Coursework for MM917
“Networks in Finance”
The goal of this coursework is that you prepare a report. It will
be 30% of your mark. There are two parts of the report. The
first part is about the analysis of research papers on
Networks in Finance and the second part is about analysis of
network data.
Part 1: Analysis of research papers on networks in finance
Please find the attached file (Papers.zip) on Myplace. The file
contains all the review papers and research papers. Each of
you needs to read three review papers and an assigned
working paper to write your report.
Review papers
The Review papers will provide you general information on
the field of financial networks. You must begin your report
with a general introduction based on the review papers
provided. Such general introduction must describe the
state-of-the-art of the use of networks in
finance and
particularly the importance of network analysis for the study
of systemic risk and contagion in financial networks.
Working papers
Each of you will read and analyze one of working papers. The
list of the randomly assigned working paper is given the table
below. You need to analyze your own working paper and
introduce your understanding of the paper to others. This
analysis should contain the following points in your language:
A. The motivation: What did the paper look for? What is the
research question and why it is important?
B. The methods: How did the authors answer to the
research question? Why they choose the methods?
C. The results: What did the authors find, measure, derive,
or analyze?
D. The implications: So, what can we learn? With the
results, what we can do?
Tips for the part 1
1) I understand that some papers may be difficult for you to
understand. BUT DO NOT WORRY too much if you do not
understand your paper perfectly. Please note that the points
of your report are:
A. To demonstrate the essence of your paper not the
technical details. This means that you need to understand the
motivation, the background, the importance, the methods, the
main results, and the implications of your paper.
B. To deliver your understanding to others in your
language.
2) One way to understand the paper better is to read the
related papers together. Normally you can find such related
papers in the introduction of the paper.
3) Read the paper with your colleagues and explain to each
other.
How to write the part 1 of Report
Write a report containing:
1) General Introduction (based on review papers)
2) Introduction (of the working paper)
3) Methods
4) Results
5) Discussion
6) Conclusions
7) References
8) Appendices (Anything you want to add)
Think like if you were explaining this paper to someone
not familiar with the subject.
In general introduction, describe the state-of-the-art of
the use of networks in finance and in particular the
importance of network analysis for the study of systemic risk
and contagion in financial networks (in one page).
In introduction, demonstrate what kind of problems the
authors consider and why
these problems are important.
Describe the method of your research paper in the
Methods in a clear and direct
language and use the authors’
mathematical expressions if necessary.
State clearly what are the results with figures and what
these results imply.
Discuss the results in a wider perspective. For example,
what could we learn from
this paper, what the results mean
in finance, and what are the limitations of the paper.
Make clear conclusions in one or two paragraphs.
Conclusions are not a list of the topics that you have
considered in your report but what you have learned from the
analysis of this paper and what are the answers the authors
asked in the introduction,
which would be interesting for the
readers.
Refer to all the literatures that you have used, including
webpages. The references
must be properly cited in the text,
e.g., using numbers. The list of references must include the
authors, title of the paper, journal, volume, year and pages.
Part 2: Analysis of network data
Please find the attached file (NetworkData.zip) on Myplace.
Each of you will have one network data set. You can analyze
this network data with your own codes or existing programs.
Specify what you use in your report. Include any program or
code used into the Appendices as well as any secondary
information. If the code is taken from a given site please
refers to it in the references. Otherwise make clear you have
created this code.
For a given network:
[1] Measure network size, number of links, average degree,
the standard deviation of degree.
[2] Plot degree distribution in linear-linear scale and in log-log
scale.
[3] Measure the average local clustering coefficient
(Newman’s) and plot the clustering coefficient as a function of
degree (i.e., local clustering coefficient of nodes with the
same degree k).
[4] Plot the average nearest neighbor’s degree as function of
degree (i.e. degree correlation function) and measure the
Newman’s assortativity coefficient (i.e., degree correlation
coefficient). We will learn about this subject on 23/02/2018.
[5] Demonstrate your codes or describe how to use your
program to analyze the data.
The following figures illustrate how to plot, as an example.
Here P(k) is the degree distribution. C(k) is the average
clustering coefficient as a function of degree k. is
the average nearest neighbor’s degree as function of degree
k.
-The network data format: 2-columns, Undirected
network):
1 2 (this means node “1” is connected to node “2”)
1 3 (this means node “1” is connected to node “3”)
34
...
-If there is 1 2 in the data, there will be also 2 1 because all
the data are undirected one. But note that you need to
consider 1 2 and 2 1 as a single link. Check your program
before use because, depending on the codes you use, some
codes may consider this case as two links.
About the programs to be used
There are some codes available for manipulating and making
calculations of parameters on the web. It is a non-exhaustive
list of sites where you can find a few of them:
a. http://strategic.mit.edu/downloads.php?page=matlab_networks
b. http://igraph.org/redirect.html (python, R, C/C++)
c.http://www.levmuchnik.net/Content/Networks/ComplexNetworksP
ackage.html
d. http://www.mathstat.strath.ac.uk/outreach/contest/toolbox.html
f.http://www.mathworks.com/matlabcentral/fileexchange/4266-grthe
ory-graph-theory-toolbox
Further suggestions
1. Please use font size 11 (or more).
2. For the part 1, do not make the first part of the report too
long (Without references
and appendices, 5~6 pages
are recommended)
3. Start early and prepare the report with sufficient time. If
you start the report just few
days before the due date,
it is really difficult to write a nice report.
4. Before you analyze the network data, test your program
with small artificial network
you make. For example. If
you have only two nodes with one link in the artificial
data, you program should report you that there are two
nodes and one link. Otherwise, your program is wrong.