Financial Cartography - PRMIA Webinar


Published on

As the financial system becomes more complex, new methods to understand the inherent risks and dynamics are needed. Kimmo Soramäki will discuss how network analysis of large‐scale financial transaction data can be used to improve our understanding systemic risk. He will also show case studies how visual analytics and accurate data driven maps of asset correlations and tail risks can enable a stronger intuition of market dynamics.

Published in: Economy & Finance
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Financial Cartography - PRMIA Webinar

  1. 1. Financial Cartography February 6, 2013 – 12 p.m. U.S. Eastern Time Kimmo Soramäki Founder and CEO of Financial Network Analytics (FNA)• Audio: Use your microphone and speakers (VoIP) or call in using your telephone.• Direct your questions to Staff via the Questions or Chat pane.• To access this webinar audio via the internet, select “Mic & Speakers” under your Audio pane.• Check that the audio on your computer is on and the volume is turned up.• For technical assistance contact the Citrix webinar utility customer number: 1-888-259-8414This material is the intellectual property of the presenterand shall not be reproduced or used without the express written permission .
  2. 2. PRMIA Webinar6 February 2013Financial Cartography Dr. Kimmo Soramäki Founder and CEO FNA,
  3. 3. “When the crisis came, the serious limitations of existingeconomic and financial models immediately became apparent.[...]As a policy-maker during the crisis, I found the availablemodels of limited help. In fact, I would go further: in the face ofthe crisis, we felt abandoned by conventional tools.” in a Speech by Jean-Claude Trichet, President of the European Central Bank, Frankfurt, 18 November 2010 3
  4. 4. We did not have maps … 4
  5. 5. Eratosthenes map of the known world5c. 194 BC
  6. 6. … but what are maps“A set of points, lines, and areasall defined both by position withreference to a coordinate systemand by their non-spatialattributes”Data is encoded as size, shape,value, texture or pattern, colorand orientation of the points,lines and areas – everything hasa meaning Political map of Europe 6
  7. 7. … but what are maps (contd.)Cartographer selects onlythe information that isessential to fulfill thepurpose of the mapMaps reducemultidimensional data intoa two dimensional spacethat is better understood byhumansMaps are intelligenceamplification, they aid indecision making and build Map by John Snow showing the clusters of cholera cases in the London epidemic of 1854intuition 7
  8. 8. I. Mapping II. MappingSystemic Risk Financial Markets 8
  9. 9. Systemic risk ≠ systematic risk News articles mentioning “systemic risk”, Source: risk that a system composed of many interactingparts fails (due to a shock to some of its parts).In Finance, the risk that a disturbance in the financialsystem propagates and makes the system unable toperform its function – i.e. allocate capital efficiently. Not:Domino effects, cascading failures, financialinterlinkages, … -> i.e. a process in thefinancial network 9
  10. 10. Network Theory can be to Financial Mapswhat Cartography is to Geographic MapsMain premise of network theory:Structure of links between nodesmattersTo understand the behavior of onenode, one must analyze thebehavior of nodes that may beseveral links apart in the networkTopics: Centrality, Communities,Layouts, Spreading and generationprocesses, Path finding, etc. 10
  11. 11. Network aspect is anunexplored dimension of data Variables Observations 11
  12. 12. First Maps Fedwire Interbank Payment Network, Fall 2001 Around 8000 banks, 66 banks comprise 75% of value,25 banks completely connected Similar to other socio- technological networksSoramäki, Bech, Beyeler, Glass and Arnold (2007), M. Boss, H. Elsinger, M. Summer, S. Thurner, ThePhysica A, Vol. 379, pp 317-333. network topology of the interbank market, SantaSee: Fe Institute Working Paper 03- 12 10-054, 2003.
  13. 13. More Maps: Federal Funds1997 - 2006 Source: Bech, M.L. and Atalay, E. (2008), “The Topology of the Federal Funds Market”. ECB Working Paper No. 986.• 2600 loans worth $335 billion per day• First Circle: 165 Second Circle: 271 Rest: 42 13
  14. 14. More Maps: Italian money market Italian (very small) Italian (small) Italian (large) Foreign Source: Iori G, G de Masi, O Precup, G Gabbi and G Caldarelli (2008): “A network analysis of the Italian overnight money market”, Journal of Economic Dynamics and Control, vol. 32(1), pages 259-278 14
  15. 15. More Maps: DebtRankAugust 2007 to April 2008 October 2008 to April 2010Nodes: Financial institutions Source: Battiston et al, NatureLinks: Impact of an institution to another Scientific Reports 2-54, 2012Nodes closer to center are more important (as are big and red) 15
  16. 16. Where are we today?Regulatory response to recent financial crisis was to strengthenmacro-prudential supervision with mandates for more regulatorydata“Big data” and “Complex Data”-> Providing tools and challenge tounderstand, utilize and operationalize the dataFinancial Networks are starting to get their own literature andmetrics different from other fields of Network Theory 16
  17. 17. Case: Oversight MonitorThe monitor will allow theidentification of systemicallyimportant banks and evaluation ofthe impact of bank failures on theinterbank payment system (network is fictional) Intraday Liquidity Network -exampleThe visualizations are available at 17
