Presentation at Payments System Oversight Course Central Bank of Bolivia, i La Paz 25 November 2011 Financial network analysis Kimmo Soramäki kimmo@soramaki.net, www.fna.fi fi
Growing interest in networks 2003 2003 2004
Is network theory the best hope for regulating systemic risk? Meltdown modeling Could agent based computer models prevent another financial crisis?... need for new and fundamental understanding of the structure and dynamics of economic networks. CFA Magazine, July 2009 Nature, August 2009 Science, July 2009
Paradigm In the face of the crisis, we felt abandoned by conventional tools. In the absence of clear guidance from existing analytical frameworks, policy-makers had to place particular reliance on our experience. Judgement and experience inevitably played a key role. Jean-Claude Trichet, (Now former) President of the ECB, Opening address at the ECB Central Banking Conference, Frankfurt, 18 November 2010
Network maps Recent financial crisis brought to light the need to look at links between financial institutions Natural way to visualize the financial system Network thinking widespread ead by regulators Mapping of the financial system has only begun Eratosthenes' map of the known world, c.194 BC.
Visualising financial networks Bech, M.L. and Atalay, E. (2008), The Topology of the Federal Funds Market. ECB Working Paper No. 986. Italian money market 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 Federal funds Unsecured sterling money market Wetherilt, A. P. Zimmerman, and K. Soramäki (2008), The sterling unsecured loan market during 2006 2008: insights from network topology, in Leinonen (ed), BoF Scientific monographs, E 42 Fedwire Soramaki, K, M.L. Bech, J. Arnold, R.J. Glass and W.E. Beyeler (2007), The topology of interbank payment flows, Physica A, Vol. 379, pp 317 333, 2007.
Europe's Web of Debt (Bill Marsh / The New York Times, 1 May 2010)
Network theory Social Network Analysis Graph & Matrix Theory Financial Network Analysis NETWORK THEORY Biological Network Analysis Network Science Computer Science
Main premise of network analysis: the structure of the links between nodes matters The properties and behaviour of a node cannot be analysed on the basis its own properties and bh behaviour alone. To understand the behaviour of one node, one mustanalyse the behaviour of nodes that maybe several links apart in the network. Financial context: network of interconnected balance sheets
Network basics Terminology node/vertex > Bank/banking group link/tie/edge/arc > Financial interlinkages, bilateral positions, exposures directed vs undirected weighed vs unweighted graph + properties = network 2 Algorithms/measures Centrality > Systemical importantance 1 Flow > Liquidity Community/pattern identification Distance, shortest paths Connectivity, clustering Cascades, epidemic spreading > Contagion 3 4
Prototypical topologies lattice complete random Ring Weighted random Scale-free 11
Homophily Birds of one feather flock together, th herd behaviour Ideas, attributes, etc tend to cluster together and enforce each other Examples: Some obvious bi (age, social ilstatus), tt others less (obesity, happiness, divorces) How about: risk appetite, portfolio decisions, etc. Spread of obesity Small world phenomenon Six degrees of separation (6.6 on MSN messenger) The shortest path between any two nodes is very short Implications for contagion? Robust yet fragile, Scale free networks The removal of "small" nodes does not alter the path structure of the remaining nodes, and thus has no impact on the overall network topology. Probability (log g) Nicholas A. Christakis, James H. Fowler New England Journal of Medicine 357 (4): 370 379 (26 July 2007) Fedwire degree distribution Degree (log) 12
Network analysis for Oversight Mostly interested in contagion process, high policy interest for measures of systemic importance A growing body of empirical research on financial networks Interbank payments: Soramäki et al (2006), Becher et al. (2008), Boss et al. (2008), Pröpper et al. (2009), Embree and Roberts (2009), Akram and Christophersen (2010) Overnight loans: Atalay and Bech (2008), Bech and Bonde (2009), Wetherilt et al. (2009), Iori et al. (2008) and Heijmans et al. (2010), Craig & von Peter (2010) Flow of funds, Credit registry, Stock trading : Castren and Kavonius (2009), Bastos e Santos and Cont (2010), Garrett et al. 2011, Minoiu and Reyes (2011), (Adamic et al. 2009, Jiang and Zhou 2011) More at www.fna.fi/blog
Possible objectives Identify important/vulnerable banks Identify contagious links Understand contagion process
Common centrality measures Degree: number of links Closeness: distance to other nodes via shortest paths Betweenness: number of shortest paths going through the node Eigenvector: nodes that are linked by other important nodes are more central, probability of a random process
Centrality depends on network process Trajectory Transmission geodesic paths, paths, trails or walks parallel/serial duplication or transfer Source: Borgatti (2004) 16
Components.. a Giant In Component (which flows into GSCC) The GWCC has a Giant Strongly Connected Component (each node can reach each other) And several disconnected components.. and a Giant Out Component (to whichgscc flows) Empirical networks often consist of a Giant Weakly Connected Component 17
Intelligence Can we algorithmically detect financial crisis? Payment data as early warning tool? Financial crisis are different and rare. The patterns to be recognized must be frequent enough for computers to learn Pattern recognition is hard for computers -> the best Go programs only manage to reach an intermediate amateur level A solution is to augment human intelligence (in contrast to AI) Intelligence amplification (William Ross Ashby 1956)
Interactive visualization examples
Ring layout
Force-directed layout
Chord layout 22
Tools Pajek, Universty of Ljublana, Slovenia Focus on social network analysis of large networks pajek.imfm.si Gephi, Gephi Foundation, France Focus on graph visualisation Like Photoshop for graphs www.gephi.org FNA, Soramaki Networks, Finland Focus on Financial/Payment Networks, time-series, dashboards www.fna.fi Many others Cytoscape, Graphviz, Network Workbench, NodeXL, ORA, Tableau, Ucinet, Visone, etc.
Thank you