The Classical Linear Model
The method outlined in Chapters 2 and 3 is not the one primarily used in practice. As noted in Ch. 1, the dominant methodology is based on the stochastic approach and on a special stochastic model used in the large majority of applications, the classical linear model (CLM). In this chapter, we review the assumptions and properties of the CLM, leaving out the proofs linking the two; the latter can be found in any mathematical statistics text, notable among which for clarity and rigor are Johnston (1984), Malinvaud (1970), Neter et al. (1985), and Theil (1971). We then examine critically these properties and their utility when the assumptions are satisfied. We consider the validity of these assumptions in Ch. 5.
KeywordsExplanatory Variable Statistical Inference Unbiased Estimator Point Forecast Interval Forecast
Unable to display preview. Download preview PDF.