# 辅导案例-QBUS6830

QBUS6830
Financial Time Series and Forecasting
Semester 1, 2020

Group Assignment

The Group Assignment will contribute 40% towards your final grade and is to be
completed in groups of 3-5 students. The due date is Monday 25th May, by 5pm AEST
via online Turnitin submission in Canvas. You will be penalised 20% for each 24-hour
period it is late. Submissions after Monday 25th May, by 5pm AEST will be penalised
instantly as if it was one day late (i.e. 20%).

There are 3 variations to this group assignment. Each group number can either be Type 1,
Type 2 or Type 3. Use this to determine which variation of the assignment your group will
use, and therefore what your group number is.

Let X be the sum of the last digit of each group members student ID.

If you have 3 people in your group, then if X is:
0 - 9 → Type 1
10 - 18 → Type 2
19 – 27 → Type 3

If you have 4 people in your group, then if X is:
0 – 12 → Type 1
13 – 24 → Type 2
25 – 36 → Type 3

If you have 5 people in your group, then if X is:
0 – 15 → Type 1
16 – 30 → Type 2
31 – 45 → Type 3

Submission Requirements

The assignment consists submission of 6 parts; Cover sheet, a Written Report, Excel
questions, a MATLAB code file, group Meeting Minutes, and a Peer Assessment forms
(collated from all members into one file).

QBUS6830, Financial Time Series and Forecasting

The cover sheet outlines all your group members, their names and signatures. I have
created an example of what your cover sheet should include. Feel free to use this one or

In the Written Report, you should provide your answers to all questions below with the title
“Written Report submission questions”. The written report has a limit of 20-pages maximum.
Only one student should submit per group.

Each group must also answer the questions listed within the Excel file. Do NOT submit an
excel file. Simply create a neat tables into your PDF submission which has all of the
“Data and Questions” of “GA1 - Data and Qs.xlsx” and “GA2 - Data and Qs.xlsx”. A template

For the MATLAB code file submission, please submit the code used to produce all your
outputs for both the Written Report and the Excel file submission as a “.m” file.

A template for the group Meeting Minutes will be placed on Canvas and should have entries
for at least 4 group meetings.

Finally, the Peer Assessment Form, requires each group member to assess the
contributions of their fellow group members and will be used to adjust marks in the case
where student contribution differs significantly across group members. Each member of the
group must create their own Peer Assessment form, and these should be collated and then
also submitted.

One person will submit all files for all group members. Only one submission will be accepted
per group. This person is responsible for submitting all the files listed above. Do NOT submit
multiple files. You must only submit one file (zipped) which has all the relevant assignment
files inside of it. If you submit multiple files, your assignment will not be considered “sent”
until all of your files are zipped and within 1 single file.

IMPORTANT:

Submission links will become available in Canvas a week before the due date. Note that the
Written Report will be submitted via Turnitin, the university’s anti-plagiarism software.

QBUS6830, Financial Time Series and Forecasting

Group Assignment Part 0

List your group members, along with their student ID. Calculate X, and therefore which
assignment variation you will be using.

QBUS6830, Financial Time Series and Forecasting

Group Assignment Part I

Obtaining Data and Excel Questions
For this assignment you will need to download the file “GA1 - Data and Qs.xlsx”. The file
can be found in Canvas by clicking on ‘Assignments’ on the left-hand menu and then clicking
in cell ‘B1’ of the first spreadsheet and this will populate the data set and the group-specific
questions for your group. Note the data are based on actual asset and market return series,
however, the dates have been artificially changed for the assignment. Question 1 –
Principle Components and Factor Analyses

Data on the percentage simple returns of 5 assets are provided in the file “GA1 - Data and
Qs.xlsx” for each group in columns D to J of spreadsheet “Data and Questions”.

Principle Components analysis

Conduct a Principle Components analysis on the covariance matrix, , of the 5 Asset simple
return series.

(7 marks) Written Report submission questions

a) (5 marks) Present the results of your analysis in a table and describe the Principle
components found. Do they have a relevant or useful interpretation?
b) (2 marks) How many components would you choose. Justify your answer.

