Some aspects of smallholder rubber production in South Sumatra, Indonesia : a production function approach

Abstract

This study attempts to highlight and quantify the major factors affecting yields on small rubber farms in Kabupaten LIOT and MURA, South Sumatra, and to indicate the adjustments which should be made to secure a better performance. The input-output data obtained from a 1977/1978 survey of smallholder rubber farms in South Sumatra is the basis of this study. An unrestricted form of the Cobb-Douglas production function is fitted to LIOT and MURA data separately. The function is estimated by two different techniques, Ordinary Least Square and Linear Programming. The function estimated by the first approach is interpreted as the average production function, and expresses the output level which an average farm can obtain from a given set of inputs. The function estimated by the second approach is interpreted as the best or frontier production function, which expresses the maximum output level that can be obtained by only the most efficient farms from the combination of factors at the existing state of technical knowledge. The major findings with regard to the production coefficients are as follows: (1) There are decreasing returns to scale in the production of Kabupaten LIOT and increasing returns to scale in the production of Kabupaten MURA. (2) For both regions, the frontier function has shifted neutrally outwards from the average function. In general for all holdings, the number of trees in tapping per hectare seems to be an important positive influence on yield. Too deep a tapping cut appears to have substantial negative effects, especially with younger trees, whilst for older trees increasing the number of cuts and maintaining the tapping panels in good condition seems to be a big advantage. The estimated production functions are used to calculate the technical efficiency of the individual farms. For each region, two different ratings of technical efficiency are derived from the average and the frontier production functions. The efficiency ratings are calculated as the ratios of observed output to the level of output estimated from the production functions. From this part of the analysis it is found that there is no significant difference in the efficiency ratings that are derived from the average or the frontier production functions. On the basis of average production function, the marginal productivities of factors of production are calculated for both the regions separately. An attempt is made to examine the factors which explain the difference between the maximum possible yield and the actual yield of both regions. The results show that in LIOT, girth is the factor responsible for reducing the extent of yield variation from the maximum possible yield, whereas depth of cut and age of trees are factors significantly widening farmer yield gap. For Kabupaten MURA, the number of cuts, condition of the tapping panel, depth of cut and education are the factors responsible for reducing the extent of yield variation from the maximum possible yield, whereas.years since tapping commenced and farm size are contributing more to the difference. Apart from their much higher rubber yields, the frontier farms in Kabupaten LIOT have bigger girthed trees and more trees in tapping. In Kabupaten MURA, they have a larger productive area surveyed, younger trees and apply more labour in maintenance. Size of farm may well have an important influence on technical efficiency and yield, but this facet must be investigated further. It is considered that many of the less technically efficient farms could be encouraged to improve their position if greater assistance could be given to them by extension services.