One thing that is missing in the contributions I saw thus far, are definitions of large farms or small farms. 25 years ago I was in Australia and found interesting statistics for Australian conditions of course. To gain an income from farming equal to average national income a beef farmer in the North needed 100 km² of land or 1000 head of beef cattle. A sugar cane farmer needed 50 ha and a farmer growing green pepper 1 ha . Things have surely changed in the mean time and Africa is not Australia, yet looking at income potential instead of at size might be a useful way of looking at farming.
Small farms (and smallholding) do have a very important role as safety net in times of crisis – and the size can range from little more than an allotment garden to a few hectares. This safety function in times of crisis was in the past very important in Europe e.g. during and after WW II, and was also important in Eastern Europe after the collapse of the communism (and the economy). And smallholder farming received very little research and development support, because the farm models that research worked for was not a small farm. For outsiders it is tedious to try to understand small holder farming, and it is also tedious to re-orient research (which holds far fewer benefits for agro-industrial companies than do large farms) to small holder farming. Definitely smallholder farms are not simply large farms at a scale of 1:100 or so, and therefore large farms ma not be able to provide many services for small holders. We also should look a small holder farms from another angle: in many African countries 60 or more %of the population still rely to a larger or smaller degree on agriculture for their livelihood. Definitely there is nothing romantic about it – smallholder farming can be very hard work, indeed. However, at present it is needed also for social security. What is needed as agricultural revolution in Africa is research and development that is pro smallholders that also takes into account the dynamics of smallholder farming, and this also has to take in institutional and regulatory systems, and at times even concepts such as varieties in plants and breeds in animals, because they may mean something different in smallholder agriculture than for large, so-called commercial farms.
Wolfgang Bayer, Technical Centre for Agricultural and Rural Cooperation