LARGE EU BANKS’ EXPOSURES TO HEDGE FUNDS

LARGE EU BANKS ’ EXPOSURES TO HEDGE FUNDSNOVEMBER 2005

In 2005 all ECB publications will feature

a motif taken from the

€50 banknote.

LARGE EU BANKS ’ EXPOSURES TO HEDGE FUNDS

NOVEMBER 2005

© European Central Bank, 2005

AddressKaiserstrasse 2960311 Frankfurt am MainGermany

Postal addressPostfach 16 03 1960066 Frankfurt am MainGermany

Telephone+49 69 1344 0

Websitehttp://www.ecb.int

Fax+49 69 1344 6000

Telex411 144 ecb d

All rights reserved.

Reproduction for educational andnon-commercial purposes is permittedprovided that the source is acknowledged.

ISBN 92-9181-862-3 (print)ISBN 92-9181-863-1 (online)

3ECB

Large EU banks’ exposures to hedge fundsNovember 2005

CONTENT SEXECUTIVE SUMMARY 5

INTRODUCTION 7

1 BANKS, HEDGE FUNDS AND FINANCIAL STABILITY 7

2 SURVEY RESULTS 102.1 Background information 10

2.1.1 Coverage 102.1.2 Definition of a hedge fund used

by banks 102.1.3 Information available to national

authorities 112.2 Banks’ exposures to hedge funds 11

2.2.1 Banks’ motivation for dealingwith hedge funds 11

2.2.2 Banks’ experience of dealingwith hedge funds and types ofservices provided 12

2.2.3 Direct exposures 132.2.4 Indirect risks 252.2.5 Other risks 25

2.3 Risk management issues 252.3.1 Internal rules for dealings with

hedge funds 252.3.2 Due diligence, credit analysis

and rating systems 262.3.3 Limits, exposure measurement

and collateralisation 272.3.4 Ongoing monitoring 352.3.5 Risks of funds of hedge funds 412.3.6 Other findings 42

3 SOME SUPERVISORY ISSUES ARISING FROMHEDGE FUND ACTIVITY 423.1 Banks as counterparties of hedge

funds 433.1.1 Risk management practices 433.1.2 Capital requirements 45

3.2 Banks as investors in hedge funds 463.2.1 Risk management practices 463.2.2 Capital requirements 46

4 CONCLUSIONS 47

ANNEX 1HEDGE FUND DEFINITIONS, AS PROVIDED BY BANKSOR SUMMARISED BY COUNTRY AUTHORITIES 50

4ECBLarge EU banks’ exposures to hedge fundsNov embe r 2005

LIST OF COUNTRY ABBREVIATIONS

AT AustriaBE BelgiumCY Cyprus*CZ Czech Republic*DE GermanyDK DenmarkEE Estonia*ES SpainFI FinlandFR FranceGR GreeceHU Hungary*IE IrelandIT ItalyLT Lithuania*LU LuxembourgLV Latvia*MT Malta*NL NetherlandsPL Poland*PT PortugalSE SwedenSI Slovenia*SK Slovakia*UK United KingdomUS United StatesEU15 European Union (15 countries, before enlargement on 1 May 2004)EU (EU25) European Union (25 countries, after enlargement on 1 May 2004 with 10 countries marked

with *)

5ECB

Large EU banks’ exposures to hedge fundsNov embe r 2005

EXECUTIVESUMMARY

EXECUTIVE SUMMARY

Due to the rapid expansion of the role of hedgefunds as participants in financial markets andcounterparties to financial institutions, monitoringtheir activities and assessing the implications forfinancial stability has become increasinglyrelevant. This study, carried out by the BankingSupervision Committee (BSC) of the EuropeanSystem of Central Banks (ESCB), with theassistance of its Working Group on Macro-Prudential Analysis, investigates the links betweenlarge EU banks and hedge funds, given theimportant role that the former play in hedge fundoperations. The report is part of the BSC’scontinued efforts to gain a better understanding ofthe implications of the rapid expansion of hedgefund activities for the European financial system.

The study is to a large extent based on thefindings of a survey that was conducted through thenational central banks and supervisory authoritiesrepresented in the BSC and which consisted of botha quantitative and a qualitative part. The qualitativepart addressed a number of risk management issuesrelevant for banks’ interactions with hedge funds.The quantitative part was dedicated primarilyto banks’ direct exposures to hedge funds,comprising their financing, investment, trading andincome exposures. As the replies to the survey weresometimes incomplete and only a limited number oflarge EU banks provided comparable quantitativedata, the conclusions presented in this reportshould be considered as preliminary and onlyindicative of existing exposures and riskmanagement practices.

The survey results revealed that the directexposures of large EU banks to hedge funds variedsignificantly across countries. In many EUcountries, investments of banks in hedge fundswere the major, and sometimes the only, direct linkbetween the two types of institutions. All types ofdirect exposures seem to be growing rapidly,although they were generally limited in relation tobanks’ balance sheets and total revenue or similarexposures undertaken by US peers. This is at leastpartially due to the fact that the global prime

brokerage market is largely dominated by USfirms. However, it is very likely that the absoluteand relative size of EU banks’ exposures to hedgefunds will increase further in line with thecontinuing expansion of the hedge fund industry.

Most of the banks extensively dealing with hedgefunds had specific guidelines and advanced riskmanagement systems for this business or were inthe process of improving them further. As a rule,large EU banks had stringent requirements forexposures to hedge funds with a strong emphasison collateralisation. Sometimes these requirementsseemed very strict for banks with lower exposuresor without a strong focus on the hedge fundbusiness. Nearly all cash lending exposures werecollateralised. However, there was also evidencethat banks quite often traded with hedge funds inOTC (over the counter) instruments on variationmargin only. In the due diligence process and incredit analysis, banks also seemed to rely ratherheavily on a manager’s track record and manybanks did not mention on-site visits as part of thisprocess. Many banks with higher financing andtrading exposures used sophisticated “potentialfuture credit exposure” measures to account for theexpected downside risks arising from theinteraction of market, credit and liquidity risks.Most of the banks also reported the use of stresstests for the evaluation of potential effects ofvolatile or illiquid markets on their exposures.

With regard to the transparency of hedge funds,some banks seemed to be content with theinformation provided to them, despite reportinglags and the diversity in existing practices. Regularinformation flows typically covered net asset valueand performance figures (changes in net asset valueper share), in many cases together with riskmanagement reports including some “Value atRisk” numbers. Transparency questions wereoften part of the due diligence process and creditrating or scoring models, although the link betweencredit terms and transparency could probably bestronger.

The survey also highlighted a number of areas withscope for further improvement, notably:


– Counterparty discipline, as applied by banks,was found to be under pressure owing to highlycompetitive market conditions. Particularly thelarger hedge funds were successful innegotiating less rigorous credit terms.

– Banks’ stress tests, in particular the regularones, included only historical scenarios andwere often applied to individual hedge fundsonly. The stress testing of collateral could alsobe further improved.

– Aggregation by banks of their hedge fundexposures across the entire financial group and/or different business areas/geographical regionswas seen as problematic.

– Hedge fund disclosures and information onleverage were, despite some progress, laggedand not always adequate. In many cases, hedgefunds still provided banks with relatively crudemeasures of leverage.

– Banks’ answers also raised questions aboutwhether they always had enough timelyinformation on the whole portfolio structure ofhedge funds or whether they took thisinformation sufficiently into account,particularly for the larger hedge funds that havefinancing and trading relationships with severalcounterparties.

Regarding recent developments, banks did not seeany systematic increase in risk-taking by hedgefunds. Leverage levels seemed to be moderate andlower than at the time of the near-default of LTCM,even though funds of hedge funds were reported tobe increasing leverage.

The report also addresses a number of supervisoryissues arising from banks as counterparts totransactions with hedge funds and banks asinvestors in hedge funds. These supervisory issuespertain mostly to banks’ risk management practicesand capital requirements tied to their hedge fundexposures. As regards risk management, bothsupervisors and the financial industry havedeveloped guidance to address the specific riskconcerns. Nevertheless, one should remain vigilant

to new developments that might require an updateof this guidance. As regards capital requirements, ithas been proven that hedge funds exposures canalso be fitted into the general solvency framework,although the current and forthcoming rules do notprovide for any specific treatment of suchexposures.

The main policy conclusions of the BSC study canbe summarised as follows. First, the survey resultsindicate that recent developments in the hedge fundindustry may not necessarily pose a direct threat tofinancial stability in the EU through banks’ directexposures to hedge funds. Nonetheless, banks mayalso be affected indirectly, for example, if hedgefund activities lead to dislocations in financialmarkets or cause strains for major non-EU primebrokers with spillover effects to EU banks. Hence,direct exposures may underestimate the true risksthat hedge funds pose to banks.

Second, the main recommendation put forward bypublic authorities in the aftermath of the LTCMcase – according to which adequate management bybanks of risks associated with hedge funds shouldbe put in place – still remains relevant for large EUbanks, as specific areas of risk management offerscope for further improvement. More generally, thesurvey evidenced the difficulties for banks toestimate hedge fund risks in an exhaustive way. Thestill limited transparency of hedge funds – takentogether with the complex interactions of credit,liquidity and market risks – makes addressinghedge fund risks by banks particularly complex. Asa minimum, however, banks should be able toaggregate their overall exposure to individualhedge funds and limit exposures to prudent levels.

Third, risk management guidance developed bysupervisors and the capital adequacy regimeprovide the appropriate framework for dealing withrisks resulting from banks’ interactions with hedgefunds. In particular the supervisory review processprovided for under Basel II allows supervisors totake the measures necessary to address such risks,including additional capital requirements.

7ECB


1 BANKS,HEDGE FUNDS

AND FINANCIALSTABILITY

INTRODUCTION

In recent years the hedge fund industry hasexpanded rapidly. Because of the important rolethat hedge funds play as participants in financialmarkets and as counterparties to financialinstitutions, especially banks, monitoring theiractivities and assessing the implications forfinancial stability has become increasinglyrelevant. In this vein, the Banking SupervisionCommittee (BSC) of the European System ofCentral Banks (ESCB), with the assistance of itsWorking Group on Macro-Prudential Analysis(WGMA), decided to investigate the nature andrelevance of links between EU banks and (funds of)hedge funds, given the important role that banksplay in hedge fund operations.

Recently there has been an initiative undertaken bythe Counterparty Risk Management Policy Group(CRMPG), an international industry group, toreview earlier recommendations that had been madeon counterparty risk management practices at thelargest financial institutions. The group’s firstreport (CRMPG I)1 was prepared in June 1999 asthe financial industry’s response to the near-collapse of LTCM in September 1998. Themotivation for an update (CRMPG II), which wasfinalised at the end of July 2005, was driven by theimpressive proliferation of hedge funds andcomplex financial instruments.2

During 2004, independent examinations ofselected banks’ exposures to hedge funds werecarried out in three EU countries (United Kingdom,France and the Netherlands). These investigationswere targeted primarily at major banks that wereknown to have extensive dealings with hedgefunds. The UK’s FSA findings on prime brokerageactivities were presented in a series of reports and atthe Financial Stability Forum (FSF) meeting inTokyo on 10 March 2005. Furthermore, the USFederal Reserve has also recently reviewed banks’management of hedge fund credit risk in relation tothe recommendations made by supervisors and theCRMPG I in 1999.3

Against this backdrop and given the substantialgeneral interest in hedge funds, the results of the

BSC survey should contribute to a betterunderstanding of EU banks’ practices in doingbusiness with hedge funds as well as the possibleimplications for financial stability.

In addition to this introductory section, the reportconsists of four more sections and is structured asfollows. Section 1 gives an overview of why theanalysis of banks’ exposures to hedge funds isimportant for financial stability. Section 2 reportson the survey results, including a descriptionof received information, an overview of thefindings on banks’ exposures to hedge funds and adiscussion on risk management issues. Section 3provides an overview of supervisory issues arisingfrom banks’ exposures to hedge funds. The reportends with some preliminary conclusions and adiscussion on possible policy implications.

1 BANKS, HEDGE FUNDS AND FINANCIALSTABILITY

It is widely acknowledged that hedge funds –through their active risk-taking, provision ofliquidity, elimination of market inefficienciesand potential enhancements to investmentdiversification – can contribute to the efficiency,integration and even stability of the global financialsystem. Moreover, hedge funds have changed theasset management industry and, according to onescenario, over time the differences between themand traditional funds may become blurred.

Nonetheless, the recent rapid growth of the hedgefund industry also raises important questions aboutpossible negative implications for financialstability. Hedge funds could affect financialstability through their potential impact on financialmarkets or via their largest creditors andcounterparties, i.e. banks. These two channels areclosely linked and a hedge fund-related triggeringevent associated with either of them could befurther reinforced by these mutual links. The

1 Counterparty Risk Management Policy Group (1999), “ImprovingCounterparty Risk Management Practices”, June.

2 Counterparty Risk Management Policy Group (2005), “TowardGreater Financial Stability: A Private Sector Perspective”, July.

3 Greenspan, A. (2005), “Risk Transfer and Financial Stability”, May.


memorable episode of the near-default of LTCM inSeptember 1998 provides the most vivid exampleof how hedge funds have the potential to disrupt thefunctioning of global financial markets. Thefinancial community was surprised by the sheersize of the leverage underlying LTCM positionsand, in particular, by the fact that banks hadfacilitated the building-up of such positions. Animportant insight made since then is that banks,because they are key hedge funds’ tradingcounterparties and lenders, constitute an importantchannel for influencing the behaviour of hedgefunds. This means that if banks manage theirexposures to hedge funds prudently, then thepotential risks for financial stability arising fromhedge fund activities should be lower.

In this report, the focus is therefore on the bankchannel. Banks’ direct exposures to hedge fundsare the most obvious way in which hedge fundscould cause financial stability concerns. Thesedirect exposures can be broadly grouped intofour categories: financing, trading, investmentand income exposures. Complex hedge fundoperations and active investment strategies requireconsiderable operational support and financing,and involve substantial trading volumes. Banks arekeen to provide such services and many of themhave developed prime brokerage platforms, which,among other services, facilitate the financing, riskmanagement, execution, clearance and settlementof transactions for hedge funds and otherprofessional market participants. For primebrokers, financing and trading exposuresconstitute the biggest source of risk, especially

given the complexity associated with themanagement of such exposures. Moreover, giventhat prime brokerage activities are concentratedamong a limited number of large global players(see Box 1), a serious mismanagement of theseexposures at an individual large bank or group ofbanks could lead to a systemic crisis via contagioneffects on other banks and ripple effects onfinancial markets.

The interplay between banks and hedge funds is notlimited to banks’ direct exposures and potentiallosses arising from such exposures. Banks alsoface indirect exposures related to their exposures toother hedge fund counterparties and the impact ofhedge fund activities in financial markets on theirproprietary trading portfolios.

Banks, investors and hedge funds themselves havelearned lessons from the near-collapse of LTCM.As a result, another high impact failure of a largehedge fund is probably less likely now, especiallyas more players have entered the market andpositions are probably much less concentrated inone or a few funds than was the case at that time.Furthermore, the largest hedge funds now usuallyhave diversified portfolios across severalstrategies.

