‘Problems in the Empirical Analysis of Agrarian Differentiation Processes’: A Reflection



The ideas in this book chapter published in 1989 began to form themselves in the 1970s. Under the guidance of cultural materialist Marvin Harris, I had written a mainly quantitative dissertation based on detailed time-budget analysis of small-peasant and landless households in a Javanese village. From 1975 to 1980, I was back in Indonesia working for the (Rockefeller-funded) Agricultural Development Council (ADC) and the Agro-Economic Survey (AES) of Indonesia, at the height of Indonesia’s ‘Green Revolution’. The young AES researchers, returning each year to the same sample villages, were concerned at the growing rural inequalities which they saw and documented in this moment of accelerated rural commodification (see for an example White and Wiradi 1989), but they had little conceptual handles to help analyse and explain them. Indonesian agricultural economics teaching – and the entire ADC staff, barring myself – were rooted in the North American neo-classical ‘farm economics’ tradition. Suharto’s ‘New Order’ regime (1966-1998) had banned the teaching and dissemination of Marxist-Leninist ideas, making impossible any reference to earlier class analyses of the Indonesian peasantry, or more general political-economy ‘differentiation of the peasantry’ frameworks and arguments. The most widely available texts on rural development were the ADC’s Getting Agriculture Moving and Geertz’ Agricultural Involution, both published in Indonesian language in 1976, presenting a vision of homogeneous, poverty-sharing peasants which clearly did not fit the realities.

I had read Chayanov’s Theory of Peasant Economy in the early 1970s, and Lenin’s ‘The Differentiation of the Peasantry’ some years later.  Emerging in 1980 from these years of field-based research into the reflective environment of the International Institute of Social Studies – my base for the next 30 years – I could not understand why Leninist ‘class-based’ and Chayanovian ‘neo-populist’ approaches were so often held up against each other in fierce polemics.

In 1983 I joined Gillian Hart and others in the first of a series of three workshops which eventually resulted in the 1989 book Agrarian Transformations: Local Processes and the State in Southeast Asia, edited by Gillian Hart, Andrew Turton and myself, in which my chapter appeared.  My methodological reflections on ‘differentiation’ were originally part of a long draft chapter on Java, later split into three chapters in the book (see references for: White 1989, Hüsken and White 1989, White and Wiradi 1989). This was the first time that I tried to suggest that class-based and Chayanovian models were more complementary than generally assumed. We need both the Leninist tradition to explain the forces and tendencies which set in motion processes of agrarian class differentiation, and the Chayanovian for the co-existing forces and tendencies which explain why in so many parts of the world, class differentiation has not culminated in class polarization, and large masses of smallholder petty commodity producers co-exist with (and often outnumber) the ‘opposing’ classes of emerging capitalist farmers and landless workers.

Careful and open-minded research in differentiated agrarian communities demonstrates the need to approach ‘differentiation’ with a non-reductionist, flexible perspective, as Marx wanted to do in his later years. Lenin (1976) also advocated this – ‘the main trends of peasant differentiation are one thing; the forms it assumes, depending on the different local conditions, are another’. However, later agrarian scholars and activists may have forgotten this.  Anyone setting out for field research in communities where smallholder farming dominates, and inspired by only class-based frameworks (as recently and brilliantly expounded by Bernstein) or only the Chayanovian tradition (as recently and brilliantly expounded by van der Ploeg) will miss half the story.

Empirical work on rural differentiation processes needs the lenses of both the Leninist and Chayanovian traditions to understand the simultaneous, co-existing ‘centrifugal’ and ‘centripetal’ forces at work within peasant communities, and to do the challenging work of disentangling them in empirical research, as in the pioneering work of Teodor Shanin and Carmen Deere and Alain de Janvry. Many studies of ‘agrarian differentiation’ go no further than presenting simple tables showing the size distribution of land ownership and/or operated farm size. Even if time-series data are available allowing comparison with earlier periods, this is only the beginning. Among the ‘landless’ group, for example, we may find young members of the village elite who have formed their own households but not yet inherited land, older elite villagers who have already passed land on to their children, village schoolteachers, nurses or police officers who may or may not own land in their own villages of origin. None of these would be correctly included as members of the landless class. Where, in class terms, should we place households where the husband is a farm labourer and the wife a small farmer who herself employs hired workers at planting and harvesting time?  Or a young adult couple who are landless share tenants on their parents’ land? Complicating the picture further, pluriactivity and plurilocality – the multi-directional movement of household members between sectors and places – means that the link between land control and rural class position becomes increasingly blurred, although certainly not absent.

Time has only strengthened my views on these methodological issues. I re-stated them most recently in the Journal of Peasant Studies theme issue celebrating Marx’s 200th birthday, and in a small book on Agriculture and the Generation ProblemNo analysis of rural differentiation processes is complete without an attempt to explore and explain the intersection of class, gender and generational differences and relationships, and the links between agrarian and non-agrarian bases of differentiation.



Hüsken, Frans and Ben White. (1989). ‘Java: social differentiation, food production, and agrarian control’. In G. Hart, A. Turton and B. White (Eds.) Agrarian Transformations: Local Processes and the State in Southeast Asia (pp. 235-265). Berkeley: University of California Press.

White, Ben. (1976). Production and Reproduction in a Javanese Village. Ph.D. dissertation, Columbia University.

White, Ben. (1989). ‘Problems in the empirical analysis of agrarian differentiation’. In G. Hart, A. Turton and B. White (Eds.) Agrarian Transformations: Local Processes and the State in Southeast Asia (pp. 15-30). Berkeley: University of California Press.

White, Ben and Gunawan Wiradi. (1989). ‘Agrarian and non-agrarian bases of inequality in nine Javanese villages’. In G. Hart, A. Turton and B. White (Eds.) Agrarian Transformations: Local Processes and the State in Southeast Asia (pp. 266-302). Berkeley: University of California Press.

1 Comment

  1. Çağrı İdiman on May 31, 2020 at 3:03 am

    Dear Professor White,

    I find your argument regarding the complementarity of Chayanovian and Leninist models an important contribution. Would you think that the union of these models can be achieved on the basis of ‘labor theory of value’? (Given the commodification of global agriculture, this might bring us closer to a Kautskian model in the Agrarfrage, than the Leninist Model in the Development of Capitalism in Russia.) Stated differently, can the internal and external relations and dynamics of peasant petty commodity production be explained on the basis of value relations? Would such an analysis be fruitful? I realize value analysis is usually employed on a macro scale i.e. for sectoral analysis, rather than the micro scale i.e the individual peasant farm, and furthermore, the production units have become much more complex and diversified – as you point out later in the text. I would really appreciate your thoughts. Thank you very much.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.