Messy hectares: questions about the epistemology of land grabbing data

Abstract

Recent research on land deals reports gigantic quantities of hectares seized, with relatively little regard for the solidity of the evidence or for considerations of scale other than area. This commentary questions the usefulness of aggregating data of uneven quality and transforming it into ‘facts’. Making claims on the basis of problematic evidence does not serve agrarian and human rights activists well, since it may undercut their legitimacy and make it difficult for them to identify their adversaries. Studying land tenure and corporate ownership is inherently complicated, with intractable legibility problems. Social scientists must subject their sources to critical scrutiny and understand the contexts of their production, preservation and dissemination. An accelerated process of dispossession is clearly in motion, but countering it effectively requires precise and accurate data, which are difficult to obtain. Oversimplified claims may not only undermine efforts to counter specific cases of land grabbing – and claims about land grabbing more generally – but may also divert attention from less publicized cases and from the actors behind the land grabbing. They also tend to reduce land grabbing to a quantitative problem rather than focusing on the social relations that it may or may not transform.

This is a much revised and expanded version of a paper presented at the Plenary Session Roundtable on Methodologies: Identifying, Counting and Understanding, International Academic Conference on Global Land Grabbing II, organized by the Land Deals Politics Initiative (LDPI), Cornell University, 17–19 October 2012. I am grateful to Jun Borras, Andrés León, Carlos Oya, Katherine Verdery and this journal's reviewers for comments that contributed to sharpening an earlier draft.

Keywords:

• land grabbing
• large-scale land acquisitions
• farm ownership
• foreign investment
• land tenure
• Latin America
• Africa

Notes

This is a much revised and expanded version of a paper presented at the Plenary Session Roundtable on Methodologies: Identifying, Counting and Understanding, International Academic Conference on Global Land Grabbing II, organized by the Land Deals Politics Initiative (LDPI), Cornell University, 17–19 October 2012. I am grateful to Jun Borras, Andrés León, Carlos Oya, Katherine Verdery and this journal's reviewers for comments that contributed to sharpening an earlier draft.

¹The original Polish title of this important work, Miary i ludzie, might more accurately be rendered in English as ‘Measures and people’, though this non-gendered translation would lack the catchy alliteration.

²GRAIN [Genetic Resources Action International] and the Oakland Institute are both small NGOs, highly critical of land grabbing and of claims about ‘win-win’ scenarios (on the Oakland Institute's approach, see Daniel and Mittal 2009). In contrast, the Land Matrix launched its database in April 2012 during the World Bank's Annual Conference on Land and Poverty. The Bank has historically argued for market-based ‘win-win’ land projects that provide returns to investors and income streams for local populations (World Bank 2007, 92). The Land Matrix effort was initially a joint project of the International Land Coalition (ILC), the Center for Development and Environment (CDE, Switzerland), the German Institute of Global and Area Studies (GIGA), the German Society for International Cooperation (GIZ) and the Center for International Cooperation in Agronomic Research for Development (CIRAD, France). It has since grown to include many other funders and ‘partner’ organizations. See Land Matrix (2012).

³The Economist article acknowledges that ‘It would be wrong to draw a line between these numbers so as to conclude that land deals have grown fourfold. Since most are secret, knowing what to count is difficult, and the figures refer to different periods’ (2011). The article errs in describing the International Land Coalition as a non-governmental organization. It is, in fact, an alliance that includes international financial institutions, such as the World Bank and the International Fund for Agricultural Development (a United Nations agency); intergovernmental institutions, such as the UN Food and Agriculture Organization; donor NGOs, such as Oxfam; and various advocacy NGOs.

