural wage labour force. During this period pay and conditions of work in fruiticulture were determined on an ad-hoc, farm by farm basis. FETAPE representatives suggested the STR aim to introduce collective bargaining to the SF valley based on a framework established in the sugar zone. They also assisted the STR in organizing regular broadcasts by the local churchbased radio station to provide information about workers’ conditions and rights in the valley. The STR cooperated with the local branch of the Ministry of Labour and began holding meetings among SF workers. A former STR president explained: Throughout 1992 and 1993 we held many meetings with rural workers explaining why we wanted collective bargaining and how we wanted to win a convensão colletiva [collective agreement] for our members. In January © 2007 The Author. Journal compilation © 2007 Blackwell Publishing Ltd, Henry Bernstein and Terence J. Byres. Journal of Agrarian Change, Vol. 7 No. 4, October 2007, pp. 526–553. 14710366, 2007, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1471-0366.2007.00155.x by Bibliothek Universität St. Gallen, Wiley Online Library on [29/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 544 1994 we convened a mass meeting with 500 workers, and later that month we were able to get the bosses [organised within VALEXPORT] to sit down with us and discuss the contents of the agreement. In February 1994 we were able to sign the first collective agreement.14 The gains made by rural trade unions in the SF valley are notable when compared to the conditions presently faced by workers across the rest of the fruiticulture sector, and to conditions on export grape farms before the implementation of the collective agreement between trade unions and the employers’ organization VALEXPORT (see below). Between the establishment of the first irrigation proje