The welfare analysis of rice pricing policies using household data for Indonesia

Abstract

A methodology with which to analyze the welfare distributional effects of food pricing policy reforms is proposed, critically investigated and refined using Indonesian household level data. The problems of welfare measurement dealt with include the need for behavioral consistency and the making of interpersonal comparisons of welfare when household specific circumstances vary. Building on recent advances in empirical tax reform analysis, an equivalent income function is used as the welfare indicator. This is constructed from the explicit indirect utility function underlying the functional form used in the econometric modelling of consumer demands for rice and housing on 1981 SUSENAS data for Java. Normalizing for spatial price differentials and for variations in demographic characteristics is found to result in a reduction in measured welfare inequality. The methodology is used to explore the welfare consequences of spatial domestic rice market integration and of across the board price level changes, as would be expected to result following an external trade liberalization of Indonesia's rice markets. The reforms are simulated under various assumptions concerning the producer and consumer income effects of the price changes. The thesis also investigates alternative public compensation schemes implemented through the input subsidies policy or by means of lump-sum transfers targeted at producers and/or consumers. The elimination of spatial rice price differentials is found to have modest potential social gains while exacerbating inequality. Rice price rises are assessed to have beneficial welfare distributional effects if accompanied by, for example, a scenario which implements cuts in input subsidies and redistributes the resulting fiscal benefits as lump-sum transfers to all households. Of fundamental concern throughout the thesis are questions about how alternative assumptions adopted in constructing the equivalent income function affect the ordinal ranking of households in the pre-reform situation and the social welfare orderings of the policy reforms. Sensitivity to the choice of functional form and reference prices, to the number of included goods and to the treatment of the regression errors are examined. The overall social welfare evaluations of policy reforms are found to be generally robust to such variations. Conversely, the welfare ranking of households appears to be quite sensitive.