Economic Aspects of Tractor and Animal Use in Luwu, South Sulawesi, Indonesia: A Quadratic Risk Programming Approach

Abstract

Quadratic risk programming was used in this study to evaluate farmers' choices between tractors and draught animals for the preparation of rice land. A case study approach was adopted, with separate models being developed for each of six farm-household units in three categories of tractor and animal use in one Indonesian village. ... Tractor use was generally inferior to animal use (in terms of utility) except when tractorised rice yields were assumed to be higher than non-tractorised rice yields. An expected yield increase of slightly more than ten per cent was necessary for fully tractorised land preparation to be optimal on all sample farms. This finding corresponded closely with the actual tractor use patterns of the case-study farmers. The three farmers who hired a tractor for land preparation in the wet season of 1983 all expected their mean rice yields to be considerably higher than if they were to use draught animals instead. The expected increase ranged from 13 to 25 per cent. On the other hand, the three farmers who did not use a tractor in 1983 all expected their mean yields to be the same with or without tractor use. These results suggest that the farmers' expectations concerning the effects of tractor use are an important determinant of their choice of land preparation technique. Since there is no clear evidence that tractor use in Luwu actually increases rice yields, it seems possible that non-optimal decisions are being mistakenly encouraged.