Structural Changes in Banking After Crisis

Committee on the Global
Financial System
CGFS Papers
No 60
Structural changes in
banking after the crisis
Report prepared by a Working Group established by the
Committee on the Global Financial System
The Group was chaired by Claudia Buch (Deutsche
Bundesbank) and B Gerard Dages (Federal Reserve Bank
of New York)
January 2018
JEL Classification: G21, G24, G32, L25

This publication is available on the BIS website (www.bis.org).
© Bank for International Settlements 2018. All rights reserved. Brief excerpts may be
reproduced or translated provided the source is stated.
ISBN 978-92-9259-131-1 (online)

Preface
The experience of the global financial crisis, the post-crisis market environment and
changes to regulatory frameworks have had a marked impact on the banking sector
globally. In response to their new operating landscape, banks have been re-assessing
and adjusting their business strategies and models. At the same time, a number of
advanced economy banking systems have to confront low profitability and legacy
problems.
Against this background, the Committee on the Global Financial System (CGFS)
mandated a Working Group chaired by Claudia Buch (Deutsche Bundesbank) and
B Gerard Dages (Federal Reserve Bank of New York) to examine trends in bank
business models, performance and market structure over the past decade, and assess
their implications for the stability and efficiency of banking markets.
The following report presents the Group’s conclusions on structural changes in
the banking sector after the crisis. The first message is that while many large advanced
economy banks have moved away from trading and cross-border activities, there
does not appear to be clear evidence of a systemic retrenchment from core credit
provision. Second, bank profitability has declined across countries, and individual
banks have experienced persistently weak earnings and poor investor sentiment,
suggesting a need for further cost cutting and structural adjustments. Supervisors
and authorities should monitor banks’ adjustment, assessing any risks that may
emerge, but also play a role in facilitating the process by removing impediments
where necessary. Third, in line with the intended direction of the regulatory reforms,
banks have significantly enhanced their balance sheet and funding resilience and
curbed their involvement in certain complex activities. Nonetheless, market
participants and authorities should not become complacent about the progress to
date and press on with the implementation of reform.
The adaption of the banking sector to the post-crisis operating landscape
warrants ongoing close attention. I hope that this work provides a useful resource by
documenting the state of ongoing structural change in banking, and providing a
starting point for further in-depth analysis.
William C Dudley
Chair, Committee on the Global Financial System

President, Federal Reserve Bank of New York
CGFS - Structural changes in banking after the crisis iii

Contents
Executive summary ................................................................................................................................. 1
1. Introduction ....................................................................................................................................... 5
2. An overview of banks’ operating environment ................................................................... 6
2.1 Background on banks’ experience in the financial crisis ....................................... 6
2.2 The post-crisis operating environment for banks .................................................... 7
3. Key trends in banking structure and performance .......................................................... 11
3.1 Banking system size ...........................................................................................................12
3.2 Banking system concentration....................................................................................... 13
3.3 Bank business models ....................................................................................................... 14
3.4 International banking ........................................................................................................24
3.5 Bank performance............................................................................................................... 26
3.6 Summary of key trends in banking .............................................................................. 29
4. Implications for the stability of the banking sector ......................................................... 30
4.1 Bank risk buffers and risk-taking .................................................................................. 30
4.2 Bank risk governance and management ................................................................... 33
4.3 Market-based indicators of banks................................................................................ 35
4.4 Future profitability and resilience of banks .............................................................. 39
4.5 Banking system-level risk and resilience ................................................................... 41
4.6 Overall assessment ............................................................................................................. 50
5. Implications for the efficiency of the banking sector ...................................................... 51
5.1 Bank credit provision to the real economy ............................................................... 51
5.2 Banks’ role in facilitating capital market activity .................................................... 58
5.3 Overall assessment ............................................................................................................. 62
6. Key messages .................................................................................................................................. 64
References ................................................................................................................................................ 70
Annex 1: Country banking data tables ..........................................................................................77
Annex 2: Individual bank data ....................................................................................................... 115
Annex 3: Business model classification methodology ......................................................... 116
Annex 4: Cost of equity estimation methodology ................................................................ 118
Members of the Working Group .................................................................................................. 119
CGFS - Structural changes in banking after the crisis v

Executive summary
The decade since the onset of the global financial crisis has brought about significant
structural changes in the banking sector. The crisis revealed substantial weaknesses
in the banking system and the prudential framework, leading to excessive lending
and risk-taking unsupported by adequate capital and liquidity buffers. The effects of
the crisis have weighed heavily on economic growth, financial stability and bank
performance in many jurisdictions, although the headwinds have begun to subside.
Technological change, increased non-bank competition and shifts in globalisation are
still broader environmental challenges facing the banking system.
Regulators have responded to the crisis by reforming the global prudential
framework and enhancing supervision. The key goals of these reforms have been to
increase banks’ resilience through stronger capital and liquidity buffers, and reduce
implicit public subsidies and the impact of bank failures on the economy and
taxpayers through enhanced recovery and resolution regimes. At the same time, the
dynamic adaptation of the system and the emergence of new risks warrant ongoing
attention.
In adapting to their new operating landscape, banks have been re-assessing and
adjusting their business strategies and models, including their balance sheet
structure, cost base, scope of activities and geographic presence. Some changes have
been substantial and are ongoing, while a number of advanced economy banking
systems are also confronted with low profitability and legacy problems.
This report by the CGFS Working Group examines trends in bank business
models, performance and market structure, and assesses their implications for the
stability and efficiency of banking markets.
The main findings on the evolution of banking sectors are as follows:
Changes in banking market capacity and structure. The crisis ended a period
of strong growth in banking sector assets in many advanced economies. Several
capacity metrics point to a shrinking of banking sectors relative to economic activity
in several countries directly impacted by the crisis. This adjustment has occurred
mainly through a reduction in business volumes rather than the exit of firms from the
market. Banking sectors have expanded in countries that were less affected by the
crisis, particularly the large emerging market economies (EMEs). Concentration in
banking systems has tended to increase, with some exceptions.
Shifts in bank business models. Advanced economy banks have tended to re-
orient their business away from trading and more complex activities, towards less
capital-intensive activities, including commercial banking. This pattern is evident in
the changes in banks’ asset portfolios, revenue mix and increased reliance on
customer deposit funding. Large European and US banks have also become more
selective and focused in their international banking activities, while banks from the
large EMEs and countries less affected by the crisis have expanded internationally.
Trends in bank performance. Bank profitability (return on equity) has declined
across countries and business model types from the historically high rates seen before
the crisis. At least in part, this reflects lower leverage induced by the regulatory
reforms. In addition, many advanced economy banks, in particular banks in some
European countries, are facing sluggish revenues and an overall cost base that has
CGFS - Structural changes in banking after the crisis 1

been resistant to cuts, including, in some cases, legacy costs associated with past
investment decisions and misconduct.
The main findings regarding the impact of post-crisis structural change for the
stability of the banking sector are related to three areas:
Bank resilience and risk-taking. Banks globally have enhanced their resilience
to future risks by substantially building up capital and liquidity buffers. The increased
use of stress testing by banks and supervisors since the crisis also provides for greater
resilience on a forward-looking basis, which should help support credit flows in good
and bad times. In addition, advanced economy banks have shifted to more stable
funding sources and invested in safer and less complex assets. Some of these
adjustments may be driven partly by cyclical factors, such as accommodative
monetary policy, and hence may diminish as conditions change. Qualitative evidence
indicates that banks have considerably strengthened their risk management and
internal control practices. Although these changes are hard to assess, supervisors
point to significant scope for further improvements, in particular because of the
inherent uncertainties about the future evolution of risks.
Market sentiment and future bank profitability. Despite a recovery in market-
based indicators of investor sentiment towards larger institutions in recent years,
equity investors remain sceptical towards some banks with low profitability.
Simulation analysis carried out by the Working Group suggests that some institutions
need to implement further cost-cutting and structural adjustments.
System-wide effects. Assessing the impact of structural change on system-wide
stability is harder than in the case of individual banks because of complex interactions
within the system. Nonetheless, a number of changes are consistent with the
objectives of public authorities and the reform process. First, banks appear to have
become more focused geographically in their international strategy and tend to
intermediate more of their international claims locally. Second, direct connections
between banks through lending and derivatives exposures have declined. Third, some
European banking systems with relatively high capacity have made progress with
consolidation. Fourth, while the effect of less business model diversity arising from
the repositioning of many banks towards commercial banking cannot be assessed
yet, this trend has been accompanied by a shift towards more stable funding sources
(such as deposits). A range of other reforms has also enhanced systemic stability
(eg money market mutual fund reforms) and further progress has been made on
resolution and recovery frameworks.
Changes in banking sector resilience have to be measured against the impact on
the services provided by the sector. The main findings regarding the impact of
changes on the efficiency of financial intermediation services are:
Provision of bank lending to the real economy. Trends in bank-intermediated
credit have been uneven over time and across countries, reflecting differences in their
crisis experience and related overhang of credit. Credit declined significantly relative
to economic activity in advanced economies that bore the brunt of the crisis, and in
most countries started to recover only from 2015. But the adjustment is still ongoing
in others, reflecting in part a legacy of problem bank assets that continues to hamper
the growth of fresh loans. By comparison, advanced economy banking systems not
significantly affected by the crisis continued to report solid loan growth,
notwithstanding tighter regulations.
2 CGFS - Structural changes in banking after the crisis

Recognising the difficulty of disentangling demand and supply drivers, the
evidence gathered by the group does not suggest a systematic change in the
willingness of banks to lend. But, in line with the objectives of regulatory reform,
lenders have become more risk-sensitive and more discriminating across borrowers.
In contrast to many advanced economies, bank lending has expanded strongly in
EMEs, raising sustainability concerns and prompting the use of macroprudential
measures and the tightening of certain lending standards more recently.
Capital market activities. Crisis-era losses combined with regulatory changes
have motivated a significant reduction in risk and scale in the non-equity trading and
market-making businesses of a number of global banks.
International banking was one of the areas most affected by the crisis.
Aggregate foreign bank claims have seen a significant decline since the crisis, driven
particularly by banks from the advanced economies most affected by the crisis,
especially from some European countries. By contrast, banks from other non-crisis
countries have expanded their foreign activities, in some cases quite substantially,
resulting in a significant change in the country composition of global banking assets.
The report highlights four key messages for markets and policymakers:
1. Post-crisis, a stronger banking sector has resumed the supply of
intermediation services to the real economy, albeit with some changes in
the balance of activities.
– Bank credit growth remains below its excessive pre-crisis pace in advanced
economies but without indications of a systematic reduction in the supply of
local credit. Lending to some sectors and borrowers has seen reductions,
however, as banks have adjusted their risk profile, and policymakers should
remain attentive to potential unintended gaps in the flow of credit.
– Experience from crisis countries underscores the benefits of acting early in
addressing problems associated with non-performing loans (NPLs).
– The withdrawal of some banks from capital markets-related business has
coincided with signs of fragile liquidity in some markets, although causality
remains an open question.
2. Longer-term profitability challenges require the attention of banks and
supervisors, as they may signal risk-taking incentives and overcapacity. Low
profitability partly reflects cyclical factors but also higher capitalisation and more
resilient bank balance sheets. As such, banks and their investors need to adapt
to a “new normal”. Market concerns about low profitability may deprive banks of
an important source of fresh capital, or encourage risk-taking and leverage by
banks, thus placing a premium on robust risk management, regulation and
supervision. In some cases, low profitability might also signal the existence of
excess capacity and structural impediments to exit for individual banks, requiring
decisive policy action to apply relevant rules.
3. Consolidation and preservation of gains in bank resilience requires ongoing
surveillance, risk management and a systemic perspective. Key indicators
show areas of improvement since the crisis, but also areas which are still a work
in progress. Authorities and market participants should not become complacent.
The system is adapting to a variety of changes, the interaction of which is difficult
to predict. Authorities should monitor the ongoing adaptation and evolution in
the nature and locus of risk-taking within the banking sector and the financial

system more broadly. In this regard, the group sees scope for the international
supervisory community to undertake a post-crisis study of bank risk
management practices. In addition, ample buffers remain critical to coping with
unexpected losses from new risks.
4. Better use and sharing of data are critical to enhanced surveillance of
systemic risk. Surveillance is crucial, given that the financial sector evolves
dynamically and because future risks will likely differ from past ones. Although
data availability has improved, there is a need to make better use of existing data
to assess banking sector structural adjustment and related risks. This effort will
likely require additional conceptual work, building on the data sets of national
authorities and the international financial institutions. Areas that warrant further
analysis include the potential for increased similarities in the exposure profile of
banks to correlated shocks, the growing role and implications of fintech, and the
migration of activity and risk to the non-bank sector.

1. Introduction
The financial crisis of 2007–09 was a watershed for the banking sector globally. It
revealed a pattern of excessive risk-taking and inadequate capital and liquidity buffers
within the industry, together with shortcomings in the prudential framework.
Regulators have responded with more demanding capital and liquidity standards,
stronger supervision, and more explicit resolution frameworks. The operating
landscape for banks has also changed markedly, reflecting a prolonged period of
private sector deleveraging, weak economic growth and historically low interest rates
in most advanced economies, accompanied by shifts in the globalisation trends of
the real economy. Stakeholder scrutiny of banks has intensified, while technology has
empowered new non-bank, challengers to banks’ businesses, thus adding to
competitive pressure. Many of these trends are ongoing and evolving.
Banks have been responding to the experience of the crisis and the post-crisis
operating environment. Globally, banks have been re-assessing and adjusting their
business strategies, including their growth plans, balance sheet positions, cost bases,
organisational structures, scope of activities and geographic presence. Adjustments
have also affected less visible aspects of their business, including governance and risk
management practices.
In light of the substantial changes in banking over the past decade, the
Committee on the Global Financial System (CGFS) established a Working Group to
examine trends in bank business models, performance and market structure, and
assess their implications for the stability and efficiency of banking markets. The Group
was also tasked with considering the drivers of trends in banking and the extent to
which the changes observed may be temporary or long-lasting.
Assessing the implications of post-crisis changes for the stability and efficiency
of banking required a focus on system-wide developments in addition to those for
individual institutions. For the purposes of its study, the Group considered the stability
of the banking sector as its ability to remain resilient and continue to provide credit
and other core intermediation services to the economy during periods of stress. The
efficiency of the banking sector was considered from the perspective of its provision
of lending and capital market services in support of growth in the real economy.
This report presents the Group’s findings, which are based on analyses of firm-
and country-level data, ongoing work at central banks and reviews of relevant
banking literature. The Group also sought out the perspectives of academics and
private sector bank analysts and reached out to bank supervisors in member countries
to get their views on post-crisis developments in banks’ risk management practices.1
The report comprises six chapters. Chapter 2 provides an overview of the post-
crisis operating environment. Chapter 3 presents the key trends in market structure,
business models and the performance of banks. Chapters 4 and 5 analyse the
implications of these trends for the stability and efficiency of the banking sector.
Issues for policymakers, including areas that warrant attention, are discussed in
Chapter 6. The report also includes a set of annex tables with time series of system-
level banking data.
1
The Group thanks Thorsten Beck, Michael Koetter and Luc Laeven for their feedback on selected parts
of this report.

2. An overview of banks’ operating environment
This chapter provides context on the conditions under which the banking sector has
evolved, including the experience of banks during the financial crisis and post-crisis
changes in banks’ operating environment, including key macro-financial, regulatory
and competitive developments.
2.1 Background on banks’ experience in the financial crisis
The period preceding the financial crisis was one of considerable exuberance,
primarily in the banking sectors of many advanced economies. Banking system assets,
credit and profits grew at a much faster pace than economic activity. Risk was often
neglected in compensation and other incentive structures – which heavily rewarded
short-term gains over long-term sustainable returns – and not properly assessed in
bank strategies. Credit standards were relaxed, and many banks relied on short-term
wholesale markets to fund activities. The inherent procyclicality of the financial system
also helped fuel credit and economic growth in mutually reinforcing ways. Banks in
some countries operated with relatively thin capital and liquidity buffers. The cross-
border business of large banks expanded sharply, as did revenue generation from
complex and opaque activities, including structured securitisations and over-the-
counter (OTC) derivatives.
Although pre-crisis developments in the banking sector were at the heart of the
crisis, other factors contributed, including misaligned incentives in the securitisation
process and in the implicit government support of banks, inadequate bank regulation
and supervision in many countries, a lack of risk discrimination in credit markets and
increased leverage in some parts the non-financial sector.2
The crisis was triggered around mid-2007 by the deflation of the US housing
boom, resulting in sizeable reported losses on US structured mortgage credit and
uncertainty about the extent of institutions’ exposures to these assets. The tightening
of financial conditions over ensuing months exposed the much broader pattern of
excessive risk-taking, maturity transformation and acute vulnerability within the
global banking industry. As some banks (and other financial intermediaries) came
under liquidity strain, central banks significantly expanded their liquidity facilities. The
closure of bank funding markets after the Lehman Brothers failure in September 2008
prompted governments to guarantee banks’ wholesale funding. Numerous banks in
Europe and the United States failed or received government capital injections, and
some were nationalised. Asset disposal schemes were set up in some countries to
help banks address their problem assets. These efforts succeeded in preventing a
collapse of the financial system and the economy. But the resultant fiscal costs from
direct banking sector financial support, output losses and increases in public debt
were very substantial, in some cases raising concerns about the solvency of
sovereigns.3
2
Since our focus is developments in the banking sector, this list is not intended to be exhaustive but
rather illustrate that other factors were involved.
3
See Laeven and Valencia (2012) for quantification of these costs.

The crisis affected all the large banking systems, although the impact varied
because of different starting cyclical conditions and structural vulnerabilities
(Graph 1). While banks in the euro area, the United Kingdom and the United States
suffered large losses at the height of the crisis, those in Australia, Canada and Sweden
fared better and did not need government capital support. EME banks were more
insulated from the turmoil given their domestic focus, relatively low use of market
funding and generally higher regulatory buffers, the latter reflecting in part lessons
of prior financial crises.
Policy responses to the crisis also contributed to the heterogeneity of outcomes.
For example, the process of bank balance sheet repair and recapitalisation in the
United States ran ahead of that in some European countries.
Another important reason for regional variation in bank outcomes over the past
decade is the impact of the euro area sovereign debt crisis that began around late
2010. To a degree, fears about sovereign creditworthiness were an extension of
banking sector problems that emerged in the 2007–09 financial crisis, such as in the
case of Ireland. However, for many euro area banks, the escalation of sovereign debt
concerns represented an additional adverse shock that dragged on their post-crisis
performance and deepened pressures for structural change.
Bank share prices and CDS spreads Graph 1
Bank share prices Credit default swap premia1

1 Jan 2007 = 100 Basis points
150 450
100 300
50 150
0 0
03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17
Australia Sweden Euro area EMEs Euro area United States
Canada United States United Kingdom United Kingdom
1
Five-year on–the-run CDS spreads; simple averages over sample of banks.
Sources: Datastream; Markit.
2.2 The post-crisis operating environment for banks
The macro-financial environment. The post-crisis macroeconomic environment has

been characterised by a prolonged period of weak economic growth and low inflation
in the large advanced economies.4 Prolonged private sector deleveraging has weighed
on aggregate investment and credit growth (Graph 2, left-hand panel). Benchmark
4
As documented in the literature, financial crises are typically associated with subsequent slower
medium-term growth. For example, see Abiad et al (2009) and Reinhart and Reinhart (2015).

nominal interest rates have been at historic lows, in some cases even reaching negative
territory as central banks have resorted to highly expansionary conventional and
unconventional monetary policies, including considerable expansion of their balance
sheets and/or negative policy rates (Graph 2, right-hand panel). Interest rates may
have also been depressed by a confluence of structural factors.5
In contrast, private sector credit has continued to expand in EMEs, despite slower
economic growth over recent years in the case of the large EMEs (Graph 2, centre
panel). In China, corporate indebtedness has risen appreciably since the onset of the
crisis. In some smaller EMEs and other countries that fared better in the crisis
(eg Australia, Canada and Sweden), housing credit has also grown at a firm pace over
the post-crisis period in association with strong rises in housing prices.
Economic growth, credit and interest rates

In per cent Graph 2
GDP Total credit-to-GDP ratio1 Two-year government bond yields
9 200 6.0
6 160 4.5
3 120 3.0
0 80 1.5
–3 40 0.0
–6 0 –1.5
03 06 09 12 15 18 21 01 03 05 07 09 11 13 15 17 01 03 05 07 09 11 13 15 17
World Advanced economies (non-crisis) United States
Advanced economies Advanced economies (crisis) United Kingdom
EMEs EMEs excluding China Japan
China Germany
1
For the private non-financial sector. Advanced economy crisis classifications are based on Laeven and Valencia (2012), see Graph 26. EMEs
include Brazil, Hong Kong SAR, India, Korea, Mexico and Singapore. Aggregates constructed based on rolling GDP PPP weights.
Sources: IMF, World Economic Outlook; Bloomberg; BIS.
Regulation and supervision. Reforms to the global regulatory framework seek

to strengthen bank resilience to adverse shocks through increased requirements for
higher-quality capital and liquidity; address risks posed by systemically important
financial institutions and reduce implicit public subsidies; and enable the effective
resolution of banks. These measures are intended to reduce the probability of default
for large internationally active banks to a low level, and significantly improve the
system’s capacity to absorb the failure of a large institution. In doing so, the reforms
go beyond enhancing the soundness of individual banks and encompass a
macroprudential or system-wide perspective of risks to financial stability (that is,
systemic risk).6 Some aspects of the reforms have already been implemented, while
5
For a discussion, see Cunliffe (2016).
6
A useful definition of systemic risk is provided by the ECB (2017a): “Systemic risk can be best
described as the risk that the provision of necessary financial products and services by the financial
system will be impaired to the point where economic growth and welfare may be materially affected.”

others are subject to transitional arrangements or are due to be introduced in coming
years. Key elements of the bank regulatory changes are noted below.
 An increase in the quantity and quality of capital through stricter risk-weighted
requirements. A non-risk-weighted ratio has been introduced as a
supplementary measure to constrain leverage and reduce model risk. Authorities
can activate a macroprudential countercyclical capital buffer for all banks in a
jurisdiction as risks build over the credit cycle and thus reduce procyclicality.
 New measures for globally systemically important banks (G-SIBs) – currently
a universe of 30 banks (see Annex 2) representing more than one third of global
banking assets – including higher capital buffers and additional requirements for
large exposures and disclosure.7 Many countries have also introduced tougher
capital and other rules for banks deemed systemically important for their
domestic economy, including stress-testing requirements.
 International prudential regulation of liquidity risk has been introduced, with
the centrepiece being two quantitative instruments – the Liquidity Coverage Ratio
(LCR) and the Net Stable Funding Ratio (NFSR). The former promotes the short-
term resilience of banks’ liquidity profiles, while the latter aims to ensure that
banks maintain a stable funding structure for their assets and off-balance sheet
activities.
 A global framework for the recovery and resolution of banks has been
established. Specifically, G-SIBs will need sufficient total loss-absorbing capacity
(TLAC), including a proportion of debt that can be “bailed in” in the event of
failure. Many countries are revising their national resolution regimes to allow
more orderly resolution of their banks, including changes to legislation and the
requirement for banks to have plans for their recovery and wind down.8
In some jurisdictions, international regulatory changes have been complemented by
national initiatives. Most notable are measures to address the structure of banks’
activities and operations, such as limits on proprietary trading and the new
Intermediate Holding Company rules in the United States, and the ring-fencing of
trading books from other banking activities in the United Kingdom. Some countries
have also placed limits on bank remuneration. Given the importance of domestic
institutional settings and the domestic costs of financial crises, macroprudential policy
remains largely a national policy task, while being coordinated internationally where
appropriate.
The global regulatory reforms also encompass elements outside banking.
Monitoring and policy frameworks have been established for systemically important
non-bank financial institutions. They have also been established for market-based
finance, to ensure that bank-like risks outside the regular banking system are subject
to appropriate oversight and regulation, including, for example, measures to reduce
money market funds’ susceptibility to runs. Other changes in most jurisdictions have
included the mandatory clearing of certain standardised OTC derivatives with central
counterparties, the introduction of capital charges on these exposures and surcharges
for positions cleared bilaterally with other banks.
7
This is a rough share of G-SIB assets in global banking assets sourced from IMF (2017).
8
However, substantial work remains to do in achieving effective resolution regimes and
operationalising plans for systemically important banks (FSB (2017a)).

Stricter regulation of banks in response to the crisis has been accompanied by
more intensive bank supervision, particularly for systemically important banks. This
adjustment has had myriad aspects, including greater engagement between
supervisors and bank boards and senior managers, wider and more stringent use of
supervisory stress tests, increased risk data requirements, more targeted (or thematic)
risk reviews, and greater attention placed on bank risk culture and crisis preparedness.
Governments have also exercised greater influence over banks in their role as
legacy shareholders from the crisis (such as by requiring specific changes to strategies
and operations). In addition, there have been sizeable regulatory fines for misconduct
at some banks in the lead-up to the crisis and subsequently.
Stakeholder scrutiny. Banks have also faced increased pressure from the
investor community, reflecting the ongoing weak performance of many banks as well
as shifts in investor risk attitudes and perceptions. (These issues are explored further
in Section 4.3 below.) This process has affected banks’ funding costs and incentivised
changes to their business models – for example, rating agencies are far more focused
on the extent of deposit funding. More generally, regulatory authorities and the
broader public have demanded an improvement in banks’ corporate culture,
remuneration practices and treatment of customers.
Development of the non-bank sector. Non-bank finance and non-bank
financial institutions have gained a greater role in financing economic activity in the
aftermath of the crisis. Corporate debt financing has increasingly shifted to capital
markets in advanced economies, with bond issuance picking up noticeably (Graph 3,
left-hand panel). The global assets of both insurance and pension funds and “other
financial intermediaries” (OFIs) (including managed funds and non-prudentially
regulated financial institutions) have also expanded relatively strongly (Graph 3,
centre panel). The OFI sectors’ share of total financial institution assets has risen in
most European countries, as well as in some countries where the banking sector fared
much better during the crisis (eg China, Korea and Mexico). Nonetheless, OFIs’ share
of lending in total financial institution lending is not as large as their corresponding
share of total assets, and they continue to represent only a small share of global
lending activity (Graph 3, right-hand panel). There are, however, clear exceptions in
some countries and markets.9
Technological change. As banking is an information-intensive industry,
innovations in information technology can have an important influence on the nature
of banking services and the structure of banks. Rapid advances in computing, internet
and mobile capabilities have underpinned a deeper digitalisation of banking services
over recent years. Nascent but growing fintech companies have emerged that provide
specialised financial services using a range of digital innovations, including those that
supply credit to the household and business sector through online platforms.10 The
9
For example, insurers and pension funds have rapidly increased their share of the mortgage market
in the Netherlands, accounting for 20% of originations in the first half of 2016 (Netherlands Bank
(2016)). Similarly, non-banks in the United States now account for a substantial share of mortgage
origination volume, representing 55% of originations among the top lenders in 2016 (based on data
from Inside Mortgage Finance).
10
To date, fintech companies have not garnered a material share of aggregate credit (CGFS-FSB (2017)).

role of technology in finance is also being assisted by certain public initiatives.11
Depending on banks’ reaction function and their capability to adapt their business
models, fintech innovations could be a competitive threat to some banks or bank
business lines, or they may provide banks with an opportunity to improve customer
experience or significantly lower their fixed cost base over the longer term.
Technological innovations might also allow banks to improve their data analytics and
compliance risk processes, yet they present new and challenging operational
complexities – including cyber and data security risks – that banks will need to
effectively manage.12
Non-bank finance and financial institutions Graph 3
Non-financial corporations’ credit1 Ratio of banks to other financial Lending assets as a share of GDP3
institutions2
Per cent of GDP Ratio Ratio Per cent of GDP
Ratio Ratio decreasing
increasing (lhs axis) (rhs axis)
120 4 (lhs axis) 8 120
90 3 6 90
60 2 4 60
30 1 2 30
0 0 0 0
CA
US
AU
BR
JP
ES
NL
CH
EA
BE
KR
GB
MX
FR
DE
IT
IN
HK
SG
CN
_____
_____
_____
_____
_____
_____
07 16 07 16 07 16 07 16 07 16 07 16 03 05 07 09 11 13 15
US EA JP GB AU SE
2007 2015 Insurance corporations OFIs
Debt securities Loans and pension funds Banks
1
For country codes, refer to Graph 4; end-year observations. 2 Based on total financial assets; banks including public financial institutions
but excluding central banks; for country codes, refer to Graph 4. 3 For the 21 jurisdictions and the euro area covered by the FSB Global
Shadow Banking Monitoring Report 2016.
Sources: Datastream; Financial Stability Board; national data; BIS.
3. Key trends in banking structure and performance
This chapter provides an overview of key trends in banking since the crisis. It identifies
common developments across countries, as well as similarities within smaller groups
of countries or types of banks that are a significant element of the overall picture. The
material provides the foundation for more targeted analysis of the key changes and
their implications in subsequent chapters.
11
For example, regulatory sandboxes, hubs or accelerators in several countries (FSB (2017b)). In the EU,
the revised Payment Services Directive (PSD2) aims to boost competition in the payments industry.
12
This report provides a brief discussion of technological change and fintech innovations as
environmental factors affecting banks over recent years, but does not seek to assess the implications
of fintech competitors on banks nor banks’ use of fintech innovations.

The chapter comprises six sections. The first section covers trends in asset size
and resource capacity of domestic banking systems across CGFS jurisdictions.13 Broad
changes in system concentration are then discussed. The third section assesses shifts
in bank business models and strategies, including asset composition, funding sources
and capital buffers. The fourth section focuses on changes in the international profile
of banking sectors, and the fifth section on the profitability of banks. The final section
concludes with a summary of key changes.
3.1 Banking system size
The crisis brought about the end of a period of fast – and in some instances excessive
– growth for many domestic banking sectors. Between 2003 and 2007, the median
annual average growth of CGFS member banking systems was about 12%, compared
with 4% from 2008 onwards. In some European banking systems, the ratio of bank
assets to GDP has fallen considerably since the crisis, whereas it has risen in other
jurisdictions (Graph 4 and Annex Table 1.1). Banking system growth in EMEs has been
less affected by the crisis and most systems have continued to grow at robust rates.
Of particular note has been the dramatic expansion of the Chinese banking system,
which grew from about 230% to 310% of GDP over 2010–16 to become the largest
in the world, accounting for 27% of aggregate bank assets across CGFS member
countries, up from about 13% in 2010 (Graph 5, left-hand panel and Annex Table 1.2).
Banking system assets to GDP1

In per cent Graph 4
Europe Other advanced EMEs (rhs axis)

400 2,200
300 1,650
200 1,100
100 550
0 0
BE CH DE ES FR GB IT NL SE EA AU CA JP US BR CN IN KR MX HK LU SG
2003 2007 2012 2016
AU = Australia; BE = Belgium; BR = Brazil; CA= Canada; CH = Switzerland; CN = China; DE = Germany; EA = euro area; ES = Spain; FR =
France; GB = United Kingdom; HK = Hong Kong SAR; IN = India; IT = Italy; JP = Japan; KR = Korea; LU = Luxembourg; MX = Mexico; NL =
Netherlands; SE = Sweden; SG = Singapore; US = United States.
1
Banking system assets are on a domestic or resident basis, except for China and Korea, which are on a consolidated basis.
Source: National data.
Slower banking system growth, together with increased competition from other
intermediation channels, has seen the banking system’s share of total financial
institution assets decline in a range of countries (see Chapter 2 and Annex Table 1.3).
13
See Annex 1 for the list of CGFS member jurisdictions.

In addition to balance sheet size, other metrics of capacity point to a shift in the
growth rate of banking sectors. The size of branch networks and the number of
employees have declined after the crisis in several European countries, a trend more
pronounced in systems that had high capacity relative to their population (eg Italy,
Spain) (Annex Tables 1.4 and 1.5). In contrast, reflecting ongoing financial deepening,
the number of branches and employees have both risen in several EMEs (China, India,
Mexico), although they remain below those for most advanced economies when
compared with population.
3.2 Banking system concentration
Adjustment to the post-crisis environment has taken place not only through changes
in the scale of activities – the intensive margin – but also through the entry and exit
of banks – the extensive margin. The number of banks has fallen in most countries
over the past decade (Annex Table 1.6 and 1.7). In some cases, this appears to be a
continuation of a consolidation process that was previously underway (eg in the euro
area, Sweden, Switzerland and the United States).
Post-crisis reductions in bank numbers have been mainly among smaller
institutions, aside from a handful of distressed large banks in the euro area and the
retreat of some international banks from specific foreign markets. Reflecting relatively
little consolidation activity among the largest institutions, the value of bank merger
and acquisition (M&A) transactions in some large advanced economies has been
subdued compared with earlier periods (Graph 5, centre panel). In the euro area, the
post-crisis decline in bank M&A has been more pronounced for cross-border deals
than domestic deals.14 This trend may suggest that banks have re-directed their
attention towards home markets and/or lifted their “hurdle rate” for cross-border
deals in particular.
Post-crisis patterns in concentration point to greater banking system
consolidation in the large advanced economies than suggested by the M&A data.
Concentration ratios – measured as the share of banking system assets held by the
largest five banks – have increased within the euro area and the United States, where
bank consolidation was partly a consequence of dealing with the effects of the crisis
(Graph 5, right-hand panel and Annex Table 1.9).15 Yet concentration has also risen in
some countries that were less affected by the crisis and where bank numbers have
continued to expand or remained steady (Australia, Brazil, Singapore). In contrast,
there has been a decline in concentration in several economies (Belgium, China, India
and Mexico). Overall, concentration in banking systems has tended to increase since
the crisis, with some clear exceptions. Nonetheless, the median concentration ratio
across countries is unchanged, and there has been little change in the dispersion of
concentration ratios across countries. Box A provides a case study on the Spanish
banking system, which has undergone substantial consolidation since the crisis.
14
See ECB (2017b).
15
Similar results can be seen when the concentration measure is based on the three largest banks
(Annex Table 1.8).