(5 marks) Excel file submission questions

Answer Q1a-Q1e in the excel file (see spreadsheet “Data and Questions” column R).

Factor Analysis

5 Asset simple returns series are provided in the excel file for each group. Estimate factor
analyses on the 5 asset simple return series using m factors where m = 1, 2, …, 5.

(6 marks) Written Report submission questions

c) (3 marks) For the 2-factor model for your group’s assets. Present the results of your
analysis in a table and describe the Principle components found. Do they have a

d) (3 marks) How many factors would you choose to include in your factor model? Justify

(6 marks) Excel file submission questions

Answer Q1f-Q1k in the excel file. (see spreadsheet “Data and Questions” column R).

Question 2 – Time series models and forecasting

Data on the percentage log returns of two assets, as well as the market index, are provided
in the file “GA1 - Data and Qs.xlsx” for each group in columns L to P of spreadsheet “Data
and Questions”.
QBUS6830, Financial Time Series and Forecasting

In question 2 you will be required to use various forecasting methods, assess their accuracy,
and use the forecasts to create portfolios in a dynamic portfolio optimization problem. Each
group will be assigned 5 forecast methods (see spreadsheet “Data and Questions”, cells
R21:R25) and should estimate suitable forecasting models for your two assets’ percentage
log return data, using the in-sample data only. You are then required to generate moving
origin horizon 1 forecasts for each observation in your forecast sample for your group’s 5
methods. Use an expanding data window for your in-sample and you should update your
model estimates daily. Each group’s in-sample and forecast sample is provided in the
spreadsheet “Data and Questions” in cells R18:R19.
Next, you should generate dynamic portfolio weights based on your group’s portfolio
strategies (see Excel file, spreadsheet “Data and Questions”, cells R39:R43) and the
forecast models mentioned above. Dynamic portfolios must be created for each combination
of forecast model and portfolio strategy (i.e. 5×5 = 25 in total). The portfolio weights must be
updated daily. Once the task is completed you are required answer the questions below.

(10 marks) Written Report submission questions

(a) (5 marks) Present the forecast accuracy measures RMSE and MAD for all forecasting
strategies in a table and discuss the performance of the different forecasting
approaches.

(b) (5 marks) Present the returns and standard deviations for all forecasting model and
portfolio strategy combinations. Discuss the performance of the various combinations
of models and strategies.

(c) (20 marks) Excel file submission questions. Answer Q2a-Q2t in the excel file
(spreadsheet “Data and Questions” column R).

QBUS6830, Financial Time Series and Forecasting

Group Assignment Part II

Obtaining Data and Excel Questions
For this assignment you will need to download the file “GA2 - Data and Qs.xlsx”. The file
can be found in Canvas by clicking on ‘Assignments’ on the left-hand menu and then clicking
in cell ‘B1’ of the first spreadsheet and this will populate the data set and the group-specific

Question 3 - Volatility Modelling and Risk estimation

In this assignment you are required to build a range of models for forecasting market risk.
Data on the Open, Low, High, Close, and percentage log returns, for a single asset are
provided in the file “GA2 - Data and Qs.xlsx” for each group in columns F:J of spreadsheet
“Data and Questions”. All cell references in what follows refer to the aforementioned
spreadsheet. Each group is assigned 5 volatility forecast models/methods (see cells L3:L7)
and must estimate forecasting models for their percentage log return series on the in-sample
data utilising information criteria where stated. Each group’s in-sample and forecast sample
periods are provided in the cells L10:L11.

For all your volatility models, examine the accuracy of 1-period ahead volatility forecasts by
comparing them to two volatility proxies across your forecast sample using MAD and RMSE.
Note that once you have chosen models based on your in-sample period you should use
the same models for all of your forecast periods. For example, if you use information criteria
to determine model 2 is an ARCH(3) model based on your in-sample data, then you should
use the ARCH(3) model for all of your model 2 forecasts. The proxies your group must use
are provided in cell L14. Use a fixed-size rolling window approach for generating these
forecasts updating your model estimates daily.

