Analysis of Structural Change Point in Multivariate Regression Models: An Application on IT Productivity Impact

Abstract

In many fields of empirical science, theories have been proposed arguing that behavioral relationships change over time. There are many possible ways in which parameters of regression models are assumed to change. With better understanding of how parameters change, estimation and testing procedures for various structural changes can be developed. This paper uses an exploratory way by augmenting the multivariate regression framework with data mining to study the less understood structural change concept by examining the effects of IT investments on industry profiles. It explores the shifting of industry boundaries over the past 17 years (1990-2006) in Australia.