Banking market structure Graph 5
Share of banking system assets Bank M&A – value of transactions Banking system concentration, share
across CGFS membership of system assets of 5 largest banks3
Per cent USD bn
80 160 85
BE
SE CA NL
MX FR
60 120 70
AU
BR
2007
40 80 CH 55
CN
GB Median ES
EA
20 40 JP More 40
IN SG concentration
US
DE IT
0 0 25
02 04 06 08 10 12 14 16 40 55 70 85 100
08 10 12 14 16
Euro area: 2016
Other advanced1 Europe United States Domestic
2
Other EMEs China Cross-border
United States
1
Australia, Canada and Japan. 2 Brazil, Hong Kong SAR, India, Korea, Mexico and Singapore. 3 For country codes, refer to Graph 4.
Median value may not reflect individual country data as shown in the graph because of the use of some confidential series.
Sources: Dealogic; national data.
3.3 Bank business models
The post-crisis period has seen a significant change in the strategic orientation of
many banking firms. Broadly speaking, banks have reassessed their earlier ventures
into trading activities as well as a growing dependence on wholesale sources of
funding, in association with stricter capital and liquidity regulation. Use of deposit
funding has increased, while businesses have tended to be repositioned towards less
complex and less capital-intensive activities, including retail banking and, in some
cases, wealth management. These patterns have been evident in the strategic
changes implemented by many banks, and in their balance sheet compositions and
revenue mix. There has also been a re-focus on home market business or core foreign
markets, which is explored separately in Section 3.4 below.
It is worth emphasising that these broad observed changes to bank business
models are far from fully consistent across the global banking sector. While many banks
have significantly adjusted their business models in the general direction outlined above,
some have been able to broadly maintain their profile and activities (in part reflecting
lower pressure for change), while others have even “swum a little against the tide”.16
From a geographic perspective, adjustments have generally been more profound for the
most globally active banks and large European banks, and less significant for EME banks.
Box B provides two prominent examples of substantial strategic and business model
shifts at European G-SIBs, RBS and UBS. In both cases, these banks have downsized their
investment banking units and repositioned to focus largely on other intermediation
activities. In addition to illustrating key trends in bank business models identified in this
report, the examples highlight the process of change at the institution level.
16
For example, some EME banks have increased their wholesale funding, as discussed below.

Box A
Consolidation in the Spanish banking system

The Spanish banking system withstood the first wave of the GFC reasonably well. However, as the crisis developed,
the imbalances built up previously (high credit growth, risk concentration in real estate and a growing reliance on
wholesale funding) placed the banking sector, in particular saving banks, in a highly vulnerable position. Loan
performance and solvency positions then deteriorated rapidly. The authorities’ initial response was to seek private
sector solutions to the required solvency and balance sheet improvements, as well as to encourage consolidation
among savings banks to address overcapacity and increase operating efficiency. In 2009, the Fund for Orderly Bank
Restructuring (FROB) was set up to allow public funds to be injected into the banking sector, subject to certain
conditions, including stricter capital requirements, strengthened corporate governance arrangements, capacity
adjustments and efficiency gains. In total, a considerable amount of state aid (€62.7 billion, including the contribution
of the Deposit Guarantee Fund and support of the European Stability Mechanism) was injected into savings banks.
As a result, the Spanish banking sector has undergone extensive restructuring and consolidation. In the 2008–16
period, the number of credit institutions (excluding foreign branches) declined from 195 to 125 (Graph A.1, left-hand
and centre panels), while total domestic banking system assets fell by around 20%. Consolidation was concentrated
among savings banks, whose numbers fell from 45 to two. Some savings banks were acquired or absorbed by
commercial banks or other savings banks, others were integrated into an Institutional Protection Scheme (IPS) or
merged, with eight of the resulting saving banks being transformed into commercial banks (following a legal reform).
In several cases, the integration processes involved more than one stage. Consolidation also occurred among
cooperatives, with the main process being the establishment of an IPS that brought together 19 credit cooperatives.
These changes have led to an increase in banking system concentration, with the domestic assets of the top five banks
accounting for 65% of system assets in 2016, up from 51% in 2008.
Structural adjustment is evident in bank distribution networks, with branch and employee numbers declining
substantially (Graph A.1, right-hand panel), and staffing costs falling by 24% from 2008 to 2016. Bank balance sheets
have been substantially strengthened, notably capital buffers, with the aggregate Tier 1 capital ratio rising from 7.8%
in 2008 to 12.8% in 2016.
Spanish banking system capacity Graph A.1
Bank assets and credit Number of deposit institutions Number of branches and employees
Per cent of GDP Total Total, in thousands Total, in thousands
250 175 260 45
200 140 240 40
150 105 220 35
100 70 200 30
50 35 180 25
0 0 160 20
1
Banks Saving Foreign- Coop. Total
01 03 05 07 09 11 13 15 17 banks controlled 02 04 06 08 10 12 14 16
subsidaries
Bank assets-to-GDP Number of employees (lhs)
Bank credit-to-GDP Number of branches (rhs)
2008 2016
1
Cooperatives.
 An Institutional Protection Scheme (IPS) is a contractual agreement among the participating institutions involving the assignment to a
central institution of the capacity to establish and implement business strategies and internal and risk control tools. Participants also agreed
to pool at least 40% of their liquidity, solvency and earnings. IPS formed in this manner became consolidated groups and were, in essence,
mergers (“cold mergers”, as they were called) since, for economic, albeit not legal, purposes, each participant lost its autonomy. Full pooling
of liquidity, solvency and results was achieved at most of the IPS formed.

Box B
Restructuring and business model adjustment at RBS and UBS
The case of RBS

RBS expanded its balance sheet rapidly in the run-up to the financial crisis, from around £750 billion in 2005 to
£2,400 billion – more than UK GDP – in 2008 (Graph B.1, left-hand panel). This pre-crisis growth was driven by an
expansion of RBS’s wholesale and derivatives business, including the acquisition of the Dutch bank ABN AMRO in
2007. The expansion was funded largely by wholesale debt (Graph B.2, centre panel). RBS failed in October 2008 and
was recapitalised by the UK government as part of a system-wide package of measures for the UK banking system.
RBS subsequently sought to shrink and strengthen its balance sheet in a plan agreed with the European
Commission, given the involvement of state aid. In 2009 RBS set up a “non-core” division to manage many of the
bank’s higher-risk assets; these assets were reduced by more than £150 billion between 2010 and 2013. Overall, the
bank’s balance sheet was strengthened through a combination of whole-business disposals, run-offs, asset sales and
write-downs rather than through capital raisings or earnings retention. As the bank made substantial losses over the
period, it was unable to generate new capital organically.
Following a review of RBS in 2013 by the UK government, its majority shareholder, RBS announced a further set
of measures to become a bank focused on lending to UK businesses and households. These included an accelerated
exit from Citizens, a US retail bank, further shrinkage of its investment banking unit and a plan to run down £38 billion
of high-risk or non-performing assets (including loans in its Irish bank subsidiary and commercial real estate lending)
over a three-year period in a new internal bad bank. RBS also planned the sale or wind-down of most of its global
footprint, from 38 countries to 13, and trade finance and cash management outside the United Kingdom and Ireland.
By 2016, RBS had narrowed its focus to become a bank serving clients in the United Kingdom and western Europe
and had achieved a CET1 capital ratio of 13.4% (compared with a core Tier 1 ratio of 6.1% at end-2008). Its revenues
have increasingly come from net interest income generated by lending, with its investment banking activity refocused
on foreign exchange and fixed-income markets (Graph B.1, right-hand panel). While RBS has already undertaken
significant cost-cutting, improving its underlying cost efficiency remains a focus. Legacy issues have also continued to
adversely affect its bottom line profitability, despite having achieved run-down targets for the internal bad bank ahead
of schedule. RBS reported an attributable loss of £7.0 billion in 2016, including litigation and conduct costs of
£5.9 billion and restructuring costs of £2.1 billion. Looking forward, RBS will remain in a period of major restructuring
through 2019 driven by the need to meet UK legislative requirements for the ring-fencing of core retail banking
activities from wholesale and investment banking activities. RBS also remains subject to state aid requirements. In
2016, RBS’s management renewed its commitment to a long-term CET1 ratio of at least 13% and to achieving a cost-
to-income ratio of less than 50%, and a 12% return on tangible equity by 2020.
The case of UBS

In the pre-crisis years, UBS’s investment banking unit contributed a little less than half of group before-tax profit (and
revenue) (Graph B.2, left-hand panel). The unit was the source of large losses during the crisis, primarily on US real
estate and other credit positions. A package of measures by the Swiss Confederation and the Swiss National Bank
was required in late 2008 to stabilise the bank.
UBS has subsequently shifted away from investment banking activities towards its other existing businesses,
specifically global wealth management and Swiss retail and corporate banking. In 2009 it announced a repositioning
of its investment bank towards a client-centric model focused on more flow and fee business. The strategic change
was accompanied by a sizeable reduction in its balance sheet, which halved between 2007 and 2010; the downsizing
of the investment banking unit accounted for most of this decline (Graph B.2, centre panel).
In 2011 and 2012, the bank presented an updated strategy, with a further diminished role for investment banking.
The intention was that the investment bank should support the bank’s new strategic focus in wealth management and
Swiss retail and corporate banking. Furthermore, UBS decided that its investment bank would exit a substantial number
of fixed income business lines, in particular complex and capital-intensive credit and interest rate products. These
changes resulted in a further shrinking of the bank’s balance sheet and a sizeable reduction of fixed income
instruments held at fair value (Graph B.2, centre panel). As a consequence of the exit process, UBS transferred some
of the investment bank’s assets to its corporate centre for run-off (legacy unit). By the end of 2014, UBS considered
its strategic transformation process to be complete.

The strategic adjustments described above are visible in the breakdown of group profit, with the investment bank
accounting for only a small share in recent years (although the share of revenue is a bit higher, at about one third). As
a result of the adjustments, domestic activities have become more important, although foreign activities continue to
play a major role, with the majority of assets and employees located abroad.
RBS balance sheet and profit Graph B.1
Asset composition Liability composition Revenue and operating expenses

GBP trn GBP trn Per cent
2.0 2.0 90
1.5 1.5 75
1.0 1.0 60
0.5 0.5 45
0.0 0.0 30
06 08 10 12 14 16 06 08 10 12 14 16 06 08 10 12 14 16
Loans and advances to customers Equity excluding minority interest Net interest income/total income
Derivatives Customer accounts Cost-to-income ratio
Debt securities, equity instruments Derivatives
& loans to banks Wholesale funding & other liabilities
Other assets
Source: RBS annual reports.
UBS balance sheet and profit Graph B.2
Profits before tax Balance sheet by segment2 Financial assets at fair value3
CHF bn CHF trn CHF trn
10 2.5 0.75
0 2.0 0.60
–10 1.5 0.45
–20 1.0 0.30
–30 0.5 0.15
–40 0.0 0.00

04 06 08 10 12 14 16 07 08 09 10 11 12 13 14 15 16 2007 2010 2016
Wealth Man., Asset Man., Wealth Man., Asset Man., Debt instruments
Retail & Corp. Retail & Corp. Equity instruments
Investment Banking Investment Banking 4
Other
1 1
Corporate Center Corporate Center
Group
1
Including legacy unit. 2
Third-party view. 3
Excluding positive replacement value. 4
Including precious metals and commodities.
Source: UBS annual reports.
 UBS (2010)  Part of the balance sheet shrinkage was due to an active reduction or compression of positions, and part was market-
driven.  Around two thirds of average interest-earning assets were attributed to foreign activities in 2016, compared with almost 90% in 2007.

To help inform its business model analysis, the Group undertook a classification
exercise for its sample of large banking groups. Graph 6 (left-hand panel) presents
trends in the share of banks falling into one of four different business model types:
retail-funded commercial banking, wholesale-funded commercial banking, trading
and processing, and universal banking. The allocation of banks into business model
types is based on a statistical methodology that uses selected balance sheet ratios to
discriminate across the four profiles (see Annex 3 for further details).17 The main
message from this analysis is that there has been an increase in the popularity of the
retail-funded profile at the expense of the wholesale-funded one and, to a lesser
extent, the universal one. While there is no major change in the popularity of the
trading model, some G-SIBs have shifted from the trading category to the universal
category after the crisis (and, as discussed elsewhere, trading activities have declined
at many banks). Relatedly, the aggregate assets of banks in the trading category have
declined substantially (Graph 6, right-hand panel).
Bank business models1 Graph 6
Popularity of business model type2 Aggregate assets of business model type3

Probability USD bn
50 750
40 600
30 450
20 300
10 150
0 0
06 07 08 09 10 11 12 13 14 15 16 06 07 08 09 10 11 12 13 14 15 16
Retail-funded Wholesale-funded Trading and processing Universal
1
Based on the sample of individual advanced economy and EME banks in Annex 2, subject to data availability. 2 Each line plots banks’
average loading on a given model in a given year. A loading is the posterior probability of a bank belonging to a given model in a given year,
as determined by discriminant analysis. For a given year, the (average) loadings sum up to 1 across models. 3 Net of derivative assets; the
solid/dashed lines correspond to advanced/emerging market economy banks.
Sources: Roengpitya et al (2017); Fitch Connect.
Asset portfolios
At a broad level, banks’ share of loans in total assets has tended to rise, although
movements vary significantly across countries, with Canada, India, Mexico and
Switzerland experiencing the largest increases (Annex Tables 1.10 and 1.11). The
composition of banks’ assets has generally shifted away from debt securities over the
17
It is important to note there is no one way to classify banks’ business models. The method used in
Roengpitya et al (2017) has the benefit of being less judgment-based than some others. It is also a
useful approach to capturing changes in business models over time, which is important given the
focus of the report on structural change. See the IMF (2017a) for a finer business model classification
of G-SIBs based on revenue shares in 2016.

post-crisis period. The main exceptions are in Italy, where holdings of government
securities have risen noticeably, and in the United States.
More important than changes in broad asset structure have been several
underlying shifts within banks’ asset portfolios, which together signal a move away
from more complex or capital-intensive assets towards assets that typically pose less
risk. The most notable post-crisis trend in this respect is a marked increase in the
share of liquid asset holdings, including in Australia, several European countries, the
United Kingdom and the United States (Graph 7, left-hand panel and Annex Table
1.12). For some countries the increase has reflected higher cash or central bank
balances, in other cases government debt holdings have increased substantially. As
the share of available-for-sale debt securities held for liquidity purposes has generally
increased, many banks have cut down their proportion of interbank assets (which is
mirrored in the reduction in interbank liabilities noted below).
Large advanced economy banks have run down their legacy or “non-core” assets
considerably – for G-SIBs they have fallen from around 12% to 5% of their assets over
2013–16.18 In addition, many large banks have reduced their exposure to trading
assets and more complex securities. The median share of trading assets in total assets
for individual G-SIBs has declined from around 20% to 12% over 2009–16 (Graph 7,
centre panel). G-SIBs’ share of more complex Level 2 and Level 3 assets have also
fallen, in particular for banks with relatively high shares of Level 3 assets (those for
which fair value cannot be determined from market prices or standard pricing
models). Another trend related to the complexity of large banks’ activities has been
the reduction in OTC derivatives – in large part because of derivatives regulatory
reforms – with this trend most evident for G-SIBs which have larger derivatives
positions relative to their assets (Graph 7, right-hand panel).
Bank asset portfolios Graph 7
Share of liquid assets in total assets, G-SIBs’ share of complex assets2 G-SIBs’ OTC notional derivatives to
in per cent1 assets
Per cent Per cent Ratio to assets
Trading/Assets L2/Assets L3/Assets (rhs)
BR IN 30 60 6 30
25 50 5 25
20 40 4 20
2006
15 30 3 15
BE JP
FR 10 20 2 10
GB More liquid
SG NL ES
assets 5 10 1 5
IT
SE DE HK US
AU 0 0 0 0
5 10 15 20 25 30 35 09 12 16 09 12 16 09 12 16 10 11 12 13 14 15 16
2016
Median 25th–75th percentile Median 25th–75th percentile
1
Cash and balances with central banks and government securities as a share of total assets; for country codes, refer to Graph 4. 2 Level 2
assets are those where fair value is determined by standard pricing models using inputs that are directly or indirectly observable. Level 3
assets are those where fair value cannot be determined by observable market prices or standard pricing models.
Sources: SNL; national data; public filings.
18
IMF (2017a).

Within banks’ lending portfolios, the share of residential mortgages has
increased across most countries, continuing a longer-run trend associated with the
considerable rise in housing prices (Graph 8, left-hand panel and Annex Table 1.13).
Increases have occurred in countries that fared better in the crisis (Australia, Canada,
Sweden and a few EMEs), but have also been evident in some European countries
(Belgium, United Kingdom). The notable exception to this trend has been the United
States, given that housing problems were at the core of its crisis. Changes in the share
of business lending in total lending have been more mixed across countries, although
movements in either direction have tended to be larger for EMEs, notably the decline
for China (Graph 8, centre panel, Annex Table 1.14).
The share of non-performing loans (NPLs) in total loans rose sharply in the euro
area, United Kingdom and the United States after the onset of the crisis, but the rise
was far more muted elsewhere (Graph 8, left-hand panel, Annex Table 1.16). NPLs
subsequently decreased at a sluggish pace in the euro area in conjunction with
relatively weak macro-financial conditions, as well as some country-specific
institutional factors, and remain comparatively high in Italy and Spain.19, 20 NPL ratios
rose for some EMEs in 2015 and 2016, particularly in India.
Funding structures
Since the crisis, advanced economy banks have responded to market and supervisory
pressures to boost stable funding by significantly shifting their funding composition
away from wholesale funding towards deposits (Graph 9, left-hand panel, Annex
Tables 1.17 and 1.18). The shift away from market-based funding has been aided by
highly accommodative monetary policies in advanced economies.21 The change has
been most pronounced for European banks given their greater use of market-based
funding prior to the crisis. Conversely, banks in some EMEs have grown their use of
wholesale funding, albeit from a relatively low level, as their access to capital markets
has improved and credit growth has remained relatively robust. There has also been
a widespread rise in equity funding as a result of higher regulatory capital standards.
Underlying the decline in many banks’ share of wholesale funding has been a
cutback in the use of short-term funding, and a contraction in interbank liabilities (Graph
9, left-hand and centre panels).22, 23 There is anecdotal evidence that some banks have
diversified their international investor base away from financial institutions and money
markets by attracting funds from other investors, such as cash-rich corporations.24
19
One institutional factor that can be important for the resolution of NPLs is the efficiency of the legal
process. A protracted debt workout can generate uncertainty about asset values, increasing the
discount on NPLs from potential buyers and thus deterring banks from disposing of them (FSI (2017)).
20
Nevertheless, for Spain, the NPL ratio calculated on a consolidated basis is considerably lower than
when calculated on a domestic basis (for 2016, 5.6% compared with 9.2%).
21
For example, accommodative monetary policy has played a role in reducing the need for banks to
trade reserves through the repo market (CGFS (2017)).
22
Also see, for example, IMF (2013) and van Rixtel and Gasparini (2013).
23
Notwithstanding the post-crisis shift to more stable funding sources for the aggregate euro area
banking system, banks in some euro area countries (Greece, Ireland, Portugal, Spain and Italy)
became reliant on central bank funding as sovereign debt concerns intensified in 2011 and 2012. The
average share of central bank funding in total funding has since fallen as sovereign debt concerns
have receded (although Greek banks have been an exception). For further details, see ECB (2016b).
24
Debelle (2017).

Bank lending composition and non-performing loans
In per cent Graph 8
Residential mortgage loans share of Business loans as a share of total Non-performing loans ratio2
total loans1 loans1
CN
80 80 6.0
CH 60 60 4.5
AU
BR JP HK
2006
2006
GB
FR 40 SEES LU 40 3.0
SE IT US MX
US CA AU FR
HK ES CH
Higher mortgage
NL 20 GB CA Higher business 20 1.5
MX JP lending share
IN IT CN DE NL SG lending share
LU IN
SG 0 0 0.0
20 40 60 80 100 20 40 60 80 100 Euro Other United Other EMEs
area Europe States advanced
2016 2016
2005 2013
2009 2016
1
For country codes, refer to Graph 4. 2 Non-performing loans are reported gross of loans provisions and can be subject to significant
measurement differences across countries. Weighted average for groups of countries, based on total loans. Euro area calculated as a weighted
average of individual reporting countries, as in Annex Table 1.16.
Bank funding structures Graph 9
Funding shares1, 2 Interbank borrowing as a share of Loan-to-deposit ratios1

total funding
Per cent Per cent Per cent
60 40 100
45 30 75
30 20 50
15 10 25
0 0 0
Euro Other United Other EMEs Euro area4 Sweden Switzerland United Euro Other United Other EMEs
area Europe States advanced Kingdom area Europe States advanced
2007: 2016: Peak 2006–09 2016 2007 2012 2016

Customer deposits
3
Short-term funding
Long-term funding
1
Based on the sample of individual advanced economy and EME banks in Annex 2, subject to data availability. 2 Relative to total liabilities,
asset-weighted average ratios. 3 Excluding customer deposits. 4 Annual peak value of weighted average based on total assets for
Belgium, France, Germany, Luxembourg, Netherlands and Spain.
Sources: Fitch Connect; SNL; national data; BIS calculations.

Greater use of deposit funding since the crisis has led to a decline in many
advanced economy banks’ loan-to-deposit ratios – an indicator of banks’ use of
wholesale markets to fund their lending (Graph 9, right-hand panel). Loan-to-deposit
ratios for EME banking systems have typically risen since the crisis, albeit from
relatively low levels, reflecting notable credit growth.
Capital positions
After the crisis, both markets and supervisors have put increased emphasis on
stronger capitalisation as being a key determinant of banks’ capacity to cope with
adverse shocks. Strengthened capital buffers at the bank level have broader benefits
at the system level because of the potential for reduced contagion between banks.
Banks’ common equity capital ratios have generally risen significantly over the
past decade (Graph 10, left-hand panel and Annex Table 1.19–1.21). Outside Europe,
most of the improvement in system-wide capital ratios has been due to increases in
capital (Graph 10, centre panel). Banks have raised higher-quality common equity
capital by issuing new equity in the market and/or retaining earnings. The
accumulation of retained earnings has been supported by dividend payout ratios that
have been cut despite the fall in returns on equity, often because of supervisory
guidance.25 Payout ratios have increased recently in some countries where
profitability has recovered to a greater extent.
Banking capitalisation Graph 10
Bank capital ratios1, 2 Adjustment strategies1, 3 Risk-weighted assets as a share of

total assets, in per cent4
Per cent Per cent
16 14 100
IN
12 7 80
IT KR BRUS
SG
2007
8 0 ES 60
MX
JP AU
4 GB More RWA
–7 40
SE CA
FR
0 –14 20
Euro Other United Other EMEs Euro Other United Other EMEs
area Europe States advanced area Europe States advanced 40 60 80 100 120
2016
CET1/RWA: CET1/total assets: Total change, Change in:

2007 2007–16 Total assets
2012 RWA density
2016 Capital
1
Based on the sample of individual advanced economy and EME banks in Annex 2, subject to data availability. 2 Median
ratios. 3 Decomposes the change in the Common Equity Tier 1 (CET1) capital ratio into additive components. The total change in the ratios
is indicated by dots. The contribution of a particular component is denoted by the height of the corresponding segment. A negative
contribution indicates that the component had a capital ratio-reducing effect. All figures are weighted averages using end-2016 total assets
as weights. 4 For country codes, refer to Graph 4.
Sources: B Cohen and M Scatigna, “Banks and capital requirements: channels of adjustment”, BIS Working Papers, no 443, March 2014; SNL;
BIS calculations.
25
Cohen and Scatigna (2015).

Declines in risk-weighted assets have also contributed to higher capital positions
in most regions (Graph 10, right-hand panel). The average risk weight (“risk density”)
on banks’ assets has fallen, reflecting a shift in the composition of credit portfolios
towards assets with lower risk weights (eg more liquid assets and mortgages in a
number of cases, less complex assets) and an improving credit environment
(including lower interest rates) (Annex Table 1.22).26 In addition, optimisation of
banks’ internal capital models and practices in the risk-based capital framework has
played a role.27 Balance sheet contraction has also been a factor in many European
banking systems – for example, it accounted for about half of the increase in euro
area banks’ capital ratios. By contrast, EME bank balance sheets have grown
substantially, modestly dampening their capital ratios.
Bank leverage has declined significantly since the crisis (Annex Table 1.23). For
example, the ratio of Tier 1 capital to total (unweighted) assets for US banks rose from
7.2% in 2008 to 9.3% in 2016, and from 3.7% to 5.8% for euro area banks over the
same period.28
Bank capital profiles and revenue mix1 Graph 11
Capital-to-assets by business model Risk density by business model2 Revenue shares

Per cent, median bank Per cent, median bank Per cent
Advanced economies EMEs Advanced economies EMEs G-SIBs Other large banks
10 75 100
8 60 80
6 45 60
4 30 40
2 15 20
0 0 0
Trading
and
Retail-
funded
Wholesale-
funded
processing
Retail-
funded
Universal
Retail-
funded
Universal
Wholesale-
funded
Trading
and
processing
Retail-
funded
Universal
Universal
05–06 10–11 15–16 05–06 10–11 15–16
Net interest
Net fees and commissions
2005–07 2010–13 2005–07 2010–13 Net securities
2008–09 2014–16 2008–09 2014–16 Other
1
Based on the sample of individual advanced economy and EME banks in Annex 2, subject to data availability, data from 2005 to 2007 for
EMEs not shown. Total assets, net of derivative positions; period ratios are simple averages across the years. 2 Ratio of risk-weighted assets
to total assets.
Sources: Roengpitya et al (2017); Fitch Connect; SNL.
26
It should be noted that average risk weights may not fully capture risks to banks associated with a
build-up of common exposures over the business cycle or other factors contributing to the build-up
of systemic risk.
27
One example is synthetic securitisations, where the ownership of the securitised assets remains on
the balance sheet of the originators, but some part of the credit risk is transferred to investors.
28
Comparisons of leverage ratios across regions can be complicated by different accounting standards,
in particular the treatment of derivatives on the balance sheet.

Capitalisation has improved across all business model types in advanced
economies, most notably for trading and universal banks, which were more highly
leveraged prior to the crisis (Graph 11, left-hand panel). A reduction in average risk
weights has contributed to higher risk-based capital positions for all business profiles
except trading banks, which have seen a significant rise in their risk-weighted assets
because of an increased focus on lending and the post-crisis introduction of stricter
minimum regulatory capital requirements for market risk and complex securities
(Graph 11, centre panel).
Revenue mix
Revenue shares can provide an indication of the type of broad activity that banks
engage in, and thus may help signal shifts in bank business models. There has been
a marked change in the revenue mix of G-SIBs and other large banks at a high level
from the pre-crisis period, with their shares of net interest revenue rising substantially
at the expense of securities revenue and “other revenue” (which includes, notably,
securitisation revenue) (Graph 11, right-hand panel). Advanced economy banks,
particularly those from Canada, the United States and larger European countries, have
driven the aggregate shift towards net interest revenue, whereas EME banks,
especially Chinese banks, exhibit the opposite trend as revenue composition has
shifted towards net fees and commissions, in part reflecting diversification beyond
their heavy emphasis on lending.
The changes in revenue mix are consistent with the above-noted repositioning
of many advanced economy banks’ business models. Other data that classify revenue
into industry segments support this trend, showing an overall shift into commercial
banking (credit intermediation) and a decline in the share of investment banking
revenue and revenue from non-traditional sources.29 They also show a narrowing of
the scope of bank activities, particularly among larger banks. There is some variation
across firms, however, with higher capitalised banks on average having widened their
revenue mix in the post-crisis period.
3.4 International banking
In the decade prior to the financial crisis, banking became increasingly global and
grew faster than world economic activity and trade.30 International banking activity
has contracted markedly since, with total consolidated bank claims on foreign
jurisdictions declining by about 16% over 2007–16 (Table 1).
From a bank nationality perspective, the global reduction in foreign banking can
be attributed to the retrenchment of European banks: their total foreign claims fell by
around 40% over this period, whereas claims of non-European banks rose in
aggregate (Annex Tables 1.24 and 1.25). Most European banks reported significantly
lower foreign claims, with Spanish banks as the notable exception.
29
Specifically, segment data based on the North American Industry Classification System (NAICS).
30
McCauley et al (2017).

Much of the reduction in European banks’ claims occurred within the European
region, in particular the troubled euro area countries (Graph 12, left-hand panel).
European banks also pulled back from other regions, including the United States.
The post-crisis adjustments of many large advanced economy banks have
translated into a substantial decline in the number of jurisdictions in which they have
operations. Banks tended to withdraw when local operations were smaller in market
share or length of tenure, or from countries that are more distant or less important
for trade.31 This trend has contributed to a reduction in the share of foreign bank
assets in a number of domestic banking systems (Annex Table 1.26).
As European banks pulled back, the foreign claims of other banking systems have
tended to rise. US banks and banks from advanced economies that were less affected
by the financial crisis – such as major banks from Australia, Canada and Japan – have
expanded their international presence to varying degrees. Banks from a number of
emerging markets, including those from larger markets (Brazil, China, India, Russia),
have also grown their international footprint, mainly within their own region. The
international expansion of the five largest Chinese banks is particularly notable: their
foreign assets have risen sixfold since 2007, to be roughly equivalent to around 6%
of global foreign bank claims at end-2016.32
In comparison to the pre-crisis global expansion of many advanced economy
banks, EME banks’ expansion has been relatively regional in nature. The increase in
EMEs’ foreign banking presence within their own regions, albeit from relatively small
positions leading into the crisis, is particularly evident in Southeast Asia and has also
occurred in Central America and the Commonwealth of Independent States
(Graph 12, centre panel).33
Consolidated foreign bank claims1 Table 1
USD billions % change Share of total (%)

2007 2016 2007–2016 2007 2016
Aggregate claims 29,302 24,624 –16 100 100
Of which, by bank nationality:
Advanced Europe 23,608 14,300 –39 81 58
Other advanced 5,147 9,046 76 18 37
EMEs 300 744 148 1 3
Of which, by claim type:
International2 18,285 14,263 –22 62 58
Local claims in local currency 11,017 10,348 –6 38 42
1
Claims for each year are as at end-December. Aggregate claims relate to that of domestic banks in reporting countries 2
Comprise cross-
border claims in all currencies and local claims in foreign currency of counterparties where booking office is located.
Source: BIS consolidated banking statistics (immediate counterparty basis).
31
Claessens and van Horen (2014).
32
Consolidated foreign claims data shown in Table 1 and elsewhere in this report do not encompass
Chinese and Russian banks. We include foreign assets figures here for Chinese banks for rough
illustrative purposes only.
33
CGFS (2014b), IMF (2015) and Remolona and Shim (2015).

Foreign banking Graph 12
European banks’ foreign claims on: Share of international claims on EME Share of local claims in local
regions by intraregional banks1 currency3
USD trn Per cent
AU ES
9 30 60
CA
GB
CH SG NL SE
2007
6 20 IT 40
US
FR
BE
3 10 20
DE
IN More
JP
local claims
0 0 BR
0
01 03 05 07 09 11 13 15 17 03 05 07 09 11 13 15 17 20 40 60 80
2016
United States Latin America Asia-Pacific2
2
Euro area Other countries Latin America and the Carribbean
Other developed Europe
1
Sum of all cross-border claims and locally extended claims in foreign currency. 2 For Asia-Pacific, sum of international claims on the region
of banks headquartered in Chinese Taipei, Hong Kong SAR, India, Korea, Singapore and the offices of banks located in the region which have
a parent institution from a non-BIS reporting country (assuming these are headquartered in Asia). For Latin America and the Caribbean, sum
of international claims on the region of regional banks (Brazil, Chile, Mexico, Panama) and the offices of banks located in the region which
have a parent institution from a non-BIS reporting country (assuming these are headquartered in Latin America and the Caribbean). 3 As a
percentage of total foreign claims; for country codes, refer to Graph 4.
Source: BIS consolidated banking statistics on an immediate counterparty basis.
Aside from changes in the geographical distribution of exposures, the most

notable trend in foreign banking is the shift away from cross-border lending towards
local lending through banks’ branches and subsidiaries. International claims declined
by more than 20% over 2007–16, whereas the decline in local currency claims was
more modest. In addition, banks’ foreign claims have been increasingly funded
locally. These aggregate trends mainly reflect changes for euro area banks, although
greater localisation of foreign banking is evident for a range of bank nationalities and
pre-existing models (Graph 12, right-hand panel). For example, Swedish banks now
largely fund their operations in the Baltics from local deposits, whereas previously
most funding was sourced from the parent using wholesale markets.
3.5 Bank performance
Overall profitability
Profitability has declined considerably in many banking systems over the past decade.
System-wide returns on equity (RoE) – that is, net profit divided by average equity –
were generally in the range of 10–15% in the few years preceding the crisis (Graph 13,
left-hand panel and Annex Table 1.27). Such returns on equity were historically high
and, in many cases, were supported by unsustainable leverage and risk-taking (see
discussion of excesses in Section 2.1).
There was a stark reversal during 2008–09, with banks in Europe and the United
States recording aggregate losses. Although RoE has since drifted higher for the US
banking sector, it has remained below shareholder return expectations, which

are tied to these banks’ estimated cost of equity (see 4.3 below). Profitability has been
even weaker within Europe, both for large and smaller banks. Banks in Australia,
Canada and EMEs performed better during the crisis, yet their rates of return have
also slowed over recent years. From a business model perspective, RoE has fallen
across all types in the post-crisis period (Graph 14, right-hand panel).
Banks’ return on assets (RoA) has followed a similar pattern across countries,
although the post-crisis decline is generally less noticeable due to relative changes in
leverage (Graph 13, left-hand panel and Annex Table 1.28). Return on risk-weighted
assets – a simple measure of risk-adjusted returns – provides a potentially more
favourable view of bank profitability trends (Annex Table 1.29). It has declined by less
than RoE for some large banking systems and is near pre-crisis levels (with the
exception of some European countries and EMEs).
Revenues and costs

The decline in non-interest income is the main driver behind the post-crisis fall in
revenue relative to assets for most advanced economy systems (Graph 14, left-hand
panel and Annex Tables 1.32–1.34). A key driver is trading revenue at large banks,
mainly reflecting a reduction in fixed income trading revenues (see Section 5.2 below).
Net interest revenue across the large advanced economies was weighed upon by
weak lending growth in the few years after the crisis, as the private sector in aggregate
deleveraged and banks rehabilitated their balance sheets (Graph 14, centre panel)
(see Section 5.1 below). Lending growth has increased over the past few years, with
the exception of the United Kingdom. Lending has expanded relatively strongly in a
number of EMEs and Canada over the post-crisis period.
Bank profitability indicators

In per cent Graph 13
Return on equity and assets1 Return on equity by business model2

Advanced economies EMEs
2 20 20
1 10 10
0 0 0
–1 –10 –10
Euro Other United Other EMEs Wholesale- Trading Retail- Universal Retail- Universal
funded and funded funded
area Europe States advanced processing
Return on assets (lhs): Return on equity (rhs):

2005–07 2008–09 2010–13 2014–16
2000–03 2000–03
2004–07 2004–07
2008–09 2008–09
2010–13 2010–13
2014–16 2014–16
1
Weighted average based on average assets; period ratios are simple averages across the years. Euro area calculated as a weighted average
of individual reporting countries, as in Annex Tables 1.27 and 1.28. 2 Based on the sample of individual advanced economy and EME banks
in Annex 2, subject to data availability; period ratios are simple averages across the years.
Sources: Roengpitya et al (2017); Fitch Connect; national data.

