Financial Network Analytics @ London School of Economics
1. London School of Economics
13 February 2012
Financial Network Analytics
Kimmo Soramรคki | www.fna.fi
kimmo@soramaki.net
2. โWhen the crisis came, the serious limitations of existing economic
and financial models immediately became apparent.
[...]
As a policy-maker during the crisis, I found the available models of
limited help. In fact, I would go further: in the face of the crisis, we
felt abandoned by conventional tools.โ
in a Speech by Jean-Claude Trichet, President of the
European Central Bank, Frankfurt, 18 November 2010
3. Network theory and related fields
Financial
Network
Analysis
Social Network Network
Analysis Science
NETWORK
THEORY
Graph & Matrix Computer
Theory Science
Biological
Network
Analysis
6. Data tsunami
โข Digital information is doubling every 1.2 years -> Big Data, Open
Data, Data Science, โChief Algorithms Officersโโฆ
โข Regulatory response to recent financial crisis was to strengthen
macro-prudential supervision with mandates for more regulatory
data
โข The challenge will be to
understand and analyze the data
โข Analytics based decision-making,
i.e. the application of computer
technology, operational research,
and statistics to support decision
making
Katsushika Hokusai. The great wave off
Kanagawa ~1830
8. Network maps
โข Natural way to visualize the financial system
โข Network visualizations are intuitive
โข Maps aid decision making
โข Mapping of the financial system
has only begun
Eratosthenes' map of the known
world, c. 194 BC.
9. Go is a board game that originated in
China more than 2,000 years ago.
10. Intelligence Amplification
โข Technology, products and
practices change constantly
โข Data is not clean, actions are not
โrationalโ
โข Algorithms donโt fare well in
periods of abrupt change
โข Algorithms do not think outside
the box
โข A solution is to augment human
intelligence
โข Intelligence Amplification vs
Artificial Intelligence
12. Main premise of network analysis:
Structure of links between nodes matters
โข The properties and behavior of a node cannot be analyzed on the
basis its own properties and behavior alone.
โข To understand the behavior of one node, one must analyze the
behavior of nodes that may be several links apart in the network.
โข Bottom up approach. Generalize and describe.
โข Financial contexts
โ Trading networks, payment networks
โ Networks of interconnected balance sheets
โ Networks of asset dependencies
13. Topology of interactions
Degree distribution
Total of ~8000 banks
66 banks comprise 75% of value Soramรคki, Bech, Beyeler, Glass and Arnold
25 banks completely connected (2006), Physica A, Vol. 379, pp 317-333.
14. The New York Times, 1 May 2010
Europe's Web of Debt
15. BBC, 18 November 2011
Eurozone debt web: Who owes what to whom?
(
)
http://www.bbc.co.uk/news/business-15748696
17. Stock correlation networks
โข Cross-asset correlation tree of
German DAX index stocks
โข Assets are linked in a backbone
correlation network
โข Clusters of asset classes clearly
visible
โข Data reduction method
20. Financial Network Analytics (FNA)
FNA is an analytics platform that helps banks and financial
regulators manage and understand financial data with network
analysis
โข Web Service (www.fna.fi) and Enterprise Software
โข Modular design, current modules for:
โ Financial Stability
โ Payment System Oversight
โ Macroprudential Supervision and
โ Cross Asset Analysis
โข Designed for operationalizing latest research
22. Demos
โข Bilateral banking sector exposures by country (2005-
2011)
โ Node = country, Link = banking system exposure
โ Force-Directed layout of network
โข Asset correlation data for DAX stocks (January 2012)
โ Node = stock, Link = coupling via correlations
โ Dynamic Force-Directed layout of spanning tree
โข Dashboards
William Ross Ashby (1956) in โIntroduction to CyberneticsโThe use of information technology in augmenting human intelligence and extending the information processing capabilities of the human mind Visual analytics is an adjacent field