Presented by Kevin J. Fox (Centre for Applied Economic Research (CAER) UNSW Business School)
The increasing availability of transactions data covering prices and quantities on a wide range of products has created new opportunities for national statistical institutes (NSIs). Price changes can now be observed across many products at high frequency, potentially allowing NSIs to publish more reliable indicators of price changes. However, conventional index number theory can break down when applied to such data, leading to massive chain drift. The use of multilateral indexes is now an accepted approach for incorporating transactions data in a price index such as the CPI, the attractiveness stemming from the ability to be able to control for chain drift bias. We examine the empirical performance of different multilateral indexes and methods for extending a prices index series when new observations become available. After extensive empirical examination using the UK data, the results suggest GEKS-Törnqvist (or “CCDI”) index with the mean splice extension method over a 25-month window is the best-performing method in the scenarios tested.
Kevin J. Fox works primarily in the field of economic measurement, with a focus on productivity and prices. After studying Japanese in Tokyo for two years, he studied economics at the University of Canterbury and the University of British Columbia. He joined the School in 1994. He chaired the 16th Series CPI Review Advisory Group in 2009-2010, he is a member of the Australian Bureau of Statistics Methodology Advisory Committee, and has been a consultant for agencies such as the Australian Treasury, Reserve Bank of New Zealand, the Swiss National Bank, the Australian Bureau of Agricultural and Resource Economics, the Asian Development Bank, and the New Zealand Treasury. He is an Associate Editor of the Journal of Productivity Analysis and President of the International Association for Research in Income and Wealth.