Premium
Latent variable modelling of price‐change in 295 manufacturing industries
Author(s) -
Georganta Zoe
Publication year - 2003
Publication title -
applied stochastic models in business and industry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.413
H-Index - 40
eISSN - 1526-4025
pISSN - 1524-1904
DOI - 10.1002/asmb.486
Subject(s) - econometrics , latent variable , variable (mathematics) , economics , regression analysis , latent variable model , econometric model , regression , structural equation modeling , manufacturing , contrast (vision) , computer science , statistics , mathematics , artificial intelligence , mathematical analysis , political science , law
Abstract In contrast to traditional regression analysis, latent variable modelling (LVM) can explicitly differentiate between measurement errors and other random disturbances in the specification and estimation of econometric models. This paper argues that LVM could be a promising approach to test economic theories because applied research in business and economics is based on statistical information, which is frequently inaccurately measured. Considering the theory of industry‐price determination, where the price variables involved are known to include a large measurement error, a latent variable, structural‐equations model is constructed and applied to data on 7381 product categories classified into 295 manufacturing industries of the USA economy. The obtained estimates, compared and evaluated against a traditional regression model fitted to the same data, show the advantages of the LVM analytical framework, which could lead a long drawn‐out conflict between empirical results and theory to a satisfactory reconciliation. Copyright © 2003 John Wiley & Sons, Ltd.