An assessment of income levels and distribution in one lowland and one highland village in West Java, Indonesia, 1979

Abstract

Absolute poverty and great inequality are still major problems in Indonesia in spite of its increase income from oil. As Indonesia has no regular statistical series on income distribution, data from surveys the main purpose of which is not income distribution have had to be used. This fact introduces many problems in addition to the important theoretical problems of measurement and of lack of other statistical series against which to test income data. The Agro-Economic Surveys have collected much information on incomes as part of their studies of farm institutions and other rural problems. This study is an assessment of income data from such a study, and examines the special problems which arise in the use of such income data. It was found that the coverage and manner of collecting data on income had resulted in a high proportion in each sample village (48 and 26 per cent, respectively) of households which reported incomes insufficient to support life, but that these reported income levels appear to be the result of anomalies in reporting. It was therefore concluded that total household income for the period of one year was not possible to compute for these households and that therefore no reliance could be placed on income distributions based on these incompletely recorded incomes. Each separate source of income was then examine in detail and tested so far as possible for consistency against other data. To do this, data on area owned and cultivated, assets, household size, education and employment were examined. It was found that the method of collection of data by season called forth conflicting responses by the farmers, particularly for land cultivated. The fact that not all data on field conditions during interview were available in Canberra (e.g., bases for imputed prices) made it impossible to check consistency of income data against yields or prices. Sophisticated methods of analysis based on such data were considered inappropriate, so only Gini Coefficients, Lorenz Curves and some regressions have been applied to the data. A concept of 'plausible' incomes is developed, in which households which report incomes insufficient to support life are considered 'implausible' and excluded from some of the analyses. Similarly, some sources of income which showed very poor internal consistency (e.g., poultry and gleaning) were considered 'implausible'. Plausible incomes were not necessarily accurate, and could also be under-reported, however. 'Plausible' income households were compared to total households. Exclusion of implausible households made a much greater difference in the lowland than in the upland village. Inequality appeared to be far greater in the lowland than in the upland village, but it is not possible to say for sure whether this conclusion arises from what appears to be much more thorough and consistent data collection procedures in the highland village, and to what extent it reflects real levels of inequality. It was found that the higher the aggregation, the greater the similarity between villages; disaggregation revealed very wide differences. The conclusion to this study is that, because income data in the SAE study were collected as a side-line to the main purpose of the study, the very difficult conceptual, theoretical and practical problems associated with income measurement have not been met, and that consequently, the data on incomes from this study are not reliable, and not suitable as a basis for policy recommendations on income levels or income distribution.