Bank revenues Graph 14
Revenue, in per cent of average Annual loan growth2 Net interest revenue by business
assets1 model, in per cent of average assets3
Per cent Per cent Per cent, median bank
IN Advanced economies EMEs
4.5 HK CA CN 50 3
MX
Change 2010–13
3.0 25 2
SG
AU
BR JP SE KR
CH FR
1.5 US 0 1
IT NL
DE
GB BE LU
ES
0.0 –25 0
Wholesale-
funded
Trading
and
processing
Retail-
funded
Universal
Retail-
funded
Universal
Euro Other United Other EMEs –10 0 10 20 30 40
area Europe States advanced Change 2014–16
2005–06: 2015–16:
Net interest revenue 2005–07 2010–13
Non-interest revenue 2008–09 2014–16
1
of individual reporting countries, as in Annex Table 1.33 and 1.34. 2 For country codes, refer to Graph 4. 3 Based on the sample of
individual advanced economy and EME banks in Annex 2, subject to data availability, period ratios are simple averages across the years.
Relative to total assets net of derivative positions.
Sources: Roengpitya et al (2017); Fitch Connect; national data.
Net interest margins (NIMs) – that is, net interest income as a percentage of
average interest-earning assets – are an indicator of the overall profitability of banks’
interest-earning activities. Trends in NIMs vary across countries. The NIM has been
broadly stable for several banking systems, including Sweden and the United
Kingdom, and has risen for the Netherlands (Annex Tables 1.30 and 1.33).34 This is
despite the decline in interest rates to very low or negative levels in these countries,
and increases in the shares of liquid assets and stable funding. In contrast, clear
declines in NIMs can be observed in a number of other countries over 2010–16,
including Brazil, Japan, Korea, Singapore and the United States.35 As a share of
average assets, net interest income has been fairly steady for advanced economy
bank types except for retail-funded banks, and has declined for EME banks (Graph
14, right-hand panel).
Cost efficiency in the large advanced economy banking systems seems to have
improved little in the post-crisis period, despite renewed focus on cost containment
(Graph 15, left-hand and centre panels and Annex Tables 1.31 and 1.35). Operating
costs have generally risen broadly in line with, or by a bit more than, operating
income. Personnel expenses have remained generally steady as a share of operating
34
NIM data are not available for a number of jurisdictions in Annex Table 1.28. Readers can also refer
to Annex Table 1.31, which presents net interest income as a per cent of average assets. These data
should be interpreted with caution since trends could be affected by shifts in the share of interest-
earning assets in total assets.
35
Historical differences in NIMs across countries could reflect numerous factors including the asset and
liability mix and competitive dynamics. We do not seek to discuss the historical level differences.

Bank costs Graph 15
Costs1 Cost-to-income ratio2 GSIBs’ legacy costs

In per cent of average assets Percentage points of equity
Europe United United

excluding Kingdom States
Average between 2005–06

60 3 80 United 4.5
DE Kingdom
CA IT CH
40 2 GB 60 3.0
BR US MX
SEAU
IN JP
CN KR
20 1 ES 40 1.5
SG
LU
Less cost
efficient
0 0 20 0.0
12 14 16 12 14 16 12 14 16
Euro Other United Other EMEs 40 60 80 100
area Europe States advanced Average between 2015–16 Litigation expenses
Restructing costs
2005–06: 2015–16:
Operating costs
Credit losses
Cost-to-income ratio (lhs)
1
of individual reporting countries, as in Annex Tables 1.31, 1.35 and 1.36. 2 For country codes, refer to Graph 4.
Sources: IMF; national data.
income, while some banks have incurred sizeable restructuring costs and costs
associated with past instances of misconduct, although these have subsided more
recently for major US banks (Graph 15, right-hand panel). In contrast to the large
advanced economies, there has been a decline in the cost-to-income ratios for some
banking systems that already had relatively low ratios, including Australia, China and
Sweden.
Significant loan losses have also contributed to the decline in banking sector
profitability (Annex Table 1.36). Banks in some euro area countries have incurred
elevated credit losses and high levels of non-performing loans during the post-crisis
period, while a rise in NPLs has weighed on the profitability of several large EME
banking systems in the past couple of years. In contrast, reductions in credit losses
from crisis peaks have provided a boost to profitability in some other countries,
notably the United Kingdom and the United States.
3.6 Summary of key trends in banking
As the crisis revealed structural weaknesses in banking systems, the analysis above
details substantial change in the global banking sector subsequently. Adjustments
have been more pronounced in Europe and the United States, which were more
affected by the crisis. Key trends are summarised below.
Market structure. Banking system asset size and other capacity metrics (the
number of bank branches and employees) have generally fallen in Europe relative to
economic activity and the population, but have expanded in countries less affected
by the crisis, particularly the large EMEs. Bank numbers have also fallen in Europe and
the United States, although banking sector adjustment through the exit of banks –
the extensive margin – has been less pronounced than through the down-sizing of

individual banks. Concentration in banking systems has tended to increase after the
crisis, with some exceptions.
Bank regulatory buffers, business models and international activities. Banks
have typically increased their capital and liquidity buffers significantly. Advanced
economy banks have tended to reorient their businesses away from trading and more
complex activities towards less capital-intensive activities and traditional commercial
banking. This pattern is evident in bank asset portfolios, revenue mix and the
increased use of customer deposit funding. Of particular note has been a significant
increase in mortgage lending across a number of banking systems. Large European
and US banks have also become more focused in their international banking activities.
European banks in aggregate have reduced their foreign exposures substantially,
whereas banks from the large EMEs and other countries that fared better in the crisis
have expanded. The relatively significant business model adjustments of European
banks and large US banks are consistent with these banks’ greater pre-crisis
expansion into new activities and jurisdictions.
Performance. Bank returns on equity have declined across countries and
business model types since the crisis, in part because of lower leverage (see discussion
in Section 4.4 below). Outcomes have been weakest in Europe, where credit or
operating costs (including misconduct costs) have remained elevated relative to
revenues. Bank profitability has declined noticeably in some EMEs more recently, as
economic growth has slowed and loan quality deteriorated.
4. Implications for the stability of the banking sector
This chapter assesses the implications of post-crisis changes in bank structure and
performance for the stability of the sector. Key trends identified in Chapter 3 are
supplemented with other information important for assessing the stability of the
banking sector.
The discussion of banking sector stability is organised as follows. First, banks’
resilience is considered, drawing on their progress in building regulatory buffers and
stress test results. The second section appraises key changes in banks’ risk
management based largely on feedback from supervisors. The third section reviews
market indicators to gauge trends in investor sentiment towards banks. The fourth
section examines the capacity of banks to generate profits under a number of
scenarios, given that accumulation of retained earnings is critical to their future
provision of intermediation activities and their resilience. The chapter then takes a
system-wide perspective of bank risk and resilience, including assessments of
changes in market structure, interconnectedness, risk correlations and international
banking. The final section concludes with an overall assessment of the implications
for banking stability.
4.1 Bank risk buffers and risk-taking
The crisis revealed that many banks were not maintaining adequate capital and
liquidity buffers to cope with stressed macro-financial conditions, let alone the
excessive financial risks that some advanced economy banks had taken on.

In response to regulatory requirements, the banking sector globally has
enhanced its resilience to adverse shocks through a substantial strengthening of
capital positions. The process of adjustment to existing Basel III capital standards is
mostly complete. As at end-2016, all large banks in the BCBS’s global monitoring
sample met their “fully phased-in” Basel III target CET1 regulatory capital ratios,
including their G-SIB ratio surcharge where relevant (Graph 16, left-hand panel).36
Moreover, large banks’ CET1 capital was 70% higher relative to aggregate risk-
weighted assets (measured on a fully phased-in basis) than in mid-2011 (and, within
this, 80% higher for G-SIBs).37 All large banks also met the minimum Basel III Tier 1
leverage ratio requirement at end-2016, with Tier 1 capital about 65% higher relative
to aggregate unweighted exposures than in 2011 (Graph 16, centre panel).
Banks’ buffers against short-term liquidity shocks have also been considerably
strengthened due to the phase-in of new regulatory liquidity requirements. As at end-
2016, the weighted average LCR was around 130%, with more than 90% of large
banks (and all G-SIBs) having a fully-phased in LCR of at least 100% – the minimum
regulatory requirement – up from 70% in 2012 (Graph 16, right-hand panel). An LCR
of 100% or more indicates that a bank has enough liquid assets to withstand its
expected net cash outflows within a 30-day period of stress.
There are also indications of a reduction in risk on both sides of advanced economy
banks’ balance sheets: many large banks have reduced their proportion of complex
assets and assets (or borrowers) that typically pose more risk, while there has been a
widespread shift away from short-term wholesale funding towards more stable
funding sources. A reduction in the risk profile of banks’ balance sheets means that,
Large banks’ fully phased-in Basel II regulatory ratios1

CET1 Leverage ratio Share of banks meeting minimum

LCR and NFSR
Weighted average Weighted average
35 15 100
28 12 80
21 9 60
14 6 40
7 3 20
0 0 0
2011 2012 2013 2014 2015 2011 2012 2013 2014 2015 2012 2013 2014 2015 2016
CET1 Min-max range Leverage ratio Min-max range LCR NSFR
1
Consistent sample of approximately 90 large banks, including G-SIBs. Dashed lines represent the minimum regulatory requirement for the
CET1 ratio (including the capital conservation buffer) and the leverage ratio.
Source: Basel Committee on Banking Supervision.
36
Fully phased-in ratios are based on the full implementation of stricter capital standards, whether or
not that is the case in practice. For further details, see BCBS (2017).
37
The (larger) increase in CET1 compared with banks’ pre-crisis balance sheets cannot be quantified
because fully phased-in risk-weighted assets under Basel III are not available before 2011.

Average bank common equity capital positions in stress test
In per cent Table 2
United States European Union United Kingdom
Stress test year Starting Stressed Starting Stressed Starting Stressed
2009 5.5 4
2011 8.9 7.7
2012 10.1 6.2
2013 11.1 7.4
2014 11.5 6.6 11.1 8.5
2015 11.9 7.1
2016 12.3 7.1 13.2 9.4 12.6 8.8
2017 12.5 7.2
1
These results should be viewed as indicative trends given that there are important differences between these stress tests (eg projection
horizon) and related disclosures (eg minimum vs ending capital), and because there have been changes over time in the severity of the tests,
in applicable capital rules, and in the population of tested banks.
Sources: Bank of England, European Banking Authority, Federal Reserve
for a given adverse shock, banks will need to draw less on their capital and liquidity
buffers. However, the assessment of bank risk is complicated by uncertainty about
banks’ activities and the macroeconomic and competitive environment in which
banks operate. Moreover, whether the observed changes represent a lasting
adjustment in bank risk will also depend on the incentives banks face, as well as their
risk management (see discussion in Section 4.2 below).
The shift to more stable funding sources is evident in banks’ NSFRs – an indicator
of the degree to which banks maintain a stable funding structure for their exposures.
As at end-2016, the weighted average Net Stable Funding Ratio (NSFR) for large
banks was around 115%, with about 95% of large banks (and all G-SIBs) meeting the
fully phased-in minimum NSFR requirement of 100%, up from 43% in 2012.
Since the crisis, bank stress tests have become an increasingly important way for
the supervisory community (and banks themselves) to assess banks’ resilience to
adverse conditions and ensure that they have adequate resources to continue
supporting the real economy throughout the credit cycle.38 Supervisors are also using
these tests to drive risk and capital management practices, as well as support market
and stakeholder confidence in banks.39 These tests tend to be forward-looking,
scenario-based and integrated into overall supervisory processes, although there are
important differences across jurisdictions.40
Recent stress test results indicate that projected capital levels under severe stress
scenarios have remained above regulatory minimums (Table 2). These stressed capital
38
Some central banks are also using stress tests to assess the sustainability of banks’ business models
in adverse environments. For example, the Bank of England introduced an exploratory scenario into
its stress testing to consider how the UK banking system might evolve if recent headwinds to bank
profitability persisted or intensified over a seven-year horizon.
39
For example, see Tarullo (2016) and Enria (2017).
40
For a comparison of selected bank stress testing frameworks, see Dent and Westwood (2016).

levels exceed capital positions that banks were maintaining prior to the crisis-related
losses (see Graph 10, left-hand panel).41
4.2 Bank risk governance and management
Inadequate governance and risk management were apparent in poor risk outcomes
at many banks during the crisis.42 They are also evident in the frequency and severity
of bank conduct risk instances and associated legal and reputational costs over recent
years. The need for robust risk management has perhaps become even more crucial
since the crisis: supervisory expectations and stakeholder scrutiny have risen, while
complexity in some aspects of risk management has grown – for example, in relation
to model risk and cybersecurity.
Yet how banks take, measure and control risks tends to be much harder to
observe than some other aspects of banking. The Group therefore reached out to
bank supervisors in member jurisdictions to obtain information on banks’ risk
management. Supervisors from 12 jurisdictions provided information on the set of
issues below. This information has been supplemented with several other official and
private sector studies.
In sum, banks appear to have improved their risk governance and risk
management practices since the crisis, although further progress is often needed to
meet supervisory expectations. It can be difficult to disentangle the drivers of
improvements: while many reflect regulatory and supervisory changes, banks have
also been adjusting as a result of their crisis experience and market pressure.
Bank risk appetite.43 Bank-wide risk appetite has declined significantly in
countries that were more affected by the crisis or where credit conditions have
deteriorated recently, whereas there has been a modest reduction or no change in
other countries. Adjustments have been more pronounced for liquidity risk, an area
where bank and regulatory frameworks were far less developed before the crisis.
Several supervisors noted that shifts in risk appetite are more apparent in investment
banking than commercial banking, as well as for activities that could entail a higher
chance of conduct or compliance risk, in particular those with a cross-border element.
Risk governance and management structures. Many banks have been
enhancing their risk appetite frameworks in response to supervisory focus. This
process has involved broadening their risk metrics, including for liquidity risk, stressed
events and, more recently, operational and reputational risks. Boards are also better
41
According to Nouy (2017), applying the capital depletion for 26 euro area banks under the 2016
stress test to these banks’ 2007 Tier 1 capital position yields an average Tier 1 ratio of 3.3% (indicating
the average bank would fall below minimum regulatory ratios), while three banks’ capital base was
completely wiped out.
42
In the immediate aftermath of the crisis, the Senior Supervisors Group (2008 and 2009) reported that
key firm-wide risk management practices – risk identification, valuations, balance sheet management,
and risk measurement and reporting – helped to differentiate better and weaker performing banks
during the crisis. Ellul and Yerramilli (2013) provide empirical evidence that the strength and
independence of risk management functions at US banks is associated with better risk outcomes.
43
Risk appetite is a broad concept that can encompass various targets set by bank boards. The target
metrics used will differ across banks but are likely to include some of the following: regulatory capital
ratios, liquidity and funding indicators, exposure or concentration limits, VaR and stress test results.

integrating risk into their strategic and capital planning. Even so, some supervisors
indicated shortcomings in the alignment of risk appetite with bank business models,
as well as the degree to which banks cover all risk dimensions and embed their
approach to risk across business units.44
Risk management expertise on boards has increased as the focus on risk has
grown. The chief risk officer role has become more prominent and has direct and
regular access to the board, which was not always the case prior to the crisis. However,
some supervisors noted a need for additional independent directors, and there
remains concern as to whether boards have sufficient expertise to effectively digest,
debate and challenge technical risk information.
Risk and compliance functions have become more important in terms of their
budget and staffing, and through their influence on decisions – for example, some
supervisors have observed more evidence of new investment proposals being
successfully challenged by risk staff. Supervisors also noted enhancements to policies
and procedures for risk and compliance issues, and the establishment of more
specialised risk committees. Banks have made front-line staff more responsible for
risk under the “three lines of defence” risk management model, such as by clarifying
responsibilities, establishing control units in the front line and increasing training.
Another development is the greater attention given to “risk culture”.45 Firms are
recognising the need to incorporate culture in their firm-wide planning and for the
CEO and CRO to influence culture through their communication and their actions.
Other strategies include the development of industry-wide codes of conduct, and the
consideration of cultural issues in recruitment and training. Banks are looking to
better understand and measure their culture through staff surveys and audits, and
specific dashboards.46 In a private sector survey, most banks acknowledge that
enhancing organisational and risk culture remains a work in progress.47
Risk management information systems. Banks have increased the frequency
and quantity of risk reporting for the board and senior officers. They have also been
upgrading their reports through greater automation and standardisation, and
increasing the granularity of information for specific business groups or risks. Systems
that feed into regulatory processes – for example, liquidity metrics, stress testing and
VaR models – have tended to be prioritised. Nonetheless, supervisors emphasised
that important gaps remain in banks’ risk data capabilities. Indeed, in March 2017 the
Basel Committee on Banking Supervision (BCBS) concluded that G-SIBs had
unsatisfactory compliance with its principles for effective risk data aggregation and
risk reporting, which were issued in 2013.48 Banks have faced challenges on several
44
For example, see ECB (2016b). In a survey of 67 large banks across 29 countries, more than half of
respondents indicated some difficulty moving their firm-wide risk appetite approach further into the
business (Ernst and Young (2015)).
45
Risk culture should be distinguished from the broader concept of non-financial risk. Risk culture can
be considered as “the norms and traditions of behavior of individuals and of groups within an
organization that determine the way in which they identify, understand, discuss, and act on the risks
the organization confronts and the risks it takes” (International Institute of Finance (2009)).
46
For example, see APRA (2016).
47
Ernst and Young (2015).
48
See BCBS (2017).

fronts, including the management of information technology and data projects,
implementation of firm-wide data policies and weakness in data quality controls.
Compensation and other incentives. In the post-crisis period, the proportion
of remuneration that is variable has fallen. The share subject to deferrals and
clawbacks has also risen for senior officers and material risk-taking staff. Risk now has
a greater weight in ex ante remuneration decisions and is better reflected in internal
capital pricing mechanisms (thus affecting pay outcomes through altered business
unit profitability).49 Measures to increase the accountability of senior executives and
boards have also been introduced by some authorities.50
4.3 Market-based indicators of banks
Market-based indicators are often used to assess bank performance and risk.
Specifically, market-based indicators give insight into market participants’ perception
of an individual bank’s solvency and risk of failure, adequacy and riskiness of returns,
and correlation with other (listed) banks. Market indicators can provide a useful
complement to accounting-based metrics and are more timely. However, they are not
without limitations as indicators of bank risk.51 Moreover, market-based indicators
provide only limited information about systemic risk, and they are not representative
for banking systems with a high share of non-publicly traded firms.
Debt and equity market indicators
Spreads on credit default swaps (CDS) represent the cost of protecting against a loss
on the debt securities that they reference. Bank CDS spreads have fallen significantly
from the highs seen during the 2007–09 crisis and the subsequent 2011–12 euro area
sovereign debt crisis, both in absolute terms and relative to the broader CDS market
(see Graph 1, right-hand panel). The distribution of spreads across G-SIBs has also
narrowed. These trends signal a marked reduction in market participants’ perception
of bank credit risk. While bank CDS spreads remain above those of the pre-crisis
period, this result can be largely attributed to a change in investor attitudes to risk
and presumably greater risk of bail-in from regulatory reforms (see below). Bank
credit risk was underpriced by investors prior to the crisis, with CDS spreads for some
banks not discernibly different from those of highly rated sovereigns during that time.
Whereas bank CDS pricing has improved over the post-crisis period, equity prices
for some banks have remained depressed (see Graph 1, left-hand panel). Attention
has focused mostly on the low level of equity prices compared with book valuations
49
Nonetheless, a survey of more than 30,000 employees at seven major Australian and Canadian banks
found that, from a risk perspective, remuneration structures are viewed significantly less favourably
than are risk frameworks, managers and training (Griffin and Sheedy (2017)).
50
For example, in the United Kingdom, a Senior Manager Regime was introduced in 2016. Under this
regime senior managers can be held accountable if they fail to take reasonable steps (including
training or ensuring proper oversight) to prevent regulatory breaches in their areas of responsibility.
51
Security prices refer to a specific claim on the cash flows (or assets) of a firm, and thus reflect not
only firm-wide risk but also the distribution of claims within the capital structure. This issue is highly
pertinent for banks because of their leveraged nature and the changes to banks’ capital structure
arising from post-crisis regulatory reforms. Furthermore, markets can misprice claims for periods of
time, with opacity of bank balance sheets perhaps making this more likely than for some other firms.

– that is, price-to-book (P/B) ratios, a common equity valuation metric. For some
banks in Europe, Japan and, to a lesser extent, the United States, P/B ratios have been
persistently below one, whereas values of more than two were common prior to the
crisis (Graph 17, left-hand panel). Over recent years, P/B ratios have tended to be
lower for G-SIBs relative to other large banks (Graph 17, right-hand panel).
Discussions with equity market analysts indicate that valuations of some G-SIBs are
below those of simpler banks that are achieving broadly comparable returns in the
same home jurisdictions, and are also below those that would be implied from valuing
some G-SIBs’ individual businesses separately.52
Price-to-book ratios1 Graph 17
By region2 G-SIBs and non-GSIBs
3.0 3.0
2.5 2.5
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
05 06 07 08 09 10 11 12 13 14 15 16 17 05 06 07 08 09 10 11 12 13 14 15 16 17
United States Other advanced3 EMEs4 G-SIB Non-G-SIB
Japan Europe
1
Simple averages across the sample. 2 Based on the sample of individual advanced economy and EME banks in Annex 2, subject to data
availability. 3 Australia and Canada. 4 Brazil, China, Chinese Taipei, Hong Kong SAR, India, Israel, Korea, Malaysia, Mexico, Qatar, Russia,
Saudi Arabia, Singapore, South Africa, Turkey and Venezuela.
Source: SNL.
P/B ratios of less than one imply destruction of bank net asset value in the future.
Put differently, returns on these banks’ equity are expected to be below those
demanded by investors – that is, their cost of equity. In fact, actual returns on equity for
some large banks in Europe and the United States have been significantly below their
estimated cost of equity over recent years (Graph 18, left-hand and centre panels).
Current depressed P/B ratios signal that markets anticipate a continuation of
inadequate returns for (at least some) banks in the period ahead. Consistent with this
view, Sarin and Summers (2016) argue that the marked reduction in P/B ratios is
symptomatic of a decline in the franchise value of banks (ie the ability of banks to
generate profits in the future). Similarly, across advanced economy banks there is a
close correlation between P/B ratios and expected future returns on equity (Graph 18,
right-hand panel). Moreover, a high share of firms with relatively weak expected returns
trade below book value.
Cost of equity can be considered an indicator of risk, since equity investors
should demand higher returns for more risky claims. Estimated cost of equity for
52
A valuation discount for financial conglomerates had been established empirically prior to the crisis
(Laeven and Levine (2007)).

banks has declined from the heights observed at the peak of the crisis, and are at
broadly similar levels than in the pre-crisis period, as a substantial fall in risk-free rates
has been offset by much higher bank equity betas (see Annex 4 for the estimation
methodology).53 Bank risk premia are higher despite the apparent reduction in risk-
taking noted above, and are likely to partly reflect investors’ reassessment of the
riskiness of banks (as observed for CDS spreads), as well as greater volatility (risk) in
bank earnings because of a confluence of bank-specific and other factors.54 It could
also be due to uncertainty about capital actions for some banks, as any new equity
issuance would dilute existing investors’ claim on future earnings. Weak valuations
may also highlight investors’ view that some banks’ efforts to restructure and alter
their business models remain a work in progress. Another possible explanation is the
diminution of implicit government subsidies due to revised bank resolution regimes.55
This latter issue is discussed below.
Bank return on equity and equity market pricing Graph 18
Cost-of-equity and actual Cost-of-equity and actual Banks’ price-to-book ratios and
return on equity return on equity projected return on equity2
Per cent Per cent
y = 3.43 +4.7x
Projected return on equity (%)

2
20 20 where R = 0.8 20
10 10 15
0 0 10
–10 –10 5
–20 –20 0
02 04 06 08 10 12 14 16 02 04 06 08 10 12 14 16 0.5 1.0 1.5 2.0 2.5 3.0
Cost of equity:
1
RoE: Cost of equity:
1
RoE: Price-to-book ratio
Japan Japan United States United States United States Other3
Europe Europe Other AEs Other AEs Europe Linear (trend line)
Japan
1
Derived from a variant of the capital asset pricing model; see Annex 4 for estimation methodology. 2 Price-to-book ratio calculated as
the average share price for 2017 to date, divided by the average book value of equity for 2015 and 2016. Projected return on equity (ROE) is
for 2018. 3 Australia, Canada, Hong Kong SAR, Korea and Singapore.
Sources: Bloomberg; Capital IQ; Datastream; Fitch Connect; SNL; Thomson Reuters; BIS calculations.
Indicators of market perceptions of government support for banks

Post-crisis reforms to bank crisis management and resolution frameworks have sought
to allow banks to be recovered or resolved in an orderly manner without reliance on
53
Cost of equity estimates should be interpreted with caution. The cost of equity is an unobservable
metric, with numerous estimation approaches.
54
Moreover, Baker and Wurgler (2015) note that higher capital requirements may raise the cost of
capital for banks because of the low risk anomaly (within the stock market, historical returns and thus
realised costs of equity are higher, not lower, for less risky equity).
55
Although implicit government guarantees should benefit debt investors more than they do equity
investors, Kelly et al (2016) find evidence that they affect stock prices.

government support. In doing so, competitive and financial stability distortions arising
from market perceptions of implicit government guarantees can be reduced.
Perceptions of government support can be inferred from market prices.
Contingent claims analysis allows the calculation of historical market-implied
expected losses on banks’ debt – in effect, contingent liabilities of the government –
based on the assumption that equity prices are less sensitive to changes in perceived
government guarantees than CDS prices because equity holders are less likely to
Market perception of government support for G-SIBs Graph 19
Contingent claims on government1 Rating grade difference between Difference in CDS spreads of junior
junior debt and senior debt2 and senior bonds
α -values Ratings upgrade Ratio
1.0 5 1.4
0.8 4 1.2
0.6 3 1.0
0.4 2 0.8
0.2 1 0.6
0.0 0 0.4
0.2
05 07 09 11 13 15 17 07 08 09 10 11 12 13 14 15 16 17 03 05 07 09 11 13 15 17
α-mean Lower bound Mean Lower bound Median Periodic average
Upper bound Upper bound
1
Alpha values are the fraction of bank losses that might become contingent government liabilities. A value of 1 implies full government
support of bank debt. 2 Moody’s credit ratings. Lower and upper bound are one standard deviation from the mean.
Sources: Bloomberg; Moody’s Investors Service.
benefit from any such support.56 “Alpha” values for a sample of G-SIBs are shown in
Graph 19 (left-hand panel), where alpha is the fraction of bank losses that might
become contingent government liabilities, varying from 0 for no government support
to 1 for full government support of bank debt. Variation in alphas should be
interpreted with care, as low values can be either signal that no government support
is expected or that bank is considered safe and no support is expected to be
necessary. The values show substantial public sector support during the financial crisis
and euro area sovereign crisis and a smaller peak during the turbulence in early 2016,
but an overall reduction in implied support over recent years.
Changes in perceptions of government support may also be inferred from
differential market assessments of junior and senior bank debt, since junior debt is
lower in the capital structure and thus more affected by revised resolution regimes. The
gap between credit ratings for G-SIBs’ junior and senior debt has widened since the
crisis, on average by one rating grade, but in some cases by two notches
56
Expected losses are inferred from CDS prices. In the case of equity, expected losses are considered
as the price of a put option written on assets, where the present value of debt is the strike price and
the value and volatility of assets determined by changes in the equity and equity options prices of
the bank. See Jobst and Gray (2013).

(Graph 19, centre panel). This suggests that a re-assessment of assumed government
support has more significantly affected the credit ratings of G-SIBs’ junior debt.57
The increase in perceived risk differentiation between junior and senior debt is
also observable in the CDS market. Market pricing of CDS spreads for G-SIBs’ junior
debt have also widened relative to those for senior debt in recent years (Graph 19,
right-hand panel). This widening appears to be a response to new terms introduced
in 2014 by the International Swaps and Derivatives Association, which included the
incorporation of government-initiated bail-in terms. This suggests that the CDS
market has begun to price in a higher expected loss arising from bail-in events.
4.4 Future profitability and resilience of banks
The capacity of banks to generate profits and accumulate capital through retained
earnings is critical to their future provision of intermediation activities and their
resilience. It is also important for access to external capital through equity markets.
Indeed, the above analysis suggests that low equity market sentiment for some banks
primarily relates to concerns over future profitability. Accordingly, this section
considers the profit outlook for individual banks using a scenario analysis. This
analysis is preceded by a decomposition of changes in post-crisis profitability. RoE is
the profit measure used because of its importance to investors and its widespread
use in banks’ own capital allocation.
As noted earlier in the report, banking sector RoE has declined across regions
and countries, as well as business model types. Within our sample of large banks,
most reported an RoE in 2015–16 which was significantly below that of a decade
earlier. Around 60% of banks had a RoE below 10%, a rough marker for banks’ cost
of equity (Graph 20, left-hand panel).58 The share is higher within the subset of
advanced economy banks, at 78%.
A DuPont analysis allows trends underlying the decline in bank RoE to be better
pinpointed. The analysis breaks down RoE into effects from financial leverage
(assets/equity), asset yield (revenue/assets) and profit margin (net profit/revenue),
with the latter two ratios being a decomposition of RoA. Lower leverage accounts for
almost half of the post-crisis decline in bank RoE (Table 3). It has pressured RoE in all
regions and for all business model types, most significantly trading banks, which were
maintaining higher leverage prior to the crisis. Profit margins have also generally been
eroded, with many European banks underperforming due to high credit or operating
expenses (including misconduct-related litigation costs). The RoE contribution from
asset revenue has fallen for most, but not all, regions and business model types, with
57
In addition, Haldane (2017) shows for major UK banks a reduction in the uplift in senior debt ratings
as a result of lesser assumed government support.
58
Cost of equity estimates for advanced economy banks are shown in Section 4.3. For illustrative
purposes, we use a 10% assumption. Given that the cost of equity should depend on a bank’s
business model and risk, it is possible that for some banks investors would be willing to accept
materially lower returns over the medium term, particularly in an environment of relatively low
interest rates. That said, in Ernst and Young (2016), 57% of large banks surveyed (and 74% of G-SIBs)
indicated that they are seeking to achieve target RoEs of 10–15% in the next three years. Furthermore,
around 70% of EU bank respondents indicated a long-term sustainable RoE of 10% or more.

pressures more acute for (retail) banks in Japan (also see Box C). Nonetheless, the
overall impact from asset yields is less than for the other components.
A scenario analysis can be used to illustrate the range of potential RoE outcomes.
The first scenario assumes that banks benefit from a more favourable cyclical
backdrop, increasing revenue and lowering credit costs to values close to the best
outcomes in their historical experience.59 Under optimistic assumptions, average
advanced economy bank RoE improves to 13%, yet around one quarter of banks post
returns below 10% and a small share of banks incurs a loss (Graph 20, centre panel).
European banks face the greatest challenges, showing a lower median and
accounting for all banks in the sample that post losses under the scenarios.
A second analysis assumes that banks’ operating efficiency adjusts to near-best
outcomes in their historical experience.60 While average advanced economy bank
ROE improves under this scenario, slightly more than half of the distribution remains
below 10%, suggesting that, without other meaningful improvements, banks would
need to improve their cost efficiency beyond levels seen in the past.
The results suggest that, across the global banking sector, the achievement of
returns that meet the estimated cost of equity will require a combination of
favourable cyclical developments and substantial improvements in cost efficiency,
particularly among European banks. Progress on efficiency may require that banks
DuPont analysis of bank profitability

Simple average, in per cent Table 3
2015–16 Change 2005–06 to 2015–16

Return on equity Return on equity Financial leverage Asset yield Profit margin
All banks 8 –7 –3 –1 –3
Of which, by region:
Euro area 3 –10 –3 0 –7
Other Europe 6 –11 –4 –1 –6
United States 9 –7 –4 –2 –2
Other advanced 12 –4 –2 –3 0
EMEs 12 –3 –3 0 0
Of which, by business model:
Retail-funded 6 –9 –2 –2 –6
Wholesale-funded 8 –6 –2 –2 –2
Universal 9 –6 –4 0 –2
Trading 9 –5 –5 1 –1
Source: SNL.
59
The analysis adjusts current ratios of net interest income and non-interest income per asset to the
90th percentiles in their historical distribution since 2005. Credit costs per revenue are adjusted to
the 10th percentile of their distribution. Net profit margins are assumed to rise to account for positive
operating leverage in a revenue growth scenario. Balance sheet size and mix are assumed to be static.
60
The analysis adjusts current ratios of operating costs per revenue to the 10th percentiles in their
historical distribution since 2005. Balance sheet size and mix are assumed to be static.

