Download Luc Laeven

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.