Format for an Econometrics Paper
An econometric paper should conform to the following generally used format:
I. Introduction and literature review
Discuss what you will attempt to do. It is often a good idea to state the question that your paper will attempt to answer. If this were a term paper, this section should contain a review of the literature of relevant theory and of other studies relating to this question. If applicable discuss any controversies which your paper will attempt to resolve. Examples, which should be expanded upon, include:
- This paper attempts to explain the determinants of investment spending by business firms.
- This paper attempts to determine the extent to which changes in the money supply affect inflation.
II. Specify the Model
The model should be stated here in general terms. For an econometric project, this section should consist of the general form of the model and an explanation of the causality of each of your independent variables. The model should include, and at least discuss the reasons for not including, variables suggested by economic theory or used by others researching this subject. A general form of the equation should look something like:
expected signs -> - + -
I = f(r, Y, X)
I = Investment Expenditures by Business Firms
r = Interest rates
Y = Incomes
X = Excess Capacity of Business firms
Interest rates are the opportunity cost of funds used for business investment expenditures. If the interest rate rises, some investment projects which were profitable may no longer be profitable. Therefore, an increase in interest rates will lead to a decline in investment expenditures, holding incomes and excess capacity constant. The expected sign is therefore negative.
As incomes grow, consumers are expected to increase their purchases. This increase in consumer spending will create excess demands for goods and services. These shortages will cause business firms to want to increase their production. To do so they will increase the amount of capital goods that can be used for production. Thus rising incomes are expected to increase the amount of business investment expenditures. The expected sign is therefore positive.
[Each of the independent variables should have a summary of the theoretical argument for its inclusion and its expected sign in this section.]
III. The Estimated Model
Data are not always perfect, the researcher must fit the model from section II to the data as best you can while remaining faithful to the theory. In this section you specify the exact
functional form of your estimating equation and a description of the (proxy) variables used represent the concepts in your model. For example, the estimating equation, using quarterly US data, might be:
I(t) = B0 + B1r(t) + B2Y(t-1) + B3X(t)
Subscripts (shown here in parentheses) indicate time periods;
t = current time period
t-1 = previous time period
t-2 = two time periods ago, etc.
I = Gross Private Non-Residential Investment ($1987)
r = Prime rate of interest less the inflation rate as measured
by the Consumer Price Index
Y = National Income (1987$)
X = 100% minus the Capacity Utilization in manufacturing (%)
- Explain the variables carefully enough so that readers will readily understand the results presented in the next section.
- Explain carefully any calculations that you performed on that data,
e.g., real interest rate = prime interest rate minus inflation rate as measured by the CPI.
- Enough detail should be provided so that a reader could locate the data that you used and replicate your results.
- Provide the exact source of your data, e.g., National Income Accounts, Survey of Current Business, etc.
IV. Present Your Results
This section contain all of the relevant statistical results of your regression in an easy to read format. Round your numbers off at reasonable points. For example, a t-statistic of 4.7654293 is not reasonable. At a minimum, you should present all of the coefficients, their t-statistics (or their standard errors or both), sample size, adjusted R2. If you are using time-series data, present the Durbin-Watson or Durbin's h-statistic. Include any other statistical information which from which you draw inferences, e.g., beta-weights, elasticities, etc. The correlation matrix should also be presented. If any correlation coefficients seem high (i.e., greater than, say, 0.75 in absolute value), VIF test results should be presented.
The presentation of the numbers, should be accompanied by a verbal interpretation of the important aspects of those results. Highlight those results that relate to the major purpose of the study as stated in the introduction. This section should state that the coefficients of interest are within reasonable ranges or make note of the fact that some or all of them are not reasonable. Do not spend time explaining less important results. Interested readers can see them for themselves. Point out any statistical shortcomings of the results, for example, is multicollinearity or serial correlation present?, is the R2 low?, is the standard error of the estimate high?. It is better that you point these things out than that the reader discover them and find no reference made to them, lest the reader think that you either did not notice them yourself or that you did not understand their implications. Readers appreciate candor on these issues.
The conclusion is not the place to introduce new information; that should have been done earlier in the paper. In this section you should (verbally) discuss your results in terms of the question asked in the introduction. Avoid using numbers in this section; you presented your numbers in the previous section.
Review for the reader: "the" question, your method for analyzing it, and your findings. Then clearly state the answer to the question asked at the beginning? Do your findings confirm or refute the theory being studied or the hypothesis being tested? What are the implications of your findings? What practical suggestions can you make for someone who wants to use your work as a beginning point and take it further?