fsa trading

8/9/2019 FSA Trading

1/126

D i s

c u s s

i o n

P

a p e r

Financial Services Authority

The prudential regimefor trading activitiesA fundamental review

August 2010

10/ 4


2/126


3/126

List o acronyms used in this paper 3

1 Overview 52 Interactions with other proposals 9

3 The current ramework or trading activities 15

4 The evolution o traded markets 27

5 Lessons rom the crisis 38

6 Valuation 52

7 Coverage, coherence and the capital ramework 59

8 Risk management and modelling 77

9 A new ramework in practice 88

Annex 1: The current ramework in practice

Annex 2: A new ramework Paradigm 1: Updated current approach

Annex 3: A new ramework Paradigm 2: Uni orm capital standards

Annex 4: A new ramework Paradigm 3: Valuation-based approach

Annex 5: A new ramework Paradigm 4: Full air valuemodelling approach

Annex 6: List o key recommendations

Annex 7: List o questions

Contents

The Financial Services Authority 2010


4/126

The Financial Services Authority (FSA) invites comments on this Discussion Paper(DP). Comments should be submitted by 26 November 2010 . This DP contains anumber o questions or respondents, which can be submitted using an electronicresponse orm. The FSA would pre er you to use this electronic orm when sending

your responses. Comments should be sent by electronic submission using the ormon the FSAs website at ( www. sa.gov.uk/Pages/Library/Policy/DP/2010/dp10_04_response.shtml ).

Alternatively, please send comments in writing to:

Derek NesbittFinancial Services Authority25 The North ColonnadeCanary Whar London E14 5HS

Telephone: 020 7066 9332Fax: 020 7066 9333E-mail: dp10_04@ sa.gov.uk

It is the FSAs policy to make all responses to ormal consultation available or publicinspection unless the respondent requests otherwise. A standard con identialitystatement in an e-mail message will not be regarded as a request or non-disclosure.

A con idential response may be requested rom us under the Freedom o In ormation

Act 2000. We may consult you i we receive such a request. Any decision we makenot to disclose the response is reviewable by the In ormation Commissioner and theIn ormation Tribunal.

Copies o this Discussion Paper are available to download rom ourwebsite www. sa.gov.uk. Alternatively, paper copies can be obtained bycalling the FSA order line: 0845 608 2372.

Simon DixonFinancial Services Authority25 The North ColonnadeCanary Whar London E14 5HS

Telephone: 020 7066 1698Fax: 020 7066 1699E-mail: dp10_04@ sa.gov.uk

or
http://www.fsa.gov.uk/Pages/Library/Policy/DP/2010/dp10_04_response.shtmlhttp://www.fsa.gov.uk/Pages/Library/Policy/DP/2010/dp10_04_response.shtmlhttp://www.fsa.gov.uk/Pages/Library/Policy/DP/2010/dp10_04_response.shtmlhttp://www.fsa.gov.uk/Pages/Library/Policy/DP/2010/dp10_04_response.shtml


5/126

A list of acronyms used in this paper is set out below:

ABS Asset Backed Security

AFS Available or Sale

BCBS Basel Committee on Banking Supervision

BIPRU Prudential sourcebook or Banks, Building Societies and Investment Firms

BIS Bank or International Settlements

CAD Capital Adequacy Directive

CCP Central Counterparty

CCR Counterparty Credit Risk

CDS Credit De ault Swap

CEBS Committee o European Banking Supervisors

CDO Collateralised Debt Obligation

CDPC Credit Derivative Product Company

CGFS Committee on the Global Financial System

CMS Constant Maturity Swaps

CP Consultation Paper

CPPI Constant Proportion Port olio InsuranceCRM Comprehensive Risk Measure

CVA Credit Valuation Adjustment

DP Discussion Paper

EAD Exposure at De ault

EU European Union

FASB Financial Accounting Standards Board

FRO Financial Risk Outlook

FSA Financial Services AuthorityFSB Financial Stability Board

List of acronyms used inthis paper

Financial Services Authority 3


6/126

4 DP10/4: The prudential regime or trading activities (August 2010)

FSF Financial Stability Forum

FVO Fair Value Option

FX Foreign Exchange

GAAP Generally Accepted Accounting Principles

HFT High Frequency TradingH T Held or Trading

IAS International Accounting Standard

IASB International Accounting Standards Board

IDRC Incremental De ault Risk Charge

IFRS International Financial Reporting Standard

IIF Institute o International Finance

IMF International Monetary Fund

IRC Incremental Risk ChargeIPV Independent Price Veri ication

IRB Internal Ratings Based

IRRBB Interest Rate Risk in the Banking Book

IT In ormation Technology

LCFI Large Complex Financial Institution

LGD Loss Given De ault

MiFID Markets in Financial Instruments Directive

MTF Multilateral Trading Facility

MtM Mark to Market

ONS O ice or National Statistics

OTC Over the Counter

P&L Pro it and Loss

PD Probability o De ault

PRDC Power Reverse Dual Currency

RMBS Residential Mortgage Backed Security

RNIV Risks Not in VaR

SEC Securities and Exchange Commission

SFT Securities Financing Transaction

SIV Structured Investment Vehicle

UK United Kingdom

US United States

VaR Value at Risk


7/126


Overview1

Introduction and purpose1.1 In July 2009, the Basel Committee on Banking Supervision (BCBS) agreed a range

o amendments to the Basel II market risk ramework, 1 targeting speci c weaknesseshighlighted by the nancial crisis. On average, these changes will increase thecapital held against trading activities in large banks to more than three timescurrent levels. Trading activities have grown enormously in recent years, and the

nancial crisis was in part triggered by losses crystallised in the trading books o large banks. It is there ore necessary to build on the changes already in progress

with a re-appraisal o the prudential approach to trading activities, dealing with thearbitrages and mis-speci cations o risk that continue to exist and complementingthe many other areas o nancial re orm currently under consideration.

1.2 We expressed this view in The Turner Review where we called or a FundamentalReview o the prudential regime or trading activities. This Discussion Paper (DP)

ollows up The Turner Review with a detailed discussion o the issues that we thinkshould orm part o the Fundamental Review which is now being developedinternationally by the BCBS.

1.3 The recently ormed Independent Commission on Banking will be addressing

questions on whether the trading activities o banks should be separated rom otheractivities. We do not comment on those questions here and the recommendationsmade in this DP will be relevant whether or not trading activities are undertakendirectly by banking entities or by other rms within regulated groups.

1.4 Many trading activities play an important role in the e ective intermediation o riskbetween real economy participants. We could increase the sa ety o the system byimposing prohibitively high standards on these activities within regulated rms, butthis would either signi cantly increase costs to the economy, or drive these activitiesoutside the regulated sector. However, where trading activities do not aid e ciency

in the economy, but simply pass risks around the regulated sector or allow economic1 Revisions to the Basel II market risk ramework nal version www.bis.org/publ/bcbs158.htm
http://www.bis.org/publ/bcbs158.htmhttp://www.bis.org/publ/bcbs158.htm


8/126


rent extraction by investment banking operations, then imposing deliberatelyconservative standards may still make the system sa er without imposing the samecosts on the wider economy.

1.5 The purpose o this DP is to stimulate debate, and the eedback we receive will eedin to the international discussions at the BCBS through the FSAs involvement inthat orum.

Key recommendations1.6 We have drawn on a number o data sources including our own internal research and

data rom the crisis period; externally produced reports; and our discussions with rms,industry groups and other regulators. Some o these ideas were discussed at aroundtable which we hosted in March 2010 with representatives rom academia,regulation, and the nancial industry. We recommend a range o actions across threekey areas (a complete list o recommendations made in the DP is contained in Annex 6):

I. Valuation

We recommend an increased ocus on valuing traded positions as an inputinto capital resources. The range o values reported between irms or similarpositions has o ten been greater than the levels o capital held against marketrisks, meaning that greater e ort should be spent ensuring that valuationpractices are robust and consistent. We also see the need or a speci icassessment o valuation uncertainty, reducing the ability o irms to add

leverage on the basis o uncertain values that can lead to pro-cyclicality.

II. Coverage, coherence and the capital framework

We recommend changing the structure o the capital ramework to bringgreater coherence and reduce opportunities or structural arbitrage within thebanking sector and the wider inancial system. During the crisis it became clearthat a number o risks that irms were exposed to were not captured in thecapital ramework, and the assumption o resiliently liquid traded markets wasseverely tested. We recommend changes to the capital ramework to address

both o these issues.

