Aim
The purpose of this test is to give you a chance to demonstrate basic competence in the use of Eviews and to provide you with the opportunity to apply the statistical and econometric concepts and techniques that we covered in the module. The test is marked out of 100. The mark allocation for each question is given below.
Due to the uncertainties surrounding Covid19 the Eviews test will be in the form of a 48 hour take-away paper. Your answers should be submitted though Turn-it-in on My Studies by 10am Friday 11th December 2020. You should all make sure that you have a copy of the student version on Eviews to carry out the tasks. Note also that you have access to MG13/10 on Tuesday morning between 10-12am during our timetabled face-to-face seminar slot.
You will be able to access the dataset and statistical tables that you will need to answer the questions on My Studies. Your answers should be word-processed using Microsoft word. You will need to present some answers using mathematical/statistical expressions. If you cannot do this in word then you are permitted to take photographs of calculations and paste them into your document where necessary. You can also paste Eviews output into the document.
The Data
The dataset ‘UK imports.wf1’ is accessible through MY Studies. It contains information on real UK imports, the real price of UK imports and real UK GDP for the years 1973 – 2003. Specifically, the following variables are contained within the dataset:
m: Real imports £bn (1990 = 100)
p: Real import prices (1990 = 100) gdp: Real income £bn (1990 = 100)
We will assume the relationship between the UK demand for imports, the UK price of imports and UK national income can be expressed:
mt = 0 + 1pt + 2gdpt + ut [1]The Task
You are expected to model UK import demand using expression [1] above and present an analysis of your results. To guide you through the task you should answer the following questions:
Give a brief explanation of how the variables contained in the dataset and the econometric specification for import demand described on the test brief conforms to economic theory. (6 Marks)
Model 1 (expression 1 above) is specified as: mt = 0 + 1pt + 2gdpt + ut. What would you expect the signs on the estimated coefficients 1 and 2 to be? (2 Marks)
Give a brief account of the means and standard deviations of each variable: m, p and gdp in their units of measurement. (5 Marks)
Compute the 95% confidence intervals (CI) for m, p and gdp. Note the formula for the CI: x ̅= ±1.96× √(σ^2⁄n) (6 Marks)
Using the graphing facility in Eviews describe the trends you observe for each variable. What is the relationship between a) real imports and real gdp, and b) real imports and real import prices? Are the relationships you describe expected? (10 Marks)you should paste your graphs into your document
a) Estimate model 1 by OLS using the data provided for the period 1973-2001 and the Eviews command ls m c p y. (3 Marks)
you should paste the Eviews output into your document
b) For this model comment on:
i) the significance of and the signs on the estimated coefficients,
ii) the impact that each variable has on import demand in their units of measurement,
iii) the goodness of fit of the model. (6 Marks)
c) i) Do the residuals suggest the presence of autocorrelation. If so why? (4 Marks)
ii) Test model 1 for first order autocorrelation using the Durbin-Watson (DW)
procedure clearly stating the null hypotheses and the critical values that you
employed to conduct the test. What do you conclude? (8 Marks)
a) Re-estimate model 1 for the period 1973-2001 in natural logs call this model ‘model 2’. (3 Marks)you should paste the Eviews output into your document
b) For this model comment on:
i) the significance of and the signs on the estimated coefficients,
ii) the impact each variable has on import demand, and
iii) the goodness of fit of the model. (6 Marks)
c) i) Do the residuals suggest the presence of autocorrelation. If so why? (4 Marks)
ii) Test model 2 for first order autocorrelation using the DW procedure, stating clearly
the null hypotheses and the critical values that you employed to conduct the test.
What do you conclude? (8 Marks)
How does model 2 compare to model 1 on econometric and theoretical grounds? (5 Marks)
Use the models that you have estimated (model 1 and model 2) to compute the percentage forecast errors for 2002 and 2003. Which models performs best in terms of its forecasting accuracy? (8 marks)
How does the presence of autocorrelation affect a time series regression model? (6 Marks)
Why might autocorrelation be present in a regression model? (4 Marks)
On the basis of the models you have estimated briefly state how you could improve the specifications? (6 Marks)It is important that you clearly indicate which question you are answering on your answer sheet.
Once you have completed the test you should submit your answers through Turnit-in
Note that this is an individual piece of work, and you should avoid collusion and plagiarism
Learning outcomes tested: L03, L04, L05, L07, L09 (see module descriptor)
Sample Solution
