Dr. Christos Argyropoulos
Dr. Eirini Bersimi
Autumn Term 2023-24 University of Essex
BE-333-6: Empirical Finance
Coursework 2
Instructions: This coursework assignment (research report) must be submitted electronically via FASER
by the due date and time. When submitting your coursework assignment, you must provide one
Microsoft Word or PDF file containing your written/text answers to the questions. In your answers to
the questions below, you should present your EViews equation estimation output as it would be in
published academic papers. (Examine the Fama-French (FF) 5- factor model paper, uploaded in the
coursework folder, to understand how to present the estimation outputs in tables. The approaches to
presentation are fairly standard.) Raw EViews output should be included only in an Appendix.
The report should not exceed 2000 words in length. It should have a clear introduction and a conclusion.
You should ensure that you have fully acknowledged the work of others in the body of the text and
include a full list of references for all articles, books and other sources (e.g. Internet sites) that have
been cited in the assignment. Coursework will be processed with plagiarism detection software. Marks
will be awarded for writing style and graphical presentation as well as content.
The data required for the coursework is contained in the excel file ` Coursework_2.xls’ in the coursework
section on Moodle. The file contains daily returns for FTSE100 from January 2000 to December 2019
that we will use to estimate and forecast the volatility of FTSE100.
Question 1 (10 points)
Consider the observations for the FTSE 100 stock index return series for the period ranging from 2000
to 2017. Test for ARCH effect the series of daily returns.
Question 2 (20 points)
Then, considering your findings from Question 1, construct an appropriate ARMA(p,q) model for the
conditional mean and test for ARCH effects the squared estimated residuals, using the model for the
conditional mean. Interpret the results between the two ARCH tests (from Question 1 and from
Question 2).
Question 3 (35 points)
Using the observations for years 2000 to 2017, estimate the following volatility models from the GARCH
family, selecting the appropriate lags:
• GARCH with normal innovations
• GJR-GARCH with normal innovations
• GARCH with Student’s t errors
• GJR-GARCH with Student’s t errors
• a specification of your choice
Notice, that for the above volatility models, you need to model the conditional mean, for which,
you should use an ARMA(p, q) model, with the specification you identified as appropriate in Question
1. For each of the models you have estimated above, forecast the daily volatility (square root of
variance) for the last two years in your sample.
Question 4 (15 points)
For each of the models you have estimated and forecasted in Question 3, plot your forecasts against
the following two proxies for the volatility: (1) (yt)2 (square returns proxy) and (2) || (absolute returns
proxy). Plot also the forecasts errors (i.e. actual – forecast), again using both proxies. Interpret the
results.
Question 5 (20 points)
Using both proxies from Question 4, compute the MSE, MAE and MAPE for all the models you have
estimated in Question 3. According to each of these criteria, which model forecasts best and which
model forecasts worst? Please discuss the reason behind the differences you observe in the forecast
evaluation.