⁴What, we might ask, is an additional three million hectares when the scale is this big or when it's not clear what's actually being aggregated? The Oxfam report appears to be seeking the largest number of hectares without any explicit methodological criteria or justification. The Land Matrix ‘database documented 1006 deals since the year 2000 amounting to 70,217,083 ha of land which equals the size of half of Western Europe’ (italics added), i.e. one-half of Oxfam's estimate if Western Europe is the point of comparison and a bit less than one-third if the hectares themselves are compared. See ILC (2012). Cotula (2012, 653) points out that the Land Matrix estimate (on which the Oxfam claim is based) isn't really dramatically greater than other estimates, since only 67 million hectares of the 227 million have been ‘verified’ and the database includes mining-, timber- and tourism-related acquisitions.

⁵Pearce (2012), writing for a mainly US and British audience, usually reports areas in acres. Since one hectare equals 2.47 acres, this has the subtle effect of citing much larger numbers.

⁶As of early 2013, the full implementation of the ProSavana plan was still under discussion and far from certain. To gloss it simply as a single export-oriented ‘venture’ farming 54,000 square miles is a gross oversimplification of a highly complex project that includes research, extension and a regional master plan (see Chichava et al. 2013). ProSavana may end up facilitating land grabbing, even though its intention is to incorporate small producers. But the pathways through which any eventual land grabbing occurs there are certain to be considerably more varied and complicated than is suggested in most of the literature on this case. For contrasting views, see UNAC (2012) and GRAIN (2012), on the one hand, and O Pais (2012), on the other.

⁷They also suggest that ‘0.7–1.75% of the world's agricultural land’ has been grabbed (Rulli et al. 2013, 892, italics added).

⁸For more nuanced views of the FAO data, see Borras, Franco, et al. (2012) and Borras, Kay et al. (2012).

⁹The World Bank concedes that ‘both media coverage and postings by users are likely to impart an upward bias’ (Deininger and Byerlee 2011, 50).

¹⁰In some cases, the grabbers themselves are not even sure of how much land they have grabbed. Pearce quotes one would-be investor in South Sudan as saying ‘The size of the land leased to us came from the Sudanese side. It wasn't a scientific figure, not well defined. For me it's not consequential whether it is really 600,000 hectares or 200,000 hectares. We can renegotiate if necessary’ (2012, 43, see also 95).

¹¹Pearce (2013) provides a succinct critique of Rulli et al.'s data on land and water grabs.

¹²This is true not just for land tenure data. The national accounts statistics used by the International Monetary Fund (IMF) – widely considered ‘objective’ and reliable – are developed through a remarkably contentious process of give-and-take between IMF missions and national governments (Harper 2000, Jerven 2013). Overzealous advocacy presents a series of problems beyond the contentious processes that may lie behind the generation of national accounts and similar statistics. Chapin, for example, in a hard-hitting review of several environmental advocates' works on deforestation in Central America, catalogued numerous instances of authors' deploying for their shock value mistaken, distorted, irrelevant or ‘strange’ data (Chapin 1995).

¹³Bräutigam, among many others, describes various cases of problematical data being recycled. For example, ‘Loro Horta's reports that Chinese interests were seeking or had acquired large land leases in Mozambique turned out to be false, but nonetheless appeared as hard data in land grabbing reports by reputable scholars and institutions, including IIED [International Institute for Environment and Development], FAO, IFAD [International Fund for Agricultural Development], IFPRI and GRAIN’ (2013, 97). World Bank researchers in Pakistan reportedly made ‘[f]ield trips to cross-check the projects cited in media reports cataloged on the GRAIN blog. In none of these cases could evidence of any investments be found’ (Deininger and Byerlee 2011, 59).

¹⁴For those whose memories don't stretch back to the early years of computing, ‘Garbage In, Garbage Out’ or ‘Shit In, Shit Out.’

¹⁵Verdery's book, The vanishing hectare, is a compelling account of on-the-ground land tenure messiness in post-socialist Romania. She also reports that smallholders sometimes expressed confusion about whether they owned an area in hectares or in ‘yokes’, an older local measure (2003, 152).

