ECB003 Coursework Brief - Introduction to Econometrics

ECB003 Coursework Task

For your assignment you will form a three-members group. Each group will have a variable(consumption,   investment,   imports,   stock   building or   unemployment)   and   a   country.   The coursework task is to estimate and present the best possible equation (such as an unbiased,consistent and efficient model) you can obtain, which explains the behaviour of your variable using the data available in the database excel sheet on Learn.

Guidance Notes

The table below gives some very general ideas about the type of functional relationship you should investigate.

In the table, underneath the functions we have indicated the usual signs for the response of the left-hand side   variable to   changes in   the right-hand   side  variable(s)  predicted   by economic theory. Note that in the case of the inventory function (stock-building), the sign is ambiguous depending on whether the increase in GDP is anticipated     or unanticipated   .These functions are a starting point only. You need to make decisions about:

1. Functional form
2. Dynamic specification

In many cases the logarithmic functional form is useful because it allows us to interpret the coefficients directly as elasticities. However, it is not always appropriate or even possible to estimate an equation in this form. For example, the logarithmic transformation cannot be applied to a data series which contains negative or zero values. You need to check your data to see if this is the case and only apply this transformation if it is possible.

The dynamic specification can also be tricky. Most time series equations will have some degree of serial correlation which can lead to misleading conclusions if not properly dealt with. You need to decide how to deal with this issue for your own particular model. Possible methods which we will cover in the lectures are differencing, the inclusion of lagged variables and the estimation   of   models   with   auto regressive   errors.   In   many   cases   all   of   these   methods   are possible   and   will   improve   the   final   specification.   It   is   up   to   you   to   decide   which   is   most appropriate for your model. You should note that it may prove impossible to eliminate serial correlation entirely from your model. If this is the case, you need to acknowledge this in your discussion and recognise the implications for the interpretation of your estimated model.

Word Count and Format

1. Your   assignment   should   be   a   maximum   of   1,500   words   in   length.   The   equation,reference list, and graphs do not count towards the word count, but in-text citations do.Please report the word count from Word at the start.
2. The first page should contain your student ID number at the top of the page. You shouldalso state the topic and country you have been allocated at the start of your assignment.
3. You must NOT include your name anywhere in your file or the filename.
4. Use 12 point Times New Roman type and single spaced text.
5. Your final estimated equation  should be accompanied by  a full  set of diagnostic test statistics. Do not copy and paste EViews output. Acceptable formats for the presentation of your results are given below.
6. A brief discussion of the economic theory underlying the model you estimate.
7. An explanation of the process by which you chose the final equation presented.
8. An assessment of the fit of your estimated equation. This should include a discussion related or supported by economic theory, the statistical fit of your estimated equation and a discussion of the diagnostic test statistics you have presented. Your discussion of the diagnostic tests should include the nature of the hypothesis being tested, the distribution of the test statistic under the null hypothesis and a brief discussion of the implications of any is specification which you detect.
9. Your assignment should be written in clear, grammatically correct English. Check your assignment for spelling and grammatical errors before you submit. A good presentation makes  it   easy   for   the   reader   to   assess   your   work   and   will   be   rewarded.   A   poor presentation will be penalised.
10. All material (academic and other) that you cite should be properly referenced in the text and in the reference list (in Harvard style) at the end of your report.

ECB003 Example 1: Regression output in equation form

US Exports Equation Sample Period: 1979Q1 to 2005Q4 (108 observations

Numbers in parentheses below coefficients are standard errors. Numbers in parentheses next to test statistics are p-values.

ECB003 Example 2: Regression output in table form

Table 1: US Exports Equation Sample Period: 1979Q1 to 2005Q4 (108 observations)

indicates that the null is rejected at the 5% level, ** indicates that the null is rejected at the 1%level