III. Risk management and modelling

Finally, we recommend speci ic measures aimed at improving irms riskmanagement and modelling standards, and ensuring that these are alignedwith regulatory objectives. The crisis revealed serious shortcomings inpractices across irms. In some cases, irms simply did not manage traded riskse ectively, while in others the shortcomings arose because the measures usedby irms were designed to ocus on narrow, irm speci ic events without due

consideration o system-wide events that regulators care about.


9/126


1.7 The recommendations we make are ocused on structural de ciencies o thecurrent regime. At this stage it is not clear whether the capital requirements needto change, or whether we will ocus on re-allocating capital requirements to bettertarget the riskiest elements o trading activities. In either case, the outcome o the

undamental review will be subject to consultation and impact assessments to

determine whether any recalibration o the regime might be necessary.1.8 In making these recommendations, we have been mind ul that regulatory capital

requirements should balance the desire to be risk sensitive, which o ten leads tocomplexity, with the need or simplicity and transparency. During the crisis, investorslost con dence that rms had a su cient quantity and quality o capital to survivesevere economic stress. This was not helped by the complexity and opacity o regulatory capital measures or the trading book.

1.9 Also underlying these recommendations is a reappraisal o the reliance onmarket-implied measures o risk. Be ore the current crisis, many markets were

trading at levels that suggested they had entered a new paradigm where risk wasstructurally lower. This proved illusory as it became clear that the market had beensystematically under-pricing risk. Trading book capital requirements currentlyplace a heavy reliance on market-implied measures o risk, and we consider waysto prevent banks being undercapitalised due to market ine ciencies.

1.10 Finally, there is an inherent confict between delivering internationally agreedregulatory standards and the need to respond quickly to market developments thatlead to erosion and/or arbitrage o capital standards over time. Firms have animportant part to play in achieving this. We require, in Principle 11 o our

Principles or Business handbook, that rms disclose to us anything relating to theiractivities or which we would reasonably expect notice. We expect rms to be readyto engage in an early and active dialogue with us relating to new productdevelopment, and innovation in product structures. Only in this way can we ensurethat our regulations remain robust whilst adapting to accommodate genuinelybene cial innovation.

The structure of this DP1.11 This DP is organised into nine chapters, including this overview:

Chapter 2 puts this DP in context. There are currently a wide range o majorinternational re orms to the regulatory ramework underway, and we considerhow this review ts in with other ongoing work-streams.

Chapter 3 describes the current UK prudential regime or trading activitiesas it applies to banks, building societies and investment rms (BIPRU rms). 2 This allows us to illustrate speci c structural de ciencies in the make-up o thecurrent ramework that may have contributed to the recent nancial crisis.

Chapter 4 describes the evolution o traded markets over the past 20 years, andpresents evidence o structural di erences between credit and other traded markets.

2 BIPRU: Prudential sourcebook or Banks, Building Societies and Investment Firms.


10/126


Chapter 5 provides a detailed analysis o what the evidence rom the crisis tellsus about designing a new ramework. This chapter gives a detailed accounto the results o our loss attribution exercise undertaken in late 2009, whichanalysed the sources o $240bn o investment banking losses across 10 rmsduring a large part o the crisis. Its ndings highlight losses both in complex

structured nance products and in much simpler areas, such as governmentbonds, demonstrating widespread risk management ailures in the lead-up toand during the crisis. The size o these losses highlights the importance o therecommendations made in the DP.

Chapter 6 begins the discussion o what changes need to be implemented. Thischapter ocuses on the importance o valuation as an input into the ramework.

Chapter 7 considers measures to address issues with the coverage and coherenceo the current regulatory capital requirements or trading activities.

Chapter 8 discusses how the regulatory regime can better address problems thatarose through rms risk management and modelling practices.

Chapter 9 draws together the recommendations rom chapters 6 to 8 andconsiders how these might be delivered in practice, highlighting key questionsthat will in orm the nal policy choices.

Who should read this DP1.12 This DP ocuses on the prudential requirements or banks and investment rms that

engage in trading activities. However, many elements could be applied more broadlyand will be o general interest in the nancial services industry, including policymakers and supervisors in other countries. There are implications or the globalregulatory ramework and global banking system, which will have clear implications

or consumers.

Next steps1.13 The discussion period ends on 26 November 2010. During this period we plan a

series o meetings to discuss the contents o this document with industry groups andother regulators to ensure a wide range o views are received and are there ore ableto eed into our discussions at the BCBS.

1.14 We expect to issue a Feedback Statement in the rst hal o 2011.


11/126


Summary2.1 The G20 has made several commitments on wide-ranging re orms o the banking

and investment banking sectors. These re orms can broadly be split into ve areas:

1. resolvability and bank structural re orm;

2. capital and liquidity re orm;

3. market in rastructure re orm;

4. development o a macro-prudential ramework; and

5. remuneration.

2.2 This chapter describes how these re orms will a ect trading activities and highlightsareas o interaction with the recommendations in this DP.

2.3 The pace o innovation in traded markets has led to the related prudential regimetaking the orm o discrete new rules added to the existing ramework. The re ormsdescribed in this chapter are part o an overall package to strengthen the nancialsystem, which could be undermined i trading activities continue to be subject to apatchwork o rules that leave opportunities or risks to grow unchecked. A coherent,simpler and more transparent regime is vital to aid longer term con dence in theresilience o rms trading operations, and we there ore see this undamental reviewas complementary to these re orms.

Resolvability and bank structural reform2.4 In April 2009, the G20s London declaration 3 rea rmed the importance o urther

work and international coordination on bank resolution arrangements. Shortly a ter,the Financial Stability Forum (FSF) published its principles or cross-border cooperationon crisis management, 4 o which Principle 8 encouraged rms to maintain contingencyplans and procedures or use in a wind-down situation. The EU Commission is now

3 G20 Declaration on strengthening the fnancial system , 2 April 20094 FSF Principles or Cross-border Cooperation on Crisis Management, 2 April 2009

Interactions withother proposals2


12/126


considering measures regarding early intervention and resolution. The UK hasimplemented a special resolution regime or banks (via the Banking Act 2009) andother EU Member States ( or example, Germany) are considering whether similarmeasures are required. Additionally, a wider debate has continued on the structure o banks, notably with the Dodd-Frank Act in the US (see Box 2.1), and the Independent

Commission on Banking set up by the UK government.2.5 The easibility o recovery and resolution will be signi cantly a ected by the nature

o any trading activities undertaken within a rm. The Financial Services Act (2010)requires us to issue rules regarding recovery and resolution plans, and we plan topublish a Consultation Paper in December 2010 on this subject. Any newrequirements could impose signi cant costs on rms who trade a range o complexor bespoke positions, or those who use complex booking practices to trans er riskacross group structures. Nevertheless, this does not prevent the need or this review,as banks and investment rms will not be prevented rom participating in various

orms o trading activities.

2.6 However, in the area o banking structural re orm, there is the potential tosigni cantly alter the scope o trading activities within regulated institutions. Even inthis area, we believe it is highly likely that trading activities will remain within thesector subject to prudential regulation, and that those activities still need to be backedby a strengthened capital ramework. There ore a strengthened prudential regime ortrading activities remains a key component in a sounder regulatory system.

Capital and liquidity reform2.7 In December 2009 the BCBS released two consultation documents that proposed

major re orms to the current Basel II ramework. 5 On 26 July 2010 the BCBSoversight body, made up o Central Bank Governors and Heads o Supervision,released a statement outlining the design o the nal package that will be agreedin September.

Raising the quality, consistency and transparency of the capital base

2.8 The package will make several changes to the composition o regulatory capital.

Two changes will speci cally a ect trading activities. First, Tier 3 capital is beingabolished, bringing into line the quality o capital that is required to be held againstmarket risks with other risks. Second, unrealised gains and losses on some available-

or-sale assets will no longer be ltered out o Tier 1 regulatory capital.

2.9 The BCBS has also committed to review the treatment o unrealised gains. Such areview would be similar in nature to some o the recommendations in Chapter 6.We make recommendations to deal with all areas o valuation uncertainty when

air values eed directly into capital resources.

5 Strengthening the resilience o the banking sector - consultative document www.bis.org/publ/bcbs164.htm , and anInternational ramework or liquidity risk measurement, standards and monitoring - consultative document www.bis.org/publ/bcbs165.htm
http://www.bis.org/publ/bcbs164.htmhttp://www.bis.org/publ/bcbs165.htmhttp://www.bis.org/publ/bcbs165.htmhttp://www.bis.org/publ/bcbs164.htm


13/126


Strengthening the risk coverage of the capital framework

2.10 The package introduces a new capital charge to better capture mark-to-marketvolatility associated with Credit Valuation Adjustments (CVA), which will directlya ect trading activities. This was highlighted by the BCBS as a particular area o weakness during the crisis, which is corroborated by our loss attribution analysisin Chapter 5.