Recently, however, concerns have also beenexpressed that, as the number and assets undermanagement of hedge funds using similarstrategies increases, the positioning of individualhedge funds is becoming more similar; there istherefore a concern that the resulting “crowding of

Box 1

PRIME BROKERAGE MARKET STRUCTURE

Based on available market information, there are at least four large EU prime brokers, namelyDeutsche Bank, Société Générale, Barclays and SEB; although some other German, French,UK, Dutch, Belgian and Scandinavian banks also have a foot in the prime brokerage market.Nevertheless, the global market for prime brokerage services is dominated by three USentities (Morgan Stanley, Goldman Sachs and Bear Stearns), which control more than half ofhedge fund capital under management (see Charts below). Two Swiss banks, namely CSFBand UBS, are important prime brokers too. Prime brokerage services, financing and trading

9ECB


1 BANKS,HEDGE FUNDS

AND FINANCIALSTABILITY

trades” could pose a significant risk for financialmarkets (and banks’ proprietary positions) in theevent of synchronous hedge fund exits.4

Sometimes banks’ trading desks use the sameinvestment strategies as hedge funds and,therefore, may also be vulnerable to the adversemarket dynamics caused by crowded trades.

Moreover, as noted in CRMPG II, crowded tradesor active and leveraged hedge fund participationcan dampen volatility and increase perceivedliquidity in certain markets, thus leading toartificially low value-at-risk (VaR)5 numbers,including liquidity-adjusted VaR. Indeed, hedgefunds are reported as accounting for 15-30% of thetrading volume in some credit markets and morethan 80% of trading in distressed debt,6 thusincreasing the perceived liquidity in these markets.However, if market participants try to unwind theirpositions in times of stress, this perceived liquiditycould easily evaporate. Moreover, in the latest

financial reports, many banks reported higher VaRvalues. Even though these values were not high inrelation to the banks’ capital, this fact raises someconcerns that banks may underestimate the truerisks of the current environment characterised bylow interest rates, low volatility and high perceivedliquidity.

with hedge funds are highly specialised activities, whereas investments in hedge funds areaccessible to a broader range of financial institutions. As a result, most EU banks that havean exposure to hedge funds have only investment exposures, whereas financing and tradingactivities are largely concentrated among a rather limited number of the largest EU bankswith significant trading activities on their own.

4 See ECB (2005), Financial Stability Review, June; UK’s FSA (2005),“Hedge Funds: a Discussion of Risk and Regulatory Engagement”,Discussion Paper, No 4, June; Counterparty Risk Management PolicyGroup (2005), “Toward Greater Financial Stability: A Private SectorPerspective”, July.

5 The VaR or Value at Risk is the estimated maximum potential loss toa portfolio over a given time period (e.g. ten trading days) at a givenlevel of confidence (e.g. 99%).

6 See Greenwich Associates (2005), “Hedge Funds: The End of theBeginning?”, January.

Prime brokerage market structure

(% of hedge fund assets under management)

Global prime brokerage league European1) hedge fund prime brokers

December 2003 January 2004 January 2005

Source: HedgeWorld’s AccreditedInvestor.

Source: EuroHedge.1) As defined by EuroHedge, does not include managed futures funds.2) Including funds which have no prime broker or have not disclosed.

0 10 20 30 40

Morgan Stanley

Goldman Sachs

CSFBDeutsche Bank

UBS

Banc of America

Other, of which:Lehman Brothers

Bear Stearns

Citigroup

Others

Merrill Lynch

23%21%

17%6%

5%3%

26%3%3%2%2%

17%

30%16%

8%7%

4%4%

31%3%3%3%2%

1%1%1%

16%

0 10 20 30 40

Morgan StanleyGoldman Sachs

CSFBDeutsche Bank

UBSJP Morgan

Other, of which:SEB

Lehman BrothersBear Stearns

CitigroupBarclays Capital

Merrill LynchFimat

Others2)

0 10 20 30 40

Morgan StanleyGoldman Sachs

CSFBDeutsche Bank

UBS

Lehman Brothers

Others2)

26%16%

7%6%

5%4% 37%


2 SURVEY RESULTS

2.1 BACKGROUND INFORMATION

2.1.1 COVERAGEIn order to investigate the links between EU banksand hedge funds, a survey was conducted under theauspices of the BSC with a dedicated focus on EUbanks. Thus, the survey excluded subsidiaries andbranches of non-EU banks. To avoid double-counting, the qualitative information andconsolidated cross-border, cross-sector data (i.e.consolidated both across borders and across thedifferent financial sectors where the bank wasactive) were reported by the country where theultimate EU parent bank was located.

Preliminary contacts with banks and supervisoryinformation had indicated that only a limitednumber of large banks – defined as banks withcross-border, cross-sector consolidated assetsabove €120 billion – could possibly haveexposures to hedge funds. These banks were alsoquite often found to be providing prime brokerageservices to hedge funds. On the basis of this priorinformation, national authorities sent out thequestionnaire to, or contacted in other ways, morethan 100 banks (some smaller countries alsoincluded smaller banks or subsidiaries of EUbanks). It turned out that of those, more than 40 hadsome direct exposures to hedge funds and theyprovided comments on their connections withhedge funds. However, the number of banks withmore significant exposures to hedge funds wasmuch smaller as many of the banks that replied hadmainly investments in hedge funds.

In total, 14 countries submitted their replies to thequalitative part of the survey (AT, DE, DK, ES, FR,GR, IT, LU, NL, PT, SE, SI, SK, UK). In theremaining EU countries, banks either did not haveexposures to hedge funds or these were considerednegligible. Even in some of the countries thatprovided qualitative inputs, banks’ exposureswere insignificant and mostly in the form ofinvestments. Only banks in DE, ES, FR, NL, SEand UK appeared to have more significantfinancing and trading links with hedge funds.Based on the coverage information provided,

35 surveyed banks that provided or were coveredin the answers to the qualitative part of the survey(including 11 smaller banks with mainlyinvestment exposures) in AT, DE, ES, FR, NL, PTand SE as a group constituted around 1%, 55% and38% of respectively the total number, consolidatedassets and Tier 1 capital7 of all eligible bankinggroups in these countries. The coverage inindividual countries ranged within 0-12% of thetotal number of institutions, 23-95% ofconsolidated assets and 21-87% of Tier 1 capital.

Reporting samples for the different sections of thequantitative part of the questionnaire variedsubstantially as banks were asked to provideinformation on a best effort basis. Somequantitative data was supplied by 22 large banksfrom seven countries (AT, DE, ES, FR, NL, PT andSE). Most banks provided only 2004 data, whereassix large banks from four countries (AT, ES, FRand SE) also provided some quantitative data for2003. Such fragmented reporting complicated theanalysis and aggregation of data and limited itspossible use for this report, as some informationreferred to a very small number of banks.

2.1.2 DEFINITION OF A HEDGE FUND USED BY BANKSFor the purposes of the survey, the followingdefinition of a hedge fund was suggested: “a fund,whose managers receive performance-related feesand generally have no or very limited restrictionson the use of various active investment strategies toachieve positive absolute returns. Such strategiesoften involve leverage, derivatives, long and shortpositions in securities or any other assets.” Thesurvey replies indicated that internal definitionsused by banks broadly corresponded with suchcharacterisation (see Annex 1 for the list ofdefinitions provided). None of the countries hadmentioned that banks had substantial difficulty inidentifying hedge funds among the wide spectrumof alternative investment vehicles, although forsome banks the lack of a precise definition couldhave posed difficulties for their risk managementon a global basis.

7 Tier 1 or core capital is regulatory capital that consists of own fundscomponents of the highest quality, such as fully paid-up capital anddisclosed reserves from post-tax retained earnings.

11ECB


2 SURVEYRESULTS

2.1.3 INFORMATION AVAILABLE TO NATIONALAUTHORITIES

Generally, countries had several sources ofinformation on banks’ links with hedge funds.First, very large investments in, or large loansextended to, hedge funds would appear insupervisory reporting of large exposures. Second,countries having credit registers have anotherpossibility of obtaining information on larger creditexposures and on loans to hedge funds. It should benoted, however, that there is a problem with the firsttwo information sources related to the difficulty ofdistinguishing hedge funds from other entities.Furthermore, information on collateralisation isnot always included in these information sources.The third information source is the supervisoryprocess itself, which provides a number ofopportunities for discussing hedge fund-relatedissues, tailored to the circumstances of a particularbank. These include regular prudential meetings,management interviews, on-site inspections,auditors’ analytical reports and ad hoc calls orsurveys. Fourth, banks are obliged to report tosupervisors their direct participations, including,for example, acquisitions of hedge fundmanagement firms, and sometimes also details ontheir investment portfolios.8

Based on the survey replies, it seems that countriesgenerally have limited information on banks’exposures to hedge funds. In most cases, the lack ofregular dedicated reporting or monitoring isjustified by the minimal size of such exposures. Inother cases, a greater reliance is placed on ad hocinformation gathering, which provides a firstindication on the linkages between banks and hedgefunds. This might later be replaced by a moresystematic approach, especially if exposures andpotential (systemic) risk arising from themcontinue to grow. As mentioned in the introduction,the latest initiative by the UK authorities and one-off investigations by authorities in FR and NL are astep in this direction. Nevertheless, it is importantto note that any possible regular data collectioninitiatives should not be misinterpreted as activeregulation where none exists.

2.2 BANKS’ EXPOSURES TO HEDGE FUNDS

2.2.1 BANKS’ MOTIVATION FOR DEALING WITHHEDGE FUNDS

According to the banks surveyed, the main reasonsfor dealing with hedge funds are the growth ofincome and the diversification of income sources,which is not surprising as dealing with andservicing hedge funds is usually a lucrativebusiness. The interest in hedge funds from aninvestment perspective was driven by theirattractive risk-return profile, uncorrelated withother major asset classes.

The second group of motives is associatedwith demand factors. Lately, high-net-worthindividuals and institutional investors have beenseeking exposure to hedge funds and, in responseto that, banks have been either offering their own orthird-party hedge funds or selling structured hedgefund products to these investors. For some banks,hedge funds have been clients for a long time andservicing them has been an important business line.However, some of the interviewed banks indicatedthat they did not have any ambition to become fully-fledged prime brokers because positions ofestablished prime brokers were too strong andentry costs were too high. The EU hedge fundindustry, encompassing hedge funds eithermanaged from or domiciled in the EU, accounts for20-30% of the global hedge fund capital undermanagement and is growing faster than the industryas a whole.9 As a result, the demand for hedge fund-related services, bank trading with and investmentsin hedge funds may be expected to increase further.

The third reason given by banks for dealing withhedge funds is that some saw trading with hedgefunds as a way of providing their proprietarytrading desks with greater trading possibilities andgreater liquidity, especially as hedge funds arewidely known as active risk-takers and significantproviders of liquidity.

8 In NL, for example, there is a quarterly reporting of bankinvestments in alternative (non-traditional) assets, although hedgefunds are not specifically singled out.

9 See Garbaravicius, T. and F. Dierick (2005), “Hedge Funds andTheir Implications for Financial Stability”, ECB Occasional PaperNo 34, August.


Finally, the last broad rationale for banks is themarket intelligence that can be obtained as part ofthe trading process, given that hedge funds havebecome increasingly significant players indomestic and international financial markets.

Other more specific reasons included the use ofhedge funds as hedges for financial derivatives orstructured products sold to customers. Some banksalso mentioned that dealing with hedge fundsinvolved the valuable transfer of technical know-how and product information.

Some hedge funds would probably welcome theopportunity to acquire all services from the sameintegrated provider, thereby inducing banks toprovide all hedge fund-related services, including,for example, prime brokerage, custody andadministration. Thus, banks with a strong base inone of these areas might logically attempt to fosterother complementing areas.10

2.2.2 BANKS’ EXPERIENCE OF DEALING WITH HEDGEFUNDS AND TYPES OF SERVICES PROVIDED

Based on the data provided, nearly half of the largebanks that provided some quantitative data hadstarted their dealings with hedge funds on apermanent and sizable basis more than six years ago(see Chart 1). However, one-third of banks wererather recent entrants into the hedge fund market,

and this could have an impact on the quality of theirrisk management of exposures to hedge fundsowing to the complexity of such exposures and thetime needed to acquire some minimum experience.

Only five out of 20 of the large banks identifiedthemselves as prime brokers. All of these fiveinstitutions were involved in equity primebrokerage, four in synthetic and derivatives primebrokerage and two of them offered fixed incomeprime brokerage.

Most of the banks invested their own money inhedge funds, and substantially more banks madeallocations to funds managed by unconnectedmanagement firms rather than by the bank or itssubsidiaries. For investors, however, more banksoffered investments in hedge funds or structuredhedge-fund products managed/created bythemselves rather than by unconnected entities.

Cash and securities lending were the two mostcommon types of services offered to hedge funds;although trade execution, clearance and settlement,custody and fund administration services were alsooften provided to hedge fund clients (see Chart 2).

10 For example, one large EU bank has a strong base in fundadministration business and wants to develop itself more in adirection of a prime broker, although currently its financing portfolioconsists mostly of lending to funds of hedge funds.

Chart 2 Types of services provided to hedgefunds

(% of total, answers not mutually exclusive; 20 large banks from7 countries)

Chart 1 Banks’ experience of dealing withhedge funds

(% of total; 20 large banks from 7 countries)

during thelast 2 years(since 2003)

35%

3-6 years ago(1999-2002)

20%

more than6 years ago

(before 1999)45%

20%

25%

30%

40%

40%

45%

50%

55%

70%

0 20 40 60 80

credit lines1)

capital introduction

risk management services

custody services

fund administration

clearance and settlement

trade execution

cash lending

securities lending

Source: BSC.1) Unsecured short-term liquidity facilities.

Source: BSC.

13ECB


2 SURVEYRESULTS

2.2.3 DIRECT EXPOSURESAs mentioned above, it is useful to distinguishbetween four different types of direct exposuresthat banks can have to hedge funds. First, financingexposures arise from lending to hedge funds(repurchase agreements and other arrangements)or from credit lines for unexpected liquidityshortages. Second, investment exposuresoccur when banks make investments in hedgefunds managed either by the bank (includingsubsidiaries) or by unconnected managementfirms. A third type of direct exposures, tradingexposures, arises from trading in OTC (overthe counter) markets and is closely related tofinancing exposures, as both types involve credit(counterparty) risk. Finally, income exposuresare associated with the dependence on revenuederived from hedge funds.

A broad picture of banks’ exposures to hedge fundscan be formed using information available incommercial hedge fund databases. In addition tovarious hedge fund characteristics, such as capitalunder management and returns, these databasesalso make available the names of various serviceproviders, including prime brokers. An example ofsuch an analysis is presented in Table 1 and Table 2,which give some indication of the magnitude,concentration and risk of exposures to hedge fundsby selected EU prime brokers.11 This analysiscould also be used by supervisors as a starting pointfor a more detailed examination of banks’ dealingswith hedge funds.

11 For the global picture, see Garbaravicius, T. and F. Dierick (2005),“Hedge Funds and Their Implications for Financial Stability”, ECBOccasional Paper No 34, August.


