What is New - in
Resources for Economists
This section contains a brief description of the various stages of an empirical study. The material here is aimed mainly, but not exclusively, for those students taking the undergraduate Applied Econometrics course (Econ. 5853) and the graduate Econometrics course (Econ. 6976). The scheme for a typically theoretical research will vary from the one presented here because the author needs not be concerned about the availability of data and the sources for collecting them. A typically applied research must nonetheless contain both the theoretical and empirical components. Often, the author must posit from the existing stock of economic theory the rationale that underlie his/her empirical formulation of the reality under study. Barring the development of an entirely new theory, which is the domain of theoretical research, the author must thus be well versed in the existing fund of economic knowledge and be sure to select the correct one (s) for the empirical work at hand.
Of course, there is no unique way to conduct an empirical study and no one particular rule can be said to apply to every empirical research situation. Thus, the framework presented here offers general guidelines and suggestions that can be adapted to many applied research situations. In the words of Ramanathan(1991) "Practice is the only way to really learn the steps involved in applied research and to develop the intuition needed to judge results and draw conclusions."
Guidelines for Carrying out Empirical Study
Conducting an empirical study involves the following basic steps:
1. Selecting a Topic
Selecting a Topic
For those of you who are employed, the problem for study is usually dictated by the requirement of the job and/or assigned by your superior. As a student I may choose to assign you a topic or you may elect to research a topic that is of primary interest to you. I prefer the latter. I suggest that you submit a one page proposal stating the topic and a brief description of the problem you want to study. After reviewing your proposal I will then make individual appointment to discuss your proposal to ensure that your ideas are consistent with received economic theory.
Knowing the topic is just only the beginning however. To get started you must formulate the right research question(s). This requires some knowledge of the existing bodies of economic theories. Recall that economics is basically concerned with the rational behavior of economic agents in their respective roles as consumers, producers, and investors as well as those of the public and private institutions in which these agents engage in daily decision making. The observed economic data therefore contain useful information for empirically assessing/measuring the behaviors of these agents in relation to the various markets in which they interplay to achieve their respective aims as well as react rationally to the changing economic conditions. Thus, having chosen a research topic, or been assigned one, it is incumbent upon you to ask yourself which of the many economic theories you have studied explain the causal relationship(s) underlying the behavior of the economic agent, phenomenon, or institution that you are about to study. Once you identify the relevant theory or bodies of economic theories the next phase is to formulate the relationship(s) empirically and subject the theories to empirical tests.
Some possible topic areas to select from include the following:
1. Macroeconomics: Estimate a consumption, or investment, or money demand function; usually using a time-series data.
2. Microeconomics: Estimate production, cost, supply, demand function. The only practical problem though is that data are generally hard to obtain. For a refresher on microeconomic theory visit Professor Daniel's MultimediaWeb Site for Imtermediate Microeconomics.
3. Urban and Regional Economics: Estimate the demand for housing, transportation, schools and other public facilities, for a city, county, or state. Measure the sensitivity of industrial location to differences across region in tax rates, energy prices, zoning laws, degree of unionization, availability of skilled workers, and so on.
4. International Economics: Estimate imports and exports functions for a given country over time or across many countries. Relate exchange rates to their determinants.
5. Development Economics: Measure the determinants of per capita income (GDP) across countries.
6. Labor Economics: Test theories on unionization, early retirement, labor force participation rates, wage, wage differential among women, minorities, young workers, and so on.
7. Industrial Organization: Measure the effects of advertising on sales and profits, or on the concentration/market share in industries. Estimate the relationship between expenditure on R&D and employee productivity. Study the relationship between industry concentration and profitability due to merger.
9. Public Finance: Estimate the relationship between local government tax revenue and its characteristics, such as population, demographic and industrial mix, wages, income, and so on. Also relate the expenditures on health, road, and education, and so on to their determinants.
10. Socioeconomics: Explain variations across cities, counties, and states in crime, poverty, divorce rates, family size, and so on.
11. Politics: Relate voter turnout to a number of characteristics of precincts, candidates, and so on. Explain the vote obtained by a politician in different districts.
A systematic approach to selecting a topic to the problem of selecting a topic is to make effective use of a classification system adopted by the Journal of Economic Literature (JEL). After selecting a topic, prepare a statement of the problem you intend to study. You can accomplish this phase painlessly by reviewing the relevant literature on the topic.
Review of Literature -- Place bookmarks to key economic journals. Or simply list key journals by JEL classification.
Formulating a General Model: Based on the literature review formulate a general model. State the original model in broad terms by identifying the dependent variable (DV) and the independent variables (IVs) for which you would like to collect data. Check your model for internal consistency - such as the presence of feedback effects (i.e; Y affects X and X affects Y). Also determine whether cross-sectional or time series data or panel data are appropriate for your stated objectives. If your goal is to explain the underlying process that makes the values of the DV vary over time, then the relevant data will be time series. On the other hand, if you wish to investigate why different groups (such as different firms, industries, employment groups, countries, counties) behave differently at a given point in time, then cross-sectional data are called for. The relationship estimated with cross-sectional data might not be stable over time, to examine this issue both time series and cross-sectional data must be pooled into panel data. Write a summary explanation why you believe the particular IV you choose to include in your model is important and justified by theory. Describe the maintained hypotheses you plan to test and the expected nature of the effect of the IV on the DV. In effect, discuss the expected signs of the coefficients of the IVs, what kind of interactions among the IVs you should look for, and whether nonlinearity might be present. The next step is collecting data on all the variables of the model.
Sources of Economic Data
Whether or not you have access to the latest generation of information technology (the Internet), you should familiarize yourself with the resources available in the Maag library. Most economic and business data can be readily found in government documents, reference materials (such as almanacs and encyclopedias), journals, and books. In practice, very little economic data are published in general-interest magazines, professional journal, or books. But remember, even if you have access to the Internet you must first of all invest time in planning the research according to the discussion above. That way you do not waste your precious time browsing the web or visiting gopher and ftp sites without a clue to a specific data series.
Assuming, however, that you have exhausted all efforts trying to locate a particular data series in the Maag library, the only recourse today is to go on the Internet and hunt for the data. For this reason, I have provided bookmarks to important data sources via their Internet addresses or Uniform Resource Locators (URLs). By simply clicking on the name of the source, you will be speeding through the information superhighway to that site. The screen interface that you see once you get there depends on whether the site is accessible by a Gopher, Telnet, Web, or an FTP protocol. In general, the Web is the most intuitive interface because it is graphical, followed by Gopher; they make searching for information much more easier and fun. You should be able to locate the data/information that you need by exploring the following sources:
Useful Sources for
Economics Data and other Subjects.
4) SHAZAM by
Professor Kenneth White of the University of British Columbia. This program
can actually be run while on-line.
Information about important Usenet Newsgroups and Mailing lists that students of
economics and business may find interesting will be provided in the near future.
Meanwhile, students can participate in Business and Economics Educational
"Talkback" discussion group, U.K.
Copyrightę 1996, Ebenge Usip, all rights reserved.