2.11 However, Chapter 5 also highlights the variation o methodologies used to calculateCVA, and we see delivering greater consistency o valuation in this area as equallyimportant as setting appropriate capital requirements. In Chapter 7, we also discussenhancing the coherence o regulatory capital and consider whether the market riskthat crystallises through CVA should be integrated with other market risks.

Asset-based leverage ratio

2.12 The leverage ratio included in the package will cover both banking and tradingactivities. The choices made in de ning total assets, especially the treatment o derivatives, are likely to have the biggest impact on trading activities. The designand calibration o the leverage ratio is intended to be tested during a parallel runperiod starting on 1 January 2013 and ending on 1 January 2017. The timing o thisprocess means we do not examine it closely in this DP.

Reducing the pro-cyclicality of the framework

2.13 The re orms package also discussed options or reducing the pro-cyclicality o theregime, ranging rom top-down measures, such as capital conservation bu ers, tobottom-up measures (e.g. orward-looking provisioning). The discussions touchedupon both capital resources and capital requirements.

2.14 A number o these measures have progressed since the original package. In particular,the BCBS has been working with the International Accounting Standards Board(IASB) on approaches to operationalise the accounting requirement or expected lossprovisioning as a way o introducing a orward-looking provisioning ramework. In

July 2010, BCBS released a consultation paper on how the proposed countercyclicalcapital bu er could work in practice. 6

2.15 The measures outlined in BCBSs December package do not consider, in any detail,ways to deal with pro-cyclicality in the trading book regime, in either capitalresources or capital requirements. 7 We there ore believe it is important that thisFundamental Review adequately deals with pro-cyclicality arising rom the regime

or trading activities. This is addressed through a number o the proposals inChapters 6 to 8.

6 Countercyclical capital bu er proposal consultative document, BCBS July 20107 The introduction o the Stressed VaR requirement in July 2009 should provide some reduction in the relative

cyclicality o the market risk capital requirements


14/126


Review of the liquidity regime

2.16 Finally, the re orms package also included new liquidity standards. These primarilydeal with unding liquidity issues, ensuring that appropriate liquid assets are held towithstand liquidity stresses. The package does not consider the implications o market liquidity in the design o a capital ramework, which is particularly relevant

or trading activities. This is covered in more detail in Chapter 7.

Box 2.1 The Dodd-Frank Act

The Dodd-Frank Wall Street Re orm and Consumer Protection Act (the act) waspassed by the US House o Representatives on 30 June 2010. Below are the parts o the act o particular relevance to trading activities.

Capital requirementsThe act requires regulators to review and establish capital levels on a consolidatedbasis or depository institutions and holding companies. Although the act doesnot provide speci ic guidance on these new capital levels, it does provide that thecapital levels currently in e ect will serve as a loor to any new capital requirements.There ore, banks and bank holding companies should expect higher requiredcapital levels in the uture. In addition, the act requires regulations so that capitallevels should increase during times o economic expansion and decrease in times o economic contraction, subject to sa ety and soundness considerations.

Asset-backed securities

The Securities and Exchange Commission (SEC) and ederal banking agencies arerequired to establish regulations to require any issuer o an Asset Backed Security(ABS) to retain ive percent (5%) o the credit risk or any asset, including certainresidential mortgage assets that the issuer trans ers, sells or conveys to a third party.

Volcker rule

The act amends the Bank Holding Company Act o 1956 to limit banking entities

ability to engage in proprietary trading and to own interests in hedge unds orprivate equity unds. A banking entity must limit its ownership interest in a hedge

und or private equity und to 3% o the total ownership interests o the und. Thetotal aggregate o all o the banking entitys interests in such hedge unds or privateequity unds may not exceed 3% o the Tier 1 capital o the banking entity.

Regulation of derivatives/swaps

The act establishes a regulatory ramework or the derivatives market and restrictsederally insured depository institutions rom participating in some o the riskiest

derivative and swap transactions. In particular, a ederally-insured depositoryinstitution must establish a separately capitalised a iliate to engage in higher-risk


15/126


swap transactions, such as uncleared credit de ault swaps. A bank must also limit itsown swap or derivatives activity to hedging, interest rate swaps or oreign exchangeswaps, or similar risk-mitigating activities directly related to a traditional banksactivities, which generally includes assets that are permissible or investment by anational bank.

While a number o these new regulations will restrict the level o trading activitieswithin certain US irms, it is clear that some trading activity will remain and animproved set o capital requirements or those activities is required.

Market infrastructure reform2.17 A key element o ongoing re orm is to review the regulatory and legislative

environment in which clearing and settlement operations take place. Speci cally, theG20 Pittsburgh statement requires greater use o Central Counterparty (CCP) clearing

or over the counter (OTC) derivatives and improvements to transparency in thesemarkets. In June, the EU Commission published a consultation paper containingproposals to strengthen the legislative ramework or CCPs and the central clearing o OTC derivatives.

2.18 As we noted in the 2010 Financial Risk Outlook :

A CCP can bring consistent and robust risk management practices, as wellas acting as circuit breaker to the de ault o a member. In addition, greateruse o CCP clearing can aid market liquidity and e iciency, be a motivating

orce behind contract standardisation, and reduce systemic risk.

2.19 Market and prudential regulators aims should be aligned in most areas, asstrengthening the system will only be achieved where participants in derivativesmarkets are backed by su cient nancial resources. Consequently, we see this DP ascomplementing these re orms, with particular synergies or encouraging increasedcontract standardisation and avoiding complexity. An example o this can be seen inthe proposed valuation uncertainty charge outlined in Chapter 6.

Development of a Macro-prudential framework2.20 The G20s declaration on strengthening the nancial system in April 2009 8 announced

that members had agreed to amend regulatory systems to ensure macro-prudentialrisks could be identi ed so that systemic risk in the nancial sector could be betteraddressed. The declaration called on the Financial Stability Board (FSB) and BCBS todevelop macro-prudential tools. The FSB has said that it will develop a policy

ramework by the end o October 2010 to address the moral hazard risks associatedwith systemically important nancial institutions.

8 G20 Declaration on strengthening the fnancial system , 2 April 2009


16/126


2.21 New macro-prudential policies are being considered to address two principal sourceso systemic risk:

the tendency or the banking system to become overly exposed to risk in theupswing o a credit cycle and overly risk averse in a downswing; and

the tendency or individual rms to take insu cient account o the spill-overe ects o their actions on risk in the rest o the system.

2.22 In the UK, the Bank o England released a discussion paper in November 2009 9 setting out possible macro-prudential tools that could be implemented. Since then,the government has announced that it will create a new Financial Policy Committeein the Bank o England with primary statutory responsibility or maintaining

nancial stability and control o macro-prudential tools to ensure that systemic risksto nancial stability are dealt with.

2.23 This undamental review o trading activities should work in tandem with the new

macro-prudential initiatives to ensure that macro-prudential objectives are notundermined by the capital regime.

Remuneration2.24 The G20 communiqu in November 2008 10 noted that inappropriate remuneration

structures contributed to the nancial crisis. Over the past two years the FSB hasbeen leading work to establish globally agreed principles or sound compensationpractices. The FSB published these principles in April 2009 11 and ollowed this up in

September 2009 with a set o implementation standards. 12 In the UK, we haveincorporated remuneration requirements into the FSA Handbook through SYSC 13 19 the Remuneration Code - (consulted on in CP09/10, Re orming remuneration

practices in fnancial services ).14

2.25 Implementing strong remuneration disciplines and controls should, i done e ectively,play a role in reducing the level o risk taken by banks in their trading activities andin infuencing rms valuation and risk measurement practices. These measuresshould complement, rather than replace, a stronger prudential ramework.

Q1: Are the most important interactions with afundamental review of prudential requirements fortrading activities covered in this chapter? If not, whatother key interactions need to be considered?

9 The role o macroprudential policy Bank o England Discussion Paper, November 2009.10 G20 communiqu Declaration Summit on Financial Markets and the World Economy , 15 November 2008.11 FSB Principles or Sound Compensation Practices , 2 April 2009.12 FSB Principles or Sound Compensation Practices Implementation Standards , 25 September 2009.13 Senior Management Arrangements, Systems and Controls section o the Handbook.14 In light o the 2010 Financial Services Act and the current European Legislative process that is implementing the

FSB principles, we will shortly be publishing a review o the FSAs Remuneration Code.


