Authorised by Academic Registrar, April 1996
Objectives On successful completion of this subject students should have a rigorous understanding of the foundations of the classical regression model and be capable of deriving the properties of least squares estimators under classical assumptions and demonstrating the consequences of common violations of these assumptions; applying tests and appropriate estimation procedures about the relationships between variables and for violations of the underlying assumptions; applying this knowledge to estimate, analyse and forecast multiple regression models with typical economic and business data.
Synopsis An introduction to linear multiple regression methods; properties of least squares estimators; probability distributions and their applications to hypothesis testing; an introduction to the generalised least squares estimator; the problems of serial correlation, heteroscedasticity and multicollinearity.
Assessment Test (30 mins): 10% + Two 2-hour open-book tests 30% + Examination (2 hours): 60%