¹⁶According to Kula (1986, 29), ‘In Europe, from the early Middle Ages until the introduction of the metric system, there were two types of measures for cultivable areas: those derived from the labor-time for plowing and those derived from the amount of seed required’. Regarding the practice of using a volume of seed to measure surface area, Kula points out that seed tends to be sown thickly on good soils and sparsely on poor ones. This observation provides a pertinent reminder for land grab analysts who fetishize hectares. ‘[I]f we appraise the matter in terms of the economic value of a piece of arable land,’ Kula declares, ‘the measuring by the amount of seed had considerable merit because the value of one hectare may be far from equal to that of another, though the two are identical in area. The seed measure would offset the differences, and two plots of unequal area might thereby be “equated”, that is, shown to have virtually the same productive potential’ (1986, 31).

¹⁷Referring to today's land rush in the Democratic Republic of the Congo, a World Bank report notes ‘[m]ultiple concessions for grants up to 1000 ha to the same investor to circumvent [the] national concession approval process’ (Deininger and Byerlee 2011, 58).

¹⁸Overlapping titles and cadastral plans continue to be a problem, for example in Paraguay (Hetherington 2012, 4). In the Brazilian Amazon, a lack of correspondence between the actual area of properties and the area specified in titles is one basis of the contemporary phenomenon of grilagem, i.e. the acquisition of properties via the registration of defective titles. In 2009, judicial authorities cancelled one title for 410 million hectares in Vitória do Xingu in Pará state, in part because its total area was three times larger than the entire state (Brito and Barreto 2011, 4).

¹⁹World Bank economists acknowledge the biases resulting from the ‘strategic use of press reports by some types of investors’ (Deininger and Byerlee 2011, 53).

²⁰The study also documented numerous land acquisitions in the 1000- to 10,000-hectare range by Costa Rican, French, US, Chinese and Italian landowners.

²¹Assertions of terra nullius – that particular lands do not belong to or are not used by anybody – have a history that predates by well over a century their current use by states that facilitate land grabbing. In Argentina's nineteenth-century ‘wars of the desert’, for example, ‘desierto’ did not refer to arid lands (some were dense humid forests), but rather to spaces that elites regarded as ‘empty geographies with enormous yet dormant economic potential defined by their absence of civilization, market relations and state presence’ (Gordillo 2004, 46–48).

²²Nor is it intended to be a single farm, as the more alarmist accounts (GRAIN 2012) repeatedly suggest (see note 6 above).

²³‘Indicative’ evidence nonetheless raises other methodological concerns. As Carlos Oya (2013) points out in this forum, there is an additional danger if problematical databases become the ‘sampling frames’ that researchers employ to identify cases for in-depth analysis. The biases in the sampling frame (e.g. the Land Matrix or GRAIN databases) will generate a biased body of case studies that will tend to focus on some cases and not on others, even though these might be at least as relevant as and possibly quite different from the ‘known’ cases.

²⁴Moreover, such analyses almost always place much more emphasis on the buyers than on the sellers.

²⁵And here is a final irony regarding sources. This maxim – widely cited in various forms – was ascribed to Twain by Rudyard Kipling (1907, 2: 282). While Kipling – the archetypal novelist of British colonialism – professed unbounded admiration for Twain – the US writer who often put his famously acid wit at the service of his anti-racist and anti-imperialist politics – the quotation should obviously be taken with at least one grain of salt.

Additional information

Marc Edelman is professor of anthropology at Hunter College and the Graduate Center of the City University of New York. His books include The logic of the latifundio (1992), Peasants against globalization (1999), The anthropology of development and globalization (co-edited 2005), Social democracy in the global periphery (co-authored 2007), and Transnational agrarian movements confronting globalization (co-edited 2008). He is currently involved in research on efforts to have the United Nations approve a declaration on the rights of peasants. Anthropology Department, Hunter College-CUNY, 695 Park Ave, New York, 10065 United States. Email: [email protected]