Bank return on equity analysis Graph 20
Actual return on equity distribution Scenario outcomes by region1 Scenario outcomes by business
model1, 2
Per cent of banks Per cent Per cent
45 30 30
30 20 20
15 10 10
0 0 0
–10 –10
_____________
_____________
_____________
______________
______________
______________
______________
Less -10 0 10 20 30 More
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
15–16
Cycl.
Exp.
Advanced economies: EMEs:
Retail Wholesale Trading and Universal
2005–06 2015–16 Euro Other United Other EMEs funded funded processing
2015–16 area Europe States Adv.
10th–90th percentile Median 10th–90th percentile
Median
25th–75th percentile 25th–75th percentile
1
Based on the sample of individual advanced economy and EME banks in Annex 2, subject to data availability. 2
For each region and
business model, actual return on equity in 2015-16 and under projected, cyclical and operating expense scenarios.
Source: SNL.
re-evaluate their cost structures more broadly. It is possible that this process could
be supported by rapid digital technological progress, which, as noted above, could
provide banks with a unique opportunity to compress their operating cost base.
Lastly, banks could seek to bring down the RoE required by investors, such as by
making their business model less opaque and easier to value. Over time, investors in
banks may also have to adjust their expectations, taking into account the effect of
structural changes that reduce the risk embedded in banks’ shares.
4.5 Banking system-level risk and resilience
This section considers banking stability from a system-wide perspective. The stability
of banking systems depends not only on individual banks’ resilience and risk-taking,
but also the properties of the system. This includes the structure of banks within the
system and the potential for distress to propagate across banks, either through direct
contractual linkages or indirectly through fire sale externalities, information contagion
and/or common exposures. The section of the report focuses on the impact of changes
in market structure, banking interconnections, risk correlations and international
banking.
Market structure
The analysis in Chapter 3 showed that there has been a reduction in the capacity of
many banking systems in Europe, as reflected in a fall in system assets and the number
of banks, as well as in a decline in bank branch and employee numbers. Although cost-
to-income ratios have generally not yet improved, lowering the cost base of the
banking sector in absolute terms can be considered a positive development for banking
stability insofar as it improves future profitability and competitive dynamics.

Another aspect of change in market structure is a tendency for higher
concentration in some banking systems, although there are some exceptions to this
trend as well as significant dispersion in concentration across countries. A more
concentrated system could have a positive effect on financial stability if bigger banks
benefit more from risk diversification and economies of scale, or derive higher
profitability and franchise value from greater pricing power. On the other hand, as
noted above, large banks can pose “too-big-to-fail” risks, and they are more complex
and difficult to monitor and, if need be, to resolve.
Across CGFS jurisdictions, there does not seem to be a clear relationship between
banking system concentration and profitability (RoA) during the crisis or since, nor is
there with operating efficiency since the crisis (Graph 21, left-hand and centre panels).61
Moreover, bank concentration ratios do not take market contestability (ease of entry)
into account, and thus do not provide a clear indication of competition in banking.62
For these reasons, the Group does not draw financial stability conclusions from post-
crisis shifts in banking system concentration. That said, future experience could differ
from the past – for instance, fixed cost pressures on banks associated with technological
deepening might raise the relative benefits of scale, all other things equal, while the
costs associated with big banks might be significantly lowered by regulatory reforms,
such as the above-noted measures to address systemically important banks.
Banking system concentration and performance Graph 21
Return on average assets Cost-to-income ratio Lerner index and Boone indicator1
y = 0.48 +0.0023x y = 53.33 +0.0738x

2
2
where R = 0.01 2.0 where R = 0.01 80 0.50 0.10
Return on average assets, %
y = 0.31 +0.0054x
Cost-to-income ratio, %
2
where R = 0.04 1.5 70 0.25 0.05
1.0 60 0.00 0.00
0.5 50 –0.25 –0.05
0.0 40 –0.50 –0.10
–0.5 30 –0.75 –0.15

30 40 50 60 70 80 90 40 50 60 70 80 90 01 03 05 07 09 11 13 15
Concentration ratio top 5 banks, % Concentration ratio top 5 banks, %
Lerner index Boone indicator
2007–09 2010–16 2010–16 (lhs): (rhs):
Median
25th–75th percentile
1
The Lerner index is a measure of market power in the banking market. It is defined as the difference between output prices and marginal
costs (relative to prices). The Boone indicator is a measure of degree of competition, calculated as the elasticity of profits to marginal costs.
Both indices cover CGFS jurisdictions included in Graph 4, except for Australia, Belgium, euro area aggregate and India.
Sources: Bureau van Dijk; World Bank; Bankscope; national data.
61
These results are indicative. Banking system concentration ratios are on a domestic basis for nearly
all countries, whereas return on asset and cost-income ratios are often on a global consolidated
group basis. This inconsistency could alter relationships for those banking sectors that have
significant foreign banking operations.
62
For example, Claessens and Laeven (2004) find no evidence that banking system concentration is
negatively related to competition.

Given the limitation of concentration ratios as indicators of banking system
competition and risk, other measures that derive from microeconomic theory are
often used. The Lerner index illustrates the ability of banks to charge a price above
marginal cost, with higher values approaching 1 indicating greater market power.63
The Boone Indicator measures the elasticity of profits to marginal costs, based on the
notion that more efficient firms gain a higher market share of profit. More negative
values indicate higher levels of competition and thus greater reallocation of profits to
more efficient firms.64 Both of these indicators suggest that competition has tended
to decrease across CGFS jurisdictions since 2000 (Graph 21, right-hand panel). While
the Lerner Index suggests that differences in the level of competition across countries
have widened, the Boone indicator points to lower dispersion over time.
Correlation in banking systems

Broadly speaking, the analysis in Chapter 3 points to bank business models becoming
somewhat more similar in the post-crisis period, with banks in advanced economies
having generally shifted towards commercial banking, particularly more traditional
retail banking business and deposit funding. While these broad changes may be
considered positive for the safety of individual banks, at a system level there is the
potential for a convergence in business models to result in risk correlations in the
banking system. In principle, the risk outcomes of banks with common exposures or
funding sources could become highly correlated: banks might behave in the same
way in the same circumstances (eg sell certain assets at the same time) or be affected
by the same shocks (eg investors in banks might withdraw funding for all banks in
response to problems within the sector).65
Similar business strategies and models might also engender excess competition
among banks, since banks are competing for the same customers. For example, two
parallel studies on relatively homogenous regional banking sectors in Germany and
Japan show that competition has been intensifying and is contributing to a greater
vulnerability in these banks (Box C).
The Group did not investigate the degree of a convergence within and across
countries from the perspective of common exposures and funding sources, which
require very granular data. It rather considered two approaches that provide some
general information on changes in risk correlations across G-SIBs. First, ∆CoVar, which
is the difference between the value-at-risk of the financial system in its median (or
typical) state and the value at risk of the financial system when a bank is in financial
distress.66 A higher value indicates the bank contributes more to the co-movement of
risk in stress times. Second, correlations of bank default probabilities are derived from
five-year CDS, adjusted for the default probabilities of each bank’s home sovereign.67
A higher value for this measure indicates that a bank’s risk co-moves more with that
of other banks. The ∆CoVaR measure shows that, after increasing sharply during the
63
Lerner (1934).
64
Boone et al (2005) and Boone (2008).
65
For example, see Wagner (2010).
66
Adrian and Brunnermeier (2016).
67
Based on the model of Abbassi et al (2017). The partial correlations across banks are the residuals of the
regression on banks’ default probabilities on default probabilities of their respective home countries.

crisis, covariance has declined over recent years to levels a bit above those seen prior
to the crisis (Graph 22, left-hand panel). In contrast, the CDS measures show a
reduction in bank correlation across countries over the post-crisis period (Graph 22,
centre left-hand panel).
Another indicator of systemic risk is the vulnerability of banks to asset fire sales.
For US bank holding companies, this vulnerability is measured as the system-wide
losses resulting from an assumed 1% decline in the prices of all bank holding
company assets.68 At the peak of the crisis in 2008, such a decline in asset values
would have caused an almost 30% fall in the aggregate equity of bank holding
companies, whereas the current figure is much smaller, at around 10% because of
lower leverage and “connectedness” among banks (Graph 22, centre right-hand
panel).69 US broker-dealers (not shown) have also seen a substantial decline in this
measure of fire sale vulnerability since the crisis.
Risk correlations Graph 22
∆CoVar Partial CDS correlation US banks’ fire sale losses German banks’ Systemic
across countries1 Liquidity Buffer2
Per cent of system equity Aggregate ratio
0.05 0.10 30 0.9
0.04 0.08 25 0.6
0.03 0.06 20 0.3
0.02 0.04 15 0.0
0.01 0.02 10 –0.3
0.00 0.00 5 –0.6

Jan 06 Apr 10 Sep 12 Sep 14
05 07 09 11 13 15 – – – – 97 02 07 12 17 04 07 10 13 16
Jul 08 Aug 12 Aug 14 Dec 16
Median
25th–75th Percentile
1
The partial correlations across banks are the residuals of the regression on banks’ default probabilities on default probabilities of their
respective home countries. 2 The Systemic Liquidity Buffer has a maximum value of 1 and a minimum value of –1.
Sources: Abbassi et al (2017); Adrian and Brunnermeier (2016); Duarte and Eisenbach (2013); Deutsche Bundesbank (2016).
An alternative indicator for the vulnerability of banks to asset fire sales is the
Systemic Liquidity Buffer (SLB).70 In this model fire sale losses result from a system-
wide funding shock, where banks cannot roll over short-term debt and liquidate
fungible assets to service their liabilities. A maturity structure with a higher share of
68
Duarte and Eisenbach (2013).
69
Connectedness measures the illiquidity concentration of the system and depends on whether illiquid
assets that are held by large and levered institutions are also widely held.
70
For further details see Deutsche Bundesbank (2016). The SLB measures the difference at the individual
bank level between assets that can be sold at short notice (eg tradable bonds) and are valued at fire
sale prices and the payment outflows from contractual obligations (such as deposits and interbank
liabilities) as expected in the event of a systemic liquidity shock. The normalised SLB is standardised
at the maximum amount of the SLB in the period under review, which means that its values are limited
to fluctuations between –1 and 1.

short-term liabilities implies higher system-wide fire sale losses. Consistent with this,
the aggregated SLB for the German banking system climbed sharply over the course
of 2008, primarily because of a significant decline in short-term liabilities (Graph 22,
right-hand panel). The positive values of the SLB after the crisis point to much more
comfortable liquidity buffers.
Box C
Regional bank business model homogeneity and competition in Germany and Japan
Regional banks in Germany (saving banks and cooperative banks) and Japan (regional banks and shinkin banks) have
highly homogenous business models. Compared with major banks, their profit sources are more concentrated on net
interest income, with a focus on traditional lending activities to households and SMEs (Graph C.1, left-hand panel). As
their financial intermediation services are relatively similar, customers can easily switch their business between regional
banks. Under these circumstances, regional banks tend to face severe competition, especially on loan interest rates.
Competitive pressures have been building for regional banks. The prolonged low interest rate environment and
intensified competition among bank branches providing similar financial intermediation services have depressed their
net interest margin. In addition, subdued loan demand due to population ageing (as well as population decline in
many areas of Japan) has also been putting downward pressure on banks’ profits.
To gauge the severity of competition facing regional banks, mark-ups (price less marginal cost) are estimated.
According to microeconomic theory, the market power of a firm can be defined in terms of the price elasticity of
demand for its outputs. Firms that have market power and maintain a competitive advantage can charge large mark-
ups, while firms that have less market power and are exposed to severe competition can only charge smaller mark-
ups. The results show that mark-ups have been declining over the long term, indicating that competition among
regional banks in both Germany and Japan has been rising (Graph C.1, centre and right-hand panels). Additionally,
the median (not shown) and mode have declined and the variance of mark-ups has shrunk, suggesting that financial
intermediation services provided by regional banks have become less profitable and even more homogeneous.
To investigate the implications of these changes for banking stability, the relationship between individual regional
banks’ mark-ups and Z-Scores is examined. The Z-score is a measure of business risk, defined as the ratio of a bank’s
loss-absorbing capacity to the volatility of its profits – the lower the score, the less stable the bank’s business is. The
results indicate that mark-ups have a statistically significant explanatory power with regard to Z-scores in both
countries. Moreover, estimated parameters indicate that the relationship between the mark-up and Z-score forms an
inverted U-shape – Z-scores are smaller at relatively low and high mark-up values. In this distribution, most German
and Japanese regional banks’ mark-ups lie in the region where their business risk increases as a result of intensified
competition. Overall, these results indicate that intensified competition among regional banks has reduced their
margins, partly because of the highly homogeneous nature of their business models but also due to low interest rates.
This situation may have the potential to negatively impact the resilience of regional banks in the long run.
The results are consistent with the theoretical literature which predicts that the effect of bank competition on
bank risk-taking is non-linear, although other studies show that the empirical relationship between competition and
(individual) bank risk highly depends on the chosen competition measure. For instance, studies applying the Lerner
Index tend to indicate that a reduction in the pricing power of individual banks due to fiercer competition leads to
increasing bank risk, whereas according to empirical analyses based on the Boone indicator a more competitive market
environment is associated with a lower level of bank risk.

Regional banks in Germany and Japan Graph C.1
Revenue components Distribution of mark-ups for Distribution of mark-ups for Japan

Germany
EUR bn JPY trn Density Ratio of banks, %
(lhs) (rhs)
30 6 100 50
25 5 80 40
20 4 60 30
15 3 40 20
10 2 20 10
5 1 0 0
0 0
Major Coop. Savings Major Regional Shinkin 0.02 0.04 0.06 0.08 0.10 -0.3 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0
banks banks banks banks banks banks Mark-up, in percentage points
2000 2010
Net interest income 2005 2015 1985 2005
Net fees and commissions 1995 2015
Other
Sources: Deutsche Bundesbank; Bank of Japan.
 The ratio of operating income to total assets is used to represent the price (P) of individual banks’ financial intermediation services.
Marginal costs (MC) are calculated based on panel estimates of individual banks’ cost function. For further details, see Bank of Japan (2017).
In a low interest environment, mark-ups over prices (such as Lerner Index) have a methodological bias because the indicator will tend to rise
as the denominator declines, even if there is no change in banks’ market power. In order to avoid this bias, mark-ups could be used as an
alternative measure of market power.  See Martinez-Miera and Repullo (2010) for theoretical literature, and for relevant empirical studies
Buch, Koch and Koetter (2013) and Kick and Pietro (2014).
Interconnectedness in banking systems

The crisis demonstrated that extensive direct financial connections between banks
can be a source of systemic risk. Several changes point to a reduction in interbank
connections since the crisis. First, as noted in Chapter 3, the use of interbank
borrowing has declined across a number of advanced economies in association with
the shift to deposit funding (although extraordinary monetary stimulus has played a
part). Second, G-SIBs have lowered their notional OTC derivative exposures relative
to the size of their balance sheets, with this change more pronounced for those with
relatively high pre-crisis exposure. Third, over recent years an increasing share of OTC
derivatives has been cleared with central counterparties, rather than bilaterally with
other banks. Most notably, around 60% of interest rate derivatives were estimated to
be centrally cleared in 2016, up from around 20% in 2008.71
The results from two simulation studies indicate a reduction in inter-
connectedness in the German and Mexican banking systems, although they do not
shed light on the underlying drivers of the reduction.
71
BIS (2017a).

The first study quantifies the reduction in aggregate banking system Tier 1 capital
of the German banking system in response to an adverse shock (Graph 23, left-hand
panel). The calculation is based on the entire network of German banks’ domestic
bilateral claims, which is constructed from the German credit register. The model
separately measures the impact of a direct effect (write-downs on direct claims) and
an indirect effect (expected loss provisions on interbank claims).72 The estimated
indirect effect stems from the impact on creditor banks’ Tier 1 capital ratio of
deterioration in the creditworthiness of its debtor banks.
The model indicates that losses from idiosyncratic bank defaults of German
banks increased during the pre-crisis period, and reached a peak during the crisis in
2008. Afterwards, the reduction of interbank exposures lowered potential
propagation effects. In addition, the steady improvement in capitalisation levels has
made the large banks more resilient to shocks at other banks in the network.
Bank interconnectedness simulations Graph 23
German simulation Mexico simulation Mexican banks’ total exposure to

foreign financial counterparties
End-2004 = 100 Per cent of assets MXN bn
175 21 350
150 18 300
125 15 250
100 12 200
75 9 150
50 6 100
25 3 50
0 0 0
06 08 10 12 14 16 09 10 11 12 13 14 15 16 17 09 10 11 12 13 14 15 16 17
Interbank exposures
Internationally triggered
Tier 1 capital
Domestically triggered
Banking system loss
Sources: Fink et al (2016); Deutsche Bundesbank; Bank of Mexico.
The second study measures the total assets of failed banks (and broker-dealers)
in Mexico resulting from a single failure of a domestic or foreign counterparty bank.
The results show a reduction over the past few years in the “worst case” outcome for
failed assets, as a percentage of total assets, from the failure of a foreign bank
(Graph 23, centre panel). This reflects a reduction in the average size of individual
links to foreign banks in the network, even though overall exposures have increased
(Graph 23, right-hand panel).
72
For more information on the model, see Fink et al (2016).

International banking risk
Broadly speaking, the financial stability implications of changes in international
banking integration will depend on trade-offs between risk-sharing and risk
spillovers. Banking integration entails diversification of risk for home country banks
and recipient countries, which could prove beneficial in the face of local shocks.73
Operating internationally may also increase bank risk-adjusted returns through
alternative (uncorrelated) income streams and funding sources. On the other hand,
international bank diversification can transmit stress between countries.74 The impact
of foreign banking relationships on home and recipient economies is likely to be more
significant where foreign bank exposures are large, or where interconnections
between foreign bank affiliates (ie funding or legal links) are strong. For host
economies, strong effects can arise if foreign banks provide critical services.
The post-crisis reduction in the international activity of some major advanced
economy banks could generally be regarded as a positive development for both
global and national financial stability. The weight of experience from the financial
crisis was that large internationally active banks tended to generate adverse
international risk spillovers rather than be a stabilising force. Most notably, cross-
border bank lending proved far less stable than activity conducted locally (either by
foreign or domestic banks). Foreign bank lending was also more prone to pullback
when funded in wholesale markets (short-term interbank and foreign currency
markets), as opposed to deposit or internal capital markets.75 In a similar vein, several
country case studies conducted under the International Banking Research Network
(IBRN) found that, when faced with a liquidity shock, lending to foreign borrowers or
by foreign bank branches was more resilient for those parent banks that made greater
use of core deposit funding and equity capital.76
A cutback in the international presence of some global banks may also help to
strengthen banks’ balance sheets by improving their profitability, because the
businesses exited tended to be small-scale or weakly performing. Indeed, some
foreign banks made disproportionate losses in their international ventures over the
past decade, including through traditional credit business or more complex
structured assets, but also conduct issues. Such problems may suggest that
international diversification involved information asymmetries or complexities that
were not well managed, or more simply that foreign offices were subject to weaker
risk management oversight than head office. As such, a reduction in the geographical
spread of global banks might make them easier to manage and monitor.
Any reduction in long-term risk from international banks needs to be weighed
against the potential for greater concentration in, or lower availability of, cross-border
and other foreign banking services, if only in the short term (see Section 5.1 below).
73
For example, by drawing on parent funding, foreign banks might be able to play a stabilising role
during times of stress in host economies (Crystal et al (2002); De Haas and Lelyveld (2010)).
74
For example, see Peek and Rosengren (1997).
75
See De Haas and van Horen (2013), De Haas and Lelyveld (2014), Reinhardt and Riddiough (2015),
McGuire and von Peter (2016).
76
Buch and Goldberg (2015). A similar result is also found in Avdjiev et al (2017).

These considerations will be affected by the changing nature of globalisation of non-
financial firms.77
Another caveat with regard to a reduction in advanced economy banks’ foreign
presence concerns the euro area banking market. Banking in the euro area became
more fragmented during the 2007–09 financial crisis and the subsequent euro area
sovereign debt crisis, and because banking flows before these events were skewed
towards interbank activity rather than more stable lending to firms and households.78
Reduced intra-region banking flows and low retail banking integration within the
euro area present policy challenges distinct from international exposures elsewhere
– private risk-sharing can help address asymmetric shocks within a monetary union
where a single monetary policy is unable to do so.79 Yet private risk-sharing could still
be facilitated by the cross-border equity funding of banks, or the greater use of cross-
border funding of non-financial sector investments through capital markets.
Regional expansion by EME banks entails the same broad risk implications as
earlier (non-regional) expansion by large advanced economy banks into EMEs, yet
there may be important nuances in the nature of risk. For instance, diversification
benefits derived for home and host parties may be smaller due to business cycles that
are already more synchronised at the regional level.80 Concentration risk could also
develop for smaller host economies since many of the expanding regional banks are
from the largest EMEs (and some advanced economy banks have retrenched at the
same time).81 On the other hand, the presence of foreign banks might be more stable
if it is grounded in deeper trade and investment links and host economy risks are
better understood and managed by regionally focused EME banks.
In addition to the geographical patterns of foreign bank presence, there have
also been important changes in the nature of banks’ foreign banking business. The
post-crisis shift away from cross-border to local intermediation for many banks
should enhance the resilience of foreign banking activity, given that local lending and
funding proved more stable during the crisis. For example, the local deposit-funded
lending of foreign bank subsidiaries in Latin America was relatively steady.82
Notwithstanding this trend and the general improvement in advanced economy
bank funding structures, the global net US dollar funding needs of foreign banks
appear to remain large. Prior to the crisis, some European banks made extensive use
of short-term foreign exchange and US dollar money markets to fund longer-term
US dollar assets, but found it unexpectedly difficult to roll over their funding when
financial sector strains emerged in 2007.83 The net USD positions of some European
banks – for example German and UK banks – have declined over recent years. Yet
positions for Japanese banks and, to a much lesser extent, Canadian banks, have
77
For a useful discussion of changes in multinational firms, see Baldwin (2016).
78
ECB (2017b).
79
Praet (2016) and ECB (2016c).
80
CGFS (2014a).
81
In this regard, Ehlers and Wooldridge (2015) find higher concentration in creditor banking systems
for borrower countries across many countries in the Asia-Pacific region than before the crisis.
82
Kamil and Rai (2010); Cull and Martinez Peria (2013).
83
McGuire and von Peter (2012).

expanded significantly as their foreign banking activity has increased.84 A possible
mitigating factor, based on discussion with banks, is that, unlike prior to the crisis,
foreign banks do not appear to be sourcing US dollar funding in FX markets for
longer-duration US dollar assets, and thus generating mismatch in the maturity of
their obligations. Additional quantitative evidence would be necessary to reach a
more definite conclusion.
4.6 Overall assessment
The analysis in this chapter considers the impact of post-crisis changes in the
structure and performance of banking for the stability of the sector. The key messages
are summarised below.
Resilience and risk-taking. Several indicators and stress test results suggest that
banks globally have enhanced their resilience to adverse shocks by significantly
building up capital and liquidity buffers. Advanced economy banks appear to have
reduced risk through the shift to more stable funding sources and assets that are less
complex or typically pose lower risk. Feedback from supervisors indicates that banks’
risk management has also considerably improved – for example, bank boards’ focus
on risk has increased and they are receiving better risk information, while risk
functions have become more influential. Even so, supervisors indicated that there
remains significant scope for improvement in banks’ risk management practices.
Moreover, surveillance of future risks to banking stability remains crucial.
Market sentiment and future profitability. While credit spreads for large
banks have declined somewhat over recent years, equity market sentiment towards
some banks remains soft. Low returns earned by some banks over recent years are
expected to persist and investors appear sceptical about the long-term viability of
certain business models. Simulations indicate that some banks would continue to
earn inadequate returns even under optimistic assumptions, suggesting more
extensive cost-cutting and structural adjustment are required, particularly among
European banks.
System-wide effects. The impact of structural change for system-level stability
is harder to assess than at the bank level because of greater uncertainties surrounding
interactions within the system. Nonetheless, with this caveat in mind, a number of
changes are consistent with the objectives of public authorities and the reform
process. First, foreign banks appear to be more focused in their geographical
presence and are conducting more of their intermediation locally. Second,
interconnections between banks through lending and derivative exposures have
declined. Third, policymakers in some European banking systems with relatively high
capacity have made progress in consolidating.
The potential effects of a decline in business model diversity (at a general level) is
of interest. The Group did not investigate if banks have been building up common
exposures or funding sources in ways that could prove problematic, as such analyses
require very granular data. Nonetheless, it is worth noting that the repositioning of
many large banks towards commercial banking is, in many respects, an intended effect
of the regulatory changes, with the increase in deposit funding a clear example.
84
BIS (2017b) and Borio et al (2017).

5. Implications for the efficiency of the banking sector
This chapter analyses the impact of structural changes in banking on the provision of
financial intermediation services at a high level. It first focuses on bank credit and
then discusses the provision of capital market related services – namely investment
banking and trading.
5.1 Bank credit provision to the real economy
Given the essential role of bank credit for the functioning of the real economy, this
section analyses how ongoing structural changes in banks have affected banks’ ability
to perform their core function in credit intermediation.85 It first covers general trends
in the provision of bank credit. It then discusses the role of supply factors (the
tightening or loosening of credit standards) in shaping these trends, and drivers of
these changes in the supply of credit.
Developments in overall bank credit since the crisis

In the years leading up to the crisis, bank credit expanded rapidly across a number of
advanced economies (see Graph 2). Lending standards deteriorated: some loan
structures involved high leverage on the assumption that asset valuations would
continue to rise, while a substantial share of lending was directed to borrowers with
relatively weak repayment capacity, such as subprime borrowers in the United States
or certain property developers in some European countries.
Trends in bank lending have diverged across the globe in the aftermath of the
crisis.86 Bank credit growth to the non-financial private sector suffered only a
temporary respite in advanced economies that were not directly affected by the crisis
– bank credit was about 20 percentage points of GDP higher in mid-2017 than at the
peak of the crisis (Graph 24, top left-hand panel, red line). By contrast, those
economies that bore the direct impact of the crisis went through a significant process
of deleveraging, as much of the growth in credit-to-GDP in the years ahead of the
crisis has been since unwound.
The importance of the crisis and the post-crisis adjustment in bank balance
sheets for overall credit (and bank performance) is evident from experiences at the
country level (Graph 24, top right-hand panel). Economies that had pronounced pre-
crisis credit booms and a large stock of non-performing loans in the wake of the crisis
(eg Ireland and Spain) have seen substantial falls in credit, while credit has also
contracted in some economies where large banks required substantial balance sheet
consolidation and repair (eg the United Kingdom). Data at the bank level show that
lending growth has been negative for numerous European banks, including European
G-SIBs, in the post-crisis period (Graph 24, bottom left-hand panel).
85
This chapter focuses on the quantity of provision of financial intermediation services. Besides
quantities, Philippon (2015) also considers the intermediation efficiency of the US banking system
and finds that the unit cost of financial intermediation has not fallen significantly between 1886 and
2012. Bazot (2017) analyses the costs of financial intermediation for European countries from 1950
to 2007, and concludes that it has increased since the late 1960s.
86
Classification of countries into crisis and non-crisis is based on the financial crises database of Laeven
and Valencia (2012).

In contrast to advanced economies, bank credit has expanded rapidly in China
over the past decade, especially in 2009 as part of the Chinese government’s policy
response to the crisis (Graph 24, top left-hand panel, purple line). It has also grown
at a firm pace across other EMEs (yellow line) throughout most of the post-crisis
period, registering a slowdown only recently.
Data on credit gaps – that is, the difference between the credit-to-GDP ratio and
its long-term trend – show that most advanced economies currently exhibit negative
credit gaps, reflecting deleveraging after the prior build-up in credit
Trends in bank credit Graph 24
Bank credit-to-GDP1, 2 Change in credit-to-GDP and return on equity, in

percentage points1, 3
September 2008 = 100
Annual change in credit-to-GDP (10–16)

140 CA 6
SG MX
HK CN
CH BE AU BR
120 JP US SE 0
KR FR
IT DE IN
GB NL
100 –6
ES
80 –12
IE
60 –18
01 03 05 07 09 11 13 15 17 5 10 15 20
Bank return on equity (2010–16)
Advanced economies: EMEs:
All countries:
Crisis Excluding China
Crisis Non-crisis
Non-crisis China
Lending growth by region and G-SIBs4 Credit-to-GDP gaps1, 5

Per cent Percentage points
24 40
20 30
16 20
12 10
8 0
Q1 2017
4 –10
0 –20
–4 –30
–8 –40
–12 –50
Euro Other United Other EMEs Non-G-SIB G-SIB –40 –30 –20 0 10–10 20 30 40 50
area Europe States advanced (excl. (excl.
EMEs) EMEs) Q1 2008
Advanced economies Europe United States
Average 25th–75th percentile 10th–90th percentile Advanced economies other EMEs
1
Credit to the private non-financial sector. 2 Advanced economies are grouped into crisis and non-crisis countries based on the Systemic
Banking Crises Database from Laeven and Valencia (2012). Non-crisis countries include Australia, Canada, Finland, Japan, Norway and New
Zealand. Crisis countries comprise Austria, Belgium, Germany, Denmark, France, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal,
Spain, Sweden, Switzerland, the United Kingdom and the United States. 3 Based on the sample of individual advanced economy and EME
banks in Annex 2; for country codes, refer to Graph 4; IE = Ireland. 4 Annual average growth in loans over 2010 to 2016. 5 Shaded areas
indicate the threshold of 10% of GDP for the absolute level of credit gap. For a derivation of critical thresholds for credit-to-GDP and property
price gaps, and their measurement, see Drehmann et al (2011).
Sources: SNL; BIS.

(Graph 24, bottom right-hand panel). In contrast, for a number of emerging market
economies, credit gaps are 10% or more of GDP, with China at more than 30%, at the
upper end of the pre-crisis experience.
Changes in the supply of bank credit since the crisis

A key question in assessing the influence of post-crisis structural changes on bank
credit intermediation is their impact on the supply of bank loans. Credit officer surveys
provide a means, albeit imperfect, to address the difficult issue of distinguishing
between shifts in credit demand and supply.
Surveys suggest that in major advanced economies bank credit markets
tightened substantially between 2009 and mid-2011 (Graph 25, left-hand and centre
panels). Tighter credit supply was evident even in countries where banks were
relatively less affected by the crisis (eg Japan). Bank lending standards eased after
mid-2011, with the exception of banks in the euro area where tightening of bank
lending continued until 2013–14 reflecting significant pressures to de-risk in the
midst of sovereign debt-related stresses.
Bank credit standards1

Diffusion index Graph 25
Non-financial corporations lending in Mortgage lending in advanced EMEs

advanced economies2 economies3
Tightening
Tightening Tightening
–60 60 –60 60 40
–40 40 –40 40 45
–20 20 –20 20 50
0 0 0 0 55
20 –20 20 –20 60
Easing
Easing Easing
40 –40 40 –40 65
02 05 08 11 14 17 02 05 08 11 14 17 10 11 12 13 14 15 16 17
Lhs (inverted): Rhs: Lhs (inverted): Rhs: Inverted scale:
Japan United States United Kingdom United States EMEs (aggregate)
United Kingdom Euro area Euro area Africa and Middle East
Latin America
Emerging Europe
Asia
1
Bank of England Credit Conditions Survey, US Senior Loan Officer Opinion Survey, euro area Bank Lending Survey and Bank of Japan Senior
Loan Officer Survey. The data have been smoothed using a weighted linear regression. 2 The Bank of England survey refers to corporate
sector lending. The ECB survey refers to terms and conditions for loans to all firms. The Federal Reserve survey refers to the net balance of
respondents which report tightening standards for C&I loans. 3 The Bank of England survey refers to changes in credit availability of
borrowers with a high LTV ratio (more than 75%). The ECB survey refers to terms and conditions for loans to households for house purchases.
The Federal Reserve survey refers to the net balance of respondents in the US Senior Loan Officer Opinion Survey reporting tightening
standards for residential mortgages.
Sources: Bank of England; Bank of Japan; European Central Bank; Federal Reserve; Institute of International Finance.
According to the credit surveys, credit supply conditions did not evolve
symmetrically across firms during the crisis, with credit to large firms more
significantly affected than that for SMEs. Nonetheless, tighter credit standards are
likely to have had a more substantial impact on SME borrowers than large firms, given

that SMEs have less access to alternative forms of finance. Indeed, there is evidence
that, as bank lending fell in the euro area, some large firms were able to increase their
bond issuance.87 More generally, non-financial corporations have sourced a greater
share of funding from bond markets across a range of countries since the crisis
(see Section 2.2 above).
Credit supply conditions to households also tightened during the crisis,
particularly for residential mortgages. In the United States, for example, standards for
many types of residential mortgage have since eased gradually, but they have
remained tight for non-traditional borrowers and borrowers with poor credit
histories.88 Available data on US bank lending by borrower risk segment suggest that
there are currently notably lower proportions of high-risk borrowers relative to 2006,
reflecting tighter credit standards, but also changes in the credit-eligible population
(Graph 26, right-hand panel). Mortgage lending to higher-risk segments – such as
loans with high loan-to-value ratios – has been tightened in parts of Europe, such as
in the Netherlands and the United Kingdom.
More generally, credit surveys indicate that easier credit standards in the major
advanced economies over recent years have been underpinned by improved macro-
financial and property market conditions and banks’ increasing risk tolerance as well
as stronger competition among banks.89
Lending growth, NPLs and US household lending Graph 26
Loan growth and NPL ratios in Change in non-performing loans and US consumer risk score distribution
advanced economies, in per cent average equity, in percentage points3 by loan type and year5
Per cent USD bn
BE NL
Annual change in peak NPLs4
30 –4 100 12
IT
1
Annual loan growth
15 –8 75 9
DE
0 FR –12 50 6
IE
US
–15 –16 25 3
GB SE
CH
–30 ES –20 0 0
_______
_______
_______
0 10 20 30 40 0 5 10 15 06 16 06 16 06 16 06 16
2
Post crisis NPL ratio Annual change in tangible equity Mortgages Cards Auto Student
(2010–16)
Euro area United States Lhs: Rhs:
Other Europe Other advanced Poor Debt level
Fair
Good
1
Annual average growth in loans from 2010 to 2016. 2 Annual average NPL ratio from 2010 to 2016. 3 Based on the sample of individual
advanced economy and EME banks in Annex 2; for country codes, refer to Graph 4; IE = Ireland. 4 Annualised reduction as a percentage of
maximum peak NPL ratio since 2007. 5 Annual average growth in loans from 2010 to 2016.
Sources: Federal Reserve Bank of New York; Equifax; SNL.
87
Kaya and Wang (2016).
88
FSOC (2016).
89
Stronger competition among banks and easing lending standards is also identified in Bassett et al
(2014), BOJ (2017) and SNB (2016).

