Specifying and Diagnostically Testing Econometric Models, 2nd Edition (2ND)

Stokes discusses—and illustrates with output from actual problems—a number of applied econometric techniques, including OLS specification tests, recursive residual analysis, limited dependent variable models, error component models, and others. His book is clearly written and copiously illustrated with equations, with follow-up analysis to show how models are built and some of their limitations. His B34S^DTM software is available and allows readers to do further research with a large number of datasets distributed with the program. A necessary resource for applied econometrics researchers in economics, finance, and in health, energy, and labor economics. This work illustrates the use of model specification and diagnostic tests applied to a variety of econometric modeling techniques. For each technique discussed the basic mathematical models are outlined. A sample problem is discussed and estimated using the B34S^DTM Data Analysis System. The output of the program is displayed in the text and discussed. Where appropriate, output from the RATS^DTM software is displayed. Follow-up models are estimated and discussed. The examples selected are taken from a variety of sources and reflect actual applied research. Complete data are given in the text to enable the reader to use these problems with other programs and techniques. It is the author's experience that applied econometric techniques are best learned by running actual problems. Since most users experiment with a limited number of techniques, their experience is limited. This book discusses a broad range of techniques and shows how they are interrelated.^L ^L The techniques discussed include the following: simple, one-equation OLS and GLS models with continuous variables on the left-hand side, which are tested with recursive residual and BLUS residual techniques. Another class of models includes restrictions on the left-hand side variables. Models studied and illustrated with data include probit, logit, multinomial l