17/126


Summary3.1 Our prudential ramework or banks, building societies and investment rms is

based on internationally agreed standards, which are codi ed into EU legislationthrough the Capital Requirements Directive. Figure 3.1 highlights key milestones inthe development o the international ramework since 1988.

3.2 There are three aspects o the ramework speci cally relevant or trading activities:

I. The valuation o assets and liabilities, which determine capital resources.

II. The trading book boundary, which de ines the positions subject to tradingbook treatment.

III. The capital requirements applied to those positions in the trading book.

3.3 In each o these areas there are structural de ciencies that need to be addressed bythis review, which are covered by the recommendations in later chapters.

The current frameworkfor trading activities3


18/126


Figure 3.1 Developments in prudential framework since 1988

1988: Basel I minimum ratio agreed. No specific charges for market risks.

1993: Europe implements the Capital Adequacy Directive (CAD), which introduced amarket risk framework in Europe.

1996: Basel I Market Risk amendment, effective end-1997. Capital requirements formarket risk arising from foreign exchange, traded debt securities, equities,

commodities and options.Subject to regulatory approval, firms can use their own internal Value at Risk (VaR)

models to calculate their regulatory capital requirements for market risk.

2005: Basel II trading book review. New requirements for modelling specific interestrate risk, the introduction of a requirement to capture default risk on traded

positions, 15 introduction of prudent valuation principles. Implemented in the EU in2007.

July 2009: Further strengthening of the trading book regime by the Basel Committee,

for implementation in 2011.

December 2009: Basel Committee consultation on range of proposals not directly relatedto market risk capital, but with an impact on trading activity

(as noted in Chapter 2).

Valuation3.4 The valuation used in the regulatory ramework is based on accounting standards,

with a series o lters and adjustments applied to generate capital resources. This isshown in Figure 3.2.

For all positions air valued on an ongoing basis, our prudent valuationramework describes limited circumstances notably illiquidity under which

we expect rms to adjust accounting valuations to better align with regulatoryobjectives. This is particularly relevant or the regulatory trading book, in whichall positions must be air valued daily.

15 IDRC requires banks to hold capital against de ault risk that is incremental to any de ault risk captured in thebanks VaR model. All UK prudentially regulated rms must now hold this, but due to the nature o the Europeanlegislation in this area it is not necessarily the case elsewhere.


19/126


For xed income securities air valued through equity, 16 we eliminate airvalue gains and losses rom capital resources. As outlined in Chapter 2, theDecember 2009 BCBS package removes this lter so that market volatility onall air valued assets eeds directly into capital resources.

Adjustments are made to air value liabilities to lter out movements inown-credit risk.

3.5 A mixed-measurement model is likely to continue in international accountingstandards. Under International Financial Reporting Standard 9 (IFRS 9), publishedin November 2009, rom 1 January 2013 nancial assets will be held either:

at air value through Pro t and Loss (P&L); 17 or

at amortised cost.

3.6 This will not signi cantly a ect the regulatory valuation model described above, butwill tend to simpli y the ramework. Following these changes, we see three areas o speci c weakness in the regulatory valuation that need to be addressed.

Inconsistency

3.7 Ensuring comparable valuations across rms can be just as important, i not moreso, as applying comparable capital requirements this is demonstrated in Box 3.1.Within the air value category, we have identi ed several areas where the absence o speci c accounting guidance has led to material variation in practice across rms these are described in Chapter 5.

3.8 Inconsistencies can also arise because rms are using di erent accounting categoriesor the same instrument type. Although trading activities tend to be subject to air

value, this is not true in all cases (see Chapter 5) and we have observed markeddi erences between UK rms regarding which positions are subject to air value.Also, IFRS 9 di ers rom the direction in which the US is moving regarding valuing

nancial instruments. In May 2010, 18 the Financial Accounting Standards Board(FASB) con rmed its intention to apply air value more widely across nancialinstruments, with limited exceptions only or short-term receivables and payables,own debt, and a small set o strictly de ned other investments.

3.9 The international nature o trading activities means it is crucial that regulators take allpractical steps to ensure consistent valuation approaches are applied internationally.

16 Gains and losses on these positions do not pass through the rms pro t and loss account.17 Strategic equity investments will be held at air value through Other Comprehensive Income.18 FASB : Proposed Accounting Standards Update Accounting or Financial Instruments and Revisions to the

Accounting or Derivative Instruments and Hedging Activities .


20/126


Figure 3.2: The current valuation framework

Accounting Valuation Framework

Regulatory Valuation

PrudentValuation Framework

Prudential Filters

Fair ValueAmortised Cost

Capital Resources

Uncertainty

3.10 The current prudent valuation ramework is designed to ensure that robust air valueseed into Tier one capital. Adjustments to accounting values may be required or:

inherently illiquid positions;

highly concentrated positions; and

complex securities/exotic derivatives whose values are based on models.

3.11 These standards are currently articulated in a reasonably high-level way and areclosely aligned to existing accounting standards. The standards were enhanced in theBCBS July 2009 package to clari y speci c circumstances under which an adjustmentwould be necessary or current illiquidity. 19 However, it is not clear that there is aconsistent view among regulators regarding how to deal with the general caseswhere there is uncertainty due to a range o plausible valuations, and how thisuncertainty should eed into capital measures. In practice, we think very little hasbeen done to address this risk on trading positions.

19 These enhancements will be supported in the EU by guidance to be issued by the Committee o European BankingSupervisors in advance o their implementation.


21/126


Pro-cyclicality

3.12 The use o air value generally and speci cally the use o mark-to-market as ameasure o air value potentially allows rms to build up leverage when marketvalues are high, urther exacerbating asset bubbles. The recent crisis highlightsspeci c methodological concerns with aggressive valuation practices during thiscrisis (especially issues concerning bid-o er adjustments and CVA calculations) andwider issues related to the cyclicality o air value. The prudential ramework shouldmoderate aggressive valuation practices where these can be identi ed and quanti ed,especially with a view to reducing pro-cyclicality in the upswing.

Box 3.1: Valuation inconsistency compared to capital requirements

In April 2008, the Bank o Englands Financial Stability Report analysed the rangeo values produced by six Large Complex Financial Institutions (LCFIs) at the endo 2007 or super-senior tranches o Collateralised Debt Obligations (CDOs). 20 These tranches were the most senior slice o CDO structures and would there orebe expected to have a AAA credit rating at inception. The chart below shows themaximum capital requirement or such a position relative to the valuation range.In all cases, the maximum capital requirement is smaller than the variation invaluations (highest valuation minus lowest valuation reported) o the tranchesproduced across the six irms.

This is important when assessing irms solvency the di erence in valuing an assetbetween irms is directly re lected in their capital resources. Where the range o valuations or an asset is larger than the capital resources required to be held against

it, a irms position in the valuation range has a more signi icant impact on assessingthe adequacy o capital resources than the capital requirement itsel . Indeed, in anextreme case, the capital requirement or a position could be covered by a irmsimply revaluing its position rom the low end o a valuation range to the high end.

Valuation range vs capital requirement (as % of notional)

0

10

20

30

40

50

60

High grade super-senior tranche

Mezzanine super-senior tranche

CDO squared super-senior tranche

Maximum capital requirement

(assuming AAArating)

Note: Maximum capital requirement assumes market value o the tranche is par, and the note is loating rate.

20 The Bank o England data analysis was done at a high level, and so the CDO tranches analysed did not necessarilyhave similar underlying collateral or overall structure. Some variation in valuation would there ore be expected.However, we consider that the degree o variation shown by the analysis is signi cantly beyond what could beaccounted or by these issues.


22/126


The trading book boundary3.13 The trading book boundary was introduced internationally in the 1996 Basel

Committee Market Risk amendment. 21 The non-trading book (banking book) is theassumed approach or all positions, with entry criteria determining positions that shouldbe subject to trading book treatment. Importantly, the boundary does not a ect thecapital requirements or oreign exchange or commodity risk, but has the most impacton credit and equity positions; those in the trading book are subject to capital chargesbased on market risk, while in the banking book these positions are subject to chargespredominantly covering de ault risk. All positions in the trading book must also besubject to air value, with gains and losses eeding directly into Tier 1 capital, whereasbanking book positions can be subject to either air value or amortised cost valuation.