Finally, generate 1-period ahead Value at Risk and Expected Shortfall forecasts, at the 5%
level, for all volatility forecast models, as well as a symmetric CAViaR model, from the end
of the data series (i.e. forecasts are for the day after the last day in your forecast sample).

When estimating a CAViaR forecast for VaRt+1, you need to have an estimate of VaRt, and
to obtain the estimate for VaRt you will need an estimate of VaRt-1, and so on. You will
obviously need to initialise this forecast at some point without the use of the CAViaR. It is
common to run the model for a large number of observations prior to your final forecast so
your first estimate of VaR has little impact on the final forecast. For this reason, you should
conduct the following steps:
a) Obtain an initial 5% VaR estimate as the 5th percentile of the 1st 1000
observations.
b) Run the CAViaR model to obtain forecasts for the last 200 periods in your data
set using a moving window of size 1000 observations updating your
parameters daily. Use the sample percentile from i) to obtain the first forecast
of these 200 forecasts. From then on iteratively use the forecasts for each
successive period to obtain the forecast for the following period up until the
c) Finally obtain the forecast for the next day following your sample period using
the forecast for the last day of your sample period and the most recent CAViaR
parameter estimates
d) For the model specified in cell L6:
QBUS6830, Financial Time Series and Forecasting

i) (5 marks) Provide your estimation results in a table and write down the
estimated model equations. Discuss the statistical significance and interpret
the estimated parameters. (Note: The model presented should be estimated
on the in-sample data used to create your final forecast for period 1201. For
example, if your in-sample data set is of length 1170 observations your
presented estimation results would use observations 31-1200.)
ii) (20 marks) Perform a thorough diagnostic analysis to assess the fit of the
model.
iii) (5 marks) Discuss any component(s) you might add to the model that might
potentially capture any model mis-specifications found; motivate your choices.
iv) (10 marks) Discuss the asymmetry of the fitted model.

e) (5 marks) Present and discuss the volatility forecast accuracy measures for all
your volatility models specified in cells L3:L7. (This does not include CAViaR
as it is not a model for volatility)
f) (5 marks) Present and discuss the 1- period ahead 5% VaR forecasts for all
models specified in cells L3:L7 plus the CAViAR model. (Note: you do not need
to assess the forecast’s accuracy or independence).
g) (5 marks) Present and discuss the 1-period ahead 5% ES forecasts for all
models specified in cells L3:L7. (Note: you do not need to assess the forecast’s
accuracy or independence). (This does not include CAViaR as it is not a model
for ES).
(h) (21 marks) Excel file submission questions. Answer questions a) - u) in the
excel file (spreadsheet “Data and Questions” column L, cells L17:L37).

QBUS6830, Financial Time Series and Forecasting

Some notes regarding group peer assessment

1. Your group will be required to document, using minute form, at least 4 group meetings.
Documentation should be in terms of attendance, discussion points, actions decided, tasks allocated
and/or completed by each member, etc. An example form for this will be distributed OR you may use

2. Peer assessment items are required to be handed in as part of the online submission process. If
you do not complete and hand these in, then you will lose marks individually.

3. At the end of the assignment, everyone will rate BOTH themselves and their other group members
in terms of participation and effort on the assignment. For each individual group member, the total
group mark will either be adjusted (i) downwards; or (ii) upwards; or (iii) remain the same, depending
on my academic judgement of the peer assessment items provided by each individual in each group
and reflecting each individual’s overall contribution to, and effort in, completing the assignment tasks.

3. Based on peer assessment and after having put in a reasonable effort on the group project, the
maximum amount a student can lose from their group mark is 10% of the total mark. However, an
exception to this is if a student has TRULY DONE NOTHING (or close enough to; i.e. not put even
close to a reasonable effort), in which case I will award a mark of 0.

4. If a group is concerned that one or more of their members is not contributing sufficiently to the
assignment please inform the course coordinator and provide any evidence (meeting minutes or
otherwise) to support your claim. If the concern appears valid a warning will be sent to the student(s)
and an immediate penalty of 10% will be imposed. Should the situation miraculously improve the
penalty may be removed later. Should the situation not improve, a mark of 0 is possible as discussed
above.

Email:51zuoyejun

@gmail.com