15ECB


2 SURVEYRESULTS

groupsub

groupsub sub sub

Distribution of leverage Distribution of size

average leverage maximum leverage

% % % % % % % % % % % % % %

Other(Multi-

Strategy)Fund of

FundsTotal

number

CR1 of sub-

strategies

CR3 of sub-

strategies

Do not use

leverage 0 ≤ 100 100-200 > 200 0 ≤ 100 100-200 > 200 ≤ $100m$100m - $1bn > $1bn

4 72 51 86 32 31 17 21 22 15 28 3 64 35 1

1 3 67 34 69 37 25 16 18 3 19 12 16 15 63 37

1 15 54 57 94 17 33 39 6 6 35 30 6 13 72 26 2

5 34 56 91 12 24 50 6 9 18 56 15 56 38 6

1 7 43 100 29 14 43 14 29 14 29 29 29 71

4 75 100 75 25 25 25 50 75 25

2 7 43 100 43 57 29 29 43 57

1 14 15 93 100 13 87 60 27 67 33

5 6 83 100 17 17 67 17 67 33 67

7 10 70 90 60 30 10 30 10 60 40

5 4 37 62 86 5 38 38 11 8 35 27 14 19 97 3

14 57 100 29 29 36 7 14 43 14 79 21

4 4 100 100 100 100 50 50

2 100 100 100 50 50 100

2 2 100 100 50 50 50 50 50

2 100 100 50 50 50 50 50 50

3 33 100 33 33 33 33 67 67 33

1 1 100 100 100 100

1 100 100 100 100 100

1 100 100 100 100 100

9 66 343 24 66 24 35 25 13 4 26 26 13 11 65 33 1

107 810 2,921 32 68 42 27 12 12 7 20 11 14 13 67 31 2

116 876 3,264 31 67 40 28 13 12 7 20 13 14 13 67 31 2

8 8 11 - - - - - - - - - - - - - -

56 23 21 - - - - - - - - - - - - - -

78 55 56 - - - - - - - - - - - - - -

4.4 - 8.1 - - - - - - - - - - - - - -


17ECB


2 SURVEYRESULTS

groupsub

groupsub sub sub

Distribution of leverage Distribution of size

average leverage maximum leverage

% % % % % % % % % % % % % %

Other(Multi-

Strategy)Fund of

FundsTotal

capital

CR1 of sub-

strategies

CR3 of sub-

strategies

Do not use

leverage 0 ≤ 100 100-200 > 200 0 ≤ 100 100-200 > 200 ≤ $100m$100m - $1bn > $1bn

0.3 11.1 57 96 18 39 14 30 25 18 34 4 15 74 11

0.2 0.4 9.0 40 72 36 20 12 26 6 22 5 9 28 15 85

0.0 2.3 6.9 51 95 20 37 27 8 7 48 22 2 9 12 60 28

2.0 6.3 53 93 28 34 16 8 13 36 13 22 9 56 35

0.9 2.4 52 100 46 6 45 3 46 6 42 6 6 94

1.9 66 100 66 34 6 5 88 45 55

0.5 1.7 40 100 63 37 32 5 7 93

0.0 1.6 1.6 100 100 46 54 40 14 12 88

1.1 1.2 86 100 14 39 47 39 47 5 95

0.8 1.2 69 99 56 34 10 34 10 22 78

0.0 0.1 1.2 78 97 1 54 31 9 5 53 24 8 14 62 38

0.9 60 100 29 29 37 5 23 39 9 52 48

0.5 0.5 100 100 100 100 13 87

0.4 100 100 100 66 34 100

0.3 0.3 100 100 14 86 86 14 86

0.2 100 100 9 91 9 91 9 91

0.2 88 100 88 3 9 3 97 12 88

0.2 0.2 100 100 100 100

0.0 100 100 100 100 100

0.0 100 100 100 100 100

1.1 10.1 47.3 27 65 25 38 15 18 5 32 15 13 15 14 72 14

26.5 105.6 415.9 27 65 40 29 10 10 11 22 11 11 16 15 65 21

27.6 115.7 463.3 27 65 38 30 10 11 10 23 11 11 16 15 66 20

4 9 10 - - - - - - - - - - - - - -

83 23 23- - - - - - - - - - - - - -

100 59 57- - - - - - - - - - - - - -

4.4 - 8.1 - - - - - - - - - - - - - -


Financing exposuresTwo forms of hedge fund financing may bedistinguished: bridge (liquidity) financing andnormal cash or security lending for gearing. Theformer is designed to allow hedge funds:

– to manage unexpected liquidity shortages ofvarious origins;

– to remain fully invested (minimising cashdrag);

– to smooth out timing mismatches of proceedsrelated to investor subscriptions andredemptions; and

– to not miss attractive investment opportunitieswhen all available funds are fully invested.

The last option is particularly important for fundsof hedge funds, as the opportunity to invest inotherwise closed funds must be accepted at shortnotice.

The survey found that some banks explicitlyprohibit outright credit to hedge funds. As lendingto hedge funds is a balance-sheet-intensive activity,it was not surprising that smaller banks or banksthat were not prime brokers usually had minorfinancing exposures.

Size and growth. At the end of 2004, for the 14 largebanks from six countries (AT, DE, ES, FR, NL andSE) cash lending to hedge funds collateralised withsecurities (e.g. via reverse repurchase agreements)amounted to 1.5% of surveyed banks’ assets and50.2% of their Tier 1 capital.12 Across countries,the ratios ranged from 0% to 5.5% of assets andfrom 0% to 224% of Tier 1 capital (see Chart 3). Themaximum possible amount of credit lines availableto hedge funds stood at 0.6% of assets and 17.3% ofTier 1 capital for the smaller sample of banks(excluding one large country), and it variedsubstantially across countries from negligiblelevels to up to 1.1% of assets and 32% of Tier 1capital. The absolute amount of collateralised cashlending to hedge funds was almost €100 billion(€99.3 billion) and large banks from two countries

clearly dominated in the sample. For the smallersamples of banks, which also provided 2003 data,the lending and maximum possible amount of creditlines had increased 1.5 and 1.3 times respectively in2004.

Maturity. Banks provided very little information onthe maturity breakdown of lending, but the datareceived from two banks confirmed that cashlending to hedge funds was very short-term andmore than 80% of obligations were to become duein less than one month.

Unsecured lending. It should also be noted that, ingeneral, banks extended no or only minimalunsecured lending and many banks had policiescompletely forbidding the extension of unsecuredlending. Cash lending was usually fullycollateralised and sometimes even significantlyover-collateralised on a marked-to-market basis.Nearly all unsecured credits were concentratedamong a few banks and did not exceed €85 millionat the end of 2004.

Chart 3 Cash lending to hedge fundscollateralised with securities

(% of Tier 1 capital and assets; 14 large banks from 6 countries)

1.5%

50.2%

0 2 4 6 8 10

TotalCountry F Country E Country DCountry C Country B Country A

TotalCountry F Country E Country D Country C Country B Country A

500 100 150 200 250

% of Tier 1 capital (higher axis)

% of assets (lower axis)

Source: BSC.

12 One large bank provided some information only for June 2005, but itsnumbers, when provided, were included in calculations referring tothe end of 2004.

19ECB


EXECUTIVESUMMARY

Box 2

PRIME BROKERAGE BASICS

Historically, prime brokerage has developed from collateralised equity and bond financing,though more often it originated from the equity side of banks’ business. Collateralisedsecurity financing remains at the core of prime brokerage operations and it involves(i) stock or bond lending and (ii) cash lending against stock or bond collateral. Cash lendingis usually implemented via reverse repurchase agreements (re-repos) and security lendingthrough repurchase agreements (repos). The maturity of reverse repurchase agreements isusually overnight and they can be continuously rolled over until the termination offinancing.

In prime brokerage, all positions are broken down into longs and shorts. Then posted cashbalances are added to the difference between the market value of longs and shorts and theresulting equity margin (or equity in account) is compared against the minimum margin(house requirement) to calculate margin excess/deficit.

+ Long market value– Short market value+ Cash balances 0

Equity margin– Minimum margin

Margin excess/deficit 0 0

There is a wide variation among prime brokers as to how they determine minimum margin,which is increasingly based on VaR calculations across all positions. In portfolio-basedmargining, the addition of a hedging (risk reducing) position would give back initial margin,although individual haircuts on products would still remain additive. In the case of marginlock-ups, certain pre-agreed parameters and correlations are fixed for a given period, so thathedge funds can calculate the margin numbers themselves.

A key feature of prime brokerage is that it provides a centralised platform and, veryimportantly, one consolidated margin (collateral) for all dealings with and services providedto hedge funds. In response to client demand, banks bring increasingly more OTC tradingand other products onto this platform, thereby effectively disbanding the silo-approachused before and combining prime brokerage with trading desks. For example, in April 2005one large EU bank announced that Global Prime Services had moved FX, credit and fixedincome products onto its platform, as hedge fund clients were looking for a fuller range ofproducts.

Types of collateral. Stocks and bonds were the twomost common types of collateral (see Box 2) inhedge fund financing. The relative shares of bondand equities varied depending on banks’ businessprofiles, but bonds seemed to dominate with an 80-100% share for a few banks that provided such

information. In the case of lending to funds ofhedge funds, the practices were less uniform, butusually the underlying investments in single hedgefunds were used as collateral and the overridingrationale behind leveraging was to amplify thediversification element of a portfolio. However,


even if funds of hedge funds were safer forinvestors, they could be riskier to lenders becauseshares of underlying hedge funds carry asubordinated credit status.

Breakdown by hedge fund size. Information fromone large prime broker on credit exposures to hedgefunds by size indicated that 60% of claims were onlarge hedge funds with more than €1 billion undermanagement. The remaining 40% were claims onsmaller hedge funds with more than €100 millionunder management. Moreover, the share ofovernight credit was significant for large hedgefunds, but negligible for smaller hedge funds, asmost of their outstanding obligations were tomature within one month.

Breakdown by strategy. Data from three countriesindicated that most credit exposures wereconcentrated in two broad strategy groups, namelymarket neutral and multi-strategy funds, whichoften have higher levels of leverage owing to thenature of their strategies (see Chart 4).

The total amount of cash lending can be comparedwith data from the Bank for InternationalSettlements (BIS) on consolidated bank claims onprivate non-bank borrowers in offshore financialcentres (see Chart 5). However, this comparison is

subject to reservations due to the fact that claims onnon-banks may comprise substantial lending tospecial purpose vehicles and other non-hedge fundentities domiciled offshore. At the end of March2005, the exposures of EU15 banks to non-banks inoffshore centres constituted more than three-quarters of their total exposures to offshorefinancial centres and nearly half of all reportingbanks’ exposures to non-bank borrowers inoffshore financial centres.

Lending spreads. In most cases, collateralisedlending spreads over LIBOR or EURIBOR variedby the type of hedge funds and by collateral.However, they were generally the result ofnegotiations between a hedge fund and the bank,whereas the risk of lending was reflected in othercredit or collateral terms. Such practice wouldimply that banks were largely price-takers in ahighly competitive market and altered non-priceterms to achieve reasonable risk-adjustedcompensation. Some banks also mentioned thatmarket competition was indeed an importantpricing factor, but collateral terms were determinedrelatively independently of price.

Banks reported various spread ranges dependingon their experience of dealing with hedge funds andthe quality and liquidity of collateral provided as a

Chart 4 Credit exposures by strategy


directionalmarket neutralmulti-strategyother

Country C

Country B

Country A

Total

Country C

Country B

Country A

Total2004

2003

0 20 40 60 80 100

Source: BSC.Note: There were no credit exposures to event driven funds andfunds of funds.

Chart 5 Consolidated bank claims on privatenon-banks in offshore financial centres

(USD billion; amounts outstanding; Q4 1995 - Q1 2005;quarterly data)

Source: BIS.

all reporting banksEU15 banks

0

200

400

600

800

1,000

1996 1997 1998 1999 2000 2001 2002 2003 2004 20050

200

400

600

800

1,000

21ECB


2 SURVEYRESULTS

protection (see Table 3). Sometimes spreads alsodepended on the overall business volume and/orthe range and complexity of services used by aparticular hedge fund or the ability to use otherlenders. Larger, more diversified funds were oftenable to command lower spreads than smaller ones.

With regard to the evolution of spreads, a number ofbanks indicated that they had actually declined over2004, especially for lending to larger hedge funds,as competition in this segment was the most intense.

Investment exposuresAs was mentioned before, in many EU countriesinvestments in (funds of) hedge funds were the mainand sometimes the only form of direct links with thehedge fund industry. Banks saw such investments asa way of gaining attractive risk-adjusted returns andimproving the diversification of their investmentportfolios. In smaller countries, risk managementassociated with these investments was quite oftentransferred to a parent institution, and frequentlychosen (funds of) hedge funds were also managedby the entities from the same financial group.

At the end of 2004, the total amount of investmentsin hedge funds by 16 large banks from six countries(AT, DE, ES, FR, NL and SE) exceeded €9.4billion,13 although most of them originated fromlarge banks in two countries.14 More than half(51%) of this amount was invested in hedge fundsmanaged by the entities of the originating banking

group, although in three countries investments inhedge funds were predominantly (sometimesexclusively) managed by firms unconnected tobanks. Total investments constituted about 0.1% ofassets and 4.3% of Tier 1 capital, although acrosscountries these ratios varied from 0% to 0.3% ofassets and from 0.1% to 8.6% of Tier 1 capital (seeChart 6). In 2004, total investments by a smallersample of banks that also provided 2003 dataincreased by 52% and allocations to unconnectedhedge funds grew much faster.

Banks highlighted performance (market), liquidityand fraud risks as the most important ones fortheir investments in hedge funds. In addition,they also mentioned a range of operationalrisks, which could be mitigated through carefuldue diligence: adequacy of risk managementsystems and business administration capabilities;technological (IT) risks; transparency; dependencyon key manager(s); legal risks in case of disputeswith hedge funds. In order to protect themselvesfrom a prolonged decrease in net asset value(NAV)15, the redemption characteristics andliquidity profile of (funds of) hedge funds werealso taken into account when conducting initial due


14 The addition of investments by smaller banks from one EU countrywould increase total reported investments by €0.9 billion.

15 Net asset value: the total value of the fund’s portfolio less liabilities;also referred to as assets (capital) under management.

Single hedge funds:

125-200 bps (basis points), depending on the type of fund and collateral.

For large, diversified hedge funds, 12 bps for high-grade bonds and 18 bps for high-yield bonds; dropped from respectively 17 and 25 bpsduring 2004.

25-75 bps for equity financing, no variation during 2004.

100-150 bps for collateralised loans and rather stable in 2004.

Average spreads for unsecured lending have increased in 2004.

50-75 bps.

10-300 bps, depending on the collateral. Individual banks reported that spreads had dropped since mid-2004, during 2004 or during the lastfew years.

Funds of hedge funds:

80-100 bps, tightened from 120 bps during 2004.

80-160 bps, spreads higher than for single hedge funds due to higher liquidity risks.

Table 3 Collateralised lending spreads and their evolution

(country summaries and individual bank replies)


diligence. Several banks were also concerned aboutcatastrophic decreases in NAV in the event ofwidespread turbulence in financial markets, whichcould create “gap” risk16 for the rebalancingof structured hedge fund products. Suddensubstantial increases in assets under managementcould also be a source of concern owing to strategycapacity limits. Some banks also underlined therisk of a “style drift”, i.e. the risk that the managermay change or abandon the stated primary strategyor strategies without informing investors.

Trading exposuresAt the end of 2004,17 for five large banks from threecountries (DE, FR and SE) the estimated gross

value18 of OTC contracts outstanding with hedgefunds in derivatives made up 2.7 % of alloutstanding banks’ OTC contracts in derivatives.In the case of OTC interest rate derivatives the sharewas 2.4% (ranges are presented in Chart 7). Whenmeasured by notional principal rather than by grossmarket value, hedge funds accounted for a muchlarger share of outstanding OTC contracts inderivatives.

