MODULE ASSESSMENT: Assignment
COLLEGE: Social Science
MODULE: Quantitative Methods for Finance II
MODULE CODE: FIN9012M
CO-ORDINATOR: Dimitrios Vortelinos
SUBMISSION DATE: 18 May 2015
INSTRUCTIONS TO STUDENTS:
QUESTIONS TO ANSWER: Answer ALL 8 Research Questions
– Provide evidence of application of the appropriate econometric
methods to all 8 research questions.
– Provide the appropriate explanation of the results for each and every of the 8 research questions.
MARKING SCHEME: – Evidence of applying the econometric methods worth 50 marks for all 8 research questions and 6.25 marks for each.
– Explanation of all 8 research questions count for 50 marks and 6.25 marks for each.
Assessment consists of a time series investigation. It should be around 3,000 words in length. It will be assessed on the grounds of the application of econometric methods to real financial data as well as presentation and interpretation of econometric results.
You are required to:
Provide answers to the following research questions. In specific:
(i) Use the appropriate econometric methods/models. Present your results in a clear and logical fashion. All EViews tables with results should be included. (50 marks)
(ii) Explain the results. (50 marks)
Research questions are equally important. The research questions are:
Conduct a test of structural stability in any one only model you are about to estimate. Such a test should use one or more dummy variables for testing different conditions.
Estimate any one only distributed lag model (any number of lagged terms of the independent variables) and explain the model output.
Estimate any one only autoregressive model (any number of lagged terms of the dependent variables) and explain the model output.
Apply the Box-Jenkins approach for the family of ARIMA models. Which model is most appropriate and why?
Estimate at least five GARCH-family models (ARCH, GARCH, GARCH-M, TGARCH, EGARCH, etc.) and comment on their strength (significance).
Test for the number of unit roots (order of integration) via the Augmented Dickey Fuller (ADF) and Phillips-Perron (PP) tests.
Test for cointegration via the Engle-Granger approach and the Johansen approach.
Estimate a VAR model, and also perform Granger causality tests for the VAR model as well as pairwise Granger causality tests.
So for each of the above research questions, most of the marks for (i) is 6.25 marks and for (ii) is also 6.25 marks. Each of the eight (8) research questions, with both (i) and (ii) best answered, worth 12.50 marks.
You may use data from any financial market you prefer either in UK or abroad.
Data can come from any data provider. However, it is better to acquire data from Datastream because of a better accuracy and completeness.
Data should be in a daily sampling frequency.
The data range should be not less than 5 years.
Close prices should be employed.
Continuously compounded returns should be calculated.