Title

A Choice of an Index under Inefficiency: Tornqvist or Fisher Index? Evidence from Simulation

Campus

Oconee

Publication date

3-10-2021

Publisher

EconPapers, hosted by Orebro University School of Business

Book or Journal Information

Economics Bulletin, 41(1), 41-47

Keywords

Index numbers, Fisher Index, Tornqvist Index, Allocative inefficiency

Abstract

This study addresses a choice of quantity index when inefficiency exists. Performing simulations, we show that 1) the finding from the previous literature that the Fisher index and the Tornqvist index approximate each other is only a special case when inefficiency is negligible, 2) the Tornqvist index begins to deviate as inefficiency increases because the Tornqvist index uses the average cost shares as a weight in the calculation, and 3) the Tornqvist index eventually suffers from the Simpson's paradox, in which the average of two higher numbers are lower than the average of two lower numbers, when the factor substitutability is large. From these results, we conclude that the Fisher index is preferred to the Tornqvist index when any inefficient use of factors is suspected.

Author Biography

Dr. Han has been teaching at the University of North Georgia for the last several years and brings application and context to the classroom from his experience as an Economist at the Bank of Korea, the central bank in Korea. He earned his Ph.D. in Economics from University of Georgia. The Munford Professorship was awarded to Dr. Han from UNG in 2014.

Link to Full Text

Share

COinS
 

A Choice of an Index under Inefficiency: Tornqvist or Fisher Index? Evidence from Simulation

This study addresses a choice of quantity index when inefficiency exists. Performing simulations, we show that 1) the finding from the previous literature that the Fisher index and the Tornqvist index approximate each other is only a special case when inefficiency is negligible, 2) the Tornqvist index begins to deviate as inefficiency increases because the Tornqvist index uses the average cost shares as a weight in the calculation, and 3) the Tornqvist index eventually suffers from the Simpson's paradox, in which the average of two higher numbers are lower than the average of two lower numbers, when the factor substitutability is large. From these results, we conclude that the Fisher index is preferred to the Tornqvist index when any inefficient use of factors is suspected.