Purchasing power parity exchange rates from household survey data: India and Indonesia

Purchasing power parity (PPP) exchange rates are extensively used by researchers and by policymakers. This paper proposes and implements a new methodology for calculating PPPs using information on unit values from household surveys. Although unit values are not identical to prices, they have compensating advantages. Large household surveys contain several million unit values, they are tied to actual transactions, and they are naturally linked to household characteristics such as income. In consequence, it is possible to calculate PPPs for different social groups, including PPPs for the poor. The paper calculates multilateral price indexes for the states and sectors of India, as well as PPPs for rural and urban Indonesia together with rural and urban India. PPPs for the poor are distinguished from general PPPs. The internal PPPs for India are not very different from previous estimates based on bilateral comparisons, but the estimated PPP between India and Indonesia is very different from the numbers calculated by either the Penn World Table or the World Bank. It implies that either India is much better-off, or Indonesia much poorer (or both) than is generally supposed.