3.14 Trading intent is critically important when determining whether a position isincluded in the trading book:

A trading book consists o positions in inancial instruments and commoditiesheld either with trading intent or in order to hedge other elements o thetrading bookPositions held with trading intent are those held intentionally

or short-term resale and/or with the intent o bene iting rom actual orexpected short-term price movements or to lock in arbitrage pro its, andmay include or example proprietary positions, positions arising rom clientservicing (e.g. matched principal broking) and market making. 22

3.15 Basing the boundary on trading intent is fawed. In buoyant markets, rmsdemonstrated trading intent or a wide range o positions. In periods o marketstress, the inability o the trading book ramework to adequately capture the risk onthese positions has been exposed in particular, the ability to hedge and/or exitthese positions within a short time horizon was undermined.

3.16 There are several urther basic requirements or positions subject to trading booktreatment including:

documented trading strategies;

actively managing positions, including either daily mark-to-market (MtM) orassessing pricing model inputs daily;

senior management in ormation on trading positions being an integral part o risk management; and

documented policies and procedures or monitoring positions against a tradingstrategy (e.g. monitoring turnover and stale positions).

3.17 Additional detailed standards are speci ed or those rms that use internal modelsto calculate market risk capital requirements (see below). There is a need to en orcee ective risk control standards across a broader range o unctions, which arede-linked rom the methodology used or calculating capital requirements andinstead act as a minimum standard to undertake trading activities.

21 In Europe this concept was introduction in 1994 through the rst Capital Adequacy Directive.22 A combination o paragraphs 685 and 687 o the Basel International Convergence o Capital Measurement and

Capital Standards June 2006 www.bis.org/publ/bcbs128.htm


23/126


Pillar 1 capital requirements3.18 Our rules allow three alternative approaches or rms to calculate Pillar 1 market

risk capital requirements. The standardised measurement method (standard rules) isthe most simple, and does not require any speci c approval. The CAD1 regime,which can only be used subject to a waiver rom us, allows more o -setting betweenproducts than standard rules or some options and interest rate risks. Finally, rmscan apply or a di erent waiver to get permission to use an internal VaR modelapproach, whereby the capital requirements are determined by the rms ownassessment o market risk.

Market risk requirements standard rules

3.19 The standard rules or market risk apply strict criteria, which generally only allowpositions to be o set with each other when risks are:

linear;23

equal (e.g. the same instrument); and

opposite (e.g. one is directionally long and one is directionally short).

In the case o options, non-linear risks are either approximated to linear risks(when the position is su iciently in the money) and there ore allowed to o -set,or are only permitted to partially o -set. In the case o securitisations, non-linearrisks (e.g. tranches) are not allowed to o -set against linear risks.

3.20 This limited o -setting can e ectively provide a barrier to entering certain traded

markets, where risks are generally hedged by adding non-matching positions to theport olio. This leads to a signi cant incentive or rms to develop an internal model

or calculating capital requirements this may not always be possible or smaller rms.

3.21 Each net position (i.e. a ter o -setting) is subject to a capital charge based on asimple look-up table that is the same or all rms. A review o these charges isnecessary, and is particularly important or the speci c risk on xed incomesecurities where the charges are linked to external credit ratings. The crisis hasshown that credit ratings are not an appropriate indicator o market risk. 24 Thisissue was compounded where the credit rating processes or certain classes o

security were revealed as seriously fawed.25

23 For a given change in the value o a re erence asset there is proportionate given change in the value o thecontract held.

24 Credit ratings measure the expected probability o de ault o a security. This is only one o a number o market andliquidity actors which a ects the value o a security that is held or trading purposes and thus is not appropriate asa sole measure o a positions market risk.

25 It is now widely accepted that the credit ratings process was materially fawed or a range o complex structuredcredit assets such as CDOs backed by mezzanine tranches sub-prime Residential Mortgage Backed Securities (RMBS).


24/126


Market risk requirements the CAD1 regime

3.22 Our CAD1 regime can be seen as a simple orm o modelling or the ollowing typeso risk:

Interest rate pre-processing allows rms to recognise more o -setting or

general interest rate risk be ore applying standard rules to the net position. Options delta calculations allows rms to produce delta-equivalent 26 measures

or option positions that can be passed through standard rules, therebya ording rms a more risk-aligned treatment or delta than is achieved by usingthe standard treatment o options.

Additional options risks covers the additional (non-delta) risk associated withoption positions. This is generally done by using matrices that apply stressesagainst movements in underlying variables. Examples typically include volatilityskew and implied correlations, but other risks will be relevant i highly exoticderivatives are covered by the CAD1 approach.

3.23 These approaches tend to deliver more risk sensitive capital allocation, withoutbeing beholden to rms internal modelling methodologies. Consideration should begiven to applying this type o approach more broadly.

Market risk requirements internal VaR models approaches

3.24 Internal models or market risk are based on the VaR concept outlined in Figure 3.3.The modelling standards are set out in the Basel Accord, but over time a range o implementation practices have developed across jurisdictions. This presents asigni cant risk o a race to the bottom, where risk migrates to those jurisdictionswhere models are treated most leniently. I this risk is not mitigated, then capitalproviders and other investors could doubt whether the regulatory capital outputs aresu ciently robust. I internal models continue to be used or regulatory capital,there must be increased international cooperation to guard against this risk.

26 The delta o an option is the rate at which its value changes relative to the movement in the instrument underlyingthe option. For example, an option on a share in Company X with a delta o 0.5 would have an increase in valueo 0.5 i the share price o Company X increased by a value o 1. The delta-equivalent measure or an option iscalculated as the value o the underlying position multiplied by the options delta.


25/126


Figure 3.3

Loss Profit

99%confidence

level

Daily VaR

Basic methodology

Estimate loss over a defined period, to a definedconfidence level

Three basic approaches to estimating loss:1. Historical simulation2. Variance-covariance3. Monte Carlo

Historical simulation most widely used (at leastfor regulatory purposes)Hold capital sufficient to cover some multiple of this Value at Risk

VaR modelling generates a portfolio measure of risk

Frequency distribution of observeddaily trading Profit/Loss

3.25 VaR models, i used appropriately, can provide a use ul aggregate measure o marketrisk and can be a use ul day-to-day management tool. However, our supervisoryexperience has highlighted several reasons why VaR modelling has ailed to deliverappropriate regulatory capital requirements. Here we note some o the key points:

The use o a low multiple o a ten-day 99% VaR measure as the primary drivero Pillar 1 capital requirements led to a material under-capitalisation o marketrisk. This represented a mis-speci cation o the market risk capital standard.

VaR measures do not speci cally measure the scale o potential losses beyondthe percentile that the VaR model is calibrated to.

Typical VaR model data windows (one to ve years) were not long enough tocapture risks that emerged during the crisis.

The multitude o market risk actors that rms are exposed to meant it was notpossible or rms to (a) uniquely identi y each o these risk actors within themodel and (b) appropriately calibrate the high percentile variance and covariancecharacteristics o each actor. Simpli cations were there ore required. However,these simpli cations led to (sometimes material) basis risks being missed.

The constant port olio assumption in the current regulatory VaR measure meansdynamic hedging costs are not adequately captured. We also think that rms

ail to hold su ciently large valuation adjustments to properly capture utureport olio hedging costs.

VaR models are not able to capture the wide range o model risks inherent inthe mark-to-model position valuations they take as model inputs.

VaR does not capture the gap risk inherent in certain trading positions, whichcan lead to material losses. This gapping was o ten associated with the absenceo secondary market liquidity, which meant that the deleveraging mechanism/ arrangements could not be enacted.


26/126


3.26 A similar range o de ciencies have been highlighted in academic literature,particularly ocusing on the pro-cyclical impacts o VaR 27 and the related build-upo systemic risk in the nancial sector. There is no single response to these issues, assome could be addressed by enhancing existing VaR techniques, while others relateto a structural weakness o VaR as a statistical risk measure. However, an important

element o this appraisal o VaR (or any other risk measure) must be a recognitionthat the objectives o individual rms or measuring risk are in many cases notaligned with regulators objectives in setting minimum capital standards that willapply to all rms across the system.

Differences between the capital approaches for market risk

3.27 The internal model approach allows a rm to imply diversi cation and o -setbetween all positions included in the model, which is vastly di erent to thephilosophy o the standard rules, where o -setting is strictly limited. This di erencemeans that, or some port olios, the level o capital delivered by standard rulesrequirements is many multiples o that required under the internal model approach.In extreme cases, the consequence o removing a rms permission to use the modelcould threaten the solvency o the rm, which can limit the set o tools realisticallyavailable to regulators when de ciencies are ound. This problem is compounded asmodel de ciencies tend to be revealed in stressed times when urther capital-raisingwould be di cult.