The structure of outstanding OTC contracts inderivatives by the type of instrument is provided inChart 8, from which it is quite difficult to discernspecific bank trading patterns with hedge funds incomparison with the overall trading structure, eventhough banks slightly tended to deal relatively morein interest rate derivatives with hedge funds andless in other, perhaps more exotic, derivatives.

One bank also provided data on trading volumesand, in 2004, trading with hedge funds accountedfor 5% of overall trading in debt securities.

16 The risk that an investment’s price will change from one level toanother with no trading in between. Usually such movements occurwhen there are adverse news announcements, which can cause aninvestment’s price to drop substantially from the previous day’sclosing price.


18 Gross value refers to the pre-netting, pre-collateral marked-to-market value as the sum of both positive and negative (without minussign) exposures.

Chart 6 Investments in hedge funds

(% of Tier 1 capital; end-2004; 16 large banks from 6 countries)

4%

3%

2%

100 2 4 6 8

Total

Country E

Country C

Country A

Country B

Country F

Country D

managed by unconnected management firmsmanaged within a group

2%

8%

4%

1%

1%

2%

Source: BSC.

Chart 7 Hedge fund share of OTC contracts in derivatives

(% of outstanding gross market value; end-2004; (% of total outstanding notional principal; end-2004;5 large banks from 3 countries) 3 large banks from 3 countries)

Country ACountry BCountry C

2% 2% 3%

1%

3%3%

2%

3%2%

4% 4%

0

2

4

6

8

10

0

2

4

6

8

10

12 12

allderivatives

interest ratederivatives

creditderivatives

otherderivatives

Country ACountry BCountry C

2%

0

2

4

6

8

10

0

2

4

6

8

10

12 12

allderivatives

interest ratederivatives

creditderivatives

otherderivatives

5%5%

3%2%

4%

5%

8%7%

8%

11%

Source: BSC.

23ECB


2 SURVEYRESULTS

However, the share of trading with hedge funds atthis bank had increased in 2004 from 12% to 21%for all derivatives and from 4% to 18% for creditderivatives.

As regards markets in credit risk transfer (CRT)instruments, data from nine large banks in threecountries suggested that hedge funds were netcredit protection buyers from banks (in eachcountry, 60%, 62% and 70% of contracts wererelated to the selling of credit protection to hedgefunds) and based on other banks’ comments it alsolooked as if banks mainly acted as net creditprotection sellers to hedge funds. However, therewere also qualitative comments that, on aggregate,the hedge fund industry was probably a net buyer ofcredit risk, i.e. net credit protection seller. Otherbanks reported that they (i) did not generally deal inCRT markets, (ii) did not trade in these productswith hedge funds at all, (iii) did not havetransactions in credit derivatives with hedge fundsor (iv) did not normally use hedge funds as acounterparty and trading volumes with them inCDS (credit default swap) markets were very low.It was also mentioned that in CDO (collateraliseddebt obligations) markets, hedge funds wereincreasingly purchasing equity and mezzanine“tranches” of securitisations underwritten bybanks.

Overall, it seems that in most cases hedge fundswere not key counterparties for banks in CRTmarkets, but some banks confirmed that theyhad become important players in these marketsand their importance was likely to increase. Despitethe relative immaturity of CRT markets, theyseemingly managed to withstand the test of GM andFord downgrades and negative effects were ratherwell-contained, although some banks and hedgefunds suffered huge losses.

Income exposuresAccording to quantitative data from nine large banksfrom four countries (AT, FR, NL and SE) banksearned nearly €0.8 billion from hedge funds in 2004.However, the share of net income derived fromhedge funds was quite modest in relation to totalnet income and its sub-components, althoughproportions were higher for net tradingcommissions (see Chart 9).19 Across countries, nettrading commissions made up the largest shareof total net income derived from hedge funds (seeChart 10). Moreover, for the smaller sample ofbanks that also provided 2003 data the growth oftotal net income and its sub-components derivedfrom hedge funds was much faster than the netincome growth from all activities and hedge fundsmade a positive contribution to the selected banks’

19 For the selected group of eight large banks, net trading commissionincome accounted for 47% of net non-interest income.

Chart 8 Distribution of OTC contracts in derivatives

(% of total outstanding gross market value; end-2004; (% of total outstanding notional principal; end-2004;9 large banks from 5 countries) 4 large banks from 4 countries)

1000 20 40 60 80

Country E

Country D

Country C

Country B

Country A

Country E

Country D

Country C

Country B

Country A

interest rate derivativesother derivativescredit derivatives

with hedge funds

all OTC contracts in derivatives

71%

87%

73%

58%

86%

85%

71%

58%

25%

12%

42%

37%

11%

9%

29%

25%

1000 20 40 60 80

interest rate derivativesother derivativescredit derivatives

Country E

Country D

Country C

Country B

Country A

Country E

Country D

Country C

Country B

Country Awith hedge funds

all OTC contracts in derivatives

87%

82%

56%

92%

76%

74%

12%

44%

23%

24%

14%

Source: BSC.


net income in 2004 (see Chart 11). This positivecontribution may further intensify banks’ efforts tofoster hedge fund-related services and to attractmore hedge fund clients, most likely putting furtherpressure on applied price and non-price businessterms.

Expectations regarding the future evolution ofdirect exposuresMany banks noted that their direct exposures tohedge funds were growing rapidly and banksanticipated further growth, although in most casesthese exposures would probably remain ratherlimited when compared to balance sheets and totalincome. Banks also noted that their internalbusiness decisions would also influence theultimate role of hedge funds in their operations.

Many banks surveyed expected that financing andtrading exposures would grow based on businessvolumes and would play an increasingly significantrole in the relationship between banks and hedgefunds. At the same time, several banks expresseda determined wish to preserve the policy ofdisallowing unsecured counterparty exposures tohedge funds. Some banks also reported anincreased hedge fund demand for credit lines forunexpected short-term liquidity shortages andgreater interest in longer-term (up to 6-12 months)funding via repurchase agreements as a means of

locking and managing their liquidity profiles.Together with more frequent requests for marginlock-ups and fixed haircuts, these attempts tolengthen the maturity of liabilities represent hedgefund efforts to minimise funding liquidity risksstemming from increasingly illiquid assets.

Some banks were confident that trading exposures,especially in OTC markets, would growsubstantially as hedge fund demand for variousplain-vanilla and sophisticated OTC contracts wasincreasing rapidly. One bank even stated that itexpected its trading exposures to hedge funds todouble over the next two years.

Chart 9 Share of net income derived fromhedge funds in 2004

(8 large banks from 4 countries)

Country ACountry BCountry CCountry DTotal

3%

6%

3%

1%0

3

6

9

12

0

3

6

9

12

total netincome

net tradingcommission

net other non-interest income

net interestincome

Source: BSC.

Chart 10 Structure of net income derived fromhedge funds in 2004


net trading commissionother net non-interest incomenet interest income

0 20 40 60 80 100

Total

Country D

Country C

Country B

Country A

57%

45%

79%

55%

34%

100%

53%

15%

9%

30%

20%

Source: BSC.

Chart 11 The growth of net income derivedfrom hedge funds during 2004


net income derived from hedge fundsall activities

total netincome

net tradingcommission

net other non-interest income

net interestincome

26%

14%

42%

-2%

3%

-1%

39%

-11%-20

0

20

40

60

-20

0

20

40

60

Source: BSC.

25ECB


2 SURVEYRESULTS

The future of investment exposures will largelydepend on the performance of existing hedge fundinvestments, particularly given concerns aboutscarcer new profit opportunities, although otherfactors might drive their evolution as well. Somebanks thought that investment exposures wouldincrease owing to the expected further “retailisation”of hedge funds, and that banks would becomeinvolved by adopting their own positions, managingclient positions or selling hedge fund-relatedproducts. Low yields elsewhere and appealingdiversification effects were mentioned as otherreasons why investments should increase further.However, one bank considered that the importanceof investments in hedge funds for diversificationpurposes would decrease, as banks would try toestablish their own hedge funds or implement similarstrategies within their trading activities.

Across sampled banks, income exposures did notseem to be very high in relation to total income andfuture growth expectations were generally moderate,even though some banks anticipated strong growthand were quite optimistic about the future prospectsof revenues derived from dealings with hedge funds,mainly because of expected higher financing andtrading volumes. Some other banks also expectedthat, over the longer term, fees and profit margins forhedge fund-related business would compress owingto strong competition and more transparency.

2.2.4 INDIRECT RISKSApart from direct exposures, banks face a numberof indirect exposures to hedge funds. As mighthave been expected, according to banks, asignificant indirect risk stemming from hedge fundactivities was banks’ credit exposure towardsfinancial institutions with large exposures to hedgefunds. Such indirect risk might not necessarilymean the default of a financial institution, as evenpayment problems involving a major prime brokercould have important contagious financial stabilityimplications for the global financial system.

The second cited important indirect risk was relatedto hedge fund activities in financial markets in theevent of a major LTCM-type hedge fund failure orunexpected international macroeconomic or creditevents, leading to the forced selling by hedge funds,

the drying-up of liquidity and spillovers to otherhedge funds (domino effect). In case of such marketdislocations, banks were concerned about thepotential impact on their trading positions or theirreputation, if they were known to be heavily relianton hedge fund business. Some banks also notedthat CRT and other derivatives markets withhighly concentrated activity could be particularlyvulnerable to such disruptions. The “crowding”of hedge funds’ and banks’ trades or similarpositioning across a number of markets was alsomentioned as an important source of indirect risk.

As regards the loss of asset management income,banks did not see this as a significant indirect riskand actually regarded it as being rather lowcompared to other indirect risks.

2.2.5 OTHER RISKSThere are also other types of risks arising frombanks’ connections with hedge funds. First, thereis a legal risk regarding the enforceability of nettingand collateralisation provisions contained invarious agreements with hedge funds. This isparticularly important given the greater use ofcross-product netting for collateral and marginpurposes in OTC derivatives. Second, hedge fundsrepresent sophisticated clients which requireextensive operational support and the possibility oftrading in a wide array of sometimes very complexinstruments, potentially challenging banks’operational and risk management capabilities.Third, the high and increasing volume of hedgefund transactions increases operational risk relatedto front, middle and back office operations (e.g.calculation of risk, valuation and settlement ofcollateral). Fourth, some banks were concernedabout reputation risk in the case of theirinvolvement in dealings with a hedge fund facingfraud issues or publicised failure.

2.3 RISK MANAGEMENT ISSUES

2.3.1 INTERNAL RULES FOR DEALINGS WITHHEDGE FUNDS

Banks surveyed with significant exposures tohedge funds usually had specific internal rules orcontrols covering their dealings with hedge funds.Other banks with less strong links to hedge funds


generally relied on various risk committees andestablished policies to oversee their dealings withhedge funds. In some countries, parent institutionswere responsible for the risk management ofinvestments in hedge funds. In some cases, thespecific policies were broader and covered either allhighly leveraged institutions or all alternativeinvestment vehicles. In those cases where specificguidelines were not in place and banks did notexpress an intention to develop them, supervisorsmight consider reviewing banks’ connections withhedge funds in order to ascertain whether the scaleof hedge fund-related activities warrants anyspecific internal rules in line with therecommendations of the Basel Committee onBanking Supervision.20

Guidelines usually included a general descriptionof (funds of) hedge funds, due diligenceprocedures, credit and counterparty risk policies,the risk monitoring framework and legaldocumentation requirements. In some cases, suchdocuments also included broader information onthe hedge fund industry and the bank’s positionwithin it or a general business strategy vis-à-vishedge funds. Some banks also had a dedicated riskmanagement unit within their risk managementcolumns.

In those cases where hedge funds were primarilyscrutinised by various risk managementcommittees in the framework of general riskpolicies, the credit or counterparty riskmanagement units were always involved beforeand after engaging in lending or tradingrelationships. New investments often had to beapproved by senior management and/or productcommittees. Moreover, some banks specificallyhighlighted that they had caps, minimumdiversification, maximum concentrationrequirements or maximum risk levels with respectto investments in hedge funds or other hedge fund-related activities.

2.3.2 DUE DILIGENCE, CREDIT ANALYSIS AND RATINGSYSTEMS

The banks surveyed shared many generalsimilarities in their due diligence procedures andminimum requirements for hedge funds, although

with some differences in the details and weightsattached to them. Some banks with less significantexposures to hedge funds, apart from standard newclient procedures, were also asking to fill in aspecific hedge fund questionnaire; whereas otherswere applying standard procedures developed fortraditional funds with some adaptations, or pointedout that they intended to set up specific duediligence guidelines for hedge funds.

At the outset of the business relationship, in creditanalysis and in rating or scoring models, usually anumber of factors were evaluated which may bebroadly grouped in the following way:

– general characteristics, such as current andminimum size as measured by NAV, capitalstructure by types of investors, regulatoryregime of a domicile, affiliation with a financialgroup and quality of a prime broker,administrator, auditor or custodian;

– management quality, which encompasses notonly the track record, but also an ability to carryout administrative duties and ensure businesscontinuity (operational capabilities, back-upsystems, etc.). Some banks would declinedealing with a hedge fund if the manager had lessthan four years experience, or three yearsexperience in a relevant strategy. Bank answersalso seemed to indicate that a manager’sexperience and past performance – in absoluteterms, on a risk-adjusted basis, throughmaximum drawdown or as proxied by the size ofcurrent NAV – had quite a heavy weight in duediligence and credit analysis processes;

– risk profile as described by the volatility of pastperformance, strategy characteristics, portfoliobreakdown by geographical areas, asset typesand liquidity of components, average andmaximum permissible leverage, funding risk(e.g. history of investor subscriptions andredemptions) and investor exit rules (e.g. lock-up periods, redemption frequency). Co-investment of own money by a hedge fundmanager was rarely mentioned in bank answers

20 For further details, see Section 3.

27ECB


2 SURVEYRESULTS

as a factor potentially reducing the motivationfor excessive risk-taking;

Nearly all banks have recently observed somelengthening of lock-up periods, particularly byhigh-profile start-ups with up to two or eventhree-year lock-up periods. Larger or successfulfunds were also becoming much more restrictivewith respect to their redemption terms and werereducing redemption frequency and/orlengthening redemption notice periods, butoften with some “gate” fees for earlyredemptions. These developments werefacilitated by strong investor demand and, atleast partly, reflected hedge funds’ efforts toaccount for the higher share of less liquidinvestments. Most banks would prefer not to belocked up for more than one year together withother redemption terms that match theunderlying investments of the fund;

– risk management quality, i.e. risk monitoringtechniques, limits and deviations from limits;

– disclosure in terms of frequency, detail andquality.

In addition to the elements listed above, some banksspecifically mentioned that they also perform on-site visits and subsequent follow-up visits to checkthe operational set-up. Certain banks alsohighlighted that, before establishing a creditrelationship, they required hedge funds to covenantnot to make any borrowings in any form from anyother party (“sole lender requirement”), therebyensuring that they would remain the sole primebroker. However, such agreements seemed to bemore the exception rather than a rule.

In the due diligence process, many informationitems were usually checked, including offeringmemorandum/prospectus, marketing materials,investor newsletters, audited annual and interimfinancial reports and available information fromthird-party sources, including commercialdatabases.