EME bank credit standards have tightened over recent years, in association with
the maturing of the credit cycle and softer commodity prices (Graph 25, right-hand
panel). Lending conditions appear to be uniformly tight for all types of lending in
emerging economies, with commercial real estate the most affected loan segment.
As noted in Chapter 3, the supply of international credit has declined significantly
since the onset of the crisis as European banks have retrenched and many
internationally active advanced economy banks have shifted their international
business towards local origination and local funding in local currency. Some areas of
cross-border credit provision – such as trade finance – have recovered, as other
market participants have stepped in to fill the breach.90 However, there has been a
significant reduction in the number of correspondent banking relationships globally
and a decline in the value of transactions in some countries over the post-crisis
period.91 As such, concentration in the supply of these cross-border banking services
has increased in some markets. These changes appear to reflect a range of factors
including banks’ concerns around regulatory compliance of correspondent
relationship and inadequate profitability, as well as changes in risk appetite.92
The findings of the Central Eastern and South-Eastern Europe Bank Lending
Survey of the European Investment Bank provide some further insight into
international bank lending. Cross-border credit extended in this region by foreign
banks grew strongly prior to the crisis, and some retail lending was denominated in
foreign currency. The survey suggests that foreign banks have often applied tighter
credit standards than domestic banks in the aftermath of the crisis, consistent with
their process of deleveraging and de-risking, and their move to a more self-sustained
and domestically funded business model. In addition, banks have generally
maintained tighter credit standards on foreign currency-denominated loans, in
association with new regulatory requirements.
Drivers of changes in the supply of bank credit

The heterogeneity in credit supply conditions across banks and countries reflects the
incidence of different cyclical and structural factors.
Macroeconomic uncertainty can be detrimental to credit growth as banks are
less willing to finance projects when their returns are volatile and less predictable.
Alessandri and Bottero (2017) find that a rise in aggregate uncertainty reduced Italian
banks’ loan approval rate, but the negative impact on credit supply is smaller for
better capitalised banks with higher liquidity buffers. Similar results are found in the
literature for other advanced and emerging market economies.93
A large stock of non-performing loans may also impede credit supply.94 Banks
exhibiting high NPL ratios in the post-crisis period have tended to reduce their
90
CGFS (2014a).
91
Correspondent banking relationships allow banks to access financial services in different jurisdictions
and provide cross-border payment services to their customers (CPMI (2016)).
92
CPMI (2016), IMF (2017b) and FSB (2017c).
93
Chi and Li (2017), Bordo et al (2016) and Buch et al (2015).
94
Bending et al (2014), Cucinelli (2015). However, Accornero et al (2017) argue that the link between
NPLs and credit in Italy may be driven by demand-side effects.

lending, particularly in the case of some euro area banks (Graph 26, left-hand panel).
Bank-level data indicate that write-offs have mechanically weighed on loan growth,
but not to a large extent – on average, net write-offs have reduced average annual
loan growth by 50 basis points over 2010–16. The mechanisms by which NPLs
negatively impact bank credit supply are likely to be more indirect: NPLs tie up bank
capital and distract management from new opportunities, such as by incentivising
weakly capitalised banks to postpone the recognition of problem loans through
“evergreening”. A body of evidence from Japan in the 1990s, as well as from Europe
more recently, indicates that evergreening can lead to misallocation and distorted
competition in the rest of the economy.95 Moreover, data show that crisis-hit
countries that have more substantially reduced their banking sector NPLs have also
rebuilt capital to a greater degree (Graph 26, centre panel).
Banks’ adaptation to global regulatory reforms also seems to have affected
credit supply, although it is not easy to isolate its impact from that of other drivers.
First, the impact would depend on the extent that individual regulations were binding
constraints for individual banks and banking systems. A number of studies point to a
negative impact from tightened capital measures on bank lending, with the impact
being larger for banks with lower capital buffers.96 However, the evidence seems to
point only to a temporary decline in supply, as bank loan growth mostly returns to
normal after a few years. It also points to a positive long-run relationship between
banks’ capitalisation and credit.97 While there is less evidence available for reforms
related to banks’ liquidity standards (such as LCR and NSFR), it suggests that the
introduction of the LCR has had only a limited impact on lending to the non-financial
sector, as banks adjusted by reducing interbank lending and increasing HQLA
holdings.98 There is also evidence specifically for the United Kingdom that tighter
liquidity regulation had no detrimental effect on lending.99
The impact of regulatory reform can also be transmitted across jurisdictions
through the operation of foreign banks. The analysis undertaken by the IBRN on the
transmission of prudential regulations from home to host countries and vice versa,
finds broad evidence that international prudential regulation can spill over to bank
lending, but with the extent and direction of spillover varying across instruments and
across banks.100 In addition, there may be a reallocation of international market shares
away from less capitalised banks towards better capitalised banks (resulting in host
country credit being supplied by more resilient banks). The case of bank lending in
Mexico is largely consistent with the evidence on international spillovers (see Box D).
Banks across the globe have increasingly focused on adopting technological
innovations in recent years, in a desire to cut costs, reduce risk and respond to new
95
See Caballero et al (2008) and Peek and Rosengren (2005) for Japan; for Europe, see Albertazzi et al
(2010), Schivardi et al (2017) and Storz et al (2017).
96
See Berrospide and Edge (2010), Bridges et al (2014), Aiyar et al (2014), Uluc and Wieladek (2017)
and Fraisse et al (2017).
97
See Bridges et al (2014) for case of the United Kingdom, and Buch and Prieto (2014) for Germany.
98
EBA (2013), EBA (2014) and EBA (2015).
99
Banerjee and Mio (2014).
100
See Buch and Goldberg (2017).

Box D
The implications of regulatory change on bank lending: the case of Mexico
This box provides some evidence of how changes in international regulation, among other factors, may affect the
lending behaviour of foreign subsidiaries. The case of Mexico helps to illuminate the effects of the regulations in
different jurisdictions for several reasons: first, the relative importance of foreign subsidiaries (more than 70% of total
assets of the whole banking system) could make this system more responsive to changes from the policies of parent
banks; second, Mexico’s definition of capital was broadly similar before and after Basel III; and third, the availability of
micro data at the bank-firm level allows the impact to be more effectively tested.
Micro data are used to test whether international regulatory changes have had a different effect on bank lending
by subsidiaries of foreign banks as compared with that of domestic banks in Mexico. From a data set that includes all
the loans from all banks to non-financial firms, a difference-in-differences regression is run using a panel that includes
only a subset of firms that receive credit from both foreign subsidiaries and domestic banks. By controlling for demand
and borrower factors, this exercise allows the identification of whether international regulatory changes have had a
significant effect on lending by subsidiaries. The regulatory changes considered in the analysis were the
implementation of Basel 2.5, Basel III and the LCR. Figure 1 shows the evolution of credit to firms in the sample using
credit levels normalised to 100 at the date of regulatory implementation. The three charts hint at the change in
behaviour 12 months after the regulatory implementation dates for Basel 2.5 and Basel III, while the effect is less clear
for the LCR. The estimation results confirm that the implementation of the LCR did not result in significantly different
behaviour, but Basel 2.5 and Basel III did have such an effect, as compared with a year before the implementation.
The use of detailed credit-level information allows us to assess whether foreign subsidiaries have behaved
differently after regulatory changes. The preliminary results suggest that changes in international regulation affect
bank lending policies. In particular, Basel III has had a significantly different effect on banks with Mexican parents
versus foreign subsidiaries: post-Basel III, foreign subsidiaries reduced their credit growth rate by more than domestic
banks did, although it is important to emphasise that lending by foreign subsidiaries still continued to grow robustly.
Impact of regulatory reforms on credit to Mexican firms1

Nominal credit in Mexican pesos; implementation date = 100 Graph D.1
Basel 2.5 Basel III LCR
200 160 130
175 140 115
150 120 100
125 100 85
100 80 70
75 60 55
50 40 40
10 11 12 13 14 15 16 10 11 12 13 14 15 16 10 11 12 13 14 15 16
Local Subsidiary
1
Vertical lines represent implementation dates: Basel 2.5 = January 2011; Basel III = January 2013; LCR = January 2015.
Source: Bank of Mexico; CNBV.
non-bank entrants. This development might be expected to drive changes in bank

credit supply of a more structural character, although empirical evidence is limited,
possibly because there is no generally accepted metric for measuring technological

innovation. Nevertheless, some studies find a positive relationship between bank
credit supply and proxies for technology adoption.101
5.2 Banks’ role in facilitating capital market activity

This section provides an overview of post-crisis developments in banks’ investment
banking and trading businesses. There is a particular focus on non-equity trading
businesses, which have experienced considerable adjustment as a result of tighter
regulation and the unwinding of pre-crisis excesses. The section then discusses the
impact of changes in capital market businesses on the market shares by region, as
well as the implications for the liquidity of fixed income markets.
Developments in banks’ investment banking and trading businesses

Banks’ capital markets businesses can be split into two broad categories: investment
banking and trading.
Investment banking refers to activities where banks earn fee revenue from raising
capital for clients in debt and equity markets, and from advisory services on M&A.
This line of business requires little direct capital and tends to generate high, albeit
volatile, returns on equity.102 Investment banking has remained a strategic focus of
the large global banks in the post-crisis period and their business units have not
changed significantly from pre-crisis years aside from some cost-cutting. Global
activity, as measured through deal volumes and fees, has at times approached pre-
crisis highs, with volumes in Asia taking on a greater share of the total (Graph 27).
Market shares of global revenue have not changed substantially over the post-crisis
period, although the top five European banks (all G-SIBs) have ceded some ground
to non-US banks over the past couple of years (see Graph 30, right-hand panel).
The second, and larger, area of bank capital market activity and revenue includes
various trading businesses, grouped broadly under equity and fixed income,
currencies and commodities (FICC). In the boom years prior to the crisis, many banks
substantially expanded their trading business, with the largest US and European
banks in particular increasing the scope, complexity and geographic reach of their
activities, often through the acquisition of independent dealers. Some major US banks
also adjusted their strategies and business models towards trading business during
the crisis by acquiring failing institutions. Meanwhile, moving in the other direction,
some trading firms changed their legal structure to become bank holding companies.
In the wake of the crisis, operating adjustments to banks’ trading businesses have
been extensive and more pronounced than for investment banking and other bank
business lines as flat revenues, rapidly evolving technology and tighter regulation
have pressured banks’ trading units. Banks have also been reversing excesses built up
in the pre-crisis boom, when trading assets and risk were increasingly warehoused on
bank balance sheets to facilitate market-making or securitisation, or to support
proprietary trading.
101
Beck et al (2016) proxy technology adoption using OECD innovation survey data on banks’ R&D
expenditure, while Koetter and Noth (2013) use banks’ IT expenditure.
102
Based on public disclosures from banks, and RoE of publicly traded standalone non-bank investment banks.

Investment banking volumes and fees Graph 27
Debt underwriting volumes and fees Equity underwriting volumes and Advisory volumes and fees
fees
USD bn USD trn USD bn USD trn USD bn USD trn
50 10 25 1.0 25 5
40 8 20 0.8 20 4
30 6 15 0.6 15 3
20 4 10 0.4 10 2
10 2 5 0.2 5 1
0 0 0 0.0 0 0
06 08 10 12 14 16 06 08 10 12 14 16 06 08 10 12 14 16
Lhs: Rhs: Lhs: Rhs: Lhs: Rhs:

Global fees Americas Global fees Americas Global fees Americas
EMEA EMEA EMEA
Asia Pacific Asia Pacific Asia Pacific
Source: Dealogic.
Equity trading, which includes prime brokerage services to hedge funds, has been
a source of strength for banks with strong franchises, as recovering equity prices have
buoyed volumes, offset margin pressures and supported revenues. The crisis did not
lead to a significant reassessment of equity trading and related risks or the
restructuring of activities. Instead, banks have focused on deepening their use of
information technology as part of the longer-run transformation of these activities
through “electronification” and algorithmic trading. These changes have perhaps
shifted the nature of banks’ risks for these activities towards harder-to-measure
operational risks.
Banks’ FICC trading businesses have suffered a drop in revenues and RoE since
the onset of the crisis (Graph 28, left-hand panel). Excesses that arguably inflated FICC
business performance prior to the crisis have given way to a challenging post-crisis
market and revenue environment. For example, private US mortgage securitisation,
trading and related derivatives – key drivers of pre-crisis growth in FICC revenues and
product complexity – contracted sharply and left many firms with illiquid, loss-making
and complex legacy positions. In addition, some large banks have faced substantial
fines associated with poor conduct in their FICC trading units (eg LIBOR and foreign
exchange collusion and rigging). Post-crisis reforms – including changing risk weight
regimes, leverage ratios, prohibition on proprietary trading (in the United States),
derivatives and securitisation reforms, and market shocks in mandated stress tests –
have affected nearly every area of FICC. The fallout among trading market participants
together with regulatory changes has resulted in broad changes in the FICC business.
Leverage has been cut back, while the scale and scope of activities has been reduced,
particularly for the top European banks, as illustrated in a decline in trading assets
(Graph 29, left-hand panel). Banks have also lowered risk in their FICC business, with
VaR falling by more than trading assets, to levels below those seen pre-crisis
(Graph 29, centre panel).

Largest 10 banks’ global trading revenues1
In billions of US dollars Graph 28
Top banks’ FICC trading revenues Top banks’ equity trading revenues
100 100
50 50
0 0
–50 –50
–100 –100
05 06 07 08 09 10 11 12 13 14 15 16 05 06 07 08 09 10 11 12 13 14 15 16
Top 5 US banks Top 5 Europe banks
1
Top 10 banks include: JP Morgan, Citigroup, Bank of America, Goldman Sachs, Morgan Stanley, Deutsche Bank, Credit Suisse, UBS, Barclays,
BNP; adjusted for mergers.
Sources: Oppenheimer (2005–09); Bloomberg (2010–16).
Largest 10 banks’ trading assets, risk and revenue1 Graph 29
Median bank trading assets Median bank VaR Trading revenue per unit of VaR
2005 = 100 2005 = 100
125 250 300
100 200 240
75 150 180
50 100 120
25 50 60
0 0 0
06 08 10 12 14 16 06 08 10 12 14 16 10 11 12 13 14 15 16
United States Europe United States Europe Average top 5 United States
Average top 5 Europe
1
For sample of banks, see Graph 28; not adjusted for mergers.
Source: Companies’ annual reports.
Banks have responded by more closely managing balance sheets, collateral, clients
and risk taking, including the adoption of quasi-agency trading models where possible,
improving revenue relative to trading assets and VaR (Graph 29, right-hand panel).103
These changes have encompassed the proprietary trading and warehousing of complex
derivatives, bonds for market-making, and loans and for securitisation. Derivatives
trading has been particularly affected, given the significant increase in the central
103
Consistent with this, Iercosan et al (2017) analyse US daily supervisory data and find that, for the
average systemically important bank, trading revenue per dollar of VaR committed has trended up
over 2011–16.

clearing of OTC derivatives as a result of regulatory reforms, while banks have also
reduced their exposures through netting and trade compression techniques.
Implications for capital markets

While some individual banks seem to have cut back their trading risk, the broader
question of impact on efficiency (and risk) also requires an examination of the impact
of post-crisis changes on the shift of activity to non-banks (including central utilities),
the implications for changing concentration of some key trading activities among
providers, and the impact on market liquidity.
In trading, the businesses are more heterogeneous, complex, and far more
impacted by post-crisis shifts. On the question of trading and market liquidity, the
January 2016 CGFS report on Fixed income market liquidity highlighted reduced
market-making capacity at dealers in the midst of an expanding bond market,
supportive monetary policy and rapid technological adoption.104 The report findings,
which focus on sovereign bond trading, indicate a “bifurcation” of liquidity – with
liquidity deteriorating most in market segments that were historically less deep to
begin with, such as credit trading, and market adjustments more through trade size
than the cost of trading. The report and market participants also indicate
apprehensions around markets having potentially become more “fragile” during times
of stress, given reduced dealer intermediation capacity. In the medium term, the report
expects measures to bolster market intermediaries’ risk-absorption capacity will
strengthen systemic stability, including through a more sustainable supply of
immediacy services. The ESRB’s October 2016 report on Market liquidity and market-
making painted a similarly mixed picture, with the researchers’ concerns focused
primarily on the European corporate bond markets.105 This topic is still an area of active
debate among academics, regulators and market participants. Box E provides a
discussion of recent research on US fixed income market liquidity. Furthermore, more
research is needed on the link between fixed income liquidity and the real economy.
The above-mentioned pressures and strategic pivots have also left an imprint on
the competitive landscape in this sector. The retrenchment of European G-SIBs,
particularly in FICC, has roughly halved the number of full-service global firms, most
of which are now US-based banks (Graph 30, left-hand and centre panels). These US
banks have taken an increasing share of US and European activity.106 A far larger
number of more specialised banks and new non-banks are focused on profitable
subsets of products, geographies or clients.107 Within this group, some niche and
regionally focused banks have increased market share, particularly in Asia, which has
been the main growth region post-crisis. Important non-banks – exchanges, clearing-
houses and trading firms – have also emerged, with many focusing on the segments
where trading is most standardised and electronic, while also operating with less
stringent leverage constraints.
104
CGFS (2016).
105
ESRB (2016).
106
Market share data for FICC trading should be treated with caution. Such information is less robust
and more subject to estimates than data on other capital market business lines for a variety of
reasons, including the OTC nature of some markets, revenue generation through earning a bid-ask
spread instead of commissions, and the participation of diverse types of listed and private market-
making and principal trading firms.
107
McKinsey & Company (2016).

Market shares in capital markets revenues1
FICC trading Equity trading Advisory and underwriting fees
50 50 50
40 40 40
30 30 30
20 20 20
10 10 10
0 0 0
10 11 12 13 14 15 10 11 12 13 14 15 10 11 12 13 14 15 16
Top 5 United States Top 5 Europe Other
1
For sample of banks, see Graph 28. Total market size estimates from McKinsey & Company.
Sources: Oppenheimer (2005–09 pro forma for mergers), Bloomberg (2010–16); McKinsey & Company.
The timing and extent of adjustment at banks has been uneven across the
industry. UBS was one of the first movers, announcing an exit from some fixed income
markets in 2009 and 2012 (see Box B). Other European G-SIBs have since announced
cuts to various degrees. As some G-SIBs have exited certain FICC segments and
products, revenues have tended to shrink faster than expenses in the short run,
resulting in little RoE improvement and prompting heightened market scrutiny of
business model sustainability.
Finally, “Brexit” has introduced new operational and market structure uncertainty,
given that London has been a global centre for managing and booking capital
markets activities. Since the Brexit vote, major trading banks have all announced
active contingency plans for shifting some staff and operations from London to new
mainland European trading hubs.
5.3 Overall assessment

Trends in bank lending have been uneven across the globe, with large drops in the
economies that bore the brunt of the crisis but continuing growth in other
jurisdictions. Credit activity at a broad level started to recover in most of the crisis-hit
advanced economies after 2015, while in selected countries the adjustment is still
ongoing, often reflecting high NPLs. By contrast, advanced economy banking systems
that were not so adversely affected by the crisis have continued to show solid loan
growth, notwithstanding tighter regulations. Bank lending has also expanded
strongly in emerging economies, raising sustainability concerns in some cases.
Across the major advanced economies, credit surveys show that bank credit
standards tightened significantly for all types of lending during the financial crisis,
and again in the euro area in 2011–13 as concerns about sovereign debt escalated.
While credit standards have subsequently eased, there are indications that they
remain much tighter for higher-risk households in the United States, loan segments
that were at the epicentre of the crisis. In contrast to the situation in advanced
economies, credit standards have been tightening in EMEs over recent years,
consistent with their differing credit cycles.

Box E
US fixed income market liquidity
Market participants’ concerns about fixed-income market liquidity in the post-crisis era, and the possible effects of
regulatory changes, have spurred significant discussion and analysis. The effects of regulation are difficult to pinpoint,
given other factors affecting liquidity in the post-crisis , including voluntary changes in dealer risk-management
era
practices, the growth of electronic trading, the evolving liquidity demands of large asset managers, and changes in
the economic environment. Moreover, assessing the status of market liquidity can be difficult due to market
fragmentation and data limitations, with limited pre-trade transparency in the corporate market and limited data on
dealer-to-customer transactions in the Treasury market.
Adrian, Fleming, Shachar and Vogt (2017) assess liquidity in the US Treasury and corporate bond markets, finding
no strong evidence of a widespread deterioration in market liquidity in the years after the crisis. As of mid-2016,
average bid-ask spreads for benchmark notes in the inter-dealer Treasury market were narrow and stable, and Treasury
market depth and price impact, although suggesting reduced liquidity, were within historical variation and far from
crisis levels. For corporate bonds, average bid-ask spreads and price impact declined after the crisis, albeit to levels
higher than those before the crisis for institutional trades (ie trades of $100,000 and above). Moreover, corporate bond
trading volume and issuance were at record highs.
Consistent with these findings, Mizrach (2015) analyses TRACE corporate bond transactions data from 2003 to
2015, and concludes that “most measures suggest a healthy market” with rising transaction volumes, narrowing bid-
ask spreads, and falling price impact of trades. Similarly, looking at price impact, round-trip costs and other measures,
Trebbi and Xiao (2015) report “a lack of any form of systematic evidence of deterioration in liquidity levels or breaks
in liquidity risk for corporate bonds”. Bessembinder et al (2016) further find lower transaction costs during the 2012–
14 Dodd-Frank phase-in period than in the 2003–07 pre-crisis period. Anderson and Stulz (2017) also report lower
average transaction costs and price impact post-crisis versus pre-crisis for all corporate bond transactions, albeit
somewhat worse liquidity for large (over $100,000) trades.
In contrast to these studies on broad liquidity trends, a number of studies document worsening liquidity along
some dimension when conditioning on stress events or on the nature of institutions providing liquidity. Bao et al
(2016) find that price impact increased among recently downgraded corporate bonds when comparing the pre- and
post-Volcker rule periods. Anderson and Stulz (2017) find that liquidity has declined after the crisis during episodes
of extreme VIX increases, but do not find evidence that liquidity has worsened for bond-specific (idiosyncratic) stress
events, such as extreme bond yield increases and downgrades from investment grade to high-yield. Adrian, Fleming,
Shachar and Vogt (2017) consider three case studies in which the resilience of market liquidity was challenged after
the crisis – the 2013 “taper tantrum”, the October 2014 “flash rally” in the Treasury market, and the liquidation of Third
Avenue's high-yield bond fund in December 2015. In all three instances, the degree of deterioration in market liquidity
was within historical norms, suggesting that liquidity remained resilient even during stress events. Focusing on a
different type of stress event, Dick Nielsen and Rossi (2017) use bond index exclusions as a natural experiment during
which index-tracking investors demand immediacy from dealers and find that the price of immediacy significantly
increased post-crisis.
Dealer-centric liquidity provision is also explored in a few other recent papers. Choi and Huh (2017) show that
dealers act as agents rather than as principals for a higher fraction of trades in the July 2012–June 2015 period as
compared with the January 2006–June 2007 period, and that transaction costs have increased for trades that cannot
be immediately matched. Furthermore, while Bessembinder et al (2016) estimate lower transaction costs after the
crisis, they document a structural break that suggests a decline in dealers’ capital commitment relative to the pre-
crisis period. Adrian, Boyarchenko, and Shachar (2017) find that there is a relationship between financial institutions'
balance sheet constraints and bond liquidity in the post-crisis period so that bonds traded by more levered and
systemic institutions, and bonds traded by institutions more akin to investment banks, are less liquid, consistent with
more stringent leverage regulation and greater regulation of dealer banks reducing institutions’ ability to provide
liquidity to the market overall.

Changes in credit supply conditions may be driven by various cyclical and
structural factors, including macro-financial conditions, bank balance sheet
constraints and regulatory changes. The literature indicates that a large stock of NPLs
may impede credit supply, while data show that banks exhibiting high NPL ratios in
the post-crisis period have tended to reduce their lending. The literature also
suggests a negative impact on credit supply from tighter regulatory capital
requirements in the short run, but a positive relationship between capitalisation and
credit provision in the long run.
Crisis-era trading losses and related bank failures subsequently motivated some
of the most pronounced operating model and strategy shifts. Returns on equity and
growth in trading of fixed income, currencies and commodities have been subpar,
pressuring overall profitability at some G-SIBs. The scale, scope and risk of activities
in this business segment have been reduced, particularly for large European banks.
This raises questions about the ability and willingness of banks to act as market-
makers, and the resultant impact on market liquidity and fragility during times of
stress, especially for fixed income market sub-segments that have been historically
less liquid. The evidence for this is mixed. Nonetheless, measures to bolster market
intermediaries’ risk-absorption capacity will strengthen systemic stability, including
through a more sustainable supply of immediacy services. Market participants and
regulators also remain watchful of the future competitive implications of the ongoing
retrenchment of some European G-SIBs from trading.
6. Key messages
1. Post-crisis a stronger banking sector has resumed the supply of

intermediation services to the real economy, albeit with some changes in
the balance of activities.
The crisis revealed substantial weaknesses in the banking system and the prudential
framework, which had led to excessive lending and risk-taking unsupported by
adequate capital and liquidity buffers. The immediate impact of the crisis was severe
for many advanced country banks and economies. Banking sectors have generally
recovered and have been adjusting to the post-crisis regulatory and macroeconomic
environment. The report derives some observations that emerge from its analysis of
the ongoing adjustment.
There is no clear evidence of systematic and long-lasting retrenchment of
banks from credit intermediation. The severity of the crisis was not uniform across
banks and systems. Weaker banks cut back credit more strongly, and riskier borrowers
saw their access to credit more tightly curtailed. In the immediate aftermath of the
crisis the response of policymakers and bank managers was also differentiated across
systems, with some moving more decisively than others to address the problems
revealed. Bank credit has since grown relative to GDP in most jurisdictions, but has
not returned to pre-crisis highs in the most affected countries, reflecting necessary
deleveraging and the unwinding of pre-crisis excesses. While disentangling demand
and supply drivers remains a challenging exercise, the evidence gathered by the
Working Group does not point to systematic change in the willingness of banks to
lend locally. In line with the objectives of post-crisis reforms, lenders have become
more sensitive to risk and more discriminating across borrowers. But healthy banks

have not changed their overall stance, notwithstanding tighter regulations. If
anything, the shift towards commercial banking activities suggests that banks are
putting more emphasis on lending than trading activities. Still, given the range of
changes in the banking sector over the past decade, policymakers should remain
attentive to potential unintended “gaps” in credit to the real economy.
Legacy asset quality problems can be an obstacle to credit growth. Excessive
pre-crisis credit growth left a legacy of problem assets, especially high levels of NPLs,
which continue to distort the allocation of fresh credit in several countries. Banks and
banking systems whose problems were addressed sooner have seen faster recovery
post-crisis. Authorities are therefore encouraged to take measures to address NPLs
expeditiously, including the stock problem. Persistently high NPLs are likely to lead
to greater ultimate losses, impede credit growth and distort credit reallocation,
potentially incentivising banks to take on more risk.
Some banks have retreated from capital market-related business. Many
banks have reduced their exposure to capital markets activity such as trading and
market-making. The shift was most clear among European G-SIBs. It is not clear how
this will play out in terms of market structure and profitability as the industry remains
in transition, while early evidence points to a shift towards new players. The key
consideration from a market functioning perspective is the impact of bank
retrenchment on liquidity. The liquidity of some market segments shows signs of
fragility and a trend towards bifurcation: traditionally liquid segments remain so, while
others have become less liquid.108 That said, causality remains an open question, and
it is therefore important to continue to monitor the impact of banks’ and market
participants’ adaptation to changes in liquidity, especially in fixed income markets.
The international banking landscape was among the areas most affected by
the crisis. Foreign bank claims registered a strong decline post-crisis and a number
of advanced economy banks, especially from Europe, have reduced the number of
foreign markets in which they provide services. By contrast, banks from EMEs and
countries less affected by the crisis have expanded their foreign activities, in some
cases, quite substantially, in a development that changes the composition of global
banking assets. While some cross-border banking services have been maintained, as
pullback from some banks has been offset by other market participants,
correspondent banking relationships have declined in some jurisdictions because of
concerns over litigation risk and inadequate profitability.
2. Longer-term profitability challenges require the attention of banks and

supervisors, as they may signal overcapacity and risk-taking incentives.
Post-crisis bank profitability has remained subdued. This reflects many factors,
including bank-specific drivers (eg business model choices), cyclical macroeconomic
drivers (eg low growth and interest rates) and structural drivers that will have a more
persistent impact. An example of this latter group includes regulatory reforms that
have implied lower leverage and the curbing of certain higher risk activities, and a
reduction of implicit subsidies for large or systemically important banks.
108
CGFS (2016).

