Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Financialization wikipedia , lookup
Securitization wikipedia , lookup
Quantitative easing wikipedia , lookup
History of the Federal Reserve System wikipedia , lookup
Systemic risk wikipedia , lookup
Fractional-reserve banking wikipedia , lookup
Interbank lending market wikipedia , lookup
Are banks too large and complex? Luc Laeven with Lindsay Mollineaux, Lev Ratnovski, Yangfan Sun, and Hui Tong (IMF Research R h Department) D t t) The views expressed here are our own and do not reflect those of the IMF or IMF Board Motivation • Large/complex banks were at center of crisis – Different from S&L crisis • Debate on optimal financial structure and TBTF policies – Are A banks b k too large l and d complex? l ? – Consequences for broader economy? Regulatory proposals • Main a – Size: Capital surcharges for SIFIs (Basel) – Scope: Activity restrictions (Vickers/Volcker/Liikanen) – Funding: Caps on wholesale funding (Basel LCR/NSFR) • But how to choose and reconcile? – Understand market failures and identify sources of systemic risk Background • Financial deregulation and innovation led to: – Concentration: Large banks grow in size – New instruments: Securitization, OTC derivatives, secured (repo) funding – Blurred boundaries between banks and markets: “securitized banking banking” (illiquid loans become tradable) tradable), scalable trading activities, wholesale funding – Increased systemic y risk? Securitization reduces bankspecific risk but increases interconnectedness Mixing of banks and markets: the rise of shadow banking Increasing share of “non-bank” non-bank activities in US BHC assets assets, 2001-2012 18 16 14 Non Bank Assets USD Trilliion 12 10 8 6 Commercial Bank Assets 4 2 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago, Federal Reserve Flow of Funds Market failures • Safety net subsidies promote excessive risk taking, especially for TBTF banks – Promotes size, complexity, and leverage – Long-standing prudential issue intensified by changing financial structure • Banks do not internalize externalities of failure • Coordination failures/asymmetric information in wholesale h l l ffunding di markets k t (“ (“repo run”) ”) Sources of systemic risk? • Size, complexity (market (market-based based activities), wholesale funding, and leverage – All grew prior to the crisis esp. for large banks – All are correlated with measures of systemic risk (CoVaR, MES) • Explain large part of variation in systemic risk – Glass half full or half empty • C Correlation l i b between b bank-specific k ifi and d systemic i risk ik is low Are large banks special? • Large banks very different from small banks – – – – Market-based business model More hard-information loans More trading assets More securitization, wholesale funding • Contribute more to systemic risk • No clear economies of scale – But funding cost advantage • Heterogeneity among large banks Increased importance of non-interest income and trading: interconnectedness with markets Ratios of the Largest US BHCs and European Banks Banks, 1994-2011 1/ 0.65 06 0.6 Non-interest Income / Total Income US 0.65 Loans / Asset Ratio 06 0.6 0.16 0 15 0.15 0.55 0.55 0.5 0.5 0.13 0.45 0.45 0.12 0.4 0.4 0.35 0.35 0.1 0.3 0.09 0 25 0.25 0 08 0.08 0.3 Europe 0 25 0.25 199419961998200020022004200620082010 US Europe 199419961998200020022004200620082010 Trading Assets / Total Assets 0.14 US 0.11 1995 1997 1999 2001 2003 2005 2007 2009 2011 Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago, and Bankscope 1/ For France, Germany, UK, top 4 banks. For US, top 50 BHCs. More repos and wholesale funding: interconnectedness with markets and other FIs Ratios of the Largest US BHCs and European Banks, 1991-2012 1/ 0.3 0.25 Repos and Fed Funds Purchased and Sold / Total Assets Top 4 US Non-Deposit Liabilities / Total Liabilities 0.8 Europe 0.7 0.6 0.2 0.5 0.15 Top 50 US 0.4 0.1 0.3 0.05 0.2 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 US 1994 1996 1998 2000 2002 2004 2006 2008 Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago, and Bankscope 1/ France, Germany, UK. US BHCs with assets in excess of $500 million. 2010 Funding cost advantage of large banks Ratio of deposit funding cost, US BHCs by size group, 1995-2012 1/ 0.04 Interest Expense on Deposits / Total Deposits 0.035 0.03 0.025 0.02 0.015 0.01 Smallll S Large (Top 50) 0.005 0 1995 1997 1999 2001 2003 2005 2007 2009 2011 Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago 1/ BHCs with over $500 million in assets Large banks contribute more to systemic risk Average ∆CoVaR, US BHCs, by size group, 1991-2010 0 Small Medium Large -0.01 -0.02 -0.03 -0.04 -0.05 -0.06 -0.07 -0.08 1991 1993 1995 1997 Top 4 BHCs 1999 2001 All Other BHCs in Top 50 2003 2005 2007 2009 All Other BHCs Bank i’s ∆CoVaR is the VaR of the banking g system y conditional on bank i being g in distress compared to when bank i is in its median state, and indicates the marginal contribution of bank i to the banking system’s