The use of some forms of rating or scoring systemsin credit analysis was widespread among banks, as

82% or 18 out of 22 large banks from sevencountries (AT, DE, ES, FR, NL, PT and SE) thatprovided such data had indicated that they usedrating or scoring models for their hedge fundclients. Some of these models were quite advanced,using probability-of-default methods or otherquantitative approaches, and banks expressed theirintentions to ensure that these models comply withBasel II requirements. Interestingly, one of thebanks has also developed a model to rate the qualityof a fund manager. In addition to quantitativecriteria, qualitative factors were also given aprominent role in rating methodologies and nearlyalways included transparency questions.

As mentioned before, the tendency to place heavyreliance on a manager’s track record raisesconcerns as other criteria may receive less attentionthan would appear appropriate. Furthermore, manybanks did not mention on-site visits as being part ofthe due diligence process and it was not alwaysclear how much importance they assigned to them.The value of such visits is underscored by the factthat quite often hedge funds fail due to operationalissues, including the misrepresentation of fundinvestments, misappropriation of investor funds,unauthorised trading and inadequate resources.21

2.3.3 LIMITS, EXPOSURE MEASUREMENT ANDCOLLATERALISATION

LimitsCredit limits in banks were generally based on theoutcome of due diligence, credit analysis and ratingor scoring models and were set in absolute amountsor in relation to the NAV or total liabilities of afund. Once again, surveyed banks’ repliesindicated that there was strong reliance on thereputation, experience and performance of amanager (especially during market stress periods),the quality of risk management, operationalcapabilities and disclosure. Limits also consideredthe maximum leverage imposed on hedge fundmanagers in hedge fund offering documents.

At all banks with higher hedge fund exposures,credit limits incorporated not only current credit

21 See, for example, Kundro, C. and S. Feffer (2004), “Valuation Issuesand Operational Risk in Hedge Funds”, Journal of FinancialTransformation, Vol. 11, Capco Institute, August, pp. 41-47.


exposures (CCE) but also potential future creditexposures (PFE) 22, 23 (see Box 3), although 19% orfour out of 21 large banks from seven countries(AT, DE, ES, FR, NL, PT and SE) that reported onthis question indicated that they only used CCE.Most banks applied PFE limits across all products(financial instruments), or planned to do so,although some banks were still using them only forderivatives business. The survey did not covernetting practices in detail, but according to banks,they were usually netting trades by product and lessfrequently across product lines (cross-productnetting) (provided appropriate legal agreementswere in place) before applying limits on a net-of-collateral or gross (pre-collateral) basis. However,the netting of trades among subsidiaries was rare(cross-affiliate or cross-entity netting). Thequantitative data provided indicated that mostbanks have been setting limits on a net-of-collateralbasis, although the use of gross or both sets oflimits was also widespread (see Chart 12). Pre-collateral limits have one advantage, namely thatthey are capable of accounting for the secondarymarket effects associated with the forcedunwinding of larger trades or collateral.

As mentioned before, some banks had total caps forall hedge funds and caps per hedge fund, per hedgefund manager, per strategy for investment,financing or trading exposures in order to avoid

concentration risk. For investment exposures,some banks also had VaR and stop-loss limits.

Based on the data provided (see Chart 13), morethan half of the banks had at least annual limitreviews. However, many banks also indicated thatthey would review limits whenever necessary.

22 Current credit exposure (CCE) is equal to the value of creditoutstanding or the replacement cost of trading positions, often afternetting by and across financial instruments and net of the marked-to-market value of collateral. By contrast, potential future creditexposure (PFE) takes into account the possible variations in the valueof CCE over the life of a contract. PFE can also incorporate portfolioeffects as well as secondary effects associated with positionreplacement and collateral liquidation. Total credit exposure isgenerally measured as the sum of current and potential futureexposure.

23 One large bank specified that the sum of loss thresholds (usuallynegotiated to be zero), minimum transfer amounts and PFEs could notexceed 5% of fund’s NAV.

Chart 12 Limit setting practice


net-of-collateral45% gross limits

25%

both methods used(two sets of limits)

30%

Source: BSC.

Chart 13 Frequency of limit reviews

(% of total banks, answers not mutually exclusive;21 large banks from 7 countries)

33%

14%

10%

10%

5%

52%

19%

0 10 20 30 40 50 60

whenever necessary, no set frequency

automatic, based on risk profile

1-31 days

1-3 months

3-6 months

6-12 months

other

Source: BSC.

29ECB


2 SURVEYRESULTS

Box 3

POTENTIAL FUTURE CREDIT EXPOSURE (PFE) MEASUREMENT

Banks’ answers provided only a broad description of their PFE measurement methodologiesand the review provided below should be interpreted as a brief snapshot of banks’ practices.Typically, PFE was based on a marked-to-market plus a product-specific add-on approach.

In the simplest case, add-ons depended upon the underlying asset and the maturity of thecontract in accordance with Basel I rules or tables of statistical estimates were used,although banks noted that more sophisticated models were in preparation.

In most other cases, various forms of the VaR approach were used, according to which PFEwas usually calculated as a peak exposure over the chosen holding period with a 95%, 97.5%or 99% confidence interval. Most banks used remaining or fixed seven or ten-day horizonsto calculate add-ons for the most types of trades, whereas some other banks used liquidity-adjusted VaR depending on the liquidity of collateral or by selecting an appropriate holdingperiod, as they thought that this was more conservative than taking a standard holdingperiod across different hedge fund strategies. For example, a one-day holding period mightbe used for a long/short equity fund client and ten or more days for a distressed debt hedgefund portfolio. In some cases, VaR-based add-ons were complemented with the results ofstress tests, in particular when setting floor values on add-ons for short-term contractsinvolving credit risk or accounting for “gap” risk in CPPI (constant proportion portfolioinsurance) structures. 0

There were also other sophisticated approaches for determining PFEs. Of those, variousMonte Carlo simulations were the most popular with 95% or 99% confidence intervals andsometimes for many different time horizons within the full lifetime of a trade. In other cases, PFEfor derivatives was modelled by assuming that financial variables, which underpin the valuesof derivatives contracts, follow a geometric Brownian motion, and by adjusting calculations,when necessary, for the forward nature of contracts or transactions with a payout structure. Ifthe underlying variables belonged to emerging markets, FX or commodity products a jumpdiffusion model combined with a Brownian motion or time-dependant volatility was employed.Furthermore, sometimes methodologies were risk-specific, as banks used Monte Carlosimulations for interest rate risk, a variance-covariance approach for FX instruments and adelta-based approach with exposure boundaries for equity derivatives.

In addition to the most common confidence levels mentioned above, some banks stated thatthey were using state-of-the-art methodologies with an extremely high (99.99%) confidencethreshold.

With respect to the advancements in PFE measurement techniques, some banks informedthat in their modelling they had started accounting for the correlation between the exposure(market risk) and the default of the counterparty (credit risk) and had improved theirunderstanding of counterparty exposure portfolio dynamics through the analysis of thesensitivity of credit valuation adjustments to the various market and credit risk parameters.Other progress was related to the ability to account for non-normal distributions, toincorporate longer time horizons or include more products in PFE models.


Stress testingIt is well known that the usage of stress tests couldsignificantly improve risk evaluation based solelyon VaR concepts, drawdown analysis or othermethodologies. Financing, trading and investmentexposures to hedge funds are so complex that thevalue of stress tests is indispensable. They could beutilised for the calculation of PFE, limit setting,accounting for “gap” risk in CPPI structures or theestimation of collateral liquidation values andassociated haircuts.

Banks with larger exposures to hedge fundsperformed stress tests on a regular, less frequent oron ad hoc basis. Most stress tests, particularly theregular ones, included historical scenarios. No bankindicated that it had tried to identify potential newstresses involving new correlations, which would bemore relevant for rapidly evolving modern financialmarkets. Banks listed a number of applied historicalscenarios: Black Monday in 1987, the UK’swithdrawal from ERM, the near-default of LTCM,the market turbulence of 1998, the tech bubblebursting in October 2000, the terrorist attacks of 11September 2001 and the Enron crisis in April 2002.Banks surveyed did not generally perform “majorplayer exit” stress tests, as only one bank reportedthat it tests the robustness of its hedge fund portfolioby simulating a sudden exit of an importantcounterparty. Another bank stated that it wouldperform market stress tests on specific componentsof its overall trading portfolio, which it views assensitive to “major player exit” scenarios. However,it will only do so on an ad hoc basis.

According to the data provided by 20 large banksfrom seven countries (AT, DE, ES, FR, NL, PT andSE), most banks were stress testing their individualexposures to hedge funds (see Chart 14), although

in the qualitative answers some banks alsomentioned that they employed stress tests on atransaction basis as well. However, the fact thatsome banks were not using stress tests at all or thatthey did not stress test consolidated exposures tohedge funds is not encouraging and these banksshould consider expanding the application of thisrisk management tool.

Some banks indicated that they used stress testsdeveloped specifically for hedge funds, whereasother banks used the same tests as for traditionalfunds. One bank was more specific, explaining thatit was stressing exposures to market risk factorsone by one and all factors at the same time. Inaddition, relevant parameters also includedstrategies, leverage, liquidity, correlation anddiversification. Another bank used stress testing toidentify shortfalls in collateral, i.e. to compare theincrease in exposure with the excess collateral held.As for the results of specific stress tests on hedgefund investments, one bank reported that, based on

Based solely on the descriptions provided, without more rigorous examinations of themethodologies applied, it is not possible, if at all possible ex ante, to judge whether bankssufficiently account for the interaction of market, credit and asset illiquidity risks. In anycase, banks should continue or sometimes strengthen substantially their efforts to modelhedge funds’ ability to meet margin calls and other liquidity demands after losses onleveraged market positions, particularly in times of stress when affected markets are lessliquid and the unwinding of hedge fund positions or liquidation of posted collateral could beproblematic.

Chart 14 Scope of stress tests

(% of total banks, answers not mutually exclusive;20 large banks from 7 countries)

10%

70%

15%

30%

0 20 40 60 80

no stress tests

individual hedge fundexposures

exposures to groups of hedgefunds (strategies)

all exposures to hedge funds

Source: BSC.

31ECB


2 SURVEYRESULTS

applied historical scenarios, directional strategiesgenerally performed better in times of stress thanother hedge fund strategies or more traditional assetclasses.

Regarding funds of hedge funds, one bank reportedthat its monthly stress tests included individualhedge fund failure, combinations of hedge fundfailures and failures of hedge fund strategies asdefined by sharp drops in NAV. The resultsshowed that the portfolio was well-diversified andlarge losses were experienced only in scenarios thatcorresponded to the complete failure of around25% of hedge funds in the portfolio. Furthermore,it reported that the largest losses were not likely toarise from market risk but from illegal activities onthe part of the fund of hedge funds manager.Furthermore, another bank used stress tests todetermine the adequacy of collateral (i.e.underlying investments in single hedge funds)posted by a fund of hedge funds.

CollateralisationThe survey revealed that banks had rather diversepractices regarding unsecured exposures, initialmargins,24 minimum transfer amounts25 and lossthresholds.26

Generally, banks had strict policies regardingunsecured exposures, as many of them had zerolimits for unsecured exposures to hedge funds,never transacted with hedge funds on an unsecuredbasis and usually did not approve loss thresholds;there was actually little pressure for higher lossthresholds or a strong reluctance from banks toapprove them at high levels. However, in somecases, there was still some leeway for unsecuredexposures or, as discussed below, trading wassometimes conducted on variation margin only.

Similar to trading in a derivatives exchange, tradingin OTC instruments usually requires initial margin(collateral) in the form of securities or cash. The useof cash as collateral has become widespread asaccording to the International Swaps andDerivatives Association (ISDA), at the beginningof 2005, cash accounted for nearly three-quartersof the collateral held to support derivativesexposures. This in turn helped to alleviate concerns

about the availability of high-quality collateralsufficient to meet the growing demands in thesemarkets.27 An initial margin is set at the inception ofa trade and then kept unchanged during the life ofthe trade or recalculated on a daily basis withtransfers of associated variation margin, ifnecessary. For transactions that fall under primebrokerage, real-time monitoring and dailymargining is common practice. Some primebrokers used portfolio-based margining, asopposed to margin charges on a trade-by-tradebasis. According to banks, portfolio-basedmargining28 will probably become the marketstandard, even though only a quarter, or five out ofthe 20 large banks from seven countries thatprovided quantitative information on this questionindicated that they used cross-margining.

As reported in Chart 15, quite a number of banksdid not actually require initial margins at theinception of trades; this raises concerns becauseinitial margins are a very important risk mitigant indealing with hedge funds. Some banks were trading

24 Initial margin – amount of collateral, which has to be posted beforetransaction.

25 Minimum transfer amount – amount of collateral below whichcounterparty is not required to transfer collateral.

26 Loss thresholds – exposures below which no collateral is posted.27 Greenspan, A. (2005), “Risk Transfer and Financial Stability”, May.28 Portfolio-based margining – margin offsets based on past

correlations of positions. Sometimes also referred to as cross-margining or the VaR-based margin.

Chart 15 The use of initial margins andtwo-way margining

(% share of hedge funds at each bank or average share of hedgefunds weighted by individual bank assets in each country;11 large banks from 4 countries)

3%0%

0 20 40 60 80 100

Country BCountry A

Bank 4Bank 3Bank 2Bank 1

Country BCountry A

Bank 4Bank 3Bank 2Bank 1

initial margin is not required at the initiation of a tradetwo-way margining is applied

65%100%

40%100%

3%81%

41%58%

100%65%

Source: BSC.


with hedge funds only in liquid local plain-vanillafixed income and FX instruments and the reliancesolely on variation margin was consideredrelatively less risky. More precisely, some banksspecified that the initial margin:

– was a function of the volatility and liquidity ofthe reference asset combined with theassessment of a fund’s creditworthiness;

– depended on the credit score of the hedge fundclient and the quality, volatility and liquidity ofunderlying collateral; 29

– was derived from standard market agreementsfor certain types of transactions, such as securitylending; and

– depended on hedge fund’s internal rating andadd-on risk.

Haircuts would also provide additionalcollateralisation, when applied to the market value ofsecurities that were posted as collateral (initialmargin) and that were less liquid or with longerduration. One of the features of dealing with hedgefunds is that both the OTC exposure and the collateralcovering it can have a negative correlation; thus,robust haircuts must be applied to protect themfrom such scenarios. Haircuts are common incollateralised financing for the securities that aredelivered as collateral.

Each day, the marked-to-market value ofoutstanding positions and collateral is recalculatedand the variation margin is transferred, subject tominimum transfer amounts. If a two-waymargining is applied, then both the bank and thehedge fund have to transfer the appropriate amountof collateral. By contrast, in a one-way marginingonly the hedge fund has to transfer collateral, andthus hedge funds cannot use excess collateral forother purposes until the end of a trade. Some banksreported that two-way margining had become (orwas already for some time) standard in the market;whereas in other cases non-domestic funds wereinsisting on it or exceptions were only granted for afew very large and well-established hedge funds(see Chart 15). With regard to minimum transfer

amounts, it was detailed that they typicallyamounted to $0.25 million or €0.1-0.5 million.

Most banks also indicated that by default, they onlyaccepted liquid collateral from single hedge funds,such as cash or cash equivalents (e.g. G10government bonds). Some banks were even stricterand accepted only cash for trading in OTCderivatives. Moreover, some replying banks wereusing VaR or stress tests to estimate possiblevariations of collateral value and to determineappropriate haircuts, but only a few of themreported the application of stress-tested collateralliquidation values.30 If underlying investments insingle hedge funds were used as collateral by a fundof hedge funds, then redemption characteristics ofthese underlying funds were also taken intoaccount. The rather limited use of stress-testedcollateral liquidation values indicates another areawhere banks need to expand the use of stresstesting.