Banks and their shareholders must adapt to a persistent reduction in
profitability, as pre-crisis levels of profitability proved unsustainably high. It is
difficult to predict future profitability, but the impact of some drivers that led to a
decline of RoE is likely to persist. For example, all else constant, lower leverage and
reduced risk-taking should reduce return on equity. Sluggish revenues have
dampened profits and, combined with low interest rates, may have contributed to the
slower progress made by some banks in dealing with legacy problem assets. At the
same time, estimates of bank equity investors’ return expectations have remained
elevated to date. Taking current market expectations as a benchmark, the Group’s
analysis suggests that, even in the event of an assumed cyclical upturn, numerous
banks’ RoE will not rise above the cost of equity currently required by their investors.
Banks’ adaptation to an environment of softer profitability could entail further
cost-cutting (eg more efficient delivery of services), diversification into new revenue
sources or, more problematically, the adoption of a riskier business profile. But where
weak profitability is a sign of overcapacity, exit from specific markets or services is an
integral component of structural adjustment in the sector.
Over time, investors in banks may have to adjust their expectations, taking into
account the effect of structural changes that reduce the risk embedded in banks’
shares. These changes relate to the greater resilience of individual institutions and of
the system as a whole, as well as to institutions’ shift towards business models with
less volatile earnings.
Supervisory vigilance is required in monitoring the adaptation of banks to
lower profitability. While excessive pre-crisis levels of profitability are not the right
point of reference, the adaptation process to more moderate levels is not without
risks. Pressure on bank management to achieve a higher RoE may incentivise banks
to take excessive risks. Alternatively, banks’ exit from low-return activities may lead to
capacity constraints for certain critical functions in the financial system. Finally, market
pressures may arise over time to increase leverage and bank risk-taking, including
through opaque channels.
Sufficient levels of capital are needed for banks to deal with unexpected shocks,
and low profitability can weaken banks’ ability to maintain sufficient buffers. Banks
that lack a steady stream of earnings to repair their capital base after an unexpected
loss will have to rely on fresh equity issuance. Yet, markets are usually an expensive
source of capital for banks, when accessed under duress. Bank stress tests are one
way to assess the impact of a prolonged period of low profitability and reduced
access to new capital. Authorities should also resist pressure resulting from the
weaker performance of banks to roll back recent regulations and to compromise on
resilience. These issues will require the continuing vigilance of supervisors and
authorities in charge of financial stability.
Authorities can facilitate adjustment of the banking sector. Whereas
adjustment must come first and foremost from banks themselves, policymakers
should ensure that it proceeds smoothly and does not threaten the stability of the
financial system. In doing so, policymakers need to assess both the root causes of
individual banks’ challenges and the impact of any structural impediments. If
overcapacity is a key driver of low profitability, institutional barriers to mergers must
be reviewed and exit regimes applied. If the problem lies with legacy assets (such as
NPLs), these should be fully addressed, which might entail a dialogue between
prudential authorities and other policymakers (eg those in charge of mechanisms
dealing with insolvency). More generally, the application of post-crisis recovery and

resolution regimes is necessary for banks that are unable to meet existing and
emerging challenges. The post-crisis period of structural change in banking has
shown that the timely and credible use of adjustment mechanisms can enhance
investor confidence in banking systems.
That said, intertemporal trade-offs do arise. The exit of financial institutions
might be politically costly in the short run, but may pay off in the longer term through
more stable banking systems, sounder lending and better allocation of resources. The
implicit subsidisation of non-viable business models might have lower short-term
costs but could lead to resource misallocation. Similarly, any assessment of
consolidation trends needs to take into account potential trade-offs between
efficiency and stability, as well as examine the nature and impact of barriers to exit
for less profitable banks.
3. Consolidation and preservation of gains in bank resilience requires ongoing

surveillance, risk management, and a systemic perspective.
The response by banks and supervisors in the aftermath of the crisis was immediate
and has brought about changes in the desired direction. However, more needs to be
done to ensure that banks and the financial system remain resilient to future risks.
Banks have made progress in dealing with pre-crisis weaknesses. A number
of trends observed in the global banking system since the financial crisis are quite
generalised and are in line with the direction of change intended by the regulatory
reform agenda. Compared with the pre-crisis period, banks are better capitalised and
have lower exposure to liquidity and funding risks. They have also reduced activities
that contributed to the build-up of vulnerabilities, such as exposure to high-risk
assets, and excessive counterparty risk through OTC derivatives and repo
transactions, among others. That said, given that markets have not yet evolved
through a full financial cycle, bank restructuring efforts remain under way. In addition,
as many relevant reforms have not yet been fully implemented, it is too early to assess
their full effect.109
Supervisors and markets should avoid becoming complacent about
progress to date. Despite progress made, risk is not static. The dynamic evolution of
the financial system, shifts in the macroeconomic environment and technological
change, create new risks to financial stability. Similarly, some trends in banking
systems that we have observed since the crisis, such as the decline in wholesale
funding, might be affected by unconventional monetary policy and may not persist.
Success in addressing prior problems does not guarantee that banks will be able
to respond to future risks. Ongoing surveillance will thus be necessary to assess the
process of adaptation and the evolution of risk-taking within the banking system and
the financial system more broadly, taking sufficient precautions with regard to
“unknown unknowns”. This entails consolidating the gains from the implementation
of regulatory reform in terms of stronger capital and liquidity buffers. Supervisors
should be wary of signs that banks’ adaptations appear to be driven by a desire to
arbitrage regulation rather than to address fundamental weaknesses.
109
It is important to note that this project has not been aimed at reviewing specific regulations and their
effects. The FSB does plan to conduct such a review through its “Proposed Framework for Post-
Implementation Evaluation of the Effects of the G20 Financial Regulatory Reforms”.

Improved governance and risk management approaches in banks will be
key to future bank resilience. Problems of bank governance and risk management
contributed to the crisis and have been a key focus of reform. Given that the sources
of future vulnerabilities are hard to predict, banks need to have robust frameworks of
risk governance and management to identify and understand emerging risks and
their potential impacts for the firm. This remains one of the most important factors
for bank resilience, particularly given ongoing changes in business lines, market
practice, and financial technology that may test banks’ governance and risk
management.
The Group did not have the resources to assess how post-crisis adjustments in
this area affected incentives and bank resilience across jurisdictions. Given the
importance of the question, it may be advisable for the international supervisory
community to follow up on the earlier report of the Senior Supervisors Group on risk
management weaknesses that contributed to the financial crisis. This follow-up work
could review the extent of subsequent risk management improvements vis-à-vis
supervisory expectations, and investigate particular questions such as how changes
in the risk environment and supervisory and regulatory frameworks have affected the
nature of and incentives for sound risk management.
The evolving nature of systemic risk requires surveillance, particularly on
certain key areas. Structural changes in the financial system, broader
macroeconomic trends, and adjustments to new regulations require continuous
surveillance of systemic risk, including the impact from shifting of activities to the
non-bank sector, and the rise of fintech.
Systemic risk assessment entails monitoring of different areas. The adaptation of
large banks’ business models, activities and systemic footprint, requires continuous
monitoring. Another area relates to tracking interconnectedness within the system.
While recent observations suggest a reduction in the interconnections between banks
(for instance, through lower interbank exposures and the use of central clearing for
many derivatives), this remains a key channel of contagion. Finally, it is critical to track
the degree to which banks become exposed to common shocks. Policymakers should
be continuously assessing which common adaptations by banks may have negative
implications for financial stability. For instance, higher reliance on more stable sources
of funding, such as deposits, may be less problematic than the potential for correlated
valuation shocks and fire sales from similarities in asset portfolios. The evolution and
impact of fintech on bank business models and services will need to remain a key
area of attention.
The shifting of some risk out of the banking sector after the crisis highlights the
need for continuing central bank investment in systemic risk analysis of the non-bank
financial sector. This includes, for example, the monitoring of CCPs’ resilience, liquidity
risks associated with the growing size of portfolios of asset managers, and the
activities of shadow banks in providing intermediation services.
4. Better use and sharing of data is critical to enhanced surveillance of

systemic risk.
The Working Group sees scope for making better use of existing and (where
necessary) new data sets to improve the ability of central banks and supervisors to
monitor systemic risk. Some concrete examples of where progress can be made relate
to the availability of information in the following areas:

Better use of the information in transaction depositories. Banks in Europe
and the United States are required to report to transaction depositories detailed data
on all derivatives transactions (EMIR and Dodd-Frank regulations). The analysis of
such granular data could prove invaluable for understanding systemic risk, and even
for better identification of microprudential risks. For example, whereas authorities
have general information relating to overall quantities of certain types of exposure,
they have less insight into counterparties, and more specifically, of counterparty
concentrations. Major technical and confidentiality obstacles to gaining access to, and
managing, large data sets are a serious practical impediment in their use.
There is scope for enhanced international cooperation on data expertise
and analysis. There appears to be considerable scope for international cooperation
by authorities in sharing data expertise and analysis (duly respecting confidentiality
and data security requirements). For instance, international cooperation between
authorities might contribute to more detailed information on interconnectedness and
common exposures among financial institutions. Several regional and international
initiatives that aim at improving the availability of data and ensuring a better use of
existing data sets are under way.110 Closely linking these initiatives to ongoing policy
work should improve surveillance of systemic risk.
110
These include EMIR derivatives transactions reporting, European Securities Holding Statistics
data, the G20 Data Gaps Initiative or the European Reporting Framework (ERF), which
aims at integrating banks‘ reporting systems. Or INEXDA, an international cooperative
project exchanging experiences on the statistical handling of granular data,
www.bundesfinanzministerium.de/Content/DE/Standardartikel/Themen/Schlaglichter/G20-
2016/iag-update-on-the-data-gaps-initiative-and-the-outcome-of-the-workshop-on-data-
sharing.pdf?__blob=publicationFile&v=3.

References
Abbassi, P, P Brownlees, C Hans and N Podlich (2017): “Credit risk interconnectedness:

What does the market really know?”, Journal of Financial Stability, vol 29, pp 1–12.
Abiad, A, R Balakrishnan, P Koeva Brooks, D Leigh and I Tytell (2009): “What’s the
damage? Medium-term output dynamics after banking crises”, IMF Working Paper,
WP/09/245.
Accornero, M, P Alessandri, L Carpinelli and A M Sorrentino (2017): “Non-performing
loans and the supply of bank credit: evidence from Italy”, Questioni di Economia e
Finanza (Occasional Papers), no 374.
Adrian, T and M Brunnermeier (2016): “CoVaR“, American Economic Review, vol 106,
no 7, pp 1705–41.
Adrian, T, M Fleming, O Shachar and E Vogt (2017): “Market liquidity after the financial
crisis“, Annual Review of Financial Economics, vol 9.
Adrian, T, N Boyarchenko and O Shachar (2017): “Dealer balance sheets and bond
liquidity provision“, Journal of Monetary Economics, vol 89, pp 92–109.
Aiyar, S, C Calomiris and T Wieladek (2014): “How does credit supply respond to
monetary policy and bank minimum capital requirements?“, Bank of England Working
Papers, no 508.
Albertazzi, U and D Marchetti (2010): “Credit supply, flight to quality and
evergreening: an analysis of bank-firm relationships after Lehman“, Bank of Italy, Temi
di Discussione, no 756.
Alessandri, P and M Bottero (2017): “Bank lending in uncertain times”, Bank of Italy,
Temi di Discussione, no 1109.
Anderson, M and R Stulz (2017): “Is post-crisis bond liquidity lower?”, NBER Working
Papers, no w23317.
Australian Prudential Regulation Authority (2016): “Risk culture”, information paper,
October.
Avdjiev, S, L Gambacorta, L Goldberg and S Schiaffi (2017): “The shifting drivers of
global liquidity”, BIS Working Papers, no 644.
Baker, M and J Wurgler (2015): “Do strict capital requirements raise the cost of capital?
Bank regulation, capital structure, and the low-risk anomaly”, American Economic
Review, vol 105, no 5, pp 315–20.
Baldwin, R (2016): The great convergence: information technology and the new
globalisation, Harvard University Press.
Banerjee, R and H Mio (2014): “The impact of liquidity regulation on banks”,
BIS Working Papers, no 470, October.
Bank of Japan (2017): Financial System Report, April.
Bao, J, M O'Hara and X Zhou (2016): “The Volcker rule and market-making in times of
stress”, Board of Governors of the Federal Reserve System, Finance and Economics
Discussion Series, no 2016-102.

Bassett, W, M Chosak, J Driscoll and E Zakrajsek (2014): “Changes in bank lending
standards and the macroeconomy”, Journal of Monetary Economics, vol 62, pp 23–40.
Basel Committee on Banking Supervision (2017): Basel III Monitoring Report,
September.
Bazot, G (2017): ”Financial consumption and the cost of finance: measuring financial
efficiency in Europe (1950–2007)”, Journal of the European Economic Association,
March.
Beck, T, T Chen, C Lin and F Song (2016): “Financial innovation: the bright and the
dark sides“, Journal of Banking and Finance, vol 72, pp 28–51.
Bending, T, M Berndt, F Betz, P Brutscher, O Nelvin, D Revoltella, T Slacik, M Wolski
(2014): “Unlocking lending in Europe”, European Investment Bank.
Berrospide, J and R Edge (2010): “The effects of bank capital on lending: what do we
know and what does it mean?”, Board of Governors of the Federal Reserve System,
Finance and Economics Discussion Series, no 2010–44.
Bessembinder, H, S Jacobsen, W Maxwell and K Venkataraman (2016): “Capital
commitment and illiquidity in corporate bonds”, SSRN.
Bank for International Settlements (2017a): “Central clearing makes further inroads”,
BIS Quarterly Review, June, pp 8–9.
——— (2017b): 87th Annual Report.
Boone, J, R Griffith and R Harrison (2005): “Measuring competition”, AIM Research
Working Paper Series.
Boone, J (2008): “A new way to measure competition”, Economic Journal, vol 118,
pp 1245–61.
Bordo, M, J Duca and C Koch (2016): “Economic policy uncertainty and the credit
channel: aggregate and bank-level U.S. evidence over several decades”, Journal of
Financial Stability, vol 26, pp 90–106.
Borio, C, R McCauley and P McGuire (2017): “FX swaps and forwards: missing global
debt”, BIS Quarterly Review, September.
Brei, M and L Gambacorta (2014): “The leverage ratio over the cycle”, BIS Working
Papers, no 471.
Bridges, J, D Gregory, M Nielsen, S Pezzini, A Radia and M Spaltro (2014): “The impact
of capital requirements on bank lending”, Bank of England Working Paper, no 486.
Buch, C, C Koch and M Koetter (2013): “Do banks benefit from internationalisation?
Revisiting the market power-risk nexus?“, Review of Finance, vol 17, no 4, pp 1401–35.
Buch, C and E Prieto (2014): “Do better capitalized banks lend less? Long-run panel
evidence from Germany”, International Finance, vol 17, no 1, pp 1–23.
Buch, C and L Goldberg (2015): “International banking and liquidity risk transmission:
lessons from across countries”, IMF Economic Review, vol 63, no 3, pp 377–410.
Buch, C, M Buchholz and L Tonzer (2015): “Uncertainty, bank lending, and bank-level
heterogeneity”, IMF Economic Review, vol 63, pp 919–54.
Caballero, R, T Hoshi, and A Kashyap (2008): “Zombie lending and depressed
restructuring in Japan”, American Economic Review, vol 98, no 5, pp 1943–77.

Committee on the Global Financial System (2017): “Repo market functioning”,
CGFS Papers, no 59.
——— (2016): “Fixed income market liquidity”, CGFS Papers, no 55.
——— (2014a): “Trade finance: developments and issues”, CGFS Papers, no 50.
——— (2014b): “EME banking systems and regional financial integration”, CGFS Papers,
no 51.
Committee on the Global Financial System and the Financial Stability Board (2017):
“FinTech credit, market structure, business models and financial stability implications”,
22 May.
Chi, Q and W Li (2017): “Economic policy uncertainty, credit risks and banks’ lending
decisions: Evidence from Chinese commercial banks”, China Journal of Accounting
Research, vol 10, no 1, pp 33–50.
Choi, J and Y Huh (2017): “Customer liquidity provision: implications for corporate
bond transaction costs”, Board of Governors of the Federal Reserve System, Finance
and Economics Discussion Series, no 2017-116.
Claessens, S and N van Horen (2014): “The impact of the global financial crisis on
banking globalization”, IMF Economic Review, vol 63, no 4, pp 868–918.
Claessens, S and L Laeven (2004): “What drives bank competition? Some international
evidence”, World Bank Policy Research Working Paper, no 3113.
Cohen, B and M Scatigna (2016): “Banks and capital requirements: channels of
adjustment”, Journal of Banking and Finance, vol 69, supplement 1, pp S56–69.
Crystal, J, B Dages and L Goldberg (2002): “Has foreign bank entry led to sounder
banks in Latin America?” Current Issues in Economics and Finance, vol 8, no 1, pp 1–6.
Committee on Payments and Market Infrastructures (2016): “Correspondent
banking”, July.
Cucinelli, D (2015): “The impact of non-performing loans on bank lending behavior:
evidence from the Italian banking sector”, Eurasian Journal of Business and Economics,
vol 8, pp 59–71.
Cull, R and M Martínez-Pería (2013): “Bank ownership and lending patterns during
the 2008–2009 financial crisis: evidence from Latin America and Eastern Europe”,
Journal of Banking & Finance, vol 37, no 12, pp 4861–78.
Cunliffe, J (2016): “Why are interest rates low?”, speech at Manchester University,
November.
Damodoran, A (2016): “Equity risk premiums (ERP): determinants, estimation and
implications”, 5 March.
De Haas, R and I van Lelyveld (2014): “Multinational banks and the global financial
crisis: weathering the perfect storm?”, Journal of Money, Credit and Banking, vol 46,
no 1, pp 333–64.
——— (2010): “Internal capital markets and lending by multinational banks
subsidiaries”, Journal of Financial Intermediation, vol 19, pp 1–25.
De Haas, R and N van Horen (2013): “Running for the exit? International bank lending
during a financial crisis”, Review of Financial Studies, Society for Financial Studies,
vol 26, no 1, pp 244–85.

Debelle, G (2017): “Recent trends in Australian capital flows”, Australian Financial
Review, Banking and Wealth Summit Sydney, April.
Dent, K and B Westwood (2016): “Stress testing of banks: an introduction”, Bank of
England, Quarterly Bulletin, no 3, pp 130–43.
Deutsche Bundesbank (2016): “Financial Stability Review”.
Dick-Nielsen, J and M Rossi (2017): “The cost of immediacy for corporate bonds”,
mimeo.
Drehmann, M, C Borio and K Tsatsaronis (2011): “Anchoring countercyclical capital
buffers: the role of credit aggregates”, International Journal of Central Banking,
December, pp 189–240.
Duarte, F and T Eisenbach (2013): “Fire-sale spillovers and systemic risk”, Federal
Reserve Bank of New York, Staff Reports, no 645, October.
European Central Bank (2016a): “Report on financial structures”, October.
——— (2016b), “SSM supervisory statement on governance and risk appetite”, June.
——— (2016c), “Financial integration and risk-sharing in a monetary union”, in
Financial integration in Europe, April.
——— (2017a): Financial Stability Review, May.
——— (2017b): “Cross-border bank consolidation in the euro area”, in Financial
integration in Europe, May.
Ehlers, T and P Wooldridge (2015): “Channels and determinants of foreign bank
lending”, in “Cross-border financial linkages: challenges for monetary policy and
financial stability”, BIS Papers, no 82, pp 29–68.
Ellul, A and V Yerramilli (2013): “Stronger risk controls, lower risk: evidence from US
bank holding companies”, Journal of Finance, vol 68, no 5, pp 1757–803.
Enria, A (2017): “Introductory remarks: new frontiers on stress testing”, EBA-IMF
Colloquium, 1 March, London.
Ernst and Young (2015): “Rethinking risk management: banks focus on non-financial
risks and accountability”, Sixth annual global EY/IIF bank risk management survey.
——— (2016): “A set of blueprints for success”, Seventh annual global EY/IIF bank risk
management survey.
European Banking Authority (2013): “Report on impact assessment for liquidity
measures under Article 509(1) of the CRR”.
——— (2014): “Second report on impact assessment for liquidity measures under
Article 509(1) of the CRR”.
——— (2015): “Report on net stable funding requirements under Article 510 of the
CRR”.
——— (2016): “Risk assessment of the European Banking System”, December.
European Systemic Risk Board (2016): “Market liquidity and market-making, October.
Fink, K, U Krüger, B Meller and L Wong (2016): “The credit quality channel: modeling
contagion in the interbank market”, Journal of Financial Stability, vol 25, pp 83–97.

Fraisse, H, M Lé and D Thesmar (2017): “The real effects of bank capital requirements”,
European Systemic Risk Board, Working Paper, no 47.
Financial Stability Board (FSB) (2017a): “Ten years on – taking stock of post-crisis
resolution reforms”, Sixth report on the implementation of resolution reforms”, 6 July.
——— (2017b): “Financial stability implications for Fintech” 27 June.
——— (2017c): “FSB Correspondent Banking Data Report”, 4 June.
——— (2016): “2016 list of global systemically important banks (G-SIBs)”,
21 November.
Financial Stability Institute (2017): “Resolution of non-performing loans – policy
options”, FSI Insights on policy implementation, no 3, October.
Financial Stability Oversight Council (2016): Annual Report.
Griffin, B and E Sheedy (2017): “Risk governance, structures, culture, and behavior: a
view from the inside”, Corporate Governance, An International Review, March.
Haldane, A (2017): “Rethinking financial stability”, speech at the Rethinking
Macroeconomic Policy IV Conference, Washington, 12 October.
International Monetary Fund (2013): “Chapter three: changes in bank funding
patterns and financial stability issues”, Global Financial Stability Report, October.
——— (2015): “Chapter two: International banking after the crisis: increasingly local
and safer?”, Global Financial Stability Report, April.
——— (2017a): “Chapter one: Is Growth at Risk?”, Global Financial Stability Report,
October.
——— (2017b): “Recent trends in correspondent banking relationships – further
considerations”, Policy Papers, 16 March.
Iercosan, D, A Kumbhat, M Ng and J Wu (2017): “Trading activities at systemically
important banks, Part 1: Recent trends in trading performance”, FEDS Notes, 10 July.
International Institute of Finance (2009): “Reform in the financial services industry:
strengthening practices for a more stable system”.
Jobst, A and D Gray (2013): “Systemic contingent claims analysis: estimating market-
implied systemic risk”, IMF Working Paper, no 13/54.
Kamil, H and K Rai (2010): “The global credit crunch and foreign banks’ lending to
emerging markets: why did Latin America fare better?”, IMF Working Paper, no
10/102.
Kaya, O and L Wang (2016): “The role of bank lending tightening on corporate bond
issuance in the Eurozone”, The Quarterly Review of Economics and Finance, vol 60, pp
1–234.
Kelly, B, H Lustig, and S Van Nieuwerburgh (2016): “Too systemic to fail: what option
markets imply about sector-wide government guarantees.” American Economic
Review, vol 106, no 6, pp 1278–319.
Kick, T and E Prieto (2013): “Bank risk-taking and competition: evidence from regional
banking markets”, Deutsche Bundesbank, Discussion Paper, no 30.
Koetter, M and F Noth (2013): “IT use, productivity, and market power in banking”,
Journal of Financial Stability, vol 9, issue 4, pp 695–704.

Laeven, L and F Valencia (2012): “Systemic banking crises database: an update”,
IMF Working Paper, no 12/163.
Laeven, L and R Levine (2007): “Is there a diversification discount in financial
conglomerates?”, Elsevier Journal of Financial Economics, vol 85, no 2, pp 331–67.
Lerner, A (1934): “The concept of monopoly and the measurement of monopoly
power”, Review of Economic Studies, vol 1, no 3, pp 157–75.
Martinez-Miera, D and R Repullo (2010): “Does competition reduce the risk of bank
failure?”, The Review of Financial Studies, vol 23, no 10, pp 3638–64.
McCauley, R, A Bénétrix, P McGuire and G von Peter (2017): “Financial deglobalisation
in banking?”, BIS Working Papers, no 650.
McGuire, P and G von Peter (2016): “The resilience of banks’ international operations”,
BIS Quarterly Review, March.
——— (2012): “The dollar shortage in global banking and the international policy
response”, International Finance, vol 15, no 2, pp 155–78.
McKinsey & Company (2016): “Capital markets and investment banking: time for real
change and bold actions”, Financial Services, September.
Mizrach, B (2015): “Analysis of corporate bond liquidity”, FINRA Office of the Chief
Economist Research Notes.
Netherlands Bank (2016): “Loan markets in motion: Larger role of pension funds and
insurers boosts financial stability”, November.
Nouy, D (2017): “Safer than ever before? An assessment of the impact of regulation
on banks’ resilience eight years on”, Bank of France Financial Stability Review, no 21,
April, pp 23–32.
Peek, J and E Rosengren (1997): “The international transmission of financial shocks:
the case of Japan”, American Economic Review, vol 87, no 4, pp 495–505.
——— (2005): “Unnatural selection: perverse incentives and the misallocation of credit
in Japan”, American Economic Review, vol 95, no 4, pp 1144–66.
Philippon, T (2015): “Has the US finance industry become less efficient? On the theory
and measurement of financial intermediation”, American Economic Review, vol 105,
no 4, pp 1408–38.
Praet, P (2016): “The importance of a genuine banking union for monetary policy”,
speech to EMU Forum 2016, Central Bank of the Republic of Austria, Vienna,
24 November.
Reinhardt, D and S Riddiough (2015): “The two faces of cross-border banking flows”,
IMF Economic Review, vol 63, no 4, pp 751–91.
Reinhart, C and V Reinhart (2015): “Financial crises, development and growth: a long-
term perspective”, World Bank Economic Review, vol 29, no 1, pp 1–24.
Remolona, E and I Shim (2015): “The rise of regional banking in Asia and the Pacific”,
BIS Quarterly Review, September.
Roengpitya, R, N Tarashev, K Tsatsaronis and A Villegas (2017): “Bank business
models: popularity and performance”, BIS Working Papers, no 682.

Sarin, N and L Summers (2016): “Understanding bank risk through market measures”,
Brookings Papers on Economic Activity, no. 1: 2016, pp 57–109.
Schivardi, F, E Sette and G Tabellini (2017): “Credit misallocation during the European
financial crisis”, School of European Political Economy, Working Paper, no 3.
Senior Supervisors Group (2008): “Observations on risk management practices during
the recent market turbulence”, 6 March.
——— (2009): “Risk management lessons from the global banking crisis of 2008”,
21 October.
Storz, M, M Koetter, R Setzer and A Westphal (2017): “Do we want these two to tango?
On zombie firms and stressed banks in Europe”, Halle Institute for Economic Research,
IWH Discussion Papers, no 13.
Swiss National Bank (2016): Financial Stability Report.
Tarullo, D (2016): “Next steps in the evolution of stress testing”, speech at the Yale
University School of Management Leaders Forum, New Haven, September.
Trebbi, F and K Xiao (2015): “Regulation and market liquidity”, NBER Working Paper,
no w21739.
UBS (2010): Transparency report to the shareholders of UBS AG, October.
Uluc, A and T Wieladek (2017): “Capital requirements, risk shifting and the mortgage
market,” Working Paper Series, no 2061, European Central Bank.
Van Rixtel, A and G Gasparini (2013): “Financial crises and bank funding: recent
experience in the euro area”, BIS Working Papers, no 406.
Wagner, W (2010): “Diversification at financial institutions and systemic crises”,
Journal of Financial Intermediation, vol 19, no 3, pp 373–86.

Annex 1: Country banking data tables
This annex contains a set of tables with time series system-level banking data across
CGFS countries. The data cover various aspects of banking:
 banking market asset size, relative to GDP and total financial sector assets
 other banking market capacity indicators, relative to the population
 banking market concentration ratios and shares of foreign banks
 asset and loan composition, as well as non-performing loan ratios
 funding composition
 capitalisation, including risk-weighted capital positions and simple leverage
 consolidated foreign bank claims
 profitability measures and key performance ratios
 profitability components, relative to assets
Data on banking market size, capacity and concentration are on a domestic (or
resident) basis (except for China and Korea), and thus include foreign bank
subsidiaries and branches in host jurisdictions but exclude the foreign operations of
banks in their home jurisdictions. Other banking data (such as capital positions and
profitability) are mainly on a consolidated group basis (ie including banks’ foreign
operations), although there are some exceptions.
Balance sheet positions and other stock data are presented for every second year,
starting 2002 and ending 2016. To reduce volatility, profitability (flow data) is instead
presented as the average of two years, starting 2002–03 and ending 2015–16.
An accompanying Excel file provides the full time series from 2000 to 2016, as
well as data sources and other metadata, where possible.111
111
The full dataset in excel is available at https://www.bis.org/publ/cgfs60/cgfs60_datasets.xlsx

Banking system assets: share of GDP1
In per cent Annex Table 1.1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 285 322 328 324 299 277
Belgium 282 306 343 359 310 280 275 261
France 238 257 326 384 392 402 395 388
Germany 292 294 300 311 324 301 268 250
Italy 157 186 212 205 217 209 200
Luxembourg 2,640 2,488 2,483 2,451 1,914 1,672 1,479 1,409
Netherlands 239 289 300 345 358 386 375 368
Spain 170 189 240 275 284 298 256 220
Other Europe
Sweden 174 193 226 289 270 282 297 287
Switzerland 192 195 191 193 207 239 250 283
United Kingdom 327 380 451 531 502 466 384 392
United States 74 78 83 94 86 87 92 91
Other advanced
Australia 127 147 166 216 196 199 229 246
Canada 88 95 98 112 125 134 117 134
Japan 144 139 141 159 169 187 196 204
Emerging markets
Brazil 60 70 73 67 100 94 90 110
China 177 195 200 195 228 243 261 304
Hong Kong SAR 462 542 552 630 692 729 816 829
India 63 67 73 83 88 89 89 86
Korea 19 18 31 55 32 33 36 39
Mexico 32 31 33 36 37 37 38 42
Singapore 611 581 643 548 541 581 586
1
Banking system assets are on a domestic or resident basis, except for Korea and China, which are on a consolidated basis.
Source: National data, for further details see https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx.

Banking system assets: share of total assets across CGFS membership1
2006 2008 2010 2012 2014 2016

Euro area 41.7 43.5 38.8 35.6 31.2 26.5
Belgium 1.9 1.8 1.4 1.2 1.1 1.0
France 10.1 10.9 9.8 9.4 8.8 7.7
Germany 12.0 11.3 10.4 9.3 8.1 7.0
Italy 4.8 4.9 4.1 3.9 3.5 3.0
Luxembourg 1.4 1.3 1.0 0.8 0.8 0.7
Netherlands 3.1 3.2 2.8 2.7 2.5 2.2
Spain 4.0 4.4 3.8 3.5 2.7 2.2
Other Europe
Sweden 1.3 1.3 1.3 1.4 1.3 1.2
Switzerland 1.1 1.1 1.2 1.4 1.4 1.5
United Kingdom 16.4 12.4 11.5 10.7 9.3 7.9
United States 14.5 14.1 12.1 12.0 13.6 14.3
Other advanced
Australia 1.7 1.9 2.5 2.6 2.6 2.5
Canada 1.6 1.6 1.9 2.1 1.7 1.7
Japan 8.0 8.4 9.5 8.3 7.2 8.3
Emerging markets
Brazil 1.0 1.2 2.1 2.0 1.9 1.7
China 7.2 9.4 13.3 17.9 23.1 27.4
Hong Kong SAR 1.4 1.4 1.5 1.6 2.0 2.2
India 0.8 0.8 1.2 1.2 1.4 1.5
Korea 1.7 1.3 1.4 1.5 1.6 1.6
Mexico 0.4 0.3 0.4 0.4 0.4 0.4
Singapore 1.1 1.2 1.3 1.4 1.5 1.4
1
Total assets across CGFS membership calculated as the sum of all non-euro area countries and the euro area aggregate; for this reason
numbers in this table do not sum up to 100%.

Share of banking system assets in total financial institution assets1
2002 2004 2006 2008 2010 2012 2014 2015

Euro area 53 58 54 53 47 45
Belgium 63 60 57 58 51 49 47 46
France 69 69 73 72 70 70 66 65
Germany 66 63 62 60 68 65 61 59
Italy 64 68 69 60 59 57 56
Luxembourg
Netherlands 35 38 32 33 29 28 26 25
Spain 69 65 66 67 65 66 64 64
Other Europe
Sweden
Switzerland 32 31 26 29 29 31 30 32
United Kingdom 58 50 49 36 38 36 34 36
United States 19 19 18 21 19 19 20 20
Other advanced
Australia 43 44 42 54 53 55 55 55
Canada 24 23 21 25 23 25 20 19
Japan 31 28 30 30 33 32 34 37
Emerging markets
Brazil 91 88 52 55 55 52 52 53
China 93 90 89 85 77 75
Hong Kong SAR 90 87 82 84 81 80 80 80
India 76 73 68 63 63 64 64 61
Korea 45 46 45 43 40 38 37 36
Mexico 55 49 47 46 43 41 39 40
Singapore 88 84 85 79 73 71 70
1
As a share of assets of financial corporations, excluding the central bank. Financial assets when available, otherwise total assets.
Sources: FSB, Global Shadow Banking Monitoring Report 2016, May 2017; national data, for further details see
https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx.

Banking system capacity indicators: bank branches per population1
Number, per 100,000 population Annex Table 1.4
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 56 54 57 58 55 52 48 44
Belgium 46 43 40 36 34 32 29
France 41 42 41 62 59 58 57
Germany 62 55 49 48 47 45 44
Italy 53 54 56 58 57 55 51 48
Luxembourg
Netherlands 27 23 21 21 17 15 11 10
Spain 92 93 97 99 92 81 68 62
Other Europe
Sweden 22 21 21 21 20 19 18 15
Switzerland 45 43 42 41 40 38 36 33
United Kingdom 23 22 21 20 19 18 17
United States 23 24 26 27 27 27 26
Other advanced
Australia 25 24 25 25 24 24 23 22
Canada 25 25 24 24 24 23
Japan 17 17 16 16 16 16 16 16
Emerging markets
Brazil 10 9 10 10 9 10 11 11
China
Hong Kong SAR 21 19 19 20 20 19 19 17
India 8 9 9 10 11 12
Korea 13 13 14 15 15 17 17 16
Mexico 8 7 8 10 10 11 11 10
Singapore 9 9 8 8 8 8 8
1
For information on the consolidation basis and coverage of these data, see https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx.