3.28 Narrowing the disparity in the capital standard across all rms should be a keyoutcome o this review.

Amendments since Basel II 2005 and 2009

3.29 In 2005 28, the BCBS reviewed trading book speci c elements o Basel II and recognisedthe need to capture de ault risk on traded credit positions that is incremental to thatcaptured by the ten-day 99% VaR model. As the nancial crisis developed, it wasagreed that this did not go ar enough and in July 2009 the BCBS consulted on apackage o amendments to the current trading book capital requirements. The keyelements o the July 2009 package are summarised in Figure 3.4.

3.30 This package o amendments represents a major change in the capital requirementsapplied to the trading book, in particular raising the capital standard or creditproducts and bringing it in line with the banking book (the Internal Ratings Basedapproach (IRB)). However, a comparable capital standard does not necessarily meancomparable capital. The IRB approach makes certain assumptions regardingport olio granularity, which may or may not apply in the Incremental Risk Charge(IRC). Firms IRC models might also use market implied data to determinecorrelations that can be applied across long and short positions, whereas the IRBapproach speci es correlations to systematic actors and applies credit riskmitigation only to exactly matching positions. These continued di erences intreating credit risk across the banking and trading books need to be reviewed.

27 See or example Risk Appetite and Endogenous Risk , Danielsson, Shin and Zigrand (2009) and Liquidty and Leverage , Adrian and Shin (2008).28 The application o Basel II to trading activities and the treatment o double de ault e ects:

www.bis.org/publ/bcbs116.htm .


27/126


3.31 In addition, the stressed VaR requirement mitigates some o the weaknesses in VaRas a capital measure highlighted above. In particular, it signi cantly increases theimplied capital standard rom the current level and helps mitigate the relatively shorttime period required by the current ramework by locking in a historical period o stress. However, it does not address some o the more structural weaknesses o VaR

described above.

Figure 3.4 Key elements of Basel July 2009 package

Proposal Positions effected Description Impact/Capital standard

Stressed VaR All modelled positionsAdditional capital chargebased on stressed calibrationof VaR model inputs

Increase in capital requirementsfor modelled positions

Reduction in relative cyclicalityof VaR capital requirement

Stressed market inputs meanhigher implied capital standard

Incremental riskcharge (IRC)

With the exception

of securitisationpositions, all creditpositions for whichthe firm modelsspecific risk

Firms must captureincremental (to VaR) creditdefault and migration risk onmodelled credit products inthe trading book

Improves risk capture on traded

credit positions Increases resilience of banking

system to credit risk shocks Comparable to banking book

approach

Comprehensiverisk measure(CRM)

Positions that arewithin a firmscorrelation tradingportfolio

Requirement to model all price risks for thesepredominantly structured CDSbooks

Improves risk capture forcorrelation portfolio

Calibrated to one year 99.9%level

Standardrules forsecuritisationpositions

Securitisationpositions that arenot deemed to becorrelation trading

positions

Removal of non-correlationbook securitisations frommodelling approaches andapplication of banking book

risk weights to net positions

Reduces scope for regulatorycapital arbitrage onsecuritisation products

Expected to lead to significant

increase in capital requirements

Counterparty Credit Risk (CCR)

3.32 CCR on all OTC derivatives and Securities Financing Transactions (SFTs) iscalculated in line with banking book credit risks. This ails to recognise the MtMnature o this risk that arises rom fuctuations in the credit quality o the derivativecounterpart captured through CVA. The BCBS December Consultation Paper 29

ocuses on this issue, requiring a new capital charge to cover the air value lossesthat arise through CVA.

3.33 However, in delivering a long-term approach to capture CVA volatility, regulatorsmust also address issues relating to CVA calculation. We see a wide range o practices, leading to signi cant divergence o valuation output. This divergenceundermines any attempts by regulators to deliver convergence in this area, and untila common set o requirements are agreed, it will remain di cult to generate acapital charge that appropriately captures this risk or all banks.

29 www.bis.org/publ/bcbs164.pd
http://www.bis.org/publ/bcbs164.pdfhttp://www.bis.org/publ/bcbs164.pdf


28/126


The current framework in practice3.34 The practical application o the current ramework described above to each asset

class that rms hold on their balance sheet is summarised in Annex 1.

3.35 Trading assets o ten make up a large percentage o a rms total assets, but commonly

account or a much smaller percentage o a rms capital requirements a gure romThe Turner Review is reproduced in Figure 3.5.

Figure 3.5 Trading book assets and capital 2007: examples

Market risk capitalrequirement as percentage of trading assets

Trading assets as percentageof total assets

Trading/Market risk capitalas percentage of total capitalrequirements

Bank 1 0.4 34.0 11.0

Bank 2 0.4 28.0 7.0

Bank 3 0.1 57.0 4.0Bank 4 1.1 27.0 7.0

Source: BIS estimates rom bank annual reports

3.36 In October 2009, the Basel Committee published interim analysis o the QuantitativeImpact Study or the 2009 amendments. 30 These results show that the averageincrease in market risk capital cross the surveyed banks would be at least 224%,which goes some way to addressing the current undercapitalisation o market risks.The ocus o this review should there ore be the structural de ciencies o the regime however, urther recalibration might be necessary subject to the results o the

ongoing prudential re orms described in Chapter 2.

30 www.bis.org/publ/bcbs163.htm


29/126


Summary4.1 This chapter is divided into two sections:

Section A summarises the growth in traded markets over the past 20 years.

This growth has increased the importance o the trading book regime to thesoundness o the banking system and rein orces the need or this review, as thecurrent ramework has not kept pace with market developments. There is alsosome evidence that, during the crisis, liquidity remained stronger in simpler andstandardised markets, potentially highlighting actors that could reasonably beused in the capital regime and that are discussed in Chapter 7B.

Section B discusses some structural di erences that are evident between thesemarkets, in particular the di erence in credit markets.

These structural di erences suggest that a di erent (non-market based) approachto credit might be justi ied on the basis that much less o the credit risk, relativeto other risks, ever passes out o the banking system despite vastly increasedtrading volumes. This leads to some key recommendations in Chapter 7A.

A. The growth of traded markets4.2 Over the past 20 years, most asset classes have seen a material increase in the level

o trading activity. This is o ten relatively easy to observe and measure throughpublicly available in ormation. A second phenomenon an increase in thecomplexity o the products traded is more di cult to track and quanti y in mostmarkets, but is something we have clearly seen evidence o .

4.3 The di culty in directly tracking this increase in the level o trading in complexproducts comes in part because complex products are o ten originated by combiningseveral more simple products, and such issuance is not routinely tracked. One areathat has seen well-documented increases in complex instruments is structured credit.However, this is not the only area o increasing complexity or example, there hasbeen signi cant trading (and signi cant losses) in areas such as equity correlation,

The evolution of traded markets4


30/126


Constant Maturity Swaps (CMS) and Power Reverse Dual Currency (PRDC) bonds some o which we discuss in this DP.

4.4 In this section we show summary data on the growth o the ve most signi cantmarkets in the nancial sector:

oreign exchange (FX); traded credit;

interest rates;

equities; and

commodities.

FX markets

4.5 The FX markets remain largely decentralised or OTC trading and are the worldslargest (on a gross market turnover measure), with a daily market turnover o $3.5trn 31 in 2007. These markets continue to be dominated by simple products,with around 90% o the OTC market comprised o traditional orward and swapcontracts (see Figure 4.1). However, given the size o these markets, even a smalllevel o complex transactions in percentage terms would represent a signi cantmarket. While pure FX risk has not been a major source o trading losses during thiscrisis, FX products have caused signi cant losses in previous crises, such as the 1997Asian nancial crisis. FX risk is also increasingly packaged with other risks incomplex products, such as PRDC notes.

Figure 4.1 FX markets growth

0

500

1,000

1,500

2,000

2,500

3,000

3,500

4,000

1992 1995 1998 2001 2004 2007

G r o s s

d a

i l y

t u r n o v e r m

i l l i o n s

Currency swaps Options* Estimated gaps in reportingForeign exchange swaps Outright forwards Spot transactions

1992 1995 1998 2001 2004 2007

P e r c e n

t a g e o

f t o t a l t r a

d e

0

20

40

60

80

100

Source: BIS*data not collected until 1995

31 Triennial Central Bank Survey o Foreign Exchange and Derivatives Market Activity in 2007 www.bis.org/publ/rp x 07t.pd gure includes traditional FX as well as options and currency swaps.
http://www.bis.org/publ/rpfxf07t.pdfhttp://www.bis.org/publ/rpfxf07t.pdf


31/126


Traded credit markets4.6 One o the most striking aspects o traded credit markets over the last 20 years has

been the explosion in the size and complexity o securitisation and structured nanceactivities and the credit derivative markets. In The Turner Review and accompanyingDP09/2, we explained our view that the demand or yield upli t, stimulated bymacro-imbalances, triggered a wave o nancial innovation, ocused on theorigination, packaging, trading and distribution o securitised credit instruments.This resulted in:

rapid growth in the value o the total stock o credit securities;

a signi cant increase in the complexity o , and in some cases embedded leveragewithin, the structured nance securities sold (particularly Asset Backed Securities(ABS)); and

a related explosion in the volume o credit derivatives, enabling investors andtraders to hedge underlying credit exposures, or to create syntheticcredit exposures.