As an additional point, some banks noted thatcollateral terms would also depend on whether thesame bank was a custodian of pledged collateral andwhether the bank was in charge of the fund’sadministration.

Regarding increasing hedge fund requests for termcommitments (and fixed haircuts) for marginfinancing, it should be noted that only a few primebrokers with high financing and trading exposuresreported that they would consider such requests,which typically involve fixing margin terms for 30or 60 (or very occasionally 120 or more) days.Other banks retained the right to change marginrequirements at any point in time. As banks makecredit terms stricter for hedge funds in tough times,hedge funds try to protect themselves from lessfavourable dealing terms at the time when suchterms would be most damaging. Hence, wider

29 In one case, the quality of collateral was estimated by averaging fiverisk components: price risk, liquidity risk, credit spread risk, issuerrisk and country risk. For a fund of hedge funds, the underlyinginvestments in single hedge funds served as collateral and “stress”rather than actual volatilities and pairwise correlations of sub-fundswere used to determine margin requirements. Moreover,concentration of investments and leverage employed by underlyingsingle hedge funds were other important model parameters.

30 One bank carried out stress tests on collateral values over a 3-10 dayliquidation period with a 95% confidence level.

33ECB


2 SURVEYRESULTS

usage of margin lock-ups or term margins, ifapplied prudently, could also be beneficial forfinancial stability, as it would lower pressures forhedge funds to liquidate investments rapidly intimes of stress owing to liquidity shortages.

Counterparty discipline and demands by largehedge fundsBased on banks’ comments, there was somegeneral sense of the erosion of counterpartydiscipline, although some banks think that thecurrent market standards should be quite similar –owing to the same legal documentation or similarinitial margin calculation methods used – and thatcredit terms were only eroded slowly. Banks alsoreported that hedge funds were, to some extent,successful in achieving more beneficial terms byselecting the most advantageous offers from a set ofbanks. Some of the largest hedge fund families havebeen able to trade with zero initial margins on someplain-vanilla instruments. Other banks reportedthat they had still escaped demands to accept lowerquality collateral or apply lower haircuts for sub-prime collateral. Depending on the country,different credit terms were subject to intensenegotiations (see Table 4).

When asked directly, banks generally deniedthat they would agree on softer limits to largerhedge funds due to competition, although somerecognised the impact of competitive pressures inthe industry and the positive relationship between ahedge fund’s business potential as proxied by itssize and relatively higher limits. In some cases,credit policy documents still gave some leewayfor individual negotiation, or qualitative analysisand/or decisions of senior credit committeeshad possibilities to overrule the outcomes ofquantitative models.

Owing to their higher NAV, larger hedge fundswould normally command higher absolute limits,but the relationship was not always automatic anddepended on the counterparty’s creditworthiness.It was also noted that sometimes larger and well-diversified (multi-strategy) hedge funds with goodtrack records and more sophisticated riskmanagement systems could represent lower risksthan smaller funds. They also quite often requiredbanks to provide a minimum amount of settlementlimits before dealing with them. Certain banks alsoindicated that recently lending spreads first of alldeclined to larger hedge funds, as competition for

• Early termination provisions, as hedge funds were insisting on:– higher NAV decline triggers, or– a definition of NAV based on NAV per share rather than on total NAV.In addition, certain hedge funds were requesting “super collateralisation” triggers (i.e. higher margin requirements) instead of earlytermination, and longer tenors.

• Margin terms for credit products.

• Initial margins are the area where competitive pressures are the most extreme.• Documentation:

– no key man clause, even if a key manager clearly exists;– “super collateralisation” rather than termination upon the breach of NAV decline triggers.

• Loss thresholds.

• Two-way margining was granted to some hedge funds.• Upward pressure on minimum transfer amounts.• One of banks stated that some institutions seemed to have extended unsecured credit facilities to gain a presence in the market.

• Initial margins, especially for equity financing.• No resistance to initial margins, but negotiations regarding applied percentages.• Bond repo haircuts.• Some of the largest hedge funds have a policy of not posting initial margin on bond repos. They also insisted on having NAV clauses

that relate only to negative performance (i.e. NAV per share).• Large US or UK-based hedge funds insisted on two-way margining, which is not common with Nordic counterparties.

Table 4 Counterparty discipline: examples of credit terms under pressure

(grouped by country; country summaries or individual bank replies)


the business with these entities was the mostintense. Moreover, in contrast to the general policy,in some cases only large hedge funds were grantedtwo-way margining, even though one bank stilleffectively avoided providing collateral in two-waymargining agreements by applying hedge fundrating-based loss thresholds.

As mentioned in the previous sub-section oncollateralisation, many banks had hedge fundclients that were on variation margin only, andalthough the size of these clients was not reported,such hedge funds were most likely the larger oneswith stronger bargaining power. Overall, therewere some indications that market discipline, as

applied by banks, was under pressure due to highlycompetitive market conditions. Even if there wereproper risk controls in place, banks’ risk appetiteseemed to be increasing.

LeverageGenerally, there are two forms of leverage: loan andderivatives-based gearing. Sometimes these twoforms are also referred to respectively as economic(debt) and financial leverage. The former is balancesheet intensive and usually provided by largerbanks, whereas the latter refers to the way in whichsmaller banks might compete for hedge fundbusiness with larger players and lies at the core ofsynthetic and derivatives prime brokerage.

Box 4

MEASURING LEVERAGE

Leverage is commonly understood to mean the ratio of total assets to equity or, in the caseof a fund, to NAV.1 The simplicity of this indicator is attractive, but such an accounting-based balance sheet measure of leverage fails to reflect the risk of the assets. Risk-basedmeasures of leverage, such as VaR/NAV or scenario (stress test) derived VaR/NAV,alleviate this shortcoming by relating market risk to the capacity to absorb it. However, risk-based leverage measures, even adjusted for potential asset illiquidity, do not capture thefunding risks arising from margin calls, redemptions or financing mismatches. The LTCMepisode has clearly underscored the role of funding liquidity in escalating the effects ofotherwise acceptable losses on market positions.

In the survey, most banks indicated that they usually used leverage as provided by hedgefunds in order to track it on a consistent basis over time, as information necessary todevelop a more complex, uniform measure for all hedge funds was not always available.Normally, hedge funds provided at least Assets/NAV ratio, although VaR/NAV was alsooften presented. Despite all VaR limitations, VaR/NAV is clearly a superior measure tosimple Assets/NAV, but VaR is usually not available for less liquid or exotic instrumentswith no historical price data as, for example, is usually the case for distressed debt. Toaccount for liquidity effects, certain banks were calculating VaR for specific holdingperiods. In other cases it was indicated that hedge funds largely preferred Longs/NAVindicator (sometimes longs included on-balance sheet assets and notional amounts of off-balance sheet long positions). Prime brokers, in contrast to other banks, were alsomonitoring VaR/(Equity in account).

One bank noted that it used different leverage ratios for different investment strategies andunderlined that leverage is best accounted for in conjunction with the market, credit and

1 There is some misunderstanding regarding the meaning of n:1 leverage, as it could mean that either total assets or liabilities are n times higherthan equity, resulting in assets to equity leverage multiple of respectively n or (n+1). More often, however, n:1 corresponds to (n+1) assets toequity leverage.

35ECB


2 SURVEYRESULTS

Regarding the prevailing level of leverage in thehedge fund industry, the most common view amongbanks was that the current level is much lower thanat the time of the near-failure of LTCM inSeptember 1998 and, according to one bank, evenas low as 10-25% of the 1998 level. However, itcannot be excluded that this view is based oninformation from market reports rather than on anown analysis.

One of the explanations for lower leverage was thathedge funds seemed to target lower but stablereturns and, as hedge fund investments in lessliquid assets had increased, they aimed at achievinga better asset-liability profile. Furthermore, otherbanks thought that this reduction could be due to thecurrent market liquidity or the growing influenceof institutional money which usually demandedmore conservative strategies. The conclusion of aninternal survey at one of the banks in 2004 was thatthe average leverage per hedge fund was rather lowand that almost half of the hedge funds did not haveany form of on-balance-sheet leverage.

By contrast, some banks felt that leverage had risenoverall over the past 18 months. Some thought thatleverage had actually increased owing to the use ofnew financial instruments with embedded leverage,such as total return swaps or options. One countryreported that leverage multiples (ranging from2x to10-15x for various fixed income funds) wereindeed significantly lower than in 1998, butpossibly slightly higher than at the end of 2004.

Judgements regarding the funds of hedge fundswere mixed. Several banks shared the view that

funds of hedge funds had reduced or were notcurrently increasing leverage, whereas most otherbanks argued that leverage had increased frompreviously low levels and that funds of hedge fundswere more leveraged in order to compensate forlower returns of single hedge funds. This opinionof increased leverage at the fund of funds levelseemed to be more prevalent.

Based on banks’ answers, the tentative inferencemight be that single hedge funds indeed operatedwith lower levels of leverage, whereas funds ofhedge funds tended to use leverage moreaggressively. There are, however, several caveatsto this conclusion. First, the absolute amount ofleverage in monetary terms is probably higher now,although it might be more widely spread out.Second, as mentioned in Box 4, figures on off-balance-sheet leverage are generally unavailable.Third, a larger number of hedge funds has beenchasing after the same profitable opportunities andthe resulting mediocre returns may have increasedincentives to employ more leverage, especiallyin an environment of very low interest rates.Moreover, overall leverage levels may also beaffected through the more widespread use ofmultiple layers of gearing at the level of investors,structured products, single hedge funds or funds ofhedge funds.

2.3.4 ONGOING MONITORING

Regular information provided by hedge fundsAfter initial due diligence and credit analysis,banks usually received regular monthlyinformation on NAV and performance (NAV per

liquidity risk of assets. However, reported leverage measures did not incorporate all theserisk factors, as for long/short equity and event driven strategies bank draw on(Longs+Shorts)/NAV; for fixed income strategies – Assets/NAV; and for global macro –VaR/NAV. Some other banks mentioned that the way they dealt with leverage datadepended on the size of credit exposures.

Another important observation is that banks generally did not have the possibility tocapture off-balance sheet leverage arising from trading in derivatives, perhaps partly owingto the silo approach used by banks and the resulting separation of prime brokerage andtrading activities. Another explanation, of course, is the inadequate disclosure of hedgefund positions.


share), in many cases together with riskmanagement reports. Quite often, all of thisinformation was obtained from monthly investorletters. In some cases weekly or even dailyestimates of hedge fund performance were alsoavailable and smaller hedge funds usually reportedweekly, whereas the larger ones reported onlymonthly. Final monthly reports were usuallyavailable 15-30 days after the end of the month. Thefrequency of audited financial statements wasusually annual, although sometimes more frequentunaudited interim accounts were also available.Many banks underlined that quarterly reporting orany other reporting less frequent than monthly, orless frequent than weekly performance estimates,was not sufficient for them, although some hedgefunds still provided only quarterly reports.

Many banks would normally refuse to dealwith hedge funds that did not provide aggregatedrisk reports with minimum predetermineddisclosure requirements. The transparency andcomprehensiveness of risk reports variedsignificantly, as some hedge funds were concernedabout revealing their trading activities to banks andcompetitors. However, information on VaR,leverage, performance attribution, main exposures,and current strategy mix or portfolio structure bygeographical area or by asset type was oftenincluded, or key and even all individual positionswere provided.

Banks nearly always relied on the informationprovided by an external third-party administrator, amanager or an advisor, and usually did not verifythe pricing of hedge fund positions themselves,bar plausibility checks or significant NAVmovements. Independence from hedge funds’managers, quality and reputation and of anadministrator together with the procedures forvaluing illiquid trades were considered during theinitial due diligence process and were reflectedin the overall rating of a hedge fund. Weeklyperformance estimates were usually received frommanagers, whereas final monthly figures camelargely from administrators.

In the case of funds of hedge funds, some banksalso attempted to look through and monitor the

performance of underlying investments, while alsounderscoring the additional layer of due diligenceprovided by fund of hedge funds managers. Themonthly information provided by funds of hedgefunds was more lagged than that from single hedgefunds owing to reporting lags by underlying hedgefunds. In one case, it was specified that weeklyreturn estimates for underlying single hedge fundswere available after three business days, whereasconfirmed monthly fund of hedge fund returnswere provided five weeks after the end of themonth.

Some banks seemed to be content with theinformation provided to them, despite reportinglags and the diversity of hedge fund disclosures interms of scope and comprehensiveness. Banks’answers also provided an indication that there stillwere funds that remained relatively opaque.Moreover, in ongoing monitoring and creditanalysis, there was, to some extent, trust and areliance on the reputation and different forms ofoversight conducted by prime brokers,administrators, auditors and other hedge fund-related parties.

Monitoring of exposuresFollow-up due diligence and scoring was normallyconducted with similar frequency as limit reviews(see Chart 13). In some cases, however, it wouldprobably be prudent to shorten regular reviewintervals to quarterly or, at least, half-yearfrequency. However, banks stated that they quiteoften phoned hedge fund managers to clarifyvarious questions, and some banks even hadagreements that hedge funds must deliverinformation that they, as counterparties, would besatisfied with and should answer questionswhenever they might occur. According toagreements with banks, hedge funds usually had tonotify banks immediately about certain importantevents, such as changes in management or strategymix. Hedge funds normally supplied theinformation requested.

In the case of an unwillingness to provide theinformation requested, some banks indicatedthat they would usually terminate businessrelationships, although other banks were more

37ECB


2 SURVEYRESULTS

flexible and sometimes preferred instead to assigna lower rating with commensurately moreconservative credit terms. The final decisiondepended on the overall risk profile of a hedge fundas well as on business considerations.

Most banks mentioned that their trading andsometimes also investment exposures to hedgefunds were included in trading portfolios and wereintegrated into the regular monitoring and reportingframework for proprietary trading portfolio risks,covering market and counterparty risks. However,in some cases trading exposures were too small tohave a more significant impact on the overalltrading portfolio.

Normally hedge fund exposures were monitored byfund, fund manager and strategy, although in somecases also by leverage, performance, region,market, hedge fund size and products (financing,fund of fund financing, derivatives, primebrokerage). However, sometimes the lack ofsufficient disclosure made it extremely difficult toprepare any meaningful reporting on leverage,strategies and some other indicators. Some bankswith mainly investment exposures examined hedgefund portfolios to analyse how concentrations bystrategy, market, asset class or currency fitted intotheir overall investment portfolios.

The qualitative information and survey datacollection process itself have revealed that somebanks with larger financing and trading exposuresto hedge funds had difficulties in aggregating theirexposures across the entire financial group and/ordifferent business areas/geographical regions. Oneof the reasons for this was the silo-based approachutilised by banks, resulting in different conceptsand risk management practices used in primebrokerage and other business areas, particularly intrading activities. Moreover, 15% of the largebanks sampled that provided relevant data (seeChart 16) indicated that they consolidated hedgefund exposures only across selected products. Oneplausible explanation might be that the scale ofexposures did not warrant a more focusedapproach. However, the fact that some banks(prime brokers) with larger financing and tradingexposures did not properly aggregate their hedge

fund exposures raises concerns. The trend to movemore trading and other products onto a centralisedprime brokerage platform (see Box 2), might, tosome extent, alleviate these concerns in the future.