  18. 18. Polling Question 1Which types of networks are most important forfinancial institutions and regulators?1) Exposure/contagion networks2) Trade/payment networks3) Supply chain networks4) Social networks 18
  19. 19. I. Mapping II. MappingSystemic Risk Financial Markets 19
  20. 20. OutlinePurpose of the maps – Identify price driving themes and market dynamics – Reduce complexity – Spot anomalies – Build intuitionThe maps: Heat Maps, Trees, Networksand Sammon’s ProjectionsBased on asset correlations or taildependenceThese methods are showcased forvisualizing markets around the collapseof Lehman Brothers 20
  21. 21. Collapse of LehmanLehman was the fourth largest investmentbank in the US (behind Goldman Sachs,Morgan Stanley, and Merrill Lynch) with26.000 employeesAt bankruptcy Lehman had $750 billion debtand $639 billion assetsCollapse was due to losses in subprimeholdings and inability to find funding due toextreme market conditionsIs seen as a divisive point in the 2007-2009financial crisis 21
  22. 22. The Data Pairwise correlations of return on 118 global assets in 4 asset classes • Stock Exchange Indices (e.g. Dow Jones) • Foreign Exchange Rates (e.g EUR/USD) • Government Bonds (e.g. Irish 10 year bond) • Corportate Bonds (e.g. EMU Corporate AAA, 1-3 years) 22
  23. 23. i) Heat Maps JanuaryCorporate 2007BondsFX RatesGovernmentBond Yields Correlation -1StockExchange 0Indices +1 23
  24. 24. January 2007 t-2 t-1CorporateBondsFX RatesGovernmentBondsStocks t+1 t+2CorporateBondsFX RatesGovernmentBondsStocks 24
  25. 25. ii) Asset TreesOriginally proposed by Rosario Mantegna in 1999Used currently by some major financial institutionsfor market analysis and portfolio optimization andvisualizationMethodology in a nutshell 1. Calculate (daily) asset returns 2. Calculate pairwise correlations of returns 3. Convert correlations to distances 4. Extract Minimum Spanning Tree (MST) 5. Visualize 25
  26. 26. Minimum Spanning TreeA spanning tree of a graph is a subgraph that:1. is a tree and2. connects all the nodes togetherLength of a tree is the sum of its links. Minimum spanning tree (MST) is a spanningtree with shortest length.MST reflects the hierarchical structure of the correlation matrix
  27. 27. Demo: Asset Trees Color of node denotes asset class: Dow Jones Size of node reflects volatility (variance) of returns Ireland 10 year Links between nodes reflect government bond backbone correlationsEMU CorporateAAA, 1-3 years - short link = high correlation - long link = low correlation EUR/USD The visualizations are available at 27
  28. 28. Correlation filtering PMFGBalance between too much and too littleinformation, signal vs noiseOne of many methods to create networksfrom correlation/distance matrices – PMFGs, Partial Correlation Networks, Influence Networks, Granger Causality, Influence Network NETS, etc.New graph, information-theory, economics& statistics -based models are beingactively developed 28
  29. 29. iii) NETS• Network Estimation for Time- Series• Forthcoming paper by Barigozzi and Brownlees• Estimates an unknown network structure from multivariate data• Based on partial correlations• Captures both comtemporenous and serial dependence (partial correlations and lead/lag effects) 29
  30. 30. iv) Sammon’s ProjectionProposed by John W. Sammon in IEEE Transactions on Computers 18: 401–409(1969)A nonlinear projection method to map ahigh dimensional space onto a space oflower dimensionality. Example: Iris Setosa Iris Versicolor Iris Virginica 30
  31. 31. Demo: Sammon Projection EMU Corporate AAA, 1-3 years Color of node denotes asset class:Dow Jones Size of node reflects volatility Ireland 10 year (variance) of returns government bond Distance between nodes reflects EUR/USD similarity of correlation profiles - close = similar - far apart = different 31The visualizations are available at
  32. 32. Tail dependence• Correlation is a linear dependence. The same visual maps can be extended to non-linear dependences.• Joint work with Firamis (Jochen Papenbrock) and RC Banken (Frank Schmielewski), see• Instead of correlation, links and positions measure similarity of distances to tail losses Tail Tree Tail Sammon (Click here for interactive visualization) (click here for interactive visualization) 32
  33. 33. Polling QuestionWhere can financial data visualization providemost value?1. Data validation and exploration2. Enhancing intuition3. Add on to statistical analysis4. Risk Management5. Trading 33
  34. 34. “In the absence of clear guidance from existing analyticalframeworks, policy-makers had to place particular reliance onour experience. Judgment and experience inevitably played akey role.” in a Speech by Jean-Claude Trichet, President of the European Central Bank, Frankfurt, 18 November 2010 34
  35. 35. Blog, Library and Demos at www.fna.fiDr. Kimmo Soramäkikimmo@soramaki.netTwitter: soramaki
  36. 36. FINANCIAL NETWORKS – LESSONS FOR RISK MANGEMENTTo learn more about this topic and attend a livetraining course with Dr. KimmoSoramaki, please click on the following coursesession for more information and how toregister. NEW YORK April 26, 2013 One-Day Training Course
  37. 37. Questions for the Presenter?Send them in now by using your Question Pane in the webinar utility panel Did you know that Sustaining Members attend thought leadership webinars at no additional cost? Find out more about Sustaining Membership at Find upcoming and recorded webinars at 37
  38. 38. Thank you for attending this PRMIA Webinar! Go to to find a full schedule of upcoming webinars 38