overall systemic risk Source: Adrian and Brunnermeier Bank-specific Bank specific ≠ systemic risk Correlation -ΔCoVaR σ(r ( E) 0 17*** 0.17*** μ(rE) 0.02 US BHCs, 1991-2010 *** denotes statistical significance at the 1% level Conclusions • Large banks are too large and complex – Create externalities (systemic risk) – Size and complexity grew over time – Trading, securitization, and wholesale funding pose significant systemic risk • Banks have no incentives to shrink – TBTF rents (cheaper funding) – Managerial M i l iincentives ti (“ (“empire i b building”) ildi ”) • Need to deal simultaneously with size, complexity, and leverage (they are related but not equivalent) Policy • Improve resolution frameworks – reduce TBTF subsidies; hard to accomplish • Just more capital – effective, but blunt: if too high / not targeted can be costly • Quantity-based tools (Volcker/Vickers/Liikanen) – hard h d to di distinguish i i hb between llending di and d trading di • Price-based tools (SIFI surcharge) – targeted t t db butt optimal ti l llevell tto b be d determined t i d • Macroprudential regulatory approach – reduce systemic risk of whole financial system Additional charts Banking assets outpaced GDP Ratios of the Largest US BHCs and European banks to GDP, 1994-2012 1/ Euro and US BHC Assets / GDP 3 2.5 Largest US BHCs / GDP 1.2 Europe 2 0.8 1.5 0.6 1 US 05 0.5 T 50 US Top 1 0.4 Top 4 US 0.2 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago, and Bankscope 1/ France, Germany, UK. US BHCs with assets in excess of $500 million. 2012 Bank concentration increased as large banks grew in size Top 4 and Top 50 concentration ratio, US BHCs, 1995-2012 1 0.9 Top 50 US 0.8 0.7 0.6 05 0.5 0.4 0.3 0.2 Top 4 US 0.1 0 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago 1/ Population is US BHCs with assets in excess of $500 million. Increase in leverage prior to crisis: more so in Europe Leverage (A/E) of largest US BHCs and European 1/ banks, banks 1994-2011 40 35 30 Europe u ope 25 20 15 10 5 US 0 Notable difference between Europe and US: Balance sheet data mask risk transferred through g securitization and exposures p reported off balance sheet Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago, and Bankscope 1/ France, Germany, UK. US BHCs with assets in excess of $500 million. Hard to close large banks (too big to fail) S&L and d subprime b i crisis i i compared d Assets of Failed US Banks Relative to US GDP, 1933-2013 Total Assets in Failed Institutions / US GDP (blue line) Assets / GDP 3.5 Number of Institutions 600 3 500 2.5 400 2 300 1.5 200 1 100 0.5 0 0 1934 1941 Source: US FDIC. 1948 1955 1962 1969 1976 1983 1990 1997 2004 2011 Absence of scale economies Cost efficiency, average across BHCs by size group, 1994-2002 Stochastic frontier model Cost function with (pos) inefficiency Time-invariant Time-varying Size group No cross- Cross- No cross- Cross(total assets) products products products products < 1 billion 0.16 0.29 0.18 0.43 1 - 2 billion 0.15 0.27 0.19 0.41 2 - 10 billion 0.14 0.25 0.18 0.38 10 - 50 billion 0.13 0.24 0.17 0.37 > 50 billion billi 0 15 0.15 0 30 0.30 0 19 0.19 0 41 0.41 Source: Staff calculations based on US BHC data Overall, large banks do not generate h h returns ffor equity investors higher Average annualized cumulative bank equity returns US BHCs, by size group, 1995-2012 0.8 0.6 L Large 0.4 Small 0.2 Medium 0 -0.2 -0.4 All Other BHCs All Other BHCs in Top 50 Top 4 BHCs Source: FFIEC Central Data Repository, Federal Reserve Bank of Chicago 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993 1992 1991 -0.6 Drivers of ΔCoVaR (systemic risk) Table 2. Correlation of ∆CoVaR and bank characteristics , US BHCs, 1994-2010 Total assets Non-interest income ratio Income diversity Non-deposit Gross repos to Asset diversity liabilities ratio assets Non-interest income ratio 0.19* 0.19 Income diversity 0.20* 0.80* Asset diversity 0.55* 0.31* 0.19* Non-deposit liabilities ab es ratio a o 0.31* 0 3 0.31* 0 3 0.20* 0 0 0.42* 0 Gross repos to assets 0.24* 0.17* 0.18* 0.38* 0.54* -ΔCOVAR 0.19* 0.25* 0.22* 0.20* 0.17* * Significant at the 1% level 0.16* Large banks enjoy support from strong sovereigns: TBTF subsidies Ordered Probit regression, regression years 2007 and 2009, 2009 international sample of stock exchange listed deposit-taking banks 2007 Dependent variable: S Support tR Rating ti Fl Floor Sovereign rating 2009 (1) (2) (3) (4) (1) (2) (3) (4) 0.16*** 0.06 -0.19** -0.17 0.16*** 0.07* -0.35*** -0.30* Tier 1 capital ratio -0.03 -0.03 Deposits to assets ratio 0.01 -0.00 Loans to assets ratio -0.00 -0.01 Ln(Assets) 0.76*** Sovereign rating * Ln(Assets) Number of banks Pseudo R-squared 129 0.07 129 0.22 0.03 0.08 0.06*** 0.06 0.06* 0.06 129 0.24 114 0.26 0.830*** -0.348 129 0.07 129 0.23 -0.19 0.11*** 0.11 0.09** 0.09 129 0.28 117 0.30 Note: ***, **, * significant at the 1%, 5%, 10% level T-test indicates that coefficient on interaction term in regressions (3) and (4) are significantly different at 1% level across two subsamples.