Preface: A note to instructors (and students)

The purpose of this module is to help instructors who wish to inject empirical analysis into macroeconomics, either at the principle or intermediate level. It may also be used to supplement classes in money and banking, introductory statistics, or econometrics. The second purpose is to illustrate a few ideas in macroeconomics which emphasize the interplay between theory and empirics. The first two chapters of the module are more economic in content, and may be skipped for anyone who wants to get right to the data. Chapter 1 is an overview of the national income and product accounts and Chapter 2 is a brief history of the leading trends in macroeconomic thought since the 1940s.

The data-based exercises of chapters 3 through 7 presented a fundamental choice between dividing the content by macroeconomic topic or by the type of quantitative analysis. Potentially, each chapter could focus on a particular area of macroeconomics (for example, the labor force and unemployment) and present a variety of problems and techniques for analysis; alternatively, each chapter could focus on a particular set of statistical techniques (for example, descriptive statistics) and use them to analyze problems in a variety of areas. My decision to group the problems by statistical technique rather than macroeconomic topic was based on the realization that statistics is the binding constraint. Students in first year macro are unlikely to have taken a statistics course, while a large proportion of students in more advanced courses have usually forgotten what they learned. Since it is likely that the primary use of this module is in a beginning or intermediate level macro class where the economics content is presented more formally, I decided to organize the book in a manner that gives both the instructor and the student a clear set of guideposts to the level of statistical analysis in use.

Chapter 3 is an overview of the data set. It emphasizes graphing and descriptive statistics which are divided into five broad categories:

GDP and its components;
The labor force and unemployment;
Prices and inflation;
Money and interest rates;
The federal budget.

Chapters 4 through 7 cover t-tests (difference in means), correlation, regression, and estimating growth rates. The coverage of macroeconomic topics is not symmetrical with respect to the topics found in a typical macro texts. Money and banking issues are covered weakly, if at all, while unemployment, growth, and inflation issues are treated more generously.

P-1 The problem of time series data

Time series data is not ideal for illustrating basic statistical analysis since it involves a variety of complex problems such as lag effects, nonstationarity, and autoregressive processes, each of which are well beyond the introductory level of quantitative analysis. Nevertheless, there are several positive features to using this particular set of time series data. First, macroeconomics is either an empirical branch of economics, or it is useless. Ultimately, everything we think we know about the macroeconomy should be able to pass empirical tests. Use of "real" data to test basic hypotheses is invaluable as a means of emphasizing this point. Second, students are familiar with the national accounts and price and labor force data. After taking a class or two in economics, these are often the only numbers get their hands on. And third, familiarity with the national accounts and related data is important to the development an understanding of the relationships between macroeconomic aggregates. Presumably, this is one of the central reasons why macroeconomics is a core course in virtually every economics program.

The potential for statistical bias when problems of nonstationarity, autoregression, etc., are not controlled formally is handled by avoiding the use of un-transformed data whenever possible. Rates of change, ratios, and first differences are used throughout the module. In most cases, this makes a kind of intuitive sense to students and should not be difficult to motivate in class. For example, we are not as interested in the number of people unemployed as we are in the unemployment rate, and we don't care about the actual level of GDP as much as we care about its growth rate.

In the regression material, I have ignored problems of autoregression and the use of the Durbin Watson statistic. This would be a natural extension for instructors that want to go beyond the relatively simple techniques presented here.

P-2 SPSS versus a spreadsheet

This module is part of a series of computer based instructional modules for the social sciences. The other modules cover topics in sociology, political science and geography. This and the other modules are tied to SPSS formatted data sets. We favor the use of SPSS because of its flexibility, its scope, its ease of use, and, not least, because the CSU has a systemwide site license which insures its availability. As an instructor, I recognize that many computer oriented faculty may not want to use SPSS, and that they would prefer a spreadsheet format such as Excel or Lotus, or an ASCII file for use with their own particular statistical software. If you prefer a format other than SPSS, it is not a problem since SPSS can save to an Excel or ASCII file. Start the program, load the file, then save it in the format you prefer.

P-3 Feedback

If you use the module, please email me (jgerber@mail.sdsu.edu) and let me know. I am convinced that this sort of exercise is of great value in the undergraduate curriculum and would like to convince my department chair and dean that developing this module was time well spent. I would also greatly appreciate any feedback and suggestions you have for improvements.

This is a first effort; undoubtedly it can be made better with changes suggested by users.