Figure 4.2a Outstanding ABS notional

0

500

1,000

1,500

2,000

2,500

3,000

1996 1998 2000 2002 2004 2006 20081997 1999 2001 2003 2005 2007 2009

U S $ b i l l i

o n

Auto Credit cardHome equity

Manufacturing housingStudent loans Other

0

20

30

40

50

10

60

70

1996 1998 2000 2002 2004 2006 20081999 2001 2003 2005 2007 2009

U S $ b i l l i

o n

Source: SIFMA Source: BIS

4.7 This nancial innovation was based on the premise that by repackaging risk it waspossible to create value, o ering investors combinations o risk, return andliquidity that were more attractive than those available rom directly originating orpurchasing the underlying credit exposures. It resulted not only in massive growth inthe importance o securitised credit, but also in pro oundly changing the nature o the credit origination. Following the crisis, there is a question about what the long-term size o this market will be.

4.8 Since the crisis broke, volumes in CDS have dropped o sharply. Some o thisreduction is due to port olio compression where o -setting trades betweencounterparties are torn up. In ABS, volumes have remained arti cially high as manybanks have issued securities to access government liquidity schemes (see Figure 4.2).

Figure 4.2b Outstanding CDS Notional


32/126


Interest rate markets

4.9 While there has undoubtedly been a huge growth in new and innovative markets inthe credit area, there has also been an increase in trading in the more standardised,longer established xed income and interest rate markets, as shown in Figure 4.3.

Figure 4.3 Average daily trading volume in US bond markets

0200

400

600

800

1,000

1,200

2002 2004 2006 20082003 2005 2007 2009

U S $ b i l l i

o n

Municipal TreasuryCorporate debt

Agency MBSFederal agency securities

Source: Federal Reserve Bank o New York, Municipal Securities Rulemaking Board, FINRA

4.10 In the OTC market, interest rate products saw a more dramatic ten- old growth inoutstanding notional in the ten years leading up to the nancial crisis, as shown inFigure 4.4. Although there has been some growth in more complex areas, such as

interest rate options, the majority o this growth has been driven by simple interestrate swap products. While the market itsel has not seen signi cant innovation,interest rate products have, however, been integral building blocks o more complexsecurities and structured products that have evolved.

Figure 4.4 Outstanding interest rate and cross-currency swap notionals

2 H 8 7

2 H 8 9

2 H 9 1

2 H 9 3

2 H 9 5

2 H 9 7

2 H 9 9

2 H 0 1

2 H 0 3

2 H 0 5

2 H 0 7

2 H 0 9

50,000

0

100,000

150,000

200,000

250,000

300,000

350,000

400,000

450,000

U S $ b

n

Source: ISDA


33/126


Equity markets

4.11 Figure 4.5 shows a general trend o growth in equity trading volumes in the last 15 to20 years however, since 2000 this growth pattern has been less evident, particularlyin more established indices such as FTSE 100 and Dow Jones. Part o this lowergrowth will be due to the increased investor interest in emerging equities markets.

4.12 Be ore the implementation o the Markets in Financial Instruments Directive(MiFID) in 2007, cash and directive equity markets were largely exchange-based.However, the introduction o MiFID has seen established regulated marketschallenged, in particular rom multilateral trading acilities (MTFs). In the 2010Financial Risk Outlook (FRO) we noted:

In 2009, MTFs accounted or around 20% o total trading in FTSE 100shares, and approximately 25% o order-book trading...while competition intrading services is welcome, potential risk rom the ragmentation o equity

trading and data have resulted and need to be appropriately mitigated.4.13 Signi cant developments in the In ormation Technology (IT) power available to

trading plat orms and participants has also seen the advent o High FrequencyTrading (HFT). HFT now makes up a much greater part o overall equity trading.The 2010 FRO suggested that HFT has grown rom virtually zero to 60-70% o trade in US markets and 30-50% o trade in EU markets over the past decade, andmay lead to urther increases in trading once the crisis has ended.

Figure 4.5 Equity trading volumes in major markets

FTSE 100

0

2,000

4,000

6,000

8,000

1996 2000 2004 2008200720052001 20031997 1998 1999 2002 2006 2009

I n

d e x

p r

i c e

T r a d e v o l u m e b i l l i o n s

Trade volume FTSE 100 high FTSE 100 low

0

10

20

30

50

40

Source: Bloomberg


34/126


Dow Jones

0

600

1,000

1,400

1,800

1,600

1,200

800

400

0

10

20

30

50

40

1993 1997 2002 2007200620031998 2000 20011994 1995 1996 1999 2004 2005 20092008

I n

d e x

p r i c e


Trade volume DJ high

Source: Bloomberg

DAX

0

3,000

6,000

9,000

I n

d e x

p r

i c e


Trade volume DAX high DAX low

1993 1997 2002 2007200620031998 2000 20011994 1995 1996 1999 2004 2005 200920080

1

2

3

5

4

Source: Bloomberg

4.14 Dealers derivative port olios have continued to grow and become more complex,exposing them to greater risk. Be ore the crisis, dealers had developed structural

positions in dividends and implied correlations. All dealers tended to be the sameway round in these exposures (long dividends and short implied correlation) as aresult o selling similar derivative products to investors. These one-way positionsare di cult to hedge and/or close out, and the change in market undamentals(e.g. reduced dividend expectations, high realised correlation) led to materiallosses at many dealers. Box 4.1 gives an example o this.


35/126


Box 4.1 Equity derivative losses

The irst hal o 2008 saw signi icant losses in equity derivatives businesses, which werepicked up by our loss analysis (see Chapter 5). Some o the key drivers o losses werespikes in volatility and correlation, combined with a drop in dividend expectations.

Ten-day volatility on the Dow Jones Eurostoxx 50 index moved rom 13.44%on 17 January 2008 to 62.89% by 30 January. This spike in volatility cruciallycoincided with a sharp rise in correlation across global markets. Add to this asigni icant reduction in dividend yields, and inancial institutions aced potentialexposure to signi icant unexpected market moves.

As a result o structured products sold to retail clients (typically relatively simpleproducts such as shorter-dated yield-enhanced products that relied on selling optionsto boost returns, or reverse convertibles) inancial institutions were signi icantlyexposed to all these movements.

The losses su ered highlighted several issues that the capital ramework is notwell-placed to capture. VaR models are not good at capturing dividends or impliedcorrelation, and the additional risks rom crowded trades are typically ignored.Signi icantly, they also highlighted instances where irms did not hedge risks thatwere seen as driving pro its.

Commodities markets

4.15 The years leading up to the nancial crisis saw signi cant growth in the notional valueoutstanding o commodity OTC derivatives. Between 2003 and 2007 the notionalvalue increased more than 500%, 32 reaching $9trn by the end o 2007. Exchangetraded commodity contracts also saw signi cant growth over the same period,although to a lesser extent than OTC trading (200% growth over the same period).

4.16 However, each commodity market has developed at a di erent rate. For example,some crude oil benchmarks are now heavily traded on exchange markets, whereas anumber o other commodity markets remain highly specialised, small andpredominantly OTC. As a result o these idiosyncrasies, it is di cult to place accurateestimates on the total size o certain individual commodity markets. In addition, the

growth rate in each o these markets has not taken place in an even or stable manner.

4.17 Bank or International Settlements (BIS) data ( gure 4.6) shows the dramatic growthin OTC commodities market trading, largely driven by trading in energy marketsand non-precious metals.

4.18 In exchange-traded markets, BIS data shows a similar pattern o increased trading.In these markets growth was driven by derivative trading, particularly in agriculturalcontracts and energy (both o which saw signi cant price rises over the same period)based on increased investor interest/development o new nancial products.

32 Based on gures on growth between 2003 and 2007 produced by IFSL Research Commodities Trading 2008, June 2008


36/126


4.19 The early stages o the crisis triggered signi cant alls in OTC commodities markets,which are particularly evident in Figure 4.6. Exchange traded contract volumesavoided a similar all, which is likely to be due to the reduced counterparty riskinherent in exchange traded products leading to a fight to quality in this market.

