FIN9008 Econometric Work Individual Assignment
FIN9008 Individual Assignment
Most firms have a foreign exchange exposure through foreign operations, sourcing of raw materials or services in foreign currency or through sales of goods, services and technology to customers in foreign currencies.
The capital market approach estimates foreign exchange exposure as the sensitivity of stock returns to movements in exchange rate(s) while controlling for market movements. An example that tests the exposure to a single exchange rate might be,
𝑆𝑇𝑂𝐶𝐾𝑅𝐹𝑡 = 𝛽1+ 𝛽2𝑀𝐾𝑅𝐹𝑡 + 𝛽3𝑋𝑡 + 𝜀𝑡
where, 𝑆𝑇𝑂𝐶𝐾𝑅𝐹𝑖 is the stock return over the risk free rate for time t; 𝑀𝐾𝑅𝐹 is the market portfolio return over the risk free rate for time t; Xt is the percentage change in the exchange rate factor for time t; 𝛽1 the intercept; 𝛽2 the market exposure; 𝛽3 the foreign exchange exposure; and 𝜀𝑡 is the error term for time t. Martin and Mauer (2005) provides a discussion of this.
In this project, you are to provide estimate(s) of the extent of foreign exchange exposure of your assigned FTSE listed company. You will be required to gather data for the closing price of an assigned listed company, the FTSE index and a selection of exchange rates; you will need to transform this data into log returns before running the analysis.
It is up to you to choose the exchange rates that you think are most appropriate given the particular stock and the scope of operations of that company. I would expect that in quite a few cases, you will include more than one exchange rate reflecting the different potential exposure that that particular company may have (but not all the exchange rates may prove statistically significant).
What your assignment should include:
The following – while not intended to be entirely prescriptive – is intended to give you some idea of what your assignment should include
An introduction that states the goals of your project – what are you setting out to explain?
An explanation of the specification of your model.
Why did you choose these particular independent variables?
Why these particular functional forms?
What a priori hypotheses do you have?
What are the expected signs, etc?
A short description identifying data sources and any issues/considerations relating to the data.
(Besides Bloomberg, suggested data sources are yahoo finance (stock indices), quandl.com
(FX), Bank of England/Kenneth French’s homepage (risk-free rates), but again there may be many other good data sources available)
Your results. (If you test a number of specifications, be sure to identify which you conclude are the best of these.)
An analysis of the results that includes a discussion of econometric problems encountered and tests that you have undertaken.
What is the significance of the results and how do they relate to the original questions posed
in the introduction?
Are they consistent with the theory?
Is the model statistically adequate in representing the data?
A short summary & conclusions that includes your major findings.
An appendix which includes all regression runs and all relevant output. (In the case of very extensive listings, you may choose to supply these electronically.)
An electronic copy of the project, code, and data used is to be uploaded to Canvas – Assignment, before 1pm on 1st February. Late submissions will not be accepted.
You should use Brooks R Guide.pdf file on Canvas to get yourself familiar with the code required to run the regression analysis in R and you should read ahead independently and use the diagnostic tests outlined here also.
All variable names that you create must end with your initials.
For example, Myron Scholes (MS), might create a variable name for UK House Prices as:
UKHPMS = read_excel(“C:/data/UKHP.xls”, col_types = c(“date”, “numeric”))
Collaboration on this assessment is strictly prohibited. Evidence of collaboration will be treated as a serious academic offence.
For plagiarism guidelines please refer to:
The project should be no more than 3,500 words in length (excluding appendices).
The following are draft outline criteria.
- Was the approach taken a reasonable one? (This would include the data used and the specifications used.)
- Were the tests undertaken and methods used appropriate? Were appropriate and reasonable modifications made in the light of results?
- Are the conclusions reached reasonable and well-argued?
Indicative checklist for economic/financial work
Your financial hypothesis:
Have you stated clearly what it is you want to test?
Have you identified what signs, etc., you expect for the equation you plan to estimate?
What do your estimates tell you about the hypothesis?
What can you conclude from your statistical investigation?
Indicative checklist for econometric work
Are your data appropriate for the work intended?
Does your overall equation have statistical significance?
Have you included all relevant variables?
Do all coefficients have the appropriate sign, i.e., as indicated by the theory?
What do the reported t-values tell you about the variables you have included?
Are any t-tests other than the ones reported appropriate? (For example, should you be testing a hypothesis about a coefficient having a value different from zero?)
Are there reasons to be concerned about multicollinearity?
Has the relationship changed significantly over time?
Is there a problem with heteroscedasticity? Do you need to do something about it?
Does serial correlation present a problem? Do you need to do something about it?