Reference no: EM132845954

25705 Financial Modelling and Analysis - University of Technology Sydney

Case Study

Section I: Company characteristics

Use "Datanalysis premium" at UTS library to provide an overview of your company, e.g. industry, main products/services, and markets (Australia, Asia, etc.). The relevant information can be found from company website and annual reports. Report the following using data from the last three financial years. Discuss the firm's business performance and outlook.

Section II: Trading volume and volatility
Report the table below containing summary statistics for weekly return, volume, and volatility: average, standard deviation, minimum, maximum, and 1st-order autocorrelation. List and explain two key features of the summary statistics. [2 marks] Test the hypotheses that the average return of your stock in 2020 was higher than that in 2015-19. Clearly state the null and alternative hypotheses.

Explain the finding in terms of the broad market factors and stock-specific factors.

Section III: Return and volatility
(1) What is the volatility feedback effect? What is the implied sign of the correlation between return and volatility? Use Marra (2015) as the initial reference and search the internet for additional explanations.
(2) Plot weekly volatility (vertical axis) against weekly return (horizontal axis) of your stock. Copy the numerical values of rt and σt to two new columns. Sort rt and σt on rt. Calculate Cor(rt,σt) when rt < 0 and Cor(rt,σt) when rt > 0. [2 marks] Are these signs different from those implied in Exhibit 4 of Marra (2015)? Explain why.

Section IV: A regression model to forecast weekly volatility.
(1) Select 4 to 6 explanatory variables to forecast weekly volatility of your stock. Explain why an explanatory variable is expected to increase or decrease future volatility.
(2) Define 5 Jan 2015 to 8 Jan 2021 as the estimation period. Estimate your model and report the results in the table below. Do the coefficients have the expected signs? If not, why?

Regression results of your model

(3) Let rt and σt be the return and volatility in week t. Estimate the following benchmark model during the estimation period:
σt = constant + β1σt-1 + β2σt-2 + β3rt-1 + et
Report the results in the table blow. Compare the performance of your model against the benchmark model.
Regression results of the benchmark model

(4) Define the 16 weeks from 11 Jan to 30 Apr 2021 as the holdout period. Use the LINEST function to conduct rolling forecasts for weekly volatility in the holdout period. Let σt and σ^t be the actual and forecasted volatility respectively. For each forecast, calculate the absolute error |σt - σ^t| and the percentage error (σt - σ^t)/σt. Tabulate weekly σt, σ^t, |σt - σ^t|, and (σt - σ^t)/σt. [2 marks] Plot the following charts:
• Chart 1: the time series of σt and σ^t over the holdout period.
• Chart 2: |σt - σ^t| in y-axis against σt in x-axis.
• Chart 3: (σt - σ^t)/σt in y-axis against σt in x-axis.
Discuss key features in the charts, including (but not limited to) whether forecasts improve over time, whether forecasts are more accurate for high or low volatility, etc.

Section V: Volatility targeting
(1) Your portfolio consists of your stock and cash. Every week you decide to put wt in the stock and (1-wt) in cash. Your stock has weekly return rt and volatility σt. It is assumed that cash has zero return, you can borrow at zero interest rate, and there is no transaction cost. Your weekly portfolio return is rp,t = wtrt, and your portfolio volatility is σp,t = wtσt.

(2) Suppose that you want to control your annual investment risk to 20%. The implied weekly target volatility is σT = 20%/√52 = 2.77%, since there are 52 weeks per year. Given your forecast σ^t, your portfolio weight is wt =σT/σ^t: if σ^t > σT, wt < 1 and you hold 1-wt in cash; if σ^t < σT, wt > 1 and you borrow cash
to invest.1

(3) Based on your weekly forecast σ^t and σT = 2.77%, calculate your weekly portfolio weight w_t = σ^T/^σ_T and weekly portfolio return r_p,t = w r during the holdout period.

Tabulate the average, standard deviation, min, and max of the weekly r_t, w_t, r_p,t, and σ_p,t. List and explain three key features of the summary statistics.

Plot weekly wt (y-axis) against weekly stock return rt (x- axis). Discuss their relation over the holdout period, e.g. whether they are positively related and why. Let X¯ be the average of Xt. Calculate the portfolio's Sharpe ratio r¯p/σ¯p and the Sharpe ratio from buying and holding the stock r¯.2/σ¯ Does the volatility targeting strategy have a higher Sharpe ratio than the buy-and-hold strategy? Explain why.

Report format and quality of writing

1. The report should have a cover page containing subject number and name, report title, student name, ID, and UTS email. It should also have a half-page Executive Summary which states the issues investigated, the forecasting variables used, and investment performance of volatility targeting.

Attachment:- Financial Modelling and Analysis.rar