Figure 4.6a Outstanding OTC derivatives notional

0

4

6

8

10

12

14

H1 00 H1 06H1 05H1 04H1 03H1 02 H1 08H1 07 H1 09

U S $ t

r i l l i o

n

2

Gold Other commodities

Source: BIS

Figure 4.6b Exchange traded contracts

0

500

1,000

1,500

2,000

2,500

1998 200820062004200220001999 20092007200520032001

0

10

20

30

40

50

60

Number of contracts traded, millions (left axis)Number of contracts outstanding, millions (right axis)

Source: BIS

B. The structure of traded markets is credit different?4.20 While all o the key markets have shown to some degree similar characteristics o

growth, there is evidence that the underlying structure o each market showssigni cant variation. This is particularly apparent in traded credit markets.

4.21 Intuitively, a hypothesis that the interaction between the banking system and creditmarkets is signi cantly di erent to the banking systems interaction with othermarkets appears reasonable. The banking system as a whole acts as a credit provider


37/126


to the real economy, but it is not a signi cant equity provider or, historically, acommodity buyer. With the exception o small proprietary trading operations,positions in those markets tend to be taken as a result o intermediary activity.Finding data to test these intuitions is di cult however, data that is available doessupport the hypothesis.

4.22 For example, or equities markets, The O ce or National Statistics (ONS) reportedthat on 31 December 2008 less than 4% o the shares (by market value) listed onUK stock exchanges were held by banks 33 (see Box 4.2).

4.23 The data or credit markets shows a di erent picture. The rise o securitisation andcredit derivative technology, leading to more commoditised credit products, mighthave led many to believe that credit risk was being trans erred out o the bankingsystem and there ore less would remain on rms balance sheets (we describe abovethe considerable increase in credit trading during the period leading up to the crisis).However, there is evidence that the overall risk o credit products largely remained

within the banking/investment banking sectors, particularly in the more complexareas o traded credit. 34

4.24 The Joint Forum report on credit risk trans er, updated in 2008, 35 qualitativelycorroborates this story. In particular, the report notes the importance o the shadowbanking sector (conduits and Structured Investment Vehicles (SIVs)) as creditpurchasers, nanced by short-term unding through Commercial Paper and MediumTerm Notes. So sponsors either kept the credit risk (through issuing liquidity lines)or passed it onto (predominantly junior) note holders. As the junior tranches o SIVs(capital notes) received ratings, these became attractive to small and medium-sized

banks. Chart 4.1 summarises our analysis undertaken in October 2007, whichshowed that banks held 76% o SIV capital notes.

Chart 4.1 Holders of SIV capital notes

BankInsuranceAsset MgmtCorporateUnknownGovt

Source: Rating agency data and FSA analysis

33 ONS Share Ownership Survey 2008 (published January 2010) showed banks held UK listed shares worth 40.6bncompared to a total value o listed shares o 1,158.4bn. Some have disputed the gures reported by the ONS,

but primarily on the basis that the size o individual and company director share ownership in much greater thanreported rather than issues with the level o bank share ownership.

34 This is supported by evidence o the increased originating and repackaging o credit risk into orms that made iteligible or inclusion in the trading book.

35 www.bis.org/publ/joint21.htm
http://www.bis.org/publ/joint21.htmhttp://www.bis.org/publ/joint21.htm


38/126


4.25 Although there is evidence to suggest that equity tranches o CDOs were passed outo the banking system (see Box 4.2), this trans er is ar short o the sum total o issuance. This is also the case in other securitisation sectors.

Figure 4.7 Customer breakdown of OTC derivative dealers revenues by

asset class

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Fixedincome

ForeignExchange

Credit Equity Commodities Fund-linked

53

31

10

6

39

36

13

12

84

646

52

12

33

3

20

70

46

1

59

3

37

Public Sector/Other

High net worth individuals

Corporates

Financial Institutions

Source: Markets Trends rom the 2007 Global Capital Markets Survey, McKinsey

4.26 These di erences between credit and other markets are not isolated to

securitisations. Figure 4.7 shows a similar situation in OTC derivatives markets.Financial institutions are much more prevalent in the credit derivatives market whencompared to OTC markets or FX, commodities, xed income and equity. This is animportant eature, 84% o credit derivatives being transacted with nancialinstitutions highlights that the CDS market acts much more as an intermediarywithin the banking sector than across the real economy. The overall level o creditrisk within the sector could be seen to broadly remain unchanged, despite there-packaging o risk into tradable ormat.


39/126


Box 4.2: Risk retained in the banking sector in CDO markets

The residual risk rom CDO markets is retained within the banking sector to agreater extent than other markets.

CDO market Significant holdings retained by banksFor investment grade portions o the CDO market, over 50% o the buyerswere banks. 36

Buyers of investment-grade portions Buyers of riskiest portion, equity, unrated

Others, 1%HedgeFunds, 3%

PensionFunds, 4%

Insurance,18%

AssetManagers

19%

Banks,PrivateBanks55%

Banks,PrivateBanks32%

AssetManagers

22%

Insurance,19%

PensionFunds,

Other, 18%

HedgeFunds10%

UK equity market Banks act as intermediary

In contrast to the CDO market, banks retain only 4% o the UK stock market.

Equity market participants

Banks

Public sector

Private non-financial companies

CharitiesOther financial institutions

Investment trusts

Unit trusts

Individuals

Pensions funds

Insurance companies

Rest of world

0 5 10 20 25 30 35 40 4515

Percentage of UK stock market owned

36 Based on 1996 2006 data, quoted in Bloomberg Markets, July 2007


40/126


Summary5.1 This chapter is divided into two sections:

Section A gathers data rom the crisis to present conclusions on areas o weakness in the current regime.

Section B brings together the conclusions o Section A with those o Chapters3 and 4 to describe the key elements that need to be addressed. These are thenmirrored in the recommendations o Chapters 6, 7 and 8.

A. What the crisis has taught usAnalysis of losses suffered in investment banking activities duringthe crisis

5.2 In 2009 we conducted a study o the losses su ered in the investment bankingoperations o major international banks operating in the UK. Its purpose was toanalyse where losses occurred and how the regulatory regime was placed to coverthe risks that crystallised. We collected data rom ten rms on signi cant lossevents37 rom January 2007 to March 2009. In total, the losses analysed amounted

to $240bn. Box 5.1 analyses the extent to which the capital held against thesepositions was su cient to cover these losses.

37 A threshold o $100m was used to de ne a signi cant loss. This level typically meant that rms produced detailedanalysis, or internal reporting purposes, o the losses at the time they were incurred.

Lessons from the crisis5


41/126


Box 5.1: Were capital requirements sufficient?

Capital sufficiency at a firm level

The need or governments to intervene to aid the re-capitalisation o banks during

the crisis can be seen as evidence that levels o minimum required capital wereinsu icient to maintain investor con idence.

However, we can also use the data rom the losses study to take a more quantitativeview on the extent to which this was the case. The total losses over the period orwhich data was received can be compared with the average level o market risk andcredit risk capital held over the period. At this level, as shown in Chart 5.1, lossesamounted to approximately 160% o the total average market and credit risk capitalheld by the banks during the period. 38 However, this analysis needs to be reviewedwith caution or the ollowing reasons:

The data does not include pro ts made during the period, which will o set theseloss events. In all cases, total net pro t/loss was covered by the average capital held.

The overall minimum capital level is designed to cover all risks to the balancesheet, not just those that arise due to the investment banking activities coveredby the study.

The loss study is biased as it only includes rms that survived the crisis,there ore by de nition the rms in the sample had su cient capital to remain asgoing concerns.

Minimum required capital is not designed to absorb losses, but to ensure thatthe rm is able to re-capitalise privately. The structure o regulatory capital(including bu ers) is being redesigned as part o the BCBS December package.

In general, we see the level o loss events exposed in this analysis as evidence thatoverall capital requirements against trading activities were insu icient.

Capital sufficiency at a product level

Data rom other sources during the crisis tends to suggest that capital was notsu icient or particular products. For example:

In April 2008, the Bank o England reported that the average valuations o six largeinancial institutions super-senior tranches (rated AAA at inception) at the end o

2007 ranged between approximately 80% o their notional amount to as low asalmost 20%. 39 This indicated losses o between 20% and 80% o notional. Thecapital required to be held against loating rate AAA notes such as these was around1.6% o notional.

38 Firms were not requested to provide their market and credit risk capital or each month during the period analysed,this was only requested or months in which a signi cant loss (as de ned or

fsa trading

Documents