Sensitivity of Purchasing Power Parity Estimates to Estimation Procedures and their Effect on Living Standards Comparisons

Abstract The motivation of this exercise is to compare the Purchasing Power Parities (PPPs) calculated using different procedures and study the sensitivity of global rankings of regions based on living standards to the PPPs used. The empirical comparison involves the GEKS, weighted CPD, GK, EWGK and the True Cost of Living Index (TCLI) based PPP estimation procedures with the Indian Rupee used as the numeraire currency. The published ICP PPPs for 2011 are used as benchmark for the non ICP PPPs obtained in this study. Evidence confirming the “Gershenkeron effect,” that affects the additive GK procedure, is provided. The results suggest that the EWGK PPPs, which are also additive, do not suffer from the extent of bias of the GK PPPs. The paper also provides evidence on the large variation in the TCLI based PPPs across expenditure quintiles originating from variation in preferences between expenditure classes. This suggests departure from the current ICP practice of providing one PPP for the entire country and points to the need to estimate PPPs by different expenditure classes. The empirical evidence points to the rich potential for the rarely used TCLI in future PPP calculations.