Asia concentrates 60 per cent of the world's population on slightly more than 20 per cent of the world's land area, thus available arable land is a major constraint. Farmers in Asian countries have access to arable lands that are 100 times smaller than lands in developed countries. With the exception of Thailand (0.5 ha) the amount of available arable land per farmer does not exceed 0.3 hectares in all the selected Asian countries, and not even 0.2 hectares in six countries.

A 2010 projection shows that the arable land ratio per farmer is not going to significantly change for Asian countries, while it is likely to further increase in developed countries, mainly due to decreases in the agricultural population. The gap in arable land availability between the most developed countries and countries with the most numerous and poorest agricultural populations is widening.

However, this data does not fully reflect the real conditions of the poor farmers. It assumes that land access is evenly distributed among agricultural households, however this is just not true. In Nepal, 44 per cent of the agricultural households operate 14 per cent of the total agricultural land area, while the top 5 per cent occupy 27 per cent. The concentration index for agricultural land is 0.54 reflecting a highly uneven distribution of farm land (Sharma, 2000). In Bangladesh, a holding size distribution from the 1996-1997 census shows that small farms increased in number by up to 83 per cent and operated 23 per cent of farmland, against respectively 75 per cent and 15 per cent in 1984 (Absan and Ahmed, 2000). In India, there are more than 105 million agricultural laborers today compared to barely 27 million in 1951. Within fifty years, the ratio of agricultural laborers to cultivators increased from 2/5 to 4/5. (Department of Agriculture and Cooperation, 2004).

Literature often mentions an inverse correlation between land size and productivity (Maxwell and Wiebe, 1998; Banerjee, 1999), arguing that smaller farms make a more rational and efficient use of resources and face fewer costs such as transaction costs. If this is true, would then the observed trend mean that Asian countries' agriculture will be more competitive thanks to this shrinking land ratio?

A regression conducted for 18 of the above-mentioned countries (no data available for Myanmar) between the agricultural value added per agricultural worker over the 1998-2000 period (World Bank, 2003) and the arable land ratio per farmer for the year 2000 shows, with a 0.83 R2 at the 95 per cent significance level and a t-test value of 8.7, the opposite situation. This means that the arable land ratio explains most of the observed variation in agricultural productivity in the selected countries. As the size of arable land further shrinks, productivity gains that were highlighted by various authors disappear. Several reasons may explain this fact. As small land area is associated with poor households, it is likely that farmers cannot afford to buy the inputs needed to increase production; they have also to engage in other activities as labourers and cannot give sufficient attention to the care of their plots. Another possibility is the overexploitation of resources over time such as land and water leading to degraded soils and lack of irrigation facilities in the most fragile areas. Actually, the same regression run for 1988-1990 data shows a somehow weaker, yet still high, R2 (0.7) confirming this temporal trend.

The key question today is: How small is too small? When the number of landless households or micro households that cannot provide livelihoods for the family members increases, rural poverty increases and productivity decreases. It also draws attention to the size factor in land redistribution policies as these are commonly advocated as useful means to increase productivity

Written by Dr. Robin Bourgeois, ISDB Pogramme Leader, UNESCAP-CAPSA, Bogor, Indonesia.