Russia Risk ConferencePRMIA and CbondsMoscow, 21 November 2012Network view tomarket riskCorrelation and Tail networks     ...
Network Theory and Financial CartographyMain premise of network theory:           Maps reduce data and encodeStructure of ...
OutlineMaps enable visual insights from          The maps: Heat Maps, Trees,complex financial data                    Netw...
i) Heat Maps                 2004-2007CorporateBonds                 151 assets in 5                 asset classes:CDS on ...
Collapse of Lehman, t=month2004-2007           t-2                   t-1   t+1             t+2                    t+3
ii) Asset TreesOriginally proposed by Rosario Mantegna in 1999Used currently by some major financial institutionsfor marke...
DemoClick here for interactive visualization   7
Correlation filteringBalance between too much and too     New graph, information-theory,little information (Tumminello,   ...
iii) Sammon’s ProjectionProposed by John W. Sammon in IEEE Transactions on Computers 18: 401–409(1969)A nonlinear projecti...
DemoClick here for interactive visualization   10
Tail dependence•   Correlation is a linear dependence. The same visual maps can be extended    to non-linear dependences.•...
Blog, Library and Demos at www.fna.fiDr. Kimmo Soramäkikimmo@soramaki.netTwitter: soramaki
Upcoming SlideShare
Loading in …5
×

Network View to Market Risk

650 views
526 views

Published on

Slides from panel presentation at Russia Risk Conference in Moscow on 21 Novemeber 2012

1 Comment
2 Likes
Statistics
Notes
No Downloads
Views
Total views
650
On SlideShare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
24
Comments
1
Likes
2
Embeds 0
No embeds

No notes for slide

Network View to Market Risk

  1. 1. Russia Risk ConferencePRMIA and CbondsMoscow, 21 November 2012Network view tomarket riskCorrelation and Tail networks Dr. Kimmo Soramäki Founder and CEO FNA, www.fna.fi
  2. 2. Network Theory and Financial CartographyMain premise of network theory: Maps reduce data and encodeStructure of links between nodes relevant data to graphical elementsmatters within a representation system + Network Theory provides the representation system for financial maps like Cartography does to geographic maps 2
  3. 3. OutlineMaps enable visual insights from The maps: Heat Maps, Trees,complex financial data Networks and Sammon’s Projections – Reduce complexity – Identify price driving themes and market dynamics – Spot anomalies – Build intuition – Aid communication of resultsThese methods are showcased forvisualizing correlations among a widerange of assets around the collapse ofLehman Brothers on 15 September2008 3
  4. 4. i) Heat Maps 2004-2007CorporateBonds 151 assets in 5 asset classes:CDS on equities, gov.GovernmentDebt bonds, corp. bonds, cds and foreignFX Rates exchangeGovernmentBond Yields Correlation +1StockExchange 0Indices -1 4
  5. 5. Collapse of Lehman, t=month2004-2007 t-2 t-1 t+1 t+2 t+3
  6. 6. ii) Asset TreesOriginally proposed by Rosario Mantegna in 1999Used currently by some major financial institutionsfor market analysis and portfolio optimization andvisualizationMethodology in a nutshell MST 1. Calculate (daily) asset returns 2. Calculate pairwise Pearson correlations of returns 3. Convert correlations to distances 4. Extract Minimum Spanning Tree (MST) 5. Visualize 6
  7. 7. DemoClick here for interactive visualization 7
  8. 8. Correlation filteringBalance between too much and too New graph, information-theory,little information (Tumminello, economics & statistics -based modelsLillo, Mantegna 1999) are being actively developed E.g.: Network Estimation for Time- Series (Barigozzi & Brownlees)One of many methods to createnetworks from correlation/distancematrices (PMFGs, PartialCorrelation Networks, InfluenceNetworks, Granger Causality, etc.) 8
  9. 9. iii) 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 9
  10. 10. DemoClick here for interactive visualization 10
  11. 11. 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 www.extreme-value-theory.com• 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) 11
  12. 12. Blog, Library and Demos at www.fna.fiDr. Kimmo Soramäkikimmo@soramaki.netTwitter: soramaki

×