ASSESSED COURSEWORK
The assessment for BE953 is by this coursework and a Final Examination. This piece of coursework is worth 50% of the overall assessment of BE953. The requirements for this coursework are as follows:
- The coursework consists of data manipulation, analysis and interpretation. Although you may discuss the project with others, the coursework must be written up individually. You may receive reduced or no marks if there are strong similarities between the work handed in by two or more people.
- All questions are to be answered.
- The word count of the project must be printed on the first page of the coursework. The maximum word count is 2000. The project should be double-spaced and word-processed.
- Your project should include a title page and a bibliography, which includes the full reference for all articles, books and other sources you have cited in the body of the text. The bibliography (and any footnotes) need not be included in the word count.
- EViews output should NOT be pasted directly into the project. You should present your EViews equation estimation output as it would be in published academic papers. (Look at some papers–sometimes output is in Tables, sometimes as estimated equations with s.e./t stats/p-values in brackets under the corresponding coefficient, together with appropriate diagnostic statistics and their p-values).
- Note that your coursework is to be submitted via FASer. The coursework should be uploaded to FASer by 09:00 on Friday 28 January. You should also upload your EViews workfiles for Questions 1 and 2.
- More information concerning late submission of coursework, can be found here:
YOU MUST READ THE INFORMATION WHICH FOLLOWS:
In submitting coursework online it must be assumed that you have read and understood the following guidelines about plagiarism. Furthermore in doing so you are agreeing to your work being monitored by the JISC Plagiarism Detection System if a lecturer should deem it necessary to do so.
Regulation 6.12 & 6.13 states that
6.12(a) It is an academic offence for a student to cheat in any examination, or in any other submitted part of his or her University work, whether or not such work is formally assessed. “To cheat” includes:
(i) to copy the work of another candidate or otherwise communicate with another candidate in an examination;
(ii) to introduce any written, printed or electronically-stored information into an examination, other than material expressly permitted in the instructions for that examination;
(iii) to use the work of others (whether in written, printed or some other form) without acknowledgement, where a judgment is made that the work has been the result of serious negligence or of intention to deceive;
(iv) to repeat work previously submitted for a different assessed assignment without full acknowledgement of the extent to which that previous work has been used.
(b) It is an academic offence for a student knowingly to assist another student to cheat in any examination, or in any other piece of work, the mark for which will count either towards the student’s result for the year, or towards his or her final degree classification.
(c) Allegations of academic offences involving cheating shall be dealt with in accordance with the Progress Procedures as determined by the Senate. Previous offences shall be taken into account.
6.13 In submitting any piece of University work (e.g. dissertation, thesis, essay or report) a student shall acknowledge any assistance received or any use of the work of others.
COURSEWORK
Data
The data for question 1 can be found on Moodle under the heading “OilPrices2018.xls”, showing the spot and futures prices for Crude Oil. For question 2, each student will be allocated a company from the FTSE 100 and will need to download price data from Yahoo Finance.
Questions
1. An estimable linear regression can be specified as:
st = b0 +b1 f t-1 +ut
where f t and st are the natural logarithms of Ft (the nearby futures price) for the Oil contract traded on NYSEX and St (the spot price). Note that ut is an error term. Import the data file “OilPrices2018.xls” into EViews. The data is sampled monthly from January 1986 to August 2018.
a) Plot Ft-1 and St on the same graph. Indicate on the graph any major economic events. Additionally, using an Augmented Dickey-Fuller test and selecting the relevant lag length using SBIC, assess whether f t and st are unit root processes.
b) Explain the term spurious regression. Why might regression (1) be spurious?
c) Using the Engle-Granger 2-step method, assess whether f t-1 and st are cointegrated. If so, what does this imply?
d) Discuss the economic rationale behind the result in 1(c) explaining terms like unbiasedness and market efficiency. Also briefly comment on how this result compares with the relevant economic literature. Some useful references (which can be downloaded electronically from the library) include:
Chow, Y-F. (2001), “Arbitrage, Risk Premium, and Cointegration Tests of the Efficiency of Futures Markets”, Journal of Business Finance and Accounting, 28: 693- 713. Kellard, N., Newbold, P., Rayner, A. and Ennew, C. (1999), “The Relative Efficiency of Commodity Futures Markets”, Journal of Futures Markets, 19: 413-432.
e) Assuming that the series, f t-1 and st , are cointegrated, formulate and estimate a general ECM (Error Correction Model) and using appropriate information criteria and diagnostic tests, ‘test down’ to a more parsimonious model. Briefly comment on the properties of your final ‘preferred’ model including discussion of short and long run elasticities. Do the standard diagnostic tests indicate any problems?
2. Read the following journal article (available on Moodle and in the library): Sun, Q. and Tong, W. (2010), “Risk and the January Effect”, Journal of Banking and Finance, 34: 965-974.
Download your allocated FTSE 100 company price data. You should choose a daily frequency from the earliest start date to 29/10/2021. Using the adjusted closing price (pt) log returns can be formed by constructing the variable:
r1=ln (p1/p1-1) r
a) Using an appropriate test, assess the possibility of ARCH effects in rt. When estimating volatility of returns, why might a GARCH type model be preferred to simply calculating the historic standard deviation?
b) Estimate an appropriate model specification for returns using (i) a GARCH (1,1) model and (ii) a GARCH-in-mean (1,1) model. Express each estimation in the appropriate equation form and interpret the coefficients.
c) Using your allocated data estimate:
(i) model (1) on page 967 of Sun and Tong (2010)
(ii) model (2) also on page 967.
If the lagged returns variable is not significant then delete it from your equations for returns. Express each estimation in the appropriate equation form and interpret the coefficients.
Given your daily data, creating a dummy variable for observations in January can be done in a number of ways; e.g.
i. Setting JAN = 0 and then editing the values for January
ii. Use a new page in workfile by:
Right clicking new page >Specify by frequency/range at bottom of workfile; Select Monthly as frequency;
Define a dummy for January, JAN, say using the @seas command;
Copy JAN;
Select original (daily) page in workfile and paste JAN.
d) On the basis of the values of appropriate information criteria, which of your 4 models (i.e., Two models from (b) and two models from (c)) would you prefer? Explain.
e) Estimate your “preferred” regression from 2(d) above using observations up to and including 31/10/. Forecast the annual volatility of returns over the next twenty-one trading days. Briefly comment on the forecast you obtain.
