Academic journal article
By Nakamura, Alice
Journal of the American Statistical Association , Vol. 87, No. 420
Ernst R. Berndt. Reading, MA: Addison-Wesley, 1991. xviii + 702 pp. $51.25.
The Practice of Econometrics: Classic and Contemporary by Ernst R. Berndt lives up to the endorsements from Robert Solow ("wonderful education"); Timothy Bresnahan ("examples and exercises really focus the student on the critical issues"); Bennett McCallum ("ambitious, unusual, and impressive"); and Harvey Rosen ("outstanding"). Between two colorful covers, the book provides exposition of relevant models and statistical/ econometric methods, historical background, data on an included diskette, computer usage instruction, and empirical exercises (the "hands-on" sections at the ends of the chapters). It is the book that we instructors teaching applied econometrics to business and economics students have dreamed of, but never thought we would have.
Chapter 1 is devoted to the history and use of computers. Section 1.1 chronicles Guy Orcutt's development in the 1940s of a prototype analog computer, with two-digit accuracy and the capacity for inputting at most 33 observations per variable! Sections 1,2 and 1.3 and two appendixes prepare students to use computers for carrying out the exercises in this book. References to supplementary material are also given. Each of the remaining chapters focuses on a specific area of empirical investigation: the capital asset pricing model (Chap. 2), learning curves and the Cobb-Douglas cost function (Chap. 3), the measurement of quality change (Chap. 4), wages and measuring wage discrimination (Chap. 5), aggregate investment forecasting (Chap. 6), the demand for electricity (Chap. 7), advertising expenditures (Chap. 8), the demands for factors of production (Chap. 9), macro-econometric research (Chap. 10), and female labor supply (Chap. 11). The book can be used in a modular fashion.
Berndt recommends the text for four types of learning environments:
1. As a supplementary textbook in an advanced undergraduate or first-year graduate econometrics and forecasting methods course
2. As the primary textbook in an applied econometrics course, for students who have already had instruction in econometric methods
3. For self-study and for training seminars and workshops
4. As a reference source for students conducting applied econometric research for term projects or theses.
The reason for recommending this book as a supplementary rather than main text in econometrics methods classes is that the presentation of econometric estimation and testing methods is less systematic and less fully and rigorously developed than the presentation in traditional econometrics texts. The Instructor's Research Guide for the text provides detailed advice for each chapter on the sections to be studied or reviewed in widely used econometrics methods books.
Nevertheless, anyone who carefully goes through the "hands-on" sections in this book will come to realize that the sequences of exercises, with interwoven questions and explanations, constitute an alternative approach to teaching econometric methodology that should prove highly effective for students who have innate potential as empirical researchers.
It is impossible to convey very much of what is "ambitious, unusual, and impressive" about this book in a review of reasonable length. I will instead concentrate on a single aspect of the book that is of general importance for instructors of econometrics: the breadth of coverage.
Applied econometrics practice consists of seven phases:
1. Choosing and specifying the research questions
2. Examining previous related research
3. Finding or collecting, and preparing, suitable data
4. Either finding or developing appropriate methods of empirical analysis
5. Assessing and choosing among the resulting estimation results
6. Drawing conclusions with respect to the specific research questions