Where has this argument come from and what are its limitations and pitfalls? The argument for agriculture-led growth of course has a long lineage, dating to the classic work of John Mellor and Bruce Johnson at Cornell and Stanford, taken up by at IFPRI and others. By emphasising expenditure linkages – especially through consumption – this work countered the argument that growth linkages are low in small-scale agriculture because of relatively low external inputs.

Using an array of techniques ranging from simple input-output and expenditure models to more complex social accounting matrices, multi-market models and village and regional CGE approaches, this research has shown that growth in agriculture does indeed generate rural non-farm growth and growth in the wider economy. Empirical studies from across Africa for example have highlighted how it is in small farm settings linked to markets that such potentials are often greatest.

For example an IFPRI study in Zimbabwe showed how in the 1990s small-scale communal farming was generating more growth than the large scale commercial sector (although some high labour input enterprises such as horticulture represented important exceptions). The policy message that emerges from this now substantial body of work is clear – increases in productivity in small-scale agriculture can result in broader economic gains to the economy, with spin-offs to the rural non-farm sector.

In time, the argument goes, this will result in a transition from a broadly subsistence based agricultural economy to one which can afford more inputs and become more commercial, specialising along the way – if directed by demand – into niche commodities and markets. As the sector’s fortunes improve, the opportunities for exit from agriculture will increase, as off-farm opportunities grow (for instance in farm labour, agro-processing, and the rural service sector). Such wider economic growth will be an economic pull – rather than the current situation of being pushed out from a failing agriculture. The end result, it is stated, will be a vibrant, fully modernised integrated economy, with a small but efficient agricultural sector continuing to generate growth and employment. Or at least that is how the standard storyline goes.

But what are the problems with this simple version, so often repeated in current policy debates? If the relentless economic logic is so powerful, why hasn’t it happened in large parts of rural Africa ? And is there really only one pathway for such a complex process?

Debates on economic growth and agriculture are huge, but some important qualifications and critiques to the standard growth narrative can be identified.

First, are the models that generate these findings sufficiently realistic? Models are of course only as good as the data and the assumptions that build them, and in the case of growth linkage models these can be open to question.

One of the most frequently used modelling approaches is the social accounting matrix (SAM). SAMs usually choose a base year and try and generate a best-estimate model of all the production, consumption, labour market and investment linkages in the economy, identifying in particular the role of agriculture (and its sub-sectors). Being based on fixed prices, where supply is perfectly elastic and responses are linear, conventional SAMs clearly cannot tell the whole story. Indeed, more complex multi-market and CGE models show how taking into account such factors is crucial.

But it is perhaps less the technical and data limitations of the models that are of concern, but the way they frame and influence the policy debate.

In Ethiopia for example, the argument for focusing new PRSP investments on ‘going for growth’, with an emphasis on boosting commercial agriculture, is based on a series of SAM modelling inputs into the policy process. Necessarily, but perhaps problematically, these simplify the debate into one that creates policy categories.

For instance, in the case of the 2001-02 Babo Gaya SAM constructed by IFPRI and associates, the agricultural economy for the nation as a whole is divided into ‘subsistence crop farming’, ‘subsistence livestock farming’ and ‘commercial (or ‘modern’) crop farming’. Showing that what is (vaguely) defined as ‘commercial’ or ‘modern’ farming generates a significant value added share and so overall growth benefits provides the ‘evidence’ for the policy proclamations now emanating from both government and donors. This confirms, in turn, a particular type of investment, and support for a particular approach to the commercialisation of agriculture.

Second, such discussion – reinforced by such models – is often wrapped up in another argument that there are somehow a defined set of ‘stages of growth’, involving singular trajectories to some desired end (usually away from a backward, subsistence form of farming towards something better, more modern and commercial). A familiar argument since Rostow is that economic development consists of ‘stages’, and that the challenge is to find the technology, institutional instruments and market incentives to push things from one stage to the next. This is central to the Mellor-style growth argument discussed above, and much current academic and policy discussion since. But such evolutionist arguments of a somehow necessary move from one stage to the next can also be questioned.

Drawing on her work in South Africa, Gillian Hart, for example, argues instead for understanding growth patterns in terms of multiple trajectories, each situated in particular spatial and historical contexts. What happens in one place and one time period, will not necessarily occur elsewhere. She critiques in particular the imposition of ‘models from elsewhere’ in our eagerness to develop blueprints for economic growth. As the structural adjustment era in Africa so harshly demonstrated, such models do not always fit, and indeed may do much harm. This is not to deny the need to generate growth and foster linkages between farm and non-farm, rural and urban. Far from it.

It is instead an argument for understanding how in particular contexts growth trajectories are different, and require different inputs, incentives and governance arrangements. Not all African farmers need to end up growing French beans and carnations for the European market, even if this is a good plan for some in some places.

There are other opportunities that need to be sought out too, and the generalised models and universalised policy prescriptions do not help in the search for these openings.

It is for this reason that the Future Agricultures Consortium work on commercialising agricultures is emphasising the diverse ways new forms of growth can be generated, and the diverse impacts this has on the agricultural economy, on wage labour and on the non-farm sector. Case studies include smallholder cotton farming in Kenya and Malawi, the horticulture/floriculture sector in Kenya and Ethiopia, and the commercialisation of small-scale staples crop production across the three countries.

Different dynamics of commercialisation are evident, each requiring different types of support, and each with different (often highly gendered) impacts on poverty reduction and rural livelihoods.