ABSTRACT
This chapter evaluates quality of model by performing diagnostics or misspecification tests and includes another misspecification test, that of parameter instability or structural breaks. The diagnostics apply both to simple and multiple regression and also to other more complicated methodologies. The field of diagnostics testing has seen ongoing developments through devising new misspecification methods and modifications to existing tests. There are alternative versions of a given test, in general the analyst has lots of choice to test for misspecification. The chapter implements the White test to examine presence of heteroscedasticity in our Hong Kong office rent model. It studies the functional form of a model and the problem of linearity in parameters. The chapter reviews the classic Durbin Watson d-test reported almost universally in diagnostics outputs, and one of most commonly used general tests, Breusch-Godfrey test. It explains the computation of diagnostics in EViews, but the exposition of most of the tests allows readers to perform these tests in Excel.
