Predicting Consumption Poverty using Non-Consumption Indicators: Experiments using Indonesian Data

Household consumption expenditure data is crucial for calculating important welfare measures such as poverty headcount rate. However, collecting such data is difficult and cumbersome. As an alternative, we experiment with three methods – consumption correlates model, poverty probability model, and wealth index principal components analysis (PCA) – to predict consumption expenditure and poverty using non-consumption indicators. The purpose is to use these alternatives for rapid monitoring and appraisal of social welfare as an early warning system. We test each method’s performance and find that the consumption correlates model is the best method to predict poverty quickly and relatively accurately. We find that education level, asset ownership, and consumption pattern are the best predictors of expenditure and poverty.     

Measuring Unemployment in Developing Countries: The Case of Indonesia

Abstract. Measuring unemployment in developing countries is not straightforward due to the presence of a large number of discouraged workers. Including them into the labor force is sometimes appropriate in order to reflect the true state of unemployment. However, the decision must be based on careful research. This study provides a case study of Indonesia, whose decision to include discouraged workers into its labor force resulted in an artificially high unemployment rates and disguised the actual post‐crisis decline in unemployment in Indonesia. The discouraged workers can be classified based on their willingness to work. If Indonesia still wants to include discouraged workers into the labor force, only those willing to work should be included.