Banking system capacity indicators: bank employees per population1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 716 686 692 697 661 637 603 575
Belgium 715 666 631 603 555 534 496
France 648 638 657 693 661 657 634 634
Germany 911 852 827 822 804 807 790
Italy 597 585 584 576 542 521 493 487
Luxembourg 5,247 4,956 5,277 5,624 5,229 5,056 4,698 4,523
Netherlands 782 724 715 707 651 619 560 505
Spain 569 557 573 586 552 495 438 407
Other Europe
Sweden 435 405 429 442 433 411 417 399
Switzerland 1,429 1,342 1,388 1,430 1,373 1,308 1,263 1,204
United Kingdom 632 644 632 644 609 567 559
United States 701 716 740 708 675 672 643 635
Other advanced
Australia 699 683 720 722 674 627 651 654
Canada 807 836 840 853 856 843
Japan 229 209 202 219 223 219 217 214
Emerging markets
Brazil 247 250 264 255 242
China 176 208 205 223 248 275 296
Hong Kong SAR
India 81 82 83 87 93 98
Korea 191 192 191 210 206 220 228 218
Mexico 102 111 124 142 144 180 188 188
Singapore 779 828 990 1,096 1,166 1,165 1,234 1,232
1

Number of banks1
Number Annex Table 1.6
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 5,590 6,097 5,714 5,395 4,828 4,385
Belgium 104 105 107 107 104 103 90
France 921 820 763 658 638 611 563 529
Germany 2,593 2,400 2,301 2,169 2,093 2,053 1,990 1,888
Italy 814 778 793 799 760 706 664 604
Luxembourg 177 162 156 152 147 141 144 141
Netherlands 113 113 106 100 87 80 56 47
Spain 274 266 271 280 274 249 219 206
Other Europe
Sweden 214 211 205 180 172 168 163 158
Switzerland 356 338 331 327 320 297 275 261
United Kingdom 333 335 345 366
United States 9,354 8,976 8,680 8,305 7,658 7,083 6,509 5,913
Other advanced
Australia 50 50 53 55 54 66 70 82
Canada 67 67 70 74 76 75 79 83
Japan 416 407 396 387 379 377 373 370
Emerging markets
Brazil 147 139 133 136 137 138 136 134
China 88,150 19,797 5,634 3,769 3,747 4,089 4,398
Hong Kong SAR 224 208 202 200 193 200 203 195
India 97 90 85 79 82 85 90 93
Korea 20 19 18 18 18 18 18 17
Mexico 30 28 31 43 41 43 45 47
Singapore 120 115 108 113 120 123 124 124
1

Banking system capacity indicators: banks per population1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 1.78 1.90 1.74 1.63 1.44 1.29
Belgium 1.00 1.00 1.00 0.98 0.93 0.91 0.79
France 1.50 1.32 1.21 1.03 0.99 0.94 0.85 0.79
Germany 3.14 2.91 2.79 2.64 2.56 2.55 2.46
Italy 1.43 1.35 1.37 1.36 1.28 1.19 1.09 1.00
Luxembourg 39.86 35.61 33.26 31.42 29.28 26.86 26.20 24.47
Netherlands 0.70 0.69 0.65 0.61 0.52 0.48 0.33 0.28
Spain 0.66 0.61 0.61 0.61 0.59 0.53 0.47 0.44
Other Europe
Sweden 2.39 2.34 2.25 1.94 1.83 1.76 1.67 1.58
Switzerland 4.87 4.56 4.41 4.25 4.07 3.69 3.34 3.10
United Kingdom 0.53 0.53 0.53 0.56
United States 3.25 3.07 2.91 2.73 2.48 2.26 2.04 1.83
Other advanced
Australia 0.26 0.25 0.26 0.26 0.24 0.29 0.30 0.34
Canada 0.27 0.26 0.27 0.27 0.27 0.26 0.27 0.28
Japan 0.33 0.32 0.31 0.30 0.30 0.30 0.29 0.29
Emerging markets
Brazil 0.08 0.08 0.07 0.07 0.07 0.07 0.07 0.07
China 6.78 1.51 0.42 0.28 0.28 0.30 0.32
Hong Kong SAR 3.33 3.06 2.93 2.87 2.74 2.79 2.80 2.64
India 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
Korea 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.03
Mexico 0.03 0.03 0.03 0.04 0.04 0.04 0.04 0.04
Singapore 2.87 2.76 2.45 2.34 2.36 2.32 2.27 2.21
1

Banking system concentration: share of system assets of three largest banks1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area
Belgium 68 70 69 67 60 49 46 47
France 54 58 58 58 56
Germany 18 21 21 20 33 31 31 28
Italy 22 19 19 29 35 35 33 34
Luxembourg 20 20 19 22 23 23 19
Netherlands 71 71 71 72 69 73 76 75
Spain 44 42 40 42 44 46 47 50
Other Europe
Sweden 64 62 61 61 60 58 58 56
Switzerland 45 48 47 45 43 39 37 40
United Kingdom 28 32 34 33 42 41 37 33
United States 21 25 30 32 33 33 33 32
Other advanced
Australia 56 53 49 57 60 61 60 60
Canada 57 52 53 52 52 54 56 54
Japan 28 28 34 36 35 36 44 43
Emerging markets
Brazil 43 45 43 51 54 53 56 57
China
Hong Kong SAR
India 34 32 31 29 27 26 26 27
Korea 41 41 39 41 41 40 39 39
Mexico 57 56 61 56 56 52 52 51
Singapore 30 28 28 30 30 33 33
1

Banking system concentration: share of system assets of five largest banks1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 43 44 47 47 48 48
Belgium 82 84 84 81 75 66 66 66
France 77 81 81 81 82
Germany 26 29 29 27 40 38 37 35
Italy 31 26 26 31 40 40 41 43
Luxembourg 30 29 27 31 33 32 29
Netherlands 82 84 84 84 82 82 86 89
Spain 53 52 49 51 57 62 63 65
Other Europe
Sweden 81 79 79 80 78 77 77 76
Switzerland 56 58 57 55 53 49 51 53
United Kingdom 41 47 50 45 53 54 51 48
United States 25 31 35 38 44 45 44 43
Other advanced
Australia 75 72 69 74 78 80 81 80
Canada 85 83 82 80 81 83 81 81
Japan 43 43 45 46 46 47 51 51
Emerging markets
Brazil 59 60 60 73 76 77 78 82
China 57 55 51 49 45 41 37
Hong Kong SAR
India 43 41 40 37 35 35 35 36
Korea 57 56 61 60 61 60 59 62
Mexico 72 74 80 78 74 70 73 70
Singapore 42 39 39 41 41 43 42
1

Bank asset structure: loans as a share of assets1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 59 60 58 56 56 52 53 55
Belgium 62 55 62 62 64 65
France 33 31 29 28 29 29 29 31
Germany 72 71 70 67 66 55 60 62
Italy 73 71 74 76 74 74 76
Luxembourg 19 17 19 22 25 24 25 28
Netherlands 81 79 79 71 69 69 70 73
Spain 69 72 73 70 69 66 65 66
Other Europe 51 42 51 52 53 54
Sweden 75 73 73 69 73 71 71 77
Switzerland 40 38 41 44 49 50
United Kingdom 53 40 51 51 52 52
United States 60 61 61 57 55 53 53 55
Other advanced 59 57 58 57 56 57 57 57
Australia 68 67 65 59 66 67 65 68
Canada 57 55 55 52 55 62 63 63
Japan 59 56 57 58 53 52 51 51
Emerging markets 63 63 59 56 59 61 61 59
Brazil 44 46 47 51 52 55 58 53
China 78 73 62 56 59 62 62 61
Hong Kong SAR 35 30 30 31 34 37 39 39
India 44 45 55 57 57 60 61 62
Korea 67 69 70 69 73 73 75 75
Mexico 42 38 38 43 41 46 48 50
Singapore 86 87 82 82 84 84 65
1
Weighted averages for groups of countries, based on total assets. Euro area calculated as a weighted average of individual countries
shown above. For information on the consolidation basis and coverage of these data, see

Bank asset structure: debt securities as a share of assets1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 22 22 23 22 23 21 23 21
Belgium 27 25 22 22 20 18 20 16
France 28 31 36 33 29 26 30 28
Germany 21 22 23 23 23 18 19 18
Italy 10 9 12 19 27 25 24
Luxembourg 23 23 25 21 23 20 20 18
Netherlands 11 14 14 19 21 18 17 16
Spain 16 16 11 11 13 15 17 15
Other Europe 24 17 20 17 17 17
Sweden 14 17 19 16 16 14 13 11
Switzerland 31 18 23 17 17 14
United States 24 23 22 22 25 26 25 24
Other advanced 21 24 23 21 27 26 23 20
Australia 10 11 11 14 16 17 18 16
Canada 17 20 22 14 27 21 20 21
Japan 22 27 26 24 31 31 25 20
Brazil 30 29 28 22 20 21 20 22
China 15 12 11 12 12 24 27 41
Hong Kong SAR 21 20 21 18 22 22 20 22
India 38 41 32 28 30 29 28 26
Korea 23 22 21 22 19 18 17 17
Mexico 23 32 36 34 31 27 27 23
Singapore 15 15 13 15 13 12 14
1
Weighted averages for groups of countries, based on total assets. Euro area calculated as a weighted average of individual countries

Bank asset structure: liquid assets as a share of assets1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 7 7 7 7 8 10 11 12
Belgium 12 12 15 15 17 19
France 10 10 10 11 13 13
Germany 6 6 5 4 6 7 8 9
Italy 6 5 5 8 11 13 15
Luxembourg 7 8 7 5 7 7 7 7
Netherlands 7 6 4 6 8 7 11
Spain 13 9 6 7 10 13 15 16
Other Europe 7 8 13 16 15 17
Sweden 3 3 3 2 3 7 7 5
Switzerland
United States 6 5 4 8 9 11 15 14
Other advanced 14 16 11 12 17 15 16 18
Australia 3 2 1 2 3 5 7 7
Canada 11 7 6 6
Japan 15 17 13 14 22 22 24 25
Brazil 32 33 30 20 25 21 18 19
China
Hong Kong SAR 6 6 5 7 10 11 9 11
India 33 38 30 30 30 27 26 25
Korea 3 6 7 6 7 6 7 7
Mexico 33 29 28 25
Singapore 6 5 6 9 9 8 10
1
Liquid assets calculated as cash and balances with central banks plus government securities; weighted averages for groups of countries,
based on total assets. Euro area calculated as a weighted average of individual countries shown above. For information on the consolidation
basis and coverage of these data, see https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx.

Bank lending: share of residential mortgages in total lending1
In per cent of total lending Annex Table 1.13
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 21 23 25 24 20 20 22 22
Belgium 25 29 32 35
France 33 38 42 42 44 46 47 47
Germany 8 8 9 9
Italy 9 11 12 10 14 14 14 15
Luxembourg 6 8 7 8 10 12 14 13
Netherlands 26 25 25 24 23 23 24 24
Spain 23 25 27 26 27 28 32 31
Other Europe 47 48 47 43 46 49 52 54
Sweden 32 39 41 41 45 45 46 48
Switzerland 56 61 63 62 63 63 63 64
United Kingdom 47 47 45 40 41 45 50 51
United States 33 33 33 29 29 28 26 26
Other advanced 24 29 30 28 32 35 34 33
Australia 55 59 57 55 62 63 63 63
Canada 33 37 34 30 32 39 37 35
Japan 19 22 22 22 24 23 22 22
Brazil 5 9 12 16 19
China 10 17 18 22 25 26
Hong Kong SAR 35 32 28 24 24 23 20 21
India 12 11 9 9 9 10
Korea 23 24 24 26
Mexico 13 12 17 16 17 16 17 16
Singapore 6 6 6 8 9 9 12
1
Weighted averages for groups of countries, based on total lending. Euro area calculated as a weighted average of individual countries

Bank lending: share of business lending in total lending1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 25 24 25 27 28 27 28 27
Belgium 32 33
France 38 35 34 35 32 32 33 34
Germany 16 15 14 15 15 16 16 17
Italy 37 37 36 35 35 34 32 31
Luxembourg 46 39 40 45 40 38 41 47
Netherlands 19 17 18 21 26 26 33 28
Spain 36 37 39 42 43 38 38 36
Other Europe 25 22 23 22 22 22 22 21
Sweden 46 42 40 43 38 38 36 34
Switzerland 31 29 26 26 26 26 25 25
United Kingdom 20 18 21 18 18 18 17 17
United States 36 34 37 40 35 36 37 38
Other advanced 54 48 45 45 42 38 36 38
Australia 29 28 30 31 27 28 28 28
Canada 29 23 24 29 22 20 21 24
Japan 60 57 55 53 52 50 47 47
Brazil 63 57 54 52 48 46 43 40
China 96 72 68 66 65 63
Hong Kong SAR 50 52 55 57 61 62 63 60
India 9 8 7 8 10 10 12 13
Korea 69 66 65 65 62
Mexico 33 36 38 46 46 44 43 46
Singapore 11 12 17 18 21 24 32
1

Bank lending: share of other lending in total lending1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 46 45 42 42 53 53 51 51
Belgium 36 32
France 29 27 24 23 24 23 20 19
Germany 77 76 75 75
Italy 54 52 52 55 51 52 53 54
Luxembourg 49 54 53 47 50 50 46 40
Netherlands 56 58 57 55 51 51 44 48
Spain 41 38 34 32 30 34 30 34
Other Europe 29 30 30 35 32 29 26 25
Sweden 22 20 19 16 17 17 18 18
Switzerland 12 10 11 13 12 11 11 10
United Kingdom 33 35 34 42 40 37 33 32
United States 31 34 30 31 36 36 37 36
Other advanced 23 23 25 27 26 27 30 29
Australia 16 13 13 14 11 9 9 9
Canada 38 41 42 41 46 41 42 41
Japan 21 21 23 25 24 27 31 31
Brazil 43 44 42 41 41
China 11 14 12 11 11
Hong Kong SAR 15 17 17 19 15 15 17 19
India 81 81 81 81 79 77
Korea 11 12 12 13
Mexico 55 53 45 37 37 39 40 39
Singapore 83 82 78 74 69 67 56
1

Bank asset performance: non-performing loan ratio1
In per cent of total loans Annex Table 1.16
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 3.0 3.3 4.4 5.4 5.8 5.0
Belgium 1.3 1.7 2.8 3.7 4.2 3.4
France 5.0 4.2 3.0 2.8 3.6 3.6 3.6 3.5
Germany 3.1 2.8 2.3 1.7
Italy 5.8 6.2 9.9 13.5 17.7 17.3
Luxembourg
Netherlands 1.5 2.6 3.0 2.9 2.5
Spain 2
0.9 0.7 0.6 3.3 5.8 10.6 12.6 9.2
Other Europe 0.9 1.8 3.8 3.4 1.7 1.1
Sweden 0.5 0.4
Switzerland 0.5 0.8 0.5 0.4 0.4 0.5
United Kingdom 1.1 2.2 4.8 4.3 2.4 1.5
United States 1.4 0.8 0.8 3.0 4.9 3.6 2.0 1.4
Other advanced 6.1 3.1 1.9 2.0 2.1 1.8 1.2 1.1
Australia 1.1 0.6 0.6 1.6 2.1 1.6 1.0 0.9
Canada 1.5 0.7 0.4 0.7 1.1 0.6 0.5 0.6
Japan 7.4 4.0 2.5 2.4 2.4 2.3 1.6 1.3
Emerging markets 16.0 8.3 4.4 2.1 1.5 1.3 1.5 2.1
Brazil 4.3 3.1 3.2 2.7 3.3 3.4 2.7 3.7
China 23.8 13.2 7.1 2.4 1.1 1.0 1.3 1.7
Hong Kong SAR 4.5 2.1 1.0 1.2 0.8 0.6 0.5 0.9
India 11.0 7.4 3.5 2.4 2.5 2.9 4.1 7.8
Korea 2.3 1.9 0.8 1.2 1.9 1.3 1.6 1.4
Mexico 4.6 2.5 2.0 3.2 2.3 2.5 3.1 2.1
Singapore 1.7 0.8 1.0 1.0 0.8 0.8 1.5
These data can be subject to significant measurement differences across countries.
1
Weighted averages for groups of countries, based on total loans. Euro area calculated as a weighted average of individual countries shown
above. For information on the consolidation basis and coverage of these data, see
https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx. 2 The NPL ratio for Spain is calculated on a domestic basis; calculation on
consolidated basis results in considerably lower figures.

Bank funding structure: share of deposits in total funding1
In per cent of total funding Annex Table 1.17
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 33 34 35 37 40 40 43 46
Belgium 48 57 62 66 69 67
France 40 37 33 31 34 37 38 41
Germany 13 14 13 12 15 16 17 17
Italy 52 50 49 48 53 49 53 59
Luxembourg 2
70 65 68 70 72
Netherlands 33 39 40 36 36 42 42
Spain 65 61 55 57 61 62 65 67
Other Europe 46 47 50 54 59 60
Sweden 31 28 28 26 27 30 30 32
Switzerland 46 54 55 61 64 66
United States 65 65 65 65 70 74 75 76
Other advanced 67 67 65 67 67 64 63 65
Australia 48 44 41 43 49 55 59 60
Canada 54 56 54 58 59 54 53 55
Japan 71 72 73 74 75 71 69 70
Brazil 59 62 65 64 62 59 59 58
China 96 96 92 89 91 91 90
Hong Kong SAR 56 54 58 56 56 56 55 57
India 89 91 91 91 93 91 91 93
Korea 66 65 61 59 64 67 65 66
Mexico 45 48 44 43 46 47 49 52
Singapore 35 41 41 43 45 45 46
1
Excludes deposits with banks. Total funding is measured as a sum of deposit funding, other borrowings and equity. Weighted averages
for groups of countries, based on total funding. Euro area calculated as a weighted average of individual countries shown above. For
information on the consolidation basis and coverage of these data, see https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx. 2 For
Luxembourg, interbank borrowing is excluded from the total funding measure.

Bank funding structure: share of other borrowing in total funding1
In per cent of total funding Annex Table 1.18
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 59 57 56 53 50 48 45 42
Belgium 48 39 32 26 23 24
France 48 51 56 56 53 51 52 47
Germany 81 81 81 81 78 76 73 73
Italy 39 41 42 42 36 40 34 27
Luxembourg 2
20 23 20 17 17
Netherlands 49 43 43 45 43 38 39
Spain 29 33 39 37 32 31 28 24
Other Europe 49 48 43 39 34 32
Sweden 67 70 71 72 71 68 68 66
Switzerland 49 41 38 32 29 26
United States 25 25 24 26 19 15 14 13
Other advanced 29 28 30 29 27 30 31 29
Australia 44 49 53 51 43 37 35 34
Canada 41 38 41 36 35 40 41 39
Japan 26 23 21 22 20 23 25 25
Brazil 26 21 19 21 22 27 29 29
China 0 2 2 4 2 1 1
Hong Kong SAR 40 42 39 41 40 40 41 38
India 11 7 8 7 6 8 8 7
Korea 29 29 32 35 28 24 26 25
Mexico 47 44 46 47 42 42 38 37
Singapore 61 56 55 52 51 51 49
1
All borrowings excluding deposits with non-banks. Total funding is measured as a sum of deposit funding, other borrowings and equity.
Weighted averages for groups of countries, based on total funding. Euro area calculated as a weighted average of individual countries
https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx. 2 For Luxembourg, interbank borrowing is excluded from other borrowing and
the total funding measure.

Bank capital positions: Common Equity Tier 1 capital ratio1
In per cent of risk-weighted assets Annex Table 1.19
2013 2014 2015 2016

Euro area 12.7 13.5 13.9
Belgium 15.3 14.7 15.4 15.7
France 13.2 11.7 12.5 13.1
Germany 14.9 15.3 15.7
Italy 12.0 12.3 11.5
Luxembourg 18.4 19.9 24.0
Netherlands 14.4 14.6 15.7
Spain 11.8 12.7 12.8
Other Europe 12.3 13.7 14.7 15.3
Sweden 15.6 17.4 18.9 20.6
Switzerland 17.5 17.4 17.3 16.6
United Kingdom 10.0 11.4 12.6 13.4
United States 11.6 12.5 12.2 12.5
Other advanced 10.4 10.6 11.2 11.4
Australia 8.6 9.0 10.1 9.8
Canada 9.3 9.8 10.0 11.1
Japan 11.4 11.7 12.1 12.1
Emerging markets 10.4 11.0 11.0 11.0
Brazil 11.5 12.3 11.8 12.7
China 9.9 10.8 10.9 10.8
Hong Kong SAR 13.2 13.7 14.6 15.4
India 9.9 10.0 10.5
Korea 11.2 10.8 10.8 12.1
Mexico 13.0 13.5 12.9 12.8
Singapore 13.8 13.6 13.7 13.9
1
Weighted averages for groups of countries, based on total assets. Euro area is as reported for the entire region. For information on the
consolidation basis and coverage of these data, see https://www.bis.org/publ/cgfs60/cgfs60_metadata.xlsx.

Bank capital positions: Tier 1 capital ratio1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 8.8 11.0 12.5 13.3 14.7
Belgium 11.5 15.5 15.9 15.3 16.2
France 11.3 13.7 15.6 13.1 14.5
Germany 9.7 11.9 13.9 15.3 16.3
Italy 8.0 8.6 7.8 7.6 9.3 11.1 12.4 12.0
Luxembourg 14.3 16.5 14.7 14.3 17.6 15.5 18.5 24.1
Netherlands 10.0 12.2 12.7 15.4 17.9
Spain 10.7 10.2 8.7 8.1 9.6 9.8 11.9 12.9
Other Europe 15.1 17.0 14.7 17.2
Sweden 7.1 7.6 7.1 9.9 13.4 15.5 19.5 23.2
Switzerland 15.9 17.7 18.1 18.6
United Kingdom 12.2 15.1
United States 10.4 10.8 10.6 9.8 12.6 13.0 13.0 12.9
Other advanced 6.4 7.5 8.3 8.9 12.2 11.7 12.0 13.0
Australia 7.7 7.6 7.4 8.2 9.7 10.8 10.7 11.6
Canada 9.1 10.4 10.4 10.1 13.0 11.1 11.3 12.7
Japan 5.8 6.9 7.9 8.7 12.8 12.3 12.8 13.6
Brazil 14.0 15.4 14.4 14.3 13.7 11.9 13.0 13.7
China 10.1 10.1 10.6 11.0 11.3
Hong Kong SAR 13.0 13.6 13.0 11.0 12.2 13.3 13.9 16.4
India 4.2 4.1 6.2 6.9 10.1 10.4 10.1 10.8
Korea 6.1 6.8 8.2 8.2 10.0 9.6 11.4 12.6
Mexico 13.5 12.8 15.1 13.3 14.9 13.8 13.9 13.2
Singapore 11.3 11.5 11.2 11.5 15.5 14.9 13.6 14.3
1

Bank capital positions: total regulatory capital ratio1
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 11.5 13.1 14.3 15.6 17.2
Belgium 60.3 80.0 79.4 71.7 18.8
France 11.6 11.5 10.8 10.6 12.5 14.0 15.3 17.4
Germany 13.5 15.2 17.3 17.9 18.8
Italy 10.8 11.4 10.5 10.7 12.4 13.8 14.7 14.2
Luxembourg 17.3 21.0 17.8 17.2 20.3 17.8 19.6 24.8
Netherlands 12.0 14.1 14.5 18.5 22.4
Spain 10.8 11.0 11.1 11.3 11.9 11.4 13.6 14.7
Other Europe 17.4 18.6 17.2 20.0
Sweden 10.4 10.1 10.0 13.4 15.4 16.7 22.5 26.9
Switzerland 18.2 19.6 20.8 20.7
United States 13.0 13.3 13.1 12.5 15.2 15.0 14.5 14.2
Other advanced 10.0 11.5 12.2 11.9 14.4 14.1 13.8 14.3
Australia 9.8 10.5 10.4 11.4 11.6 12.1 12.4 13.8
Canada 13.0 14.4 13.9 12.9 15.6 14.1 13.4 14.9
Japan 9.5 11.0 12.2 11.7 15.0 14.9 14.6 14.2
Brazil 16.6 18.1 17.8 17.7 16.9 16.5 16.7 17.2
China 12.2 12.3 13.6 13.5 13.3
Hong Kong SAR 15.7 15.4 14.9 14.7 15.8 15.7 16.8 19.2
India 6.2 6.5 8.2 9.8 14.5 14.2 13.0 13.3
Korea 11.4 12.2 12.8 11.9 13.1 12.9 14.0 14.9
Mexico 15.4 14.1 16.1 15.3 16.8 16.0 15.8 14.9
Singapore 16.9 16.2 15.4 14.7 18.6 18.1 15.9 16.5
1

Bank capital positions: average risk weight1
Risk-weighted assets as a per cent of total assets Annex Table 1.22
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 42 42 38 39 40
Belgium 34 32 34 35 36
France 64 46 36 33 35 30 32 33
Germany 39 37 30 34 34
Italy 62 76 67 66 57 52 51
Luxembourg 25 22 24 24 28 35 32
Netherlands 36 36 36 30 30
Spain 65 64 64 57 54 44 46 46
Other Europe 51 47 47 36 35 33 34 32
Sweden 54 47 49 35 31 26 21 21
Switzerland 30 30 32 34
United Kingdom 51 47 46 36 39 36 37 33
United States 73 72 74 74 69 67 70 72
Other advanced 56 53 52 45 41 39 39 38
Australia 69 58 57 43 43 41 40 43
Canada 44 40 39 37 34 34 37 34
Japan 57 55 55 48 43 41 39 39
Brazil 68 67 71 74 75 85 75 66
China 30 30 31 63 63
Hong Kong SAR
India 101 101 100 100 69 71 71 74
Korea 69 68 73 75 75 74 66 62
Mexico 43 52 58 70 65 68 67 63
Singapore 63 60 65 63 54 53 55 59
1

Bank capital positions: simple Tier 1 unweighted capital ratio1
Tier 1 regulatory capital as a per cent of assets Annex Table 1.23
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 3.7 4.7 4.8 5.3 5.8
Belgium 3.9 5.0 5.3 5.4 5.9
France 3.8 4.9 4.6 4.2 4.9
Germany 3.8 4.4 4.2 5.3 5.6
Italy 5.3 5.9 5.1 6.1 6.3 6.5 6.2
Luxembourg 3.6 3.7 3.5 3.4 5.0 5.4 5.9 6.3
Netherlands 3.7 4.5 4.6 4.6 5.3
Spain 7.0 6.5 5.6 4.6 5.2 4.3 5.5 5.9
Other Europe 4.5 4.8 4.7 5.2
Sweden 3.8 3.6 3.5 3.5 4.1 4.0 4.0 4.8
Switzerland 3.3 3.6 4.7 5.3 5.8 6.3
United States 7.6 7.8 7.9 7.2 8.7 8.7 9.1 9.3
Other advanced 3.4 3.8 4.2 3.7 4.7 4.5 4.5 4.8
Australia 5.3 4.4 4.2 3.6 4.2 4.4 4.3 4.9
Canada 4.0 4.2 4.0 3.7 4.4 3.8 4.1 4.3
Japan 3.1 3.6 4.3 3.8 4.9 4.8 4.8 4.9
Brazil 9.6 10.3 10.2 10.6 10.3 10.1 9.8 9.1
China 3.0 3.0 3.3 7.0 7.1
Hong Kong SAR
India 4.3 4.1 6.2 6.9 6.9 7.3 7.1 8.0
Korea 4.2 4.6 6.0 6.1 7.4 7.1 7.5 7.8
Mexico 5.8 6.6 8.8 9.3 9.6 9.4 9.4 8.4
Singapore 7.1 6.9 7.2 7.3 8.4 7.9 7.5 8.5
1

Banks’ foreign claims by home nationality
In billions of US dollars Annex Table 1.24
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 1
5,375 8,320 10,805 12,328 10,241 9,236 8,854 8,235
Belgium 548 781 1,111 957 385 261 212 190
France 1,042 1,700 2,619 3,659 3,138 2,663 2,728 2,487
Germany 2,182 3,204 3,542 3,604 2,998 2,683 2,453 2,087
Italy 308 398 422 1,026 897 843 756 701
Luxembourg 30 38 39 154 100 38
Netherlands 922 1,475 2,085 1,737 1,333 1,245 1,210 1,252
Spain 343 723 988 1,190 1,389 1,503 1,496 1,517
Other Europe 2,811 4,540 6,152 6,312 6,497 6,582 6,201 5,069
Sweden 174 515 597 665 743 915 863 688
Switzerland 1,282 1,909 2,458 2,003 1,750 1,604 1,740 1,399
United Kingdom 1,354 2,115 3,096 3,645 4,004 4,063 3,599 2,982
United States 742 1,013 1,334 1,463 2,868 3,358 3,092 2,990
Other advanced 1,469 2,127 2,826 3,435 4,188 5,089 5,386 6,056
Australia 262 333 415 588 729 757 712
Canada 327 383 628 710 898 1,139 1,227 1,403
Japan 1,142 1,482 1,865 2,310 2,702 3,221 3,402 3,941
Brazil 14 26 30 42 81 98 124 125
China
Hong Kong SAR
India 21 20 33 41 56 61 68 118
Korea 110 142 171
Mexico 3 5 8 6 5 6 10
Singapore 79 109 156 190 259 363 482 436
Memos:
All remaining advanced 555 862 1,456 1,723 1,481 1,157 1,003 841
All remaining EMEs 92 122 190 203 245 304 399 340
Global 2
11,191 17,187 23,072 25,868 26,045 26,569 26,035 24,624
1
Euro area total includes only countries shown in the table. Data for other euro area reporting countries is included in “All remaining
advanced.” 2 Global data includes restricted data, whereas the values reported by countries for other data items are based on only
unrestricted data.

Shares of banks’ foreign claims by home nationality
In per cent of global total Annex Table 1.25
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 1
48.0 48.4 46.8 47.7 39.3 34.8 34.0 33.4
Belgium 4.9 4.5 4.8 3.7 1.5 1.0 0.8 0.8
France 9.3 9.9 11.3 14.1 12.0 10.0 10.5 10.1
Germany 19.5 18.6 15.4 13.9 11.5 10.1 9.4 8.5
Italy 2.8 2.3 1.8 4.0 3.4 3.2 2.9 2.8
Luxembourg 0.3 0.2 0.2 0.6 0.4 0.1
Netherlands 8.2 8.6 9.0 6.7 5.1 4.7 4.6 5.1
Spain 3.1 4.2 4.3 4.6 5.3 5.7 5.7 6.2
Other Europe 25.1 26.4 26.7 24.4 24.9 24.8 23.8 20.6
Sweden 1.6 3.0 2.6 2.6 2.9 3.4 3.3 2.8
Switzerland 11.5 11.1 10.7 7.7 6.7 6.0 6.7 5.7
United Kingdom 12.1 12.3 13.4 14.1 15.4 15.3 13.8 12.1
United States 6.6 5.9 5.8 5.7 11.0 12.6 11.9 12.1
Other advanced 13.1 12.4 12.2 13.3 16.1 19.2 20.7 24.6
Australia 1.5 1.4 1.6 2.3 2.7 2.9 2.9
Canada 2.9 2.2 2.7 2.7 3.4 4.3 4.7 5.7
Japan 10.2 8.6 8.1 8.9 10.4 12.1 13.1 16.0
Brazil 0.1 0.2 0.1 0.2 0.3 0.4 0.5 0.5
China
Hong Kong SAR
India 0.2 0.1 0.1 0.2 0.2 0.2 0.3 0.5
Korea 0.4 0.5 0.7
Mexico 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Singapore 0.7 0.6 0.7 0.7 1.0 1.4 1.9 1.8
Memos:
All remaining advanced 5.0 5.0 6.3 6.7 5.7 4.4 3.9 3.4
All remaining EMEs 0.8 0.7 0.8 0.8 0.9 1.1 1.5 1.4
1
Euro area total includes only countries shown in the table. Data for other euro area reporting countries is included in “All remaining
advanced.”

