Project description
The objective of the assignment is to give the students a realistic portfolio management experience by carrying out asset allocation and performance evaluation. By conducting quantitative analysis on actual market data using Matlab, the students will gain knowledge on portfolio optimisation theories and their implementation issues, and acquire Matlab programming skills.
All the relevant data will be provided in the separate excel file, ‘summative assignment data.xlsx’.
You are encouraged to use Matlab for quantitative analysis, but you can also choose another program provided you have a good reason to do so. The source codes, e.g., Matlab m-files, should be included as appendix in the report. The words in the source code will not be counted.
You may employ quantitative methods other than those covered in the class. Whatever method you choose, a proper reason should be given.
Assessment will not be based on the performance of your portfolio but on the analysis and discussion throughout the portfolio management process.
Your report must address the points below.
Write a report including the following:
1. Calculate monthly return of each asset in the investment pool and the benchmark index, S&P
500, during the sample period. The data will be provided in the excel file.
2. Carry out initial inspection on the return time series. This may include visual inspection of
the time series data, test of autocorrelation, etc.
3. Refine the investment pool: You may want to eliminate some stocks based on the initial
inspection and/or some other reasons such as ethical reasons.
4. Estimate expected return and covariance matrix of the stocks.
5. Construct an optimal risky portfolio (tangent portfolio). Explain why you choose the particular
objective function, constraints, and/or optimisation algorithm. You should provide the details
of implementation such as, but not limited to, objective function and constraints.
6. Construct an optimal portfolio by mixing the tangent portfolio with a risk-free asset.
7. Repeat 4-6 using different methods to obtain 4 optimal portfolios to compare. Different methods
may mean any of the followings:
Different constraints, different sampling schemes (in sample period, etc), different optimisation
methods (Classical Mean-Variance, Robust optimisation, Treynor-Black, etc.), different input
parameter estimation methods (Sample mean/covariance, Black-Litterman, Bayes-Stein, etc.)
Just imagine that youre a portfolio manager and want to test several methods to determine
the final model for your data. What is important is to justify why you choose those methods.
Using more advanced methods will guarantee extra credits.
8. Evaluate the portfolios and compare them with one another and also with the benchmark.
Discuss the results. Choose the final model and explain why it is your choice of model.
9. Reflect on the modern portfolio theory based on your experience through this assignment.
you will have to attached the sources in your report as appendix. The appendix should contain:
1. The overall structure of the programs
1
Software packages: If you are using multiple software packages, describe what is used
for what purpose.
Matlab source list: List the Matlab sources (scripts/functions/data files) with a short
description for each source. Draw a diagram to show the relationship between the sources.
Program logic: For each script and function, describe the logic. For example, for the
main scripts I created for the lectures, it may look like
(a) Load data files.
(b) Estimate expected return and covariance matrix by calling function AAA.
(c) Set constraints.
(d) Draw the efficient frontier by calling function BBB.
(e) Find the optimal portfolio by calling function CCC.
(f) Calculate out-of-sample return.
(g) Evaluate out-of-sample performance.
You can give more details than the above example.
2. Program sources.
Copy and paste the source codes. Avoid screen capture whenever possible.
3 Use of the Sample Programs form the Class
You are free to use the sources for your assignment. However, try to write your own program
and borrow some codes from the samples only when it is necessary. It is, though allowed, not
recommended to use the samples as it is and change only a few lines for your assignment.
4 Multiple Software Packages
It is absolutely possible to use several programs if necessary. For example, you can use Eviews for
autocorrelation test or draw efficient frontiers in Excel (after you get the efficient portfolios from
Matlab). You can also calculate asset returns in Excel and then upload to Matlab, or calculate
out-of-sample returns in Matlab and do the performance evaluation in Excel. Just explain how the
overall analysis is done in appendix.
5 Collabration
It is encouraged to collaborate with your friends. However, students should write their own report.
If several students work together to develop the Matlab program, acknowledge it in your report.
6 Transaction cost
If you plan to consider transaction cost in your analysis, use the values:
To Buy: 20 bp
To Sell: 50 bp
———-
Added on 23.04.2015 19:30
Some models can be regarded either as an input estimation method or as a portfolio optimisation model.
If this is your case, just choose for different methods for portfolio optimisation as a whole.
Use the order calculator below and get started! Contact our live support team for any assistance or inquiry.
[order_calculator]