1 NBS8185 MATLAB for Finance, final project, 2019-2020 Instructions Word limit: 2,000 words, not including references section and appendix. Download three data sets: Data set 1 Using Yahoo Finance, download daily stock price data for any four U.S. companies for the period 31/12/97-31/12/08. Data set 2 Using Yahoo Finance, download monthly stock price data for any four U.S. companies for the period 1997:12-2007:12. Data set 3 Using the Bank of England’s statistics database, download monthly data on the end-of- month GBP/USD spot exchange rate (1998:01-2018:12) and the GBP/USD one-month forward exchange rate (1998:01-2018:12). When you have downloaded your data, answer the following questions using MATLAB for all computations. Report your results using tables and graphs where necessary. Include a references section and an appendix section containing the MATLAB programs used. Neither of these sections is included in the word count. 2 Questions Q1. (a) Using either Dickey-Fuller tests or variance ratio tests, analyse whether the natural logarithm of the stock price data in data set 1 is consistent with the weak- form efficient market hypothesis. 10% (b) Pick one of the stocks from data set 1 and compute the log-returns for this stock. Using this data, for each day in 2008 compute a one-day ahead log-return forecast from an appropriate ARIMA(p,d,q) model. Starting with $1,000, assume that depending on the direction of each log-return forecast you invest all available funds in the stock (if the forecast is positive), or sell all of your holdings and earn a return of 0 (if the forecast is negative or zero). Analyse the performance of this trading strategy and discuss whether your results are consistent with the weak-form efficient market hypothesis. 30% Q2. Using data set 3 and linear regression, investigate how the statistical evidence on the forward rate unbiasedness (FRU) hypothesis for the GBP/USD changes over the sample period. 10% Q3. Using data set 2 and assuming a one-year investment horizon, no borrowing and no short sales, compute the optimal risky portfolio weights and the optimal final portfolio weights for a portfolio of the four stocks and a single risk-free asset. Produce results for three hypothetical values of the risk-free rate (low, moderate, and high) and compare the three sets of results. 10% Q4. Using data set 1: (a) For each stock and employing the delta-normal (DN) approach, compute and graph the one-day ahead return-VaR over the period 02/01/08-31/12/08. Conduct backtesting and critically discuss the results. 20% (b) Using the DN approach, compute and graph the one-day ahead return-VaR for an equally weighted portfolio of the four stocks over the period 02/01/08- 31/12/08. Conduct backtesting and critically discuss the results. 20%
