BEA653 Applied Quantitative Finance
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
EXAMINATIONS FOR DEGREES AND DIPLOMAS
October - November 2022
BEA653 Applied Quantitative Finance
Instructions:
This examination consists of ONE (1) Section.
Section A: Students must answer SIX (6) questions. Question 1-4 are worth 20 marks and Questions 5,6 are worth 10 marks each. A total is 100 marks.
Total marks for this paper: 100 marks.
This exam constitutes 40 percent of your overall mark.
SECTION A – Answer SIX (6) questions. Total marks for Section A - 100.
Question 1
Brailsford and Faff (1997) argue that the conditional CAPM can be improved by taking into consideration time-variation in volatility. They apply a GARCH-M model
to the Australian stock data set in which Tmt is an excess return on the market in period t and Gm(2)t represents variance of excess returns on the market in period t .
Moreover, Brailsford and Faff (1997) report the following results:
Table 1. Coefficient estimates and t-statistics in brackets are reported.
a Carefully interpret the coefficients in Table 1. What is the main conclusion about the performance of CAPM? [5 marks]
b Write the null hypothesis confirming the conditional CAPM is ‘true’ . Using the results from Table 1 show in which cases this hypothesis is not rejected. [5 marks]
c Use the provided data set (log returns 9 series.wf1 contains daily returns for different
markets) for the US stock market and report the results similar to Table 1 but for the US data at daily and weekly frequency. Discuss if the results are similar for the US and Australian markets. Support your answer using outputs from EViews. (Hint: compute weekly returns as an average across 5 daily returns). [10 marks]
[Total Question 1= 20 marks]
Question 2
ARIMA models can be used as a powerful tool for forecasting. Suppose you would like to predict a series of APPLE (AAPL) stock prices in EViews.
a Plot the daily closing prices for AAPL stock, along with the ACF and PACF.
Explain what each plot shows and which ARIMA model is the most appropriate. [5 marks]
b Consider the model you selected in part a. Generate a forecast over the last 10 days
in the sample and compare it with observed data. Is your forecast accurate? Why? [5 marks]
c Plot forecasts from an ARIMA(1,0,0) model with drift and compare these to part b. [5 marks]
d Can you conclude from your findings that a price is well represented by a random walk? Explain why. [5 marks]
[Total Question 2= 20 marks]
Question 3
Assidenou (2011) shows that major capital market indices (in OECD, Pacific, and Asia regions) are strongly interrelated. This finding is confirmed by results in Table 2.
Table 2. Pacific group cointegration test results.
a Explain which test is used to obtain results in Table 2 in detail. What implications the results from Table 2 have for the global market integration? [5 marks]
b A similar test for OECD countries identifies a zero cointegration rank. Does this imply stronger integration between the markets? Why? [5 marks]
c The long-run parameter estimate of the VECM, F, has the following form:
1.000000
0.000000
0.000000
0.000000
1.000000
0.000000
=
0.000000
0.000000
0.000000
0.000000
-0.984151 (0.15466) [-6.36344]
-1.032363
(0.01984)
[-52.0387]
Interpret the parameters and explain how these estimates are related to a level of market integration. [5 marks]
d Do you agree that a VECM allows capturing the short-run dynamics between the variables? How? [5 marks]
[Total Question 3= 20 marks]
Question 4
Engle, Ito and Lin (1990) (paper is attached) explain volatility clustering in exchange rates and formulate meteor shower and heat wave hypotheses. The heat wave hypothesis is that the volatility has only country-specific autocorrelation. Alternatively, the meteor shower is a phenomenon of intra-daily volatility spillovers from one market to the next.
a Formally introduce and explain GARCH models associated with meteor shower and
heat wave hypotheses. Which coefficients can be tested to verify each of these hypotheses? [4 marks]
b Table 2 in Engle, Ito and Lin (1990) provides result for testing meteor
showers and heat waves in the Pacific region, Tokyo, Europe and the United States. Ignore the results for the Pacific region and provide a similar Table using more recent data for the bond market from log returns 9 series.wf1. Can you broadly confirm that the volatility patterns in the bond market are similar to the foreign exchange market? [6 marks]
c Estimate News Impact Curves (NICs) for Japan, Europe and US and plot them.
Explain how different they are. Do you observe any issues with these NICs? [6 marks]
d A relationship between the risk and expected return of an asset depends upon
attitudes toward risk of asset holders. Which volatility model would you choose to capture this relationship? [4 marks]
[Total Question 4 = 20 marks]
Question 5
Consider an MGARCH framework. Suppose you intend to check if volatility series in 3 different European markets are interconnected. Use log returns 9 series.wf1 data set to answer the following questions
a Which MAGRCH specification you would choose and why? Formally introduce this
specification and discuss which parameters are significant. [5 marks]
b Plot conditional volatility and covariance series obtained from the specification
discussed in the previous part. Based on these plots can you identify periods when markets are more connected? [5 marks]
[Total Question 5 = 10 marks]
Question 6
Consider the Capital Asset Pricing Model that relates the excess return of a security to the market portfolio. The relevant data set (CAPM.xls) contains prices of different US stocks, S&P 500 index and the series of the US Treasury bills.
a) Estimate CAPM betas for Microsoft and interpret the parameters in the regression. [4 marks]
b) Check if the CAMP model for Microsoft suffers from heteroscedasticity. Use a formal test to check that. What are the consequences if a model suffers from heteroscedasticity? [3 marks]
c) For the model estimated in part a) formally test if the residuals are correlated? How would you solve the problem of autocorrelation? [3 marks]
[Total Question 6 = 10 marks]
2022-11-09