Essays on some measurement issues in welfare economics

In applied economics, the correct measurement of policy targeted variables is of utmost importance. In this dissertation, we address some measurement problems that arise owing to partial information.; First, we deal with the issue of using small changes in national income at constant prices to measure the change in the level of welfare of a country in the absence of information regarding the preference orderings of the individuals. We show that the national income criterion for assessing changes in social welfare may be inherently flawed.; Next, we deal with measurement issues that arise from using direct indicators such as calorie intake, housing conditions etc., to calculate the level of welfare in a society. Often in this multi-dimensional framework we may know the distribution of achievements in terms of separate attributes such as housing, health etc. in a society, but we may not know, for any given individual, the achievements in terms of the different attributes. Under this circumstance, instead of the standard procedure of first aggregating over the attributes for each individual and then aggregating over all the individuals, we are forced first to aggregate over the individuals separately for each attribute and then to aggregate over all the attributes to assess the overall social deprivation (or standard of living). We focus on this second procedure. We impose mild regularity conditions on the aggregation rules used at different stages and also a consistency condition that requires the social deprivation index reached by the procedure based on aggregate data to be exactly the same as the intuitively satisfactory procedure that is typically used with full information. We show that all these requirements can be met only if aggregation rules involved are linear. This is unfortunate since it rules out more ethically sensible aggregation rules.; Last, we undertake an empirical exercise to calculate the housing deprivation in village in India in a multi-dimensional framework. We consider housing deprivation over 15 different attributes. The qualitative information has been systematically quantified using the Borda rule. Here we have contributed empirically to the multi-dimensional analysis of housing deprivation in India.