Answer the following questions and briefly explain your reasoning: [30 marks] a) According to a friend, finance theory should be concerned with modelling prices of assets, since asset prices are used for buying and selling. (i) Comment on why one usually works with asset returns rather than prices in time series analysis. [4 marks) (ii) The definition of log return is defined below P 11 = In = In P, In P-1 P-1 Explain why one usually uses the log return in empirical time series work in finance and discuss one disadvantage of using log returns. [6 marks] b) (i) What do skewness and kurtosis of a variable imply for its distribution? What are the values of skewness and kurtosis of a normal distribution? [6 marks] (ii) Please name two tests for normality. Select one test for normality and discuss the intuition behind it. [4 marks] c) (i) Please use regressions to explain the difference between autocorrelation and par- tial autocorrelation functions. How can we use these two statistics to identify the autoregressive and moving average components of a time series? [7 marks] (ii) Please discuss one test for autocorrelation coefficients. [3 marks] Answer the following questions and briefly explain your reasoning: [30 marks] a) According to a friend, finance theory should be concerned with modelling prices of assets, since asset prices are used for buying and selling. (i) Comment on why one usually works with asset returns rather than prices in time series analysis. [4 marks) (ii) The definition of log return is defined below P 11 = In = In P, In P-1 P-1 Explain why one usually uses the log return in empirical time series work in finance and discuss one disadvantage of using log returns. [6 marks] b) (i) What do skewness and kurtosis of a variable imply for its distribution? What are the values of skewness and kurtosis of a normal distribution? [6 marks] (ii) Please name two tests for normality. Select one test for normality and discuss the intuition behind it. [4 marks] c) (i) Please use regressions to explain the difference between autocorrelation and par- tial autocorrelation functions. How can we use these two statistics to identify the autoregressive and moving average components of a time series? [7 marks] (ii) Please discuss one test for autocorrelation coefficients. [3 marks]
