Premium
Lancaster's characteristics approach revisited: product selection using non‐parametric methods
Author(s) -
FernandezCastro Angel S.,
Smith Peter C.
Publication year - 2002
Publication title -
managerial and decision economics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.288
H-Index - 51
eISSN - 1099-1468
pISSN - 0143-6570
DOI - 10.1002/mde.1048
Subject(s) - data envelopment analysis , selection (genetic algorithm) , judgement , computer science , product (mathematics) , frontier , operations research , economics , parametric statistics , management science , mathematical economics , microeconomics , econometrics , mathematical optimization , mathematics , artificial intelligence , political science , geometry , statistics , law
Abstract In consumer theory, the principles of Lancaster's characteristics approach and hedonic pricing appear to offer the most promising insight into choice when qualitative aspects are important. The paper reconciles these principles with the family of non‐parametric frontier estimation methods known as data envelopment analysis. It is shown that, with some straightforward adjustments, DEA is entirely consistent with the characteristics view of consumer choice found in the economics literature. In making Lancaster's ideas operational, the paper also addresses the theoretical concern voiced by Lancaster about combining indivisible products. The principles are illustrated with a case study involving the comparison of diesel cars. The paper concludes that the user will ultimately have to apply some judgement in choosing between competing efficient products. However, the analysis should help to restrict the number of products to be assessed to manageable proportions. Copyright © 2002 John Wiley & Sons, Ltd.