The Properties of Agroecosystems 133Indonesia’s rice strategy has also had dramatic effects on labour use and equi-
tability (Conway and McCauley, 1983). Prior to the introduction of the new rice
varieties the harvest was open to all members of the village, the poor and landless
being able to retain a fixed proportion of the rice they harvested. But the growing
number of landless has made it difficult for farmers to control the harvesting so
that now it is carried out by contract groups (Collier et al, 1973). The traditional
hand-milling of rice by the women of the village, who received grain in lieu of
money payment, has also disappeared, replaced by mechanical hulling (Timmer,
1973). And, despite self-sufficiency, malnutrition persists, even in some of the
most agriculturally productive villages.
The final question is whether the national effort is sustainable. Indonesia’s oil
and gas boom is now over and the budget surpluses which subsidized the agricul-
tural revolution are gone. Inevitably, there will be bad years due to pest and disease
outbreaks or poor rainfall. These will place a heavy demand on central resources
and agencies who will have to deliver what in the past was the function of com-
munal self-help arrangements and the evolved diversity of traditional agro-
ecosystems.
Agroecosystem Analysis and Development (AAD)
Until recently the various agricultural regions of the world differed from each
other in the various weights accorded the agroecosystem properties. In the 1930s,
for example, conservation agriculture was being practised in Britain while the Dust
Bowl was being created in the midwest of the US. But today agriculture is facing a
common worldwide challenge. The pursuit of high productivity in both the devel-
oped and less developed countries has brought with it declines in sustainability
and equitability that increasingly are being regarded as undesirable. Our present
priority is for policy research, practical analytical tools and development packages
aimed at increasing agricultural sustainability and rectifying undesirable inequi-
ties.
One step in this direction has been the development of Agroecosystem Analy-
sis (Conway, 1985a, 1986). This is a technique of multidisciplinary analysis which
may be used at any level in the agroecosystem hierarchy, and generates a set of
research and development priorities that explicitly take account of the trade-offs
between the system properties. The technique has been used so far to determine
research priorities for university and government teams (Gypmantasiri et al, 1980;
KKU-Ford, 1982a, 1982b; KEPAS, 1985a, 1985b), development priorities in
project design (Limpinuntana and Patanothai, 1984; Conway et al, 1985) and to
evaluate project performance and recommend corrective actions (Conway and
Sajise, 1986).
A second component of AAD (Conway, 1985b) is the further development
and refinement of a variety of packages that promise high productivity without