Foreign-owned banks’ share of domestic banking system1
Share of total assets, in per cent Annex Table 1.26
2002 2004 2006 2008 2010 2012 2014 2016

Euro area 16.0 17.0 18.2 18.3 17.2 17.1
Belgium 43.1 27.6 28.4 30.1 65.8 64.3 64.8 63.7
France 3.9 3.3 3.3 5.9 5.5
Germany 1.6 1.3 1.7 2.2 2.6 3.9 3.0 4.3
Italy 7.7 17.2 17.9 15.3 14.7 13.8 14.1
Luxembourg 83.1 84.1 86.6 87.7 88.4 87.8 86.4
Netherlands 10.7 12.4 13.4 12.4 14.2 10.4 7.2 6.8
Spain 9.2 10.8 11.7 10.7 9.6 8.9 8.7 6.1
Other Europe
Sweden 6.0 7.0 9.0 9.0 8.0 8.0 9.0 10.0
Switzerland 5.9 5.4 6.2 7.7 8.5 13.3 10.5 9.4
United Kingdom 47.6 51.1 50.5 52.6 46.2 45.8 48.2 50.4
United States 13.3 12.5 15.1 16.7 16.7 19.8 20.3 18.4
Other advanced
Australia 17.0 19.7 21.3 20.0 16.0 13.8 13.3 13.5
Canada 7.0 9.0 10.0 11.0 9.0 7.0 7.0 6.0
Japan 5.1 5.4 6.8 4.0 3.2 2.9 4.0 3.8
Emerging markets
Brazil 27.4 23.4 24.0 24.2 21.2 19.8 20.1 17.3
China 1.6 1.9 2.2 2.0 1.9 1.7 1.4
Hong Kong SAR
India 7.6 7.1 7.5 8.9 7.6 7.5 7.4 6.9
Korea 7.6 9.1 8.3 8.1 7.3 6.4 5.4 4.6
Mexico 61.2 73.9 79.8 83.2 74.1 70.9 69.6 67.7
Singapore 74.1 74.0 74.1 69.5 68.6 65.4 63.4
1

Bank profitability: return on equity1
Net profit as a per cent of average equity Annex Table 1.27
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 6.3 5.1 10.7 3.6 3.6 0.2 3.2 4.4
Belgium 11.8 14.7 20.4 –11.6 4.0 1.9 6.8 9.2
France 5.5 5.9 3.6 4.7 6.3
Germany 3.7 0.1 7.1 0.1 –0.7 2.3 2.4 3.1
Italy 7.8 8.3 10.5 6.7 2.7 –3.3 –3.9 –1.7
Luxembourg 3.2 8.9 9.0 7.8 7.0
Netherlands –9.0 3.8 5.0 4.4 7.4
Spain 14.2 14.9 16.1 14.7 8.3 –12.4 6.3 5.8
Other Europe 16.3 3.9 5.8 4.2 4.5 3.9
Sweden 19.7 17.1 7.8 12.0 12.5 12.7
Switzerland 17.1 –3.3 7.7 4.7 5.0 4.0
United Kingdom 15.7 4.9 4.8 2.6 2.7 2.2
United States 13.7 14.3 12.5 3.4 2.2 8.2 9.2 9.3
Other advanced –10.6 4.1 14.4 5.4 8.0 9.5 10.5 9.2
Australia 14.7 15.3 17.2 15.6 11.7 13.8 14.3 12.1
Canada 11.3 16.2 17.6 15.0 12.1 14.9 14.9 13.6
Japan –15.1 1.0 13.3 0.5 6.2 7.0 8.1 7.3
Brazil 10.3 16.1 22.6 21.9 16.3 14.6 13.1 13.3
China 5.6 58.0 19.5 21.1 16.7 21.5 19.8 16.0
Hong Kong SAR
India 15.8 20.0 14.9 14.9 13.7 13.5 11.2 6.3
Korea 11.8 9.4 16.5 10.9 6.5 7.3 3.3 1.9
Mexico 7.5 14.5 22.2 16.9 13.1 13.3 14.1 12.6
Singapore 8.1 9.9 11.5 11.1 9.4 11.5 11.2 10.2
1
Weighted averages for groups of countries, based on average assets. Euro area calculated as a weighted average of individual countries
shown above. Weighted averages for groups of countries may not reflect individual country data as shown in the table because of the use
of some confidential series. For information on the consolidation basis and coverage of these data, see

Bank profitability: return on assets1
Net profit as a per cent of average assets Annex Table 1.28
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 0.3 0.2 0.5 0.2 0.2 0.0 0.1 0.2
Belgium 0.4 0.5 0.6 –0.5 0.2 0.1 0.4 0.6
France 0.2 0.3 0.2 0.3 0.4
Germany 0.1 0.0 0.2 0.0 0.0 0.1 0.1 0.1
Italy 0.8 0.5 0.2 –0.4 –0.5 –0.2
Luxembourg 0.4 0.4 0.6 0.2 0.4 0.3 0.5 0.5
Netherlands –0.3 0.2 0.2 0.2 0.4
Spain 0.8 0.8 1.0 0.9 0.5 –0.7 0.4 0.4
Other Europe 0.8 0.2 0.3 0.2 0.3 0.2
Sweden 0.7 0.8 0.7 0.3 0.5 0.6 0.6
Switzerland 0.8 –0.1 0.4 0.3 0.3 0.3
United Kingdom 0.8 0.3 0.2 0.1 0.2 0.1
United States 1.2 1.4 1.3 0.4 0.2 0.9 1.0 1.0
Other advanced –0.4 0.3 0.7 0.3 0.4 0.5 0.6 0.6
Australia 1.0 1.0 1.1 0.9 0.7 0.8 0.9 0.8
Canada 0.9 1.2 1.3 1.0 1.0 1.2 1.4 1.6
Japan –0.6 0.0 0.5 0.0 0.3 0.3 0.3 0.3
Brazil 1.0 1.7 2.4 2.4 1.8 1.6 1.4 1.3
China 0.2 0.6 0.6 1.1 0.9 1.3 1.3 1.0
Hong Kong SAR 1.2 1.3 1.4 1.2 1.0 1.0 1.2 1.2
India 0.8 1.2 1.0 1.1 1.1 1.1 0.9 0.5
Korea 0.6 0.5 1.2 0.8 0.5 0.6 0.3 0.1
Mexico 0.5 1.2 1.9 1.6 1.4 1.4 1.5 1.3
Singapore 0.8 1.0 1.2 1.1 1.0 1.1 1.0 1.0
1

Bank profitability: return on risk-weighted assets1
Net profit as a per cent of average risk-weighted assets Annex Table 1.29
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 0.5 0.4 0.0 0.4 0.7
Belgium –3.8 0.6 0.3 1.1 1.7
France 0.6 0.8 0.5 0.8 1.1
Germany –0.7 –0.1 0.3 0.3 0.4
Italy 0.8 1.0 1.2 0.9 0.4 –0.7 –0.9 –0.5
Luxembourg
Netherlands –1.0 0.4 0.6 0.6 1.4
Spain 1.2 1.2 1.4 1.4 0.9 –1.6 0.9 0.9
Other Europe 1.6 0.7 0.8 0.7 0.8 0.8
Sweden 1.5 1.6 1.7 1.0 1.9 2.5 3.1
Switzerland 1.8 0.9 1.0 0.8
United Kingdom 1.6 0.5 0.6 0.4 0.4 0.4
United States 1.6 1.9 1.7 0.5 0.4 1.4 1.5 1.4
Other advanced –0.7 0.4 1.2 0.6 0.9 1.2 1.3 1.2
Australia 1.3 1.6 1.7 1.5 1.5 1.8 1.9 1.7
Canada 1.2 1.8 2.1 1.8 1.9 2.4 2.4 2.2
Japan –1.1 0.1 0.9 0.0 0.6 0.8 0.9 0.8
Brazil 1.2 2.7 3.6 3.1 2.5 1.9 1.6 1.7
China 2.4 3.4 1.9 1.4
Hong Kong SAR
India 1.1 1.0 1.0 1.0 1.0 1.0 0.7
Korea 0.9 0.8 1.7 1.1 0.7 0.8 0.4 0.2
Mexico 1.4 2.8 3.6 2.7 2.1 2.1 2.3 2.0
Singapore 1.0 1.1 1.5 1.4 1.3 1.4 1.2 1.2
1

Bank profitability: net interest margin1
Net interest income as a per cent of average interest-earning assets Annex Table 1.30
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 1.1 1.0 0.9 1.2 1.2 1.2 1.2
Belgium 1.1 1.4 1.5 1.6 1.7
France 0.7 1.3 1.2 1.1 1.1
Germany 1.1 1.1 1.0 0.9 1.1 1.2 1.2 1.3
Italy
Luxembourg
Netherlands 0.9 1.1 1.1 1.2 1.3
Spain
Other Europe 2.5 2.1 2.1 2.0 2.0 2.1
Sweden 1.5 1.3 1.3 1.2 1.2 1.2 1.1
Switzerland
United Kingdom 2.7 2.2 2.2 2.2 2.1 2.3
United States 3.9 3.7 3.4 3.3 3.6 3.5 3.2 3.1
Other advanced 1.5 1.6 1.4 1.4 1.4 1.4 1.3 1.2
Australia 3.5 2.3 2.1 2.2 2.1 2.1 2.0
Canada
Japan 1.5 1.4 1.3 1.2 1.3 1.1 1.0 0.9
Brazil 8.6 9.4 10.4 8.7 8.3 7.3 6.1 5.9
China 2.7 2.7 2.4
Hong Kong SAR 2.1 1.8 1.7 1.9 1.4 1.3 1.4 1.3
India 3.6 3.6 3.7 3.3 3.1 3.5 3.3 3.2
Korea 7.4 6.3 6.2 6.8 5.5 5.8 4.3 3.3
Mexico 4.0 4.1 5.2 6.0 5.6 5.7 6.1 5.8
Singapore 3.6 3.1 4.4 4.5 2.9 2.7 2.5 2.6
1

Bank profitability: cost-to-income ratio1
Operating expenses as a per cent of revenue Annex Table 1.31
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 76.6 71.0 69.4 76.6 66.4 68.6 68.1 69.3
Belgium 72.9 70.7 68.9 61.0 58.5
France 72.6 65.0 68.0 68.9 68.6
Germany 84.4 78.2 81.2 81.3 78.3 78.5 80.7 82.7
Italy 63.2 67.3 69.4 70.0 64.5 72.4
Luxembourg 39.8 42.1 38.5 39.1 43.6 53.8 48.4 47.8
Netherlands 186.5 66.3 63.2 63.2 57.7
Spain 55.7 52.7 47.4 43.9 44.8 52.3 50.8 54.1
Other Europe 62.0 74.7 63.7 69.1 70.1 73.4
Sweden 57.0 53.6 52.9 53.5 53.0 49.6 46.8
Switzerland 64.0 99.2 70.1 73.8 70.6 74.4
United Kingdom 62.2 68.6 63.0 70.1 73.8 78.0
United States 57.2 57.5 58.0 65.3 60.3 62.3 62.4 59.7
Other advanced 55.5 54.1 54.6 60.2 57.0 55.7 56.3 58.4
Australia 59.9 58.0 53.4 54.9 53.1 49.2 47.6 47.6
Canada 68.1 67.2 65.6 66.4 61.1 59.3 59.0 59.7
Japan 53.8 51.9 52.9 60.3 57.2 56.8 58.8 61.6
Brazil 69.2 63.0 57.0 56.5 52.4 52.0 54.4 52.8
China 52.5 46.3 47.6 42.5 46.3 33.3 32.3 30.9
Hong Kong SAR
India 53.0 46.7 50.2 49.5 45.4 45.7 47.0 48.3
Korea 40.5 41.3 47.9 44.7 45.4 44.9 53.7 58.3
Mexico 80.1 67.2 57.7 66.6 67.6 69.3 68.4 68.3
Singapore 40.8 40.3 40.3 41.5 39.7 41.3 42.3 44.2
1

Bank profit components: revenue1
In per cent of average assets Annex Table 1.32
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 1.8 1.8 1.9 1.8 1.9 1.8 1.9 2.0
Belgium 1.4 1.6 1.4 2.0 2.2
France 1.9 2.2 2.1 2.1 2.1
Germany 1.4 1.5 1.4 1.2 1.4 1.2 1.3 1.4
Italy 2.9 2.4 2.0 2.0 2.1 2.1
Luxembourg 1.2 1.1 1.2 1.3 1.3 1.2 1.5 1.5
Netherlands 0.6 1.7 1.7 1.7 1.8
Spain 4.1 3.6 3.2 3.1 2.9 2.8 3.2 3.1
Other Europe 3.8 2.4 2.6 2.4 2.5 2.4
Sweden 2.2 2.1 1.8 1.7 1.6 1.6 1.6
Switzerland 2.8 1.8 2.4 2.2 2.5 2.5
United Kingdom 4.3 2.7 2.7 2.6 2.6 2.6
United States 5.7 5.6 5.2 4.6 5.0 4.8 4.5 4.4
Other advanced 2.2 2.3 2.4 2.2 2.2 2.1 2.1 2.0
Australia 3.7 4.0 3.8 3.3 3.1 2.7 2.7 2.4
Canada 6.5 5.9 5.4 4.6 4.9 4.8 5.2 5.8
Japan 1.6 1.6 1.6 1.4 1.4 1.3 1.2 1.0
Brazil 9.4 10.4 11.2 9.4 8.9 8.0 7.0 6.8
China 2.2 2.6 2.7 3.4 3.0 3.3 3.1 2.8
Hong Kong SAR 2.8 2.7 2.7 2.8 2.3 2.2 2.3 2.1
India 4.8 5.3 4.8 4.4 4.3 4.3 4.1 4.1
Korea 2.9 3.1 2.8 2.7 2.3 2.5 2.0 1.8
Mexico 4.2 5.1 6.2 5.3 5.1 5.4 5.4 5.2
Singapore 2.7 2.6 2.6 2.6 2.6 2.5 2.3 2.4
1

Bank profit components: net interest revenue1
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 1.2 1.2 1.1 1.0 1.2 1.1 1.1 1.2
Belgium 0.9 1.2 1.1 1.4 1.5
France 0.7 1.2 1.1 1.0 1.0
Germany 1.0 1.0 0.9 0.8 1.0 0.9 0.9 1.0
Italy 1.5 1.4 1.1 1.1 1.0 1.0
Luxembourg 0.6 0.6 0.6 0.7 0.7 0.6 0.6 0.6
Netherlands 1.0 1.1 1.1 1.3 1.3
Spain 2.4 2.0 1.6 1.7 1.8 1.7 1.8 1.9
Other Europe 1.5 1.2 1.1 1.1 1.1 1.2
Sweden 1.2 1.1 1.0 1.0 0.9 0.9 0.9
Switzerland 0.6 0.7 0.7 0.7 0.8 0.9
United Kingdom 1.7 1.4 1.2 1.3 1.3 1.3
United States 3.4 3.2 3.0 2.9 3.1 3.0 2.8 2.8
Other advanced 1.6 1.5 1.4 1.4 1.4 1.4 1.4 1.3
Australia 2.1 1.9 1.9 1.7 1.8 1.8 1.8 1.7
Canada 3.2 2.9 2.5 2.4 2.6 2.6 2.8 3.1
Japan 1.3 1.2 1.1 1.1 1.1 1.0 0.9 0.8
Brazil 7.4 8.0 8.8 7.4 6.9 6.3 5.2 5.0
China 2.0 2.3 2.3 2.9 2.4 2.6 2.4 2.2
Hong Kong SAR 1.9 1.7 1.6 1.7 1.2 1.2 1.2 1.1
India 2.9 3.1 3.2 2.9 2.7 3.1 2.9 2.8
Korea 6.0 5.4 5.3 5.7 4.6 4.5 3.7 2.8
Mexico 3.4 3.6 4.6 5.2 4.9 5.0 5.2 5.0
Singapore 3.2 2.6 3.5 3.6 2.3 2.2 2.1 2.2
1

Bank profit components: non-interest revenue1
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 0.6 0.6 0.8 0.7 0.6 0.6 0.7 0.7
Belgium 0.5 0.4 0.4 0.7 0.7
France 0.7 0.7 0.7 0.7 0.7
Germany 0.4 0.4 0.5 0.4 0.4 0.3 0.4 0.4
Italy 1.4 1.0 0.8 0.9 1.1 1.2
Luxembourg 0.6 0.5 0.7 0.6 0.6 0.5 0.9 0.9
Netherlands –0.4 0.5 0.5 0.4 0.5
Spain 1.7 1.6 1.6 1.5 1.1 1.1 1.4 1.1
Other Europe 2.3 1.2 1.5 1.3 1.3 1.3
Sweden 1.0 1.0 0.8 0.8 0.7 0.7 0.8
Switzerland 2.2 1.1 1.7 1.5 1.6 1.6
United Kingdom 2.5 1.4 1.5 1.3 1.4 1.2
United States 2.4 2.4 2.2 1.7 1.9 1.8 1.7 1.6
Other advanced 0.7 0.8 1.0 0.8 0.7 0.7 0.7 0.7
Australia 1.6 2.1 1.9 1.6 1.3 0.9 0.9 0.8
Canada 3.3 2.9 3.0 2.2 2.3 2.2 2.4 2.6
Japan 0.3 0.4 0.5 0.3 0.3 0.3 0.3 0.3
Brazil 2.0 2.4 2.4 2.0 1.9 1.7 1.8 1.8
China 0.2 0.3 0.3 0.5 0.6 0.7 0.7 0.6
Hong Kong SAR 0.9 1.0 1.1 1.1 1.1 1.0 1.0 0.9
India 1.8 2.2 1.6 1.5 1.6 1.2 1.2 1.3
Korea 3.9 3.0 2.8 3.5 2.6 2.4 1.9 1.2
Mexico 1.6 2.0 2.2 1.9 2.1 1.8 1.9 1.7
Singapore 0.9 1.0 1.1 0.9 1.0 1.0 0.9 0.9
1
Weighted average for groups of countries, based on average assets. Euro area calculated as a weighted average of individual countries

Bank profit components: operating expenses1
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 1.3 1.2 1.3 1.2 1.2 1.2 1.3 1.3
Belgium 1.0 1.1 1.0 1.2 1.3
France 1.4 1.4 1.4 1.4 1.5
Germany 1.2 1.1 1.1 1.0 1.1 0.9 1.1 1.1
Italy 1.8 1.6 1.4 1.4 1.4 1.5
Luxembourg 0.5 0.5 0.5 0.5 0.6 0.6 0.7 0.7
Netherlands 1.1 1.1 1.0 1.1 1.0
Spain 2.3 1.9 1.5 1.4 1.3 1.5 1.6 1.7
Other Europe 2.4 1.7 1.6 1.6 1.8 1.8
Sweden 1.3 1.1 1.0 0.9 0.8 0.8 0.8
Switzerland 1.7 1.6 1.7 1.6 1.7 1.9
United Kingdom 2.7 1.8 1.7 1.8 1.9 2.0
United States 3.3 3.2 3.0 3.0 3.0 3.0 2.8 2.6
Other advanced 1.3 1.3 1.4 1.3 1.2 1.2 1.2 1.1
Australia 2.2 2.3 2.0 1.8 1.6 1.3 1.3 1.2
Canada 4.4 3.9 3.6 3.0 3.0 2.8 3.1 3.4
Japan 0.9 0.8 0.8 0.9 0.8 0.7 0.7 0.6
Brazil 6.5 6.6 6.4 5.3 4.6 4.2 3.8 3.6
China 1.2 1.2 1.3 1.5 1.4 1.2 1.1 1.0
Hong Kong SAR 1.2 1.1 1.1 1.2 1.2 1.0 1.0 0.9
India 2.5 2.5 2.4 2.2 1.9 2.0 1.9 2.0
Korea 1.2 1.3 1.3 1.2 1.0 1.1 1.1 1.0
Mexico 3.4 3.4 3.6 3.5 3.5 3.7 3.7 3.5
Singapore 1.1 1.1 1.1 1.1 1.1 1.0 1.0 1.1
1

Bank profit components: loan impairment expense1
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area 0.4 0.3 0.3 0.4 0.4 0.5 0.4 0.3
Belgium 0.5 0.4 0.3 0.2 0.2
France 0.2 0.4 0.4 0.2 0.2
Germany 0.4 0.3 0.3 0.4 0.3 0.1 0.1 0.1
Italy 0.2 0.2 0.4 0.5 0.9 0.8
Luxembourg 0.2 0.1 0.0 0.5 0.2 0.1 0.1 0.1
Netherlands 0.4 0.3 0.3 0.3 0.1
Spain 0.6 0.4 0.4 0.6 0.9 1.7 1.0 0.8
Other Europe 0.3 0.4 0.6 0.4 0.3 0.2
Sweden 0.0 0.0 0.1 0.3 0.1 0.1 0.1
Switzerland 0.0 0.1 0.1 0.1 0.3 0.1
United Kingdom 0.4 0.6 0.8 0.5 0.3 0.2
United States 0.6 0.4 0.3 1.1 1.6 0.5 0.2 0.3
Other advanced 0.8 0.5 0.1 0.3 0.3 0.1 0.1 0.1
Australia 0.2 0.2 0.1 0.3 0.4 0.2 0.1 0.1
Canada 0.8 0.3 0.2 0.3 0.6 0.3 0.3 0.4
Japan 0.9 0.5 0.1 0.3 0.1 0.1 0.0 0.0
Brazil 1.8 1.6 2.0 2.1 2.4 2.2 1.9 2.2
China 0.5 0.5 0.4 0.5 0.3 0.3 0.3 0.3
Hong Kong SAR 0.4 0.1 0.0 0.1 0.1 0.0 0.0 0.1
India 2.7 2.4 1.6 0.9 0.9 1.1 1.4 1.5
Korea 1.0 1.2 0.3 0.4 0.7 0.6 0.5 0.5
Mexico 0.8 0.5 0.7 1.8 1.8 1.4 1.7 1.6
Singapore 0.4 0.2 0.1 0.3 0.4 0.2 0.1 0.2
1

Bank profit components: other expenses1
2001–02 2003–04 2005–06 2007–08 2009–10 2011–12 2013–14 2015–16

Euro area –0.2 0.1 –0.1 0.0 0.1 0.2 0.1 0.1
Belgium 0.3 –0.1 0.1 0.2 0.2
France 0.1 0.1 0.2 0.2 0.1
Germany –0.3 0.0 –0.2 –0.1 0.0 0.0 0.0 0.0
Italy 0.0 0.0 0.0 0.4 0.3 0.0
Luxembourg 0.2 0.2 0.1 0.1 0.1 0.1 0.1 0.2
Netherlands –0.5 0.1 0.1 0.2 0.2
Spain 0.4 0.5 0.3 0.2 0.2 0.3 0.2 0.2
Other Europe 0.3 0.1 0.1 0.2 0.2 0.2
Sweden 0.2 0.2 0.1 0.2 0.2 0.2 0.2
Switzerland 0.3 0.2 0.2 0.2 0.1 0.2
United Kingdom 0.4 0.1 0.0 0.1 0.2 0.2
United States 0.6 0.7 0.6 0.2 0.2 0.4 0.5 0.5
Other advanced 0.5 0.3 0.3 0.3 0.2 0.3 0.2 0.2
Australia 0.3 0.5 0.5 0.3 0.3 0.3 0.4 0.4
Canada 0.3 0.5 0.4 0.2 0.3 0.4 0.4 0.4
Japan 0.6 0.2 0.2 0.3 0.2 0.2 0.2 0.1
Brazil 0.2 0.5 0.5 –0.4 0.0 0.0 –0.1 –0.3
China 0.3 0.4 0.4 0.4 0.4 0.5 0.5 0.5
Hong Kong SAR
India –1.4 –0.8 –0.2 0.2 0.4 0.2 –0.1 0.1
Korea 0.2 0.1 –0.1 0.3 0.1 0.2 0.1 0.0
Mexico –0.5 –0.1 0.0 –1.7 –1.6 –1.1 –1.5 –1.2
Singapore 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.1
1

Annex 2: Individual bank data
The individual bank sample used in this report comprises 168 firms (Annex Table 2).
The sample includes consolidated institutions with total assets greater than $100
billion as of year-end 2016, as well as select other domestic systemically important
banks. Additions or exclusions of institutions from the sample were made in some
instances based on member discretion. Some graphs and tables contain only a subset
of these banks because of data availability. Geographic determination is based on the
ultimate parent’s jurisdiction. In a few instances, foreign subsidiaries are included in
certain jurisdictions with large foreign banking presence (eg Mexico). G-SIB status is
based on the FSB’s 2016 designation.112
Bank financial report data are sourced from SNL Financial, except for a few
funding indicators and the ratios underlying the business model classification and
some related graphs, which are sourced from Fitch Connect.
Sample of individual banks Annex Table 2
Region/Country #Banks #G-SIBSs Region/Country #Banks #G-SIBSs

Euro area 42 8 EMEs 65 4
Austria 2 0 Brazil 3 0
Belgium 3 0 China 17 4
Finland 1 0 Chinese Taipei 6 0
France 6 4 Hong Kong SAR 2 0
Germany 13 1 India 5 0
Ireland 2 0 Israel 2 0
Italy 6 1 Malaysia 2 0
Netherlands 3 1 Mexico 7 0
Spain 6 1 Qatar 1 0
Other Europe 18 7 Russia 2 0
Denmark 2 0 Saudi Arabia 1 0
Norway 1 0 South Africa 1 0
Sweden 4 1 South Korea 1 0
Switzerland 5 2 Turkey 2 0
United Kingdom 6 4 United Arab Emirates 2 0
United States 19 8 Venezuela 2 0
Other advanced 24 3
Australia 4 0
Canada 7 0
Japan 13 3
Total 168 30
112
See FSB (2016).

Annex 3: Business model classification methodology
The identification and characterisation of the bank business models is based on

Roengpitya et al (2017). This Annex provides a short description of the methodology
and the way it has been applied to this report.
Roengpitya et al (2017) use a three-stage approach to define distinct business
profiles across banks that is primarily data-driven but also incorporates judgment in
a systematic way. At the first stage, they select a set of bank balance sheet ratios that
hold the promise of differentiating business profiles and they produce a large number
of alternative classifications of banks into business models using cluster analysis (the
agglomerative Ward algorithm). In the other two stages, they sort through various
possible variants and the associated cluster allocations in order to narrow them down
using a goodness-of-fit metric and the judgmental criteria of parsimony and clear
discrimination. In the third stage, they combine cluster and discriminant analyses with
the criterion of stability of the results in order to further narrow down the variants to
a classification that includes four business models.
They use eight candidate input variables that are all ratios expressed in terms of
balance sheet size. They are evenly split between the asset and the liability sides of
the ledger: Loans-to-Assets, Trading Book-to-assets, Trading Assets-to-Total Assets,
Interbank Lending-to Assets, Interbank Borrowing-to-assets, Deposits-to-Assets,
Wholesale Funding-to-Assets, and the Stable Funding Ratio. Because the banks in our
panel are subject to different accounting standards that result in differences in the
reporting of derivative positions, these positions are netted out in calculating the
measure of total assets. They interpret these input variables as reflecting strategic
managerial choices, as opposed to other variables from the income statement of
banks, which they consider as not chosen by the banks’ management but rather as
representing outcomes of strategic choices, managerial skill and market conditions.
The final sorting of banks into business models is based on three of these ratios (total
loans, interbank loans and wholesale debt) but the characterisation is based on the
scores of banks on all eight ratios (Annex Table 3).
The adaptation of the results of Roengpitya et al (2017) to the data in this report
makes use of the discriminant analysis results produced in the third stage of these
authors’ paper. Discriminant analysis takes a clustering allocation, estimates a
mapping from input variables to that allocation, and summarises this mapping in the
probability that a particular observation (a bank) belongs to a given cluster. Using the
input ratios for the sample of banks in this report and the estimated discriminant
function of Roengpitya et al (2017), these banks can then be classified as belonging
to one or the other business model. This classification underpins the further analysis
on model characteristics and performance.

Descriptive statistics1
In per cent Annex Table 3
Retail-funded Wholesale-funded Trading Universal

Gross loans 67.9 73.2 17.2 48.9
[60.5; 75.6] [62.8; 83.3] [0.0; 33.0] [38.7; 56.6]
Trade 18.9 16.1 53.1 33.2
[10.6; 28.3] [6.9; 26.2] [23.1; 91.1] [17.0; 47.5]
Trading book 3.3 5.3 17.4 9.3
[0.1; 9.2] [0.5; 11.3] [0.7; 38.6] [0.3; 22.0]
Interbank lending 5.6 6.0 25.3 13.4
[0.9; 11.5] [1.2; 10.8] [3.3; 40.2] [4.5; 23.2]
Interbank borrowing 7.8 10.3 18.9 13.4
[0.5; 18.7] [2.4; 20.0] [0.0; 42.3] [2.7; 28.0]
Wholesale debt 10.0 36.1 21.0 14.6
[3.4; 18.2] [23.4; 51.3] [4.9; 38.8] [2.3; 34.0]
Stable funding 76.0 68.4 48.0 63.2
[62.9; 86.4] [57.2; 81.5] [15.4; 85.4] [42.2; 80.5]
Deposits 68.8 41.0 35.3 53.8
[50.2; 82.8] [27.0; 56.1] [0.0; 76.6] [26.1; 78.2]
No of bank-year pairs 556 344 171 532
1
Average values; 10th and 90th percentiles in brackets. Statistics associated with the input variables for a given trial, in bold.
Sources: Roengpitya et al (2017); Bankscope.

Annex 4: Cost of equity estimation methodology113
The cost of equity is estimated for 75 globally active banks using the CAPM (Capital
Asset Pricing Model).114 According to the CAPM, excess expected equity return (ie in
excess of the risk-free rate) linearly depends on excess expected market return: the
steepness of the relationship is captured by CAPM-beta, the systematic
(undiversifiable) risk of the equity.115 Formally:
Requity  Rrisk free   (Rmarket  Rrisk  free ) (1.1)
This implies that the expected equity return (ie the cost of equity) of a stock can
be decomposed into two main components: (1) the risk-free rate, and (2) the equity
risk premium, which is further a function of market excess return (market risk premium
or price of risk) and the equity beta (quantity of risk, ie the portion of risk in an equity
investment that cannot be diversified away by the marginal investor) as shown below:
Requity  Rrisk  free   equity *( Rmarket  Rrisk  free ) , (1.2)

market _ risk _ premium
where the beta coefficient measures the systematic co-movements between the
bank stock and the stock index.
Specifically, we estimate the cost of equity in two steps. First, cost of equity is
estimated for the market as a whole to infer the market risk premium. Our main
assumption here is that history is the best, albeit certainly imperfect, guide for the
future. Hence, first we calculate the market-level cost of equity for the United States
for each year t as the observed geometric average of historical annual returns from
1920 to t-1. We then calculate expected return for other markets (which tend to have
shorter time series available than the United States does) with the help of US
estimates and differences in country riskiness (following ideas explained, for instance,
in Damodoran (2016)). More concretely, we add to the expected US returns the
difference between the CDS spread of the country in question and that of the United
States to obtain the expected return for the country index.
Second, we calculate institution-level betas for all banks in our sample. The betas
are estimated by obtaining the linear correlation coefficient between observed excess
(ie over risk-free) bank stock returns and excess market index returns, as indicated by
equation 1.1.116 In our analysis, we use daily return over a year-long (more precisely,
250 trading days) window to obtain daily beta estimates.117, 118 Using the above
methodology we calculate cost of equity for the 2001–early-2017 period. For
graphical representation, we group the estimates by major geographic regions.
113
This annex was prepared by Bilyana Bogdanova, Ingo Fender and Elöd Takáts (BIS)
114
The sample of banks comprises the stock exchange-listed banks from Brei and Gambacorta (2014).
115
For a discussion of the CAPM, see Damodoran (2016).
116
Banks and markets are matched. For example, US bank returns are measured against the S&P 500.
117
While our choice of using daily returns over an annual horizon allows for the straightforward
computation of daily betas, it could potentially introduce additional noise through the closing bid-
ask spread. Hence, to ensure robustness, we also estimate betas using (i) daily returns over a two-
year window and (ii) weekly returns over two- and five-year windows.
118
As a final step, in order to obtain daily expected return estimates, we interpolate the yearly country
specific market risk premium estimates across trading days.

Members of the Working Group
Co-Chairs Claudia Buch

B Gerard Dages
National Bank of Belgium Janet Mitchell
Central Bank of Brazil Frederico Torres
Bank of Canada Sheryl King
European Central Bank Sándor Gardó
Bank of France Sébastien Frappa
Deutsche Bundesbank Benedikt Fritz

Ulrich Krüger
Hong Kong Monetary Authority Eric Wong

Kelvin Ho
Reserve Bank of India Saurav Sinha
Bank of Italy Vincenzo Cuciniello

Arianna Miglietta
Bank of Japan Yoshiyuki Fukuda

Wataru Hirata
Bank of Korea Cheolwon Park
Central Bank of Luxembourg Francisco Nadal de Simone
Bank of Mexico Fabrizio Lopez-Gallo

Pascual O'Dogherty
Netherlands Bank Herwin Loman
Monetary Authority of Singapore Evelyn Chen
Bank of Spain Soledad Nuñez
Sveriges Riksbank Reimo Juks
Swiss National Bank Pascal Towbin
Bank of England Jeremy Leake
Federal Reserve Bank of New York Nicola Cetorelli

Paul Licari
Afshin Taber
Board of Governors of the Federal Nicholas Coleman

Reserve System
Bank for International Settlements Kostas Tsatsaronis
Grant Turner (Secretary)
Anamaria Illes (Research assistance)

Structural Changes in Banking After Crisis

Uploaded by

Copyright:

Structural Changes in Banking After Crisis

Uploaded by

Document Information

Copyright

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Structural Changes in Banking After Crisis

Uploaded by

Copyright:

Committee on the Global

JEL Classification: G21, G24, G32, L25

ISBN 978-92-9259-131-1 (online)

Chair, Committee on the Global Financial System

CGFS - Structural changes in banking after the crisis iii

Executive summary ................................................................................................................................. 1

2. An overview of banks’ operating environment ................................................................... 6

2.1 Background on banks’ experience in the financial crisis ....................................... 6

2.2 The post-crisis operating environment for banks .................................................... 7

3. Key trends in banking structure and performance .......................................................... 11

3.1 Banking system size ...........................................................................................................12

3.2 Banking system concentration....................................................................................... 13

3.3 Bank business models ....................................................................................................... 14

3.4 International banking ........................................................................................................24

3.5 Bank performance............................................................................................................... 26

3.6 Summary of key trends in banking .............................................................................. 29

4. Implications for the stability of the banking sector ......................................................... 30

4.1 Bank risk buffers and risk-taking .................................................................................. 30

4.2 Bank risk governance and management ................................................................... 33

4.3 Market-based indicators of banks................................................................................ 35

4.4 Future profitability and resilience of banks .............................................................. 39

4.5 Banking system-level risk and resilience ................................................................... 41

4.6 Overall assessment ............................................................................................................. 50

5. Implications for the efficiency of the banking sector ...................................................... 51

5.1 Bank credit provision to the real economy ............................................................... 51

5.2 Banks’ role in facilitating capital market activity .................................................... 58

5.3 Overall assessment ............................................................................................................. 62

6. Key messages .................................................................................................................................. 64

Annex 1: Country banking data tables ..........................................................................................77

Annex 2: Individual bank data ....................................................................................................... 115

Annex 3: Business model classification methodology ......................................................... 116

Annex 4: Cost of equity estimation methodology ................................................................ 118

Members of the Working Group .................................................................................................. 119

CGFS - Structural changes in banking after the crisis v

CGFS - Structural changes in banking after the crisis 1

2 CGFS - Structural changes in banking after the crisis

CGFS - Structural changes in banking after the crisis 3

4 CGFS - Structural changes in banking after the crisis

CGFS - Structural changes in banking after the crisis 5

2.1 Background on banks’ experience in the financial crisis

6 CGFS - Structural changes in banking after the crisis

Bank share prices and CDS spreads Graph 1

Bank share prices Credit default swap premia1

2.2 The post-crisis operating environment for banks

The macro-financial environment. The post-crisis macroeconomic environment has

CGFS - Structural changes in banking after the crisis 7

Economic growth, credit and interest rates

GDP Total credit-to-GDP ratio1 Two-year government bond yields

Regulation and supervision. Reforms to the global regulatory framework seek

8 CGFS - Structural changes in banking after the crisis

CGFS - Structural changes in banking after the crisis 9

10 CGFS - Structural changes in banking after the crisis

Non-bank finance and financial institutions Graph 3

3. Key trends in banking structure and performance

CGFS - Structural changes in banking after the crisis 11

3.1 Banking system size

Banking system assets to GDP1

Europe Other advanced EMEs (rhs axis)

12 CGFS - Structural changes in banking after the crisis

3.2 Banking system concentration

CGFS - Structural changes in banking after the crisis 13

3.3 Bank business models

14 CGFS - Structural changes in banking after the crisis

Consolidation in the Spanish banking system