Early warning signals and termination triggersAccording to banks’ answers, two sets of warningsignals could be distinguished. The first set relatesto factors that would raise concerns, lead tocontacts with hedge fund managers or a review oflimits, but would not give a bank the legal rights toterminate transactions. By contrast, the secondgroup of events, in addition to common events ofdefault in legal agreements, could lead to thetermination of hedge fund positions and the seizureof collateral held, if a bank opted to do so.Inevitably, the two sets overlap to some extent, asdescribed below.

For the first group, all banks highlighted sudden orprolonged declines in NAV caused by either poorperformance or investor redemptions, even thoughmany other factors that were relevant forinvestment exposures and have already beenmentioned before were also listed as a source ofconcern. More precisely, changes in key personnel,style drift, changes of fund rules, stressedconditions in certain markets in which a hedge fundis involved, persistent underperformance relativeto peer group or relevant index, breaches ofminimum NAV or established risk limits, low fundliquidity, spikes in VaR or higher volatility would

Chart 16 Banks’ exposure aggregationpractices


only acrossselectedproducts

15%

across the entirefinancial group

75%

no aggregation,fragmented information 10%

Source: BSC.


also attract attention and could lead to lower limits.If only performance data is available, thenamplified performance volatility could beindicative of problems in risk control, higher VaRor even a style drift. One bank was also followingstrategy-specific drawdowns. It indicated that adecline in NAV of more than 20% from a peakwould require explanation from hedge fundmanagers. One bank stated that limits could also belowered owing to the insufficient level of businessactivity.

All of the abovementioned factors were usuallymonitored either monthly or less frequently, asprovided in regular hedge fund reporting packages.Prime brokers focused mainly on VaR and dailyinformation was available. Timely payment ofmargins was checked daily, too.

Regarding termination events which give banksthe legal right to end a business relationship,NAV decline triggers were frequently mentioned(see Table 5), in addition to a failure to meetmargin calls or departure of key principals.NAV decline triggers usually related to bothnegative performance and the impact of investorredemptions, thus a widespread run on hedge fundsby investors would give banks the opportunity toterminate all transactions and, if exercised, couldescalate financial instability, not to mention havingdisastrous consequences for the funds affected.However, hedge fund trading strategies are ratherdiverse and the likelihood of widespread sharp

declines in NAVs and associated runs in investorwithdrawals is relatively low.

It was also sometimes mentioned that larger hedgefunds could negotiate larger NAV decline triggers,even though size alone was not usually the primarydeterminant in negotiations, as smaller funds withmore diversified portfolios, stronger riskmanagement systems or good track records couldbe awarded higher limits as well. Moreover, onebank also reported that larger hedge funds could beallowed NAV decline triggers that relate only tonegative performance, i.e. excluding the impact ofredemptions.

Reporting to senior managementBased on the answers provided by banks, thefrequency and content of risk reports to the seniormanagement varied from bank to bank. Usually thesenior management and various risk managementcommittees received information on hedge fundexposures on a monthly or quarterly basis. Area(division) heads were provided with daily orweekly updates. It was also indicated that seniormanagers would receive ad hoc information intimes of distress, related, for example, to thesignificant NAV or VaR movements or whenhedge fund performance deviated considerablyfrom expectations. Several banks highlighted inparticular the fact that the exceeding of limits wouldbe instantly reported to senior managers or therelevant risk management committees. At somebanks there were also regular meetings with seniormanagers to discuss risks related to dealings withhedge funds.

Reports usually contained figures for NAV,performance and VaR. In one case it was specifiedthat VaR was reported for each individualinvestment and at aggregate level in the form ofnon-diversified VaR. Some other banks havedeveloped and utilised special risk (e.g. usingCornish-Fisher expansion) or performancemeasures (e.g. Omega, Sortino ratios) in regularreports to senior managers. For those banks withhigher financing and trading exposures to hedgefunds, regular reports also included various othermarket (e.g. expected tail loss) and credit risk (e.g.credit VaR) measures. Only one bank indicated that

1 month 3 months 12 months

15% 25% 35%10-20% 20-30% 50%

15% 25% 35%15% 25% 35%

20-25% 25-30% 35-40%2)

15% - 45%20%3) 30%3) 40%3)

Source: BSC.1) Normally including redemptions.2) From the last fiscal year end.3) All for larger funds, smaller funds might have lower triggers.

Table 5 NAV decline triggers on a rollingbasis 1)

(rows represent individual bank replies)

39ECB


2 SURVEYRESULTS

its senior management was provided with stresstest results on a quarterly basis. The analysisincluded a comparison of the increase in exposureunder stress scenarios with the excess collateralheld. This finding might indicate that banks tend toplace too much emphasis on VaR metrics withoutsupplementing such information with variousstress test results. Given the complex nature ofhedge fund exposures, stress tests could providevery valuable information to senior managementand assist in setting prudent risk tolerance levels.

Analysis of hedge fund balance sheetsOnly a few of the banks surveyed with onlyinvestments in hedge funds mentioned specificallythat they examined the structure of hedge funds’investment portfolios on an aggregate basis inorder to spot various concentrations and todetermine how the underlying hedge funds’portfolios fit into the banks’ overall investmentportfolio. Some banks actively monitored theirinvestments in hedge funds and even used to askhedge fund managers to adjust their portfoliostructure if individual investments did not satisfythe needs of banks’ own investment portfolios. Inother cases, some banks clearly indicated that theyhad refrained from analysing in greater detail hedgefunds’ portfolios; while other banks reported thatgenerally they relied on summary risk reports, asthe usefulness of information provided by hedgefunds for their risk management systems seemed tobe limited or because risk analysis based on thedetailed data from administrators would requirededicated staff. On the other hand, informationavailable to sole prime brokers generally made itpossible to build a rather comprehensive picture ofthe total portfolios of individual hedge funds. In thecase of exposures to funds of hedge funds, someinformation on underlying single hedge funds wasusually available, but banks did not go as far as tolook through the portfolios of those single hedgefunds.

Box 5 highlights the importance of individual andaggregate hedge funds’ portfolio analysis anddescribes the potential transmission of stress that

could result from inadequate monitoring of hedgefunds’ portfolios by banks.

The possibility of initiating an aggregatedmonitoring of hedge funds’ portfolios wouldprobably only be relevant for a limited number ofthe largest global prime brokers, given the ratherconcentrated prime brokerage market structure,and for banks extensively trading with hedge fundsin OTC derivatives markets. Thus, it would onlyapply to about ten large EU banks from BE, DE, ES,FR, NL, SE and UK. Market participants, includingbanks, already monitor developments in and thefinancial standing of various corporate sectors andfinancial institutions, therefore the deeper analysisof hedge funds’ balance sheets does not look socontroversial, especially as banks are in the bestposition to do that in order to safeguard their owncommercial interests.31

Use of several prime brokersIn addition to not always having adequateinformation, the use of multiple prime brokersfurther complicates the ability of banks to obtainhigh quality information about the fund as a whole.In that respect, one prime broker reported that theaverage number of prime brokers of its hedge fundclients was 3.3 in 2004. Two other surveyed banksspecified that only 30% and 15% of hedge fundclients used them as sole prime brokers.

According to the banks, they identified the numberand identities of prime brokers, borrowing capacityas well as trading counterparties used by a hedgefund in the course of a due diligence process.Nonetheless, one bank underscored that it tended tobe more difficult to obtain information on thenumber of regular counterparties, even thoughhedge funds would probably disclose how manylegal trading agreements they had.

31 Admittedly, by trying to implement this, banks would face substantialobstacles due to the diversity of hedge fund reporting practices. Thiscould be partly overcome if the popularity of risk managementsoftware developed by banks increased among their hedge fundclients or hedge funds (their administrators) were able to provideuniform inputs into banks’ risk management systems.


Box 5

SIMPLIFIED HYPOTHETICAL CRISIS SCENARIO

In the left panel of the diagram below, there are one large and a number of small hedge fundsfollowing similar investment strategies. The large fund has two similar individual positions(trades 1 and 2) in market A. Two banks (banks A and B) finance or are counterparties tothese two medium-sized trades 1 and 2. Both banks monitor their trades individually withoutpaying attention to or not having enough information about the structure of the total hedgefund’s portfolio, and to each of them trades look manageable and liquid in isolation.Moreover, banks A and B also take proprietary trading positions in market A (trades 3 and 4).Outstanding trades 1 to 4 together constitute a significant share of market A, leading to highmarket concentration and increased vulnerability to the sudden withdrawal or insolvency ofa major player.

As such, the constellation described above is not unlike that preceding the near-collapse ofLTCM. Since 1998, however, the hedge fund industry has grown to more than $1 trillion inassets under management with a large number of smaller players entering the market. Thedecreasing concentration of the hedge fund industry could be seen as beneficial from afinancial stability perspective. However, there are some indications that in the context of theglobal search for yield a larger number of hedge funds could be pursuing similar strategies,thereby leading to the “crowding” of trades. If the concentration of trades is not detected bybanks, in the case of market turmoil and mass unwinding of positions, crowded trades mayhave a similar impact as the exit of a large hedge fund.

Bank Afinances (is counterparty to) and

monitors individual Trade 1

Bank Bfinances (is counterparty to) andmonitors individual Trade 2 andsome individual crowded trades

Financialinstitutions

Large HF(LTCM)

Smaller HFs

HF Strategy

Market A

Market B

Spillover

Spillover

Trade 1

Trade 2

Trade 3

Trade 4

Crowded trades

41ECB


2 SURVEYRESULTS

Most banks noticed a tendency among hedge funds,particularly larger ones, to use several primebrokers, each with specific product or geographicalexpertise, in order to diversify their counterpartybase and ensure pricing efficiency. Many of themwere employed for particular strategies although insome cases, more than one prime broker was usedfor the same strategy. One bank noted that macroand fixed income arbitrage funds used one or veryfew prime brokers for their clearing and settlementbut had always used multiple tradingcounterparties. Several other banks also mentionedthat long/short equity funds tended to use severalprime brokers in order to secure stock or bondborrowings for short selling.

Banks also pointed out that this tendency couldhave an adverse effect on risk management owingto a lack of transparency and increased competitionamong banks, thus putting pressure on them tooffer more favourable pricing and marginrequirements. On the other hand, one bank alsostressed that it would draw comfort from the factthat a particular fund was not dependent on one solesource of liquidity/leverage, which could adverselyaffect the fund in times of distress.

However, under portfolio-based margining the useof several prime brokers is not practical for smallerhedge funds, as working with only one primebroker provides substantial margin savings.Moreover, in these cases banks, as sole primebrokers, benefit from a high degree of transparencyin a fund’s positions. Thus the increasingpopularity of portfolio margin requirements willact against the trend of using more prime brokers.

2.3.5 RISKS OF FUNDS OF HEDGE FUNDSFor some banks with mainly investment exposures,investments in funds of hedge funds (FOHFs)were the main type of investments in hedge funds,whereas other banks stated that they wereinterested only in single hedge funds. Quite ofteninvestments in FOHFs were related to soldstructured products or used as hedges forderivatives.

Generally, FOHFs were deemed less risky thansingle hedge funds. All banks surveyed mentioned

higher diversification and associated lowervolatility of returns as key advantages of FOHFscompared to single hedge funds. Some banks notedthat relative safety was further enhanced by theirpractices of looking into the characteristics ofunderlying single hedge funds. Furthermore,FOHFs managers were usually expected to monitorunderlying single hedge funds more actively andhave deeper market knowledge together with widercontacts. Quite often, FOHFs also offered betterredemption terms, thereby taking more of theliquidity management burden from end-investorsupon themselves. FOHFs may have morerestrictions on leverage, short-selling and otherrisk-taking activities, particularly if registered incountries where their activities are regulated.Moreover, FOHFs were often used in structureswith capital protection, therefore posing lowerreputation risks for banks arising from possibleinvestor complaints.

However, some banks also highlighted that in somecases FOHFs could pose higher risks, ascorrelations among underlying funds andstrategies increase in times of stress. One banknoted that diversification benefits were usuallyoverstated due to style drift. Moreover, theincreasing crowding of trades may further lowerdiversification among underlying funds followingsimilar strategies. Other banks were also worriedabout the increasing use and greater availability ofleverage for FOHFs, particularly if based on theperception that these vehicles were safer than singlehedge funds. The second layer of leverage orleverage on leverage could increase riskssignificantly, especially if coupled with themismanagement of funding liquidity. Indeed, thelatest tendency of longer lock-ups at the singlehedge fund level may pose higher funding liquidityrisks for FOHFs, as they usually offer morefrequent redemption possibilities than underlyingsingle hedge funds. In addition, FOHFs are riskierfor lenders as they pledge shares of underlyinginvestments in single hedge funds, which carry asubordinated credit status. Finally, banks thatinvest or lend to FOHFs do not have direct (micro)control of underlying investments in single hedgefunds.


According to banks, they tried to mitigate theserisks by careful due diligence and ongoingmonitoring. Many banks required full transparencyof FOHF portfolios, including information onliquidity, leverage and other risk parameters. Banksgenerally had extensive diversificationrequirements for FOHFs in terms of minimum/maximum allocations to underlying individualsingle hedge funds, managers and strategies. Theywere also investigating track record (performanceand experience), reputation, size, investmentphilosophy, risk management systems, investmentand monitoring processes and minimumrequirements for targeted single hedge funds.Similar due diligence and ongoing monitoring wasapplied to sub-funds of multi-strategy hedge fundsas well, because diversification and cross-contamination issues are more important for suchfunds.

2.3.6 OTHER FINDINGSBased on the experience of the banks surveyed,hedge funds’ responses to losses varied from fundto fund, although on balance it seemed as if hedgefunds were becoming more cautious, i.e. reducingexposures or applying different strategies afterperiods of underperformance. The increase in risk-taking was largely seen as a function of theopportunity set available. However, in mid-2004some banks witnessed a higher use of leverage,albeit from relatively low levels, by their hedgefund clients employing fixed income strategies inresponse to the difficult trading conditions seenearlier in the year. One bank also mentioned thatmany institutional investors, including itself, werelooking for hedge fund managers who reduce ratherthan increase risk-taking in times of stress orunderperformance.

Regarding the implications of the forthcomingBasel II requirements, banks that provided answerson this question did not generally think that the newcapital adequacy rules would have a material impacton the risk management of exposures to hedgefunds, as exposures were collateralised and wouldbe treated as corporate risk. However, one bankcomplained that the consultative paper on thetrading book review had not yet fully recognisedcross-product netting according to the ISDA

Master Agreement and therefore could pose abusiness constraint. Another bank expected that ifit were allowed to use the Internal Ratings Based(IRB) Advanced Approach, it would be able toassign a high recovery rate to exposures secured byhedge fund collateral. In this way, Basel II couldreduce regulatory capital requirements. A thirdbank argued that VaR was not necessarily anappropriate measure for the market risk posed byhedge funds, due to illiquidity, fat tails and relativenon-transparency. A fourth bank reported thatthere was no clear treatment of hedge funds in BaselII documentation, particularly regarding theboundaries between banking and trading books. Afifth bank anticipated that banks would break downhedge fund exposures to primary risk factors,although it thought that such efforts would not bevery successful on a cost-benefit basis and becausehedge funds are dynamically managed. Moreover,the largest hedge funds would probably beunwilling to offer full transparency of theirportfolios to entities that could potentially takeadvantage of this information.

3 SOME SUPERVISORY ISSUES ARISING FROMHEDGE FUND ACTIVITY

This section reviews a number of supervisory issuesrelated to the findings of the BSC survey. Hence, itfocuses on sound risk management practices andcapital requirements for banks’ interaction withhedge funds as well as on supervisory actions to thisend, rather than on the issue of a possible directregulation of hedge funds. Sound risk managementand capital requirements contribute to the widerpolicy aim of safeguarding financial stability, whichis also at the core of the debate regarding the possibleregulation of hedge funds. Indeed, by managing theirrisks prudently, banks can limit their potential lossesfrom troubled hedge funds. Banks acting prudentlyas counterparties of hedge funds can also exertmarket discipline, thereby influencing the risktaking of hedge funds. In line with the approach takenin the survey, the supervisory implications areinvestigated for the main direct links that existbetween banks and hedge funds, i.e. banks ascounterparties to hedge funds, and banks asinvestors in hedge funds.

43ECB


3 SOMESUPERVISORY

ISSUES ARISINGFROM HEDGE

FUND ACTIVITY

3.1 BANKS AS COUNTERPARTIES OF HEDGE FUNDS

3.1.1 RISK MANAGEMENT PRACTICESImmediately after the near-default of LTCM in1998 and the market turbulence surrounding thisevent, a number of international initiatives werelaunched to address potential systemic riskconcerns arising from the activity of hedge funds.Starting in 1999, the Basel Committee on BankingSupervision (BCBS) and the InternationalOrganization of Securities Commissions (IOSCO)published a number of papers on banks’interactions with “highly leveraged institutions”(HLIs), foremost hedge funds. Furthermore, theJoint Forum, IOSCO and the Financial StabilityForum undertook a detailed analysis of the policylessons learnt from the LTCM episode.

The BCBS document of 1999 on the soundpractices for banks’ interactions with HLIs32 stillconstitutes a sound basis for the supervisoryreview of banks’ counterparty exposures to hedgefunds. This document addresses some of the majorrisk management failures that became apparent inthe LTCM episode, i.e. an over-reliance on thecollateralisation of marked-to-market exposuresand the insufficient weight placed on the in-depthcredit analyses of HLIs. Thereafter, the BCBS

focused on monitoring the implementation of itsrecommendations and published in 2000 in thisrespect a review.33 This review outlined a series ofissues relating to HLIs which required furtherattention from banks, supervisors and internationalfora. It also proposed continued collaborationbetween bank and security firm supervisors and anongoing dialogue with the financial industry,particularly in challenging technical areas, such asthe measurement of potential future credit exposureand stress testing. This led in 2001 to a joint reviewby the BCBS and IOSCO of issues related to HLIs(see Box 6).34 These various initiatives targetingspecifically the interaction between regulated firmsand hedge funds should be seen in conjunction withthe more general guidance developed by the BCBSin the different areas of banks’ risk management, inparticular concerning credit risk.35

32 Basel Committee on Banking Supervision (1999), “Sound Practicesfor Banks’ Interactions with Highly Leveraged Institutions”,January.

33 Basel Committee on Banking Supervision (2000), “Banks’Interactions with Highly Leveraged Institutions: Implementation ofthe Basel Committee’s Sound Practices Paper”, January.

34 Basel Committee on Banking Supervision and the InternationalOrganization of Securities Commissions (2001), “Review of issuesrelating to Highly Leveraged Institutions (HLIs)”, March.

35 Basel Committee on Banking Supervision (2000), “Principles of theManagement of Credit Risk”, September.

Box 6

THE JOINT REVIEW OF ISSUES RELATED TO HLIs BY THE BCBS AND IOSCO

Overall, the joint BCBS-IOSCO report of 2001 was encouraged by firms’ continued progressin implementing the recommendations in the BCBS “sound practices” document. Asachievements, the report particularly mentions strengthened senior managementoversight, clearer definitions of overall risk appetites, improved internal reporting andefforts to improve information flows from hedge funds and progress in the credit duediligence process.

However, a number of areas where additional progress was needed were also identified,foremost the need to enhance exposure measurement methodologies. In this regard, progressin conducting regular and comprehensive stress testing was considered rather slow.Furthermore competitive pressures were noted to affect firms’ ability to insist on the full rangeof risk mitigants, including initial margin. The report called for individual firms to haveadequate “packages” of mitigants and risk management techniques for their risk exposures. Italso argued that supervisors should remain alert to the risks attaching to HLI counterpartiesand to exercise judgement about the way in which the elements of the package are combined.


Given the weight of these concerns, the financialindustry also took several initiatives to addressthe concerns of authorities. The most prominentones include the recommendations of theCounterparty Risk Management Policy Group(CRMPG), a group of senior officials frommajor, internationally active banks and securitiesfirms. In 1999 the CRMPG published a set ofrecommendations that aimed at improving practicesregarding counterpart risk management36, andwhich included areas such as information-sharingbetween counterparties, disclosure practices andindustry documentation, risk assessments andstress-testing.

In 2005 the CRMPG released a follow-up report toanalyse the implementation of the recommendationsin the 1999 report, to study the impact of newdevelopments such as CRT instruments on riskmanagement practices and to provide a conceptualanalysis of some “emerging issues”.37 TheCRMPG II goes beyond credit and market risk,dealing now also with operational and reputationalrisk. Though many of its recommendations wouldfurther enhance or refine initiatives alreadyunderway, the CRMPG II takes the view thateven where this is the case, the enhancementsand refinements are substantive and material.Furthermore, its recommendations and guidingprinciples should be seen as forward-looking and asan integrated framework of initiatives. Most of theaction points of the CRMPG II relate to measures thatare within the control and reach of individualinstitutions. Others, by contrast, entail collectiveactions by institutions or their representative bodies.

Supervisors as well as market participants shouldremain vigilant to new developments in the hedgefund industry which may confront them with

situations requiring risk management practicesgoing beyond the abovementioned internationallyagreed standards. One such area is, for example, thedevelopment of the CRT markets. These marketsallow the transfer of credit risk from banks to othermarket participants such as institutional investors,non-financial firms and, increasingly, also hedgefunds.38 Although the BSC survey indicated that ingeneral hedge funds were not key counterparties tobanks in CRT instruments, hedge funds havebecome important players in these markets and theirimportance is generally expected to increasefurther.

In view of this development there are a number ofspecific risk management concerns to be kept inmind. Banks purchasing credit protection need tobe aware not only of the residual risks that canresult from the contractual terms and enforceabilityof CRT instruments but also of the risks posed bycounterparties providing credit protection, such ashedge funds. In that respect, it should be noted thatthere is a different treatment of credit protectionpurchased by banks, depending on whether theinstrument is allocated to the banking book or thetrading book.39 As this can result in markedly

The survey conducted by the BSC revealed that some of the areas identified by the BCBSand IOSCO five years ago as offering scope for further improvement are still relevant. Theseareas include the timeliness and comprehensiveness of the information provided by hedgefunds to their bank counterparts, the aggregation of exposures on hedge funds across thewhole (banking) group, the impact of competition on the use of risk mitigants (in particularinitial margins) and the application of stress-testing (also in relation to the collateral taken).

36 Counterparty Risk Management Policy Group (1999), “ImprovingCounterparty Risk Management Practices”, June.

37 Counterparty Risk Management Policy Group (2005), “TowardGreater Financial Stability: A Private Sector Perspective”, July.

38 See, for example, Fitch (2005), “Hedge Funds: An Emerging Forcein the Global Credit Markets”, July.

39 In the banking book, only protection from providers that have a lowerrisk weight than the original borrower reduces the capitalrequirement. This is, for example, the case if a 100% risk-weightedexposure is protected by a 20% risk-weighted bank (in contrast to ahedge fund as a potential counterparty), a low default risk, broadlydiversified protection provider. If a credit risk position in the tradingbook is hedged by a credit derivative, the reduction of specific riskcharges would be independent of the protection provider’s creditquality. The latter would be accounted for by a separate counterpartycredit risk capital requirement.

45ECB


3 SOMESUPERVISORY

ISSUES ARISINGFROM HEDGE

FUND ACTIVITY

different capital charges, supervisors shouldmonitor prevailing bank practices, though financialinnovations make it increasingly difficult to make aclear-cut distinction between the two types ofportfolios. Other concerns related to CRT marketsand hedge funds’ growing presence are therelatively small set of market participants that arefully active (hence the need to monitor credit riskconcentration), the performance of CRT marketsunder stress conditions (hence the need foradequate stress-testing) and the long settlementlags (hence the need to further improve operationalefficiency in the back-office functions).

3.1.2 CAPITAL REQUIREMENTSAs regards minimum capital requirements, there isforemost the issue of the appropriate risk weight forcredit exposures to hedge funds. Compared to a“plain-vanilla” corporate borrower, most hedgefunds at least aim to run low exposures tosystematic, non-diversifiable risk, which, taken inisolation, might justify sub-average capitalrequirements. However, a different issue iswhether this investment objective is effectivelyachieved as some hedge funds seem to engage in“long only” strategies similar to traditionalinvestment funds. Moreover, the low correlation togeneral market risk may not hold under stressconditions. Other financial characteristics of hedgefunds, such as their potentially high leverage andrelative opacity, by contrast, would argue forincreased risk weights. The 100% risk weight forunrated corporates under both the currentframework and forthcoming Basel II StandardisedApproach40 obviously does not really reflect thisconsideration.

However, devising more adequate reflections ofhedge funds’ default risk under the Internal RatingsBased (IRB) Approach41 of Basel II is also aconsiderable challenge. As large and complexfinancial institutions (LCFIs) will most likelyapply the IRB Approach, supervisors may need topay particular attention to the modelling of therequired risk parameters of hedge fund exposures.The rating criteria for corporates or financialinstitutions will therefore need to be adapted to thespecificities of hedge funds counterparties. Thelack of sufficient information and adequate

transparency are in that respect particularlychallenging. It should be noted that transparencyand leverage are already important elements to beconsidered by banks in their rating process.Furthermore, the complex risk structure of hedgefunds’ assets may not be particularly suitable forrating models, which have typically beendeveloped to estimate the credit risk of “plain-vanilla” corporate or interbank exposures. Theacademic literature on this topic has not yetproduced satisfactory answers and publicinformation about practitioners’ approaches is sofar unavailable. For smaller banks that apply lesssophisticated approaches to manage credit andmarket risk, the challenges relating to themonitoring and managing of hedge fund exposuresmay even be more intense.

The BSC survey showed that banks, at least fortheir financing exposures, require high degrees ofcollateralisation from their hedge fundcounterparties, which takes the form of financialcollateral. Where this collateral is not of the highestquality, there may be a high correlation between thecollateral value and the default risk of the borrower.This implies that there is only limited additionalprotection from the collateral in situations of stress.Consequently, where collateral of lesser qualityis used, and which under Basel II’s moresophisticated approaches is now also recognisedfor regulatory capital purposes, the degree of over-collateralisation may need to reflect potential stressscenarios. In that respect, it should be recalled thatthe BSC survey found only a limited use of stresstests to determine the liquidation value of collateral.

Where netting agreements are used as creditrisk mitigation, the interaction of market risk andcredit risk also deserves consideration. This isparticularly true if previously thin or non-existentmarkets receive liquidity from new hedge funds’activity. In such a scenario, the bank may be able to

40 Under the Standardised Approach, the bank can use external ratings(if available) provided by rating agencies to calculate its regulatorycapital requirements for credit risk.

41 Under the Internal Ratings Based Approach, a bank can use its owncredit assessments to calculate its regulatory capital requirementsfor credit risk. Depending on the risk factors the bank is allowed toestimate itself, a distinction is made between a Foundation IRB andan Advanced IRB.


close out its OTC transactions with the defaultedcounterparty effectively, but it could be left withopen unhedged market risk positions it hadconsidered as hedged by the OTC derivativesbefore. This may pose problems for the bank’s riskmanagement if these positions are illiquid anddifficult to re-hedge because there is only a limitednumber of counterparties in the market.

Finally, under the supervisory review process(“Pillar II”) of Basel II the management of the bankhas to make sure that the institution has adequatecapital available to support its risks, andsupervisors should take appropriate action whenthis is not the case. Such action could include, forexample, requiring the bank to strengthen its riskmanagement, improve internal controls, increaseprovisions and, ultimately, even increase capital.The supervisory review process therefore providesa useful framework to ensure that the bankadequately addresses the risks resulting from itsinteractions with hedge funds. For smaller bankswith hedge fund exposures, though, the challengefor supervisors is to have sufficient resourcesavailable to cover also them adequately under thePillar II review on a regular basis.

3.2 BANKS AS INVESTORS IN HEDGE FUNDS

3.2.1 RISK MANAGEMENT PRACTICESThe BSC survey highlighted that in many EUcountries, investments of banks in hedge funds arethe major, and sometimes the only, direct linkbetween the two types of institutions. The banks inthe survey identified a range of operational risksthat could potentially affect their investments. Suchrisks can be mitigated through a careful duediligence process. Adequate internal processesshould therefore be in place, first, for scrutinisingnew investment instruments and reporting theresults to senior management before risks areincurred; and second, to manage the incurred riskson an ongoing basis. Less sophisticatedinstitutions in particular should not makesignificant investments in hedge funds withoutfully understanding the risk they entail. Theseinternal processes should be regularly reviewed byexternal auditors as well as supervisors.

3.2.2 CAPITAL REQUIREMENTSUnder the current capital requirements rules,investments in hedge funds are treated as plain equityand consequently risk weighted at 100%. Given theabove considerations of the risk weights for creditexposures and given equity’s even higher risk, thisrisk weight is probably too low for most of the hedgefunds and may need to be adjusted upward in certaincases. The Standardised Approach of Basel II givesin that respect supervisors the option of assigning a150% risk weight to certain high risk asset categorieswhich may constitute a somewhat more adequatetreatment.

Under Basel II’s IRB Approach, the bank has twopossibilities: either it applies fixed and rather high-risk weights of 400%42 to equity investments; orinstead, it will use its own estimates for riskweights that apply to equity positions. The ownestimates can be based either on a VaRmethodology or on the probability of default thebank assigns to the hedge funds and the loss givendefault associated with its equity (the “PD/LGDmethod”).43

The VaR methodology for equity exposures will beinadequate for most hedge funds because thehistoric returns it relies upon would need to bederived from the NAV calculations of the funds.These will often constitute a bad proxy for futurevolatility of hedge funds’ returns, in particular ifevent or credit risk plays an important role. Furtherchallenges to this modelling approach are skewedreturns distributions and data availability. Thus itmay be expected that supervisors will in most casesnot authorise the use of VaR approaches. The PD/LGD method, on the other hand, faces the sameproblems that the IRB modelling of creditexposures to hedge funds entails. Furthermore,only for LCFIs with a large number of hedge fundsexposures, the authorisation of the PD/LGDmethod should be considered. Hence, it is likelythat all other IRB banks will need to take recourse tothe fixed risk weights.

42 For publicly traded equity, this is reduced to 300%.43 The probability of default is the likelihood that a borrower will

default within a certain time period (generally one year). The lossgiven default is the loss, measured as a percentage of the exposureat default, which is likely to occur in the case a borrower defaults.

47ECB


4 CONCLUSIONS

4 CONCLUSIONS

Hedge funds are playing an increasingly importantrole in the financial landscape and thereforecontinue to attract the attention of both authoritiesand the financial community. