Regression Analysis and Multicollinearity: Two Case Studies

In this chapter, we explore two applications of regression modeling: the question of regression-weighting of GNP forecasts and the issue of estimating models associated with security totals returns. We examine the forecasting of GNP by major econometric firms and the modeling of security returns as a function of well-known investment variables and strategies. We illustrate regression analysis and problems with highly correlated independent variables. We will refer to the correlation among independent variables as multicollinearity.