Network analysis of the e-Mid interbank e-MID market: implications for systemic risk Dataset Market composition reshuffling Vasilis Hatzopoulos and Giulia Iori Entropy City University London Network Metrics The research leading to these results has received funding from the European Union, Seventh Framework Programme FP7/2007-2013 under grant agreement FET Open Project FOC, Nr. 255987. Latsis Symposium 2012, Zurich Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 1 / 54
Introduction • In normal times, interbank markets are among the most e-MID liquid in the financial sector and the financial literature has Dataset Market composition historically devoted a relatively low consideration to the reshuffling interbank market due to the short term nature of the Entropy exchanged deposits. Network Metrics • During the 2007-2008 financial crisis though liquidity in the interbank market has considerably dried up, even at short maturities, and an increasing dispersion in the credit conditions of different banks has emerged. • These events have triggered a new interest in interbank markets. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 2 / 54
Effect of the crisis: market freeze 4 Daily volumes avaraged within a month Daily trades avaraged within a month x 10 3 700 Aggressor to quoter volumes Banks proposing to borrow e-MID Quoter to aggressor volumes Banks proposing to lend 600 2.5 Dataset 500 Market composition 2 Daily volume / 10 6 Euros reshuffling 400 Daily trades 1.5 Entropy 300 Network Metrics 1 200 0.5 100 0 0 J99 J00 J01 J02 J03 J04 J05 J06 J07 J08 J09 D09 J99 J00 J01 J02 J03 J04 J05 J06 J07 J08 J09 D09 Month Month Figure: Left: monthly average of daily volumes. Left: monthly average of daily trades. In both cases trades have been separated into borrow initiate deals and lending initiated deals). Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 3 / 54
Effect of the crisis: dispertion of credit spreads borrowers lenders mean of cross−sectional variance of daily spread e-MID 0.04 0.04 Dataset Market composition 0.03 0.03 reshuffling Entropy 0.02 0.02 Network Metrics 0.01 0.01 0.00 0.00 J99 J01 J03 J05 J07 J09 J99 J01 J03 J05 J07 J09 month month Figure: Montly average of the cross-sectional variance of spreads for borrower (left) and lender(right). Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 4 / 54
The literature has presented two main explanations for the volume collapse (or freeze) in the money market and for the e-MID Dataset raise in spreads during the recent turmoil: Market composition • liquidity hoarding: banks were hoarding liquidity in order reshuffling to anticipate additional money demand, both for internal Entropy needs, and from external operators. Network Metrics • Trust evaporation: banks, rationally or irrationally, perceive an increase in the counterparty-risk and became reluctant to lend. Our analysis is an attempt to quantify the second effect. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 5 / 54
Randomness in matching and Trust Evaporation e-MID Dataset • The question we address is whether banks behaviour Market composition regarding the choice of counter parties in a trade changed reshuffling before and during the subprime crisis. Entropy • In particular we try and quantify the level of randomness in Network Metrics the weights distribution across the links of the credit network. • We interpret this randomness as a proxy of the level of trust among credit institution. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 6 / 54
e-MID: electronic Market for Interbank Deposits e-MID • This market is unique in the Euro area in being a screen Dataset based fully electronic interbank market. Outside Italy Market composition interbank trades are largely bilateral or undertaken via reshuffling voice brokers. Entropy • The central system is located in the office of the SIA and Network Metrics the peripherals on the premises of the member participants. • The names of quoting banks are visible next to their quotes to facilitate credit line checking. A transaction is finalized if the ordering bank accepts a listed bid/offer. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 7 / 54
e-MID Both italian banks and foreign banks can exchange funds. Market players are 246 members from 29 EU countries and the US, of which: e-MID • 30 central banks acting as market observers Dataset Market composition • 2 Ministries of Finance reshuffling • 108 domestic banks Entropy • 106 international banks Network Metrics The number of transactions and the volume increased systemically until the beginning of the financial crisis, with an average of 450 transactions each day and an average exposure of about 5.5 million euros per transaction. According to European Central Bank (2007) e-MID accounted, before the crisis, for 17% of total turnover in unsecured money market in the Euro Area. In the last report on money markets (European Central Bank, 2010), it recorded 12% of the total overnight turnovers. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 8 / 54
e-MID Types of trade: e-MID • Overnight (O/N): Trades for a transfer of funds to be Dataset effected on the day of the trade and to return on the Market composition reshuffling subsequent Business Day; Entropy • Tomorrow next (T/N): Trades for a transfer of funds on the Network Metrics first Business Day following the day of the trade and to return on the second Business Day following that of the trade; • Spot next (S/N): Trades for a transfer of funds on the second Business Day following the day of the trade and to return on the third Business Day following that of the trade; Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 9 / 54
e-MID e-MID • Time Deposits: Trades for an initial transfer of funds and to Dataset return at a predetermined maturity (from 1 week to 12 Market composition months); reshuffling • Broken Date Deposit: Trades with freely agreed Initial Entropy Value Date and Final Value Date between parties without Network Metrics standardization obligations provided that both dates do not coincide with the previous ones and that the two days are not separated by a period superior to a calendar year We only look at ON and ONL transactions!! Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 10 / 54
Dataset • The data base is composed by the records of all transactions registered in the period 01/1999–12/2009 for e-MID a total of 1.523.510 transactions. Dataset • For each contract we have information about the date and Market composition reshuffling time of the trade, the quantity, the interest rate and the Entropy encoded name of the quoting and ordering bank. Network Metrics • The banks are reported together with a code representing their country and, when the bank is Italian, a final label that indicates the class of capitalization (major, large, medium, small, minor) • The aggregate characteristics of the entire set of transactions can thus be studied in terms of the statistical and topological properties of this HIGHLY HETEROGENEOUS network. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 11 / 54
System Heterogeneity: number of banks per group 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 MA 6 7 7 6 6 6 6 6 6 4 4 e-MID GR 12 12 8 7 9 7 7 9 9 6 5 ME 26 26 26 23 17 14 13 13 12 12 10 Dataset PI 68 64 74 61 57 52 53 58 54 53 51 MI 76 59 33 31 28 26 26 14 16 20 19 Market composition FB 2 13 20 31 48 54 60 59 62 58 39 reshuffling Table: Banks active as borrowers per group per year. Entropy Network Metrics 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 MA 6 7 7 6 6 6 6 6 6 4 4 GR 12 12 8 7 9 7 7 9 9 6 4 ME 26 26 26 23 18 14 13 13 11 11 10 PI 75 65 76 66 60 58 55 62 58 54 55 MI 91 70 44 39 34 35 33 17 20 22 22 FB 3 11 21 32 48 57 61 66 70 70 45 Table: Banks active as lenders per group per year. Giulia Iori Network analysis of the e-Mid interbank market: implications for systemic risk September 11-14, 2012 12 / 54
Recommend
More recommend
