Book review: Mathematical Statistics for Economics and Business, Ron C. Mittelhammer, Springer-Verlag, New York, 1996, xviii+723 pp. $57·95, hardback only; solutions manual free to instructors who adopt the text.
✍ Scribed by B. D. McCullough
- Book ID
- 101285652
- Publisher
- John Wiley and Sons
- Year
- 1999
- Tongue
- English
- Weight
- 76 KB
- Volume
- 14
- Category
- Article
- ISSN
- 0883-7252
No coin nor oath required. For personal study only.
✦ Synopsis
Recent years have seen an increasingly sophisticated set of estimation, testing and evaluation methods enter the applied economist's toolkit. The level of statistical competency required of the applied economist has increased concomitantly. If these statistical topics are covered in an econometrics course, either econometrics or statistics receives insucient attention. Taking the standard ®rst-year graduate course in the statistics department is not an ecient option: too much time is spent on topics not relevant to econometrics and not enough time is spent on those topics which are relevant. Therefore, some PhD programmes oer a semester of `statistics for econometrics' (SFE) prior to the ®rst econometrics course. Perusing their syllabi makes clear the need for a book such as the one reviewed here.
These SFE courses all rely on a standard, introductory graduate statistics textbook such as Hogg and Craig (1994). Such textbooks often presume a degree of mathematical sophistication not possessed by many ®rst-year graduate students in economics. Only a portion of the textbook is relevant to econometrics, so many of the topics in the textbook are not covered. That portion which is relevant is not treated in nearly enough detail. There are many statistical topics relevant to econometrics which are not covered in the text. Therefore, many supplementary readings may be required. Naturally, the text, its exercises, and possibly the supplementary readings are not written with econometrics in mind. Consequently, as far as economics is concerned, the material is neither uni®ed nor integrated and may be judged marginally adequate, at best. Even upon successful completion of such a course, the student may well wonder what relevance his newly acquired knowledge has for econometrics. All this can make it dicult not only to take such a course but to teach such a course as well. Given these twin diculties, it is not surprising that SFE is not oered by more schools, but Mittelhammer's contribution can ameliorate the situation. This ®ne text largely remedies many of the above defects, and will as much improve the taking of the course as the teaching of it.
The preface begins This book is designed to provide beginning graduate students with a rigorous and accessible foundation in the principles of probability and mathematical statistics underlying statistical inference in the ®elds of business and economics'. My impression is that in the ®elds of business and economics' is a bit too broad. I doubt that many MBA programmes require such a thorough knowledge of statistics. Moreover, ANOVA is not covered, which might trouble professors from a business curriculum but not economics professors. However, I am certain that in the ®eld of econometrics' is not too narrow, and that this book is well suited to the needs of economists who aspire to do competent econometric work. Disconcertingly, there are some typos and misprints, and not all of them are innocuous. Some of them aect formulae and chapter exercises. Fortunately, Mittelhammer has an errata list on his homepage (www.agecon.wsu.edu/faculty/mittelhammer/index.htm). Also, the book does have a grammatical ¯aw. While the gender-neutral use of the masculine pronoun has been almost completely eliminated in favour of the feminine pronominal form (Campbell and Campbell-Wright, 1995), I did ®nd one use of the word he' ( p. 168). These minor quibbles aside, Mittelhammer admirably attains his stated goals.
The book is rigorous and accessible. Only rarely are proofs omitted, (e.g., completeness in the exponential class, a few central limit theorems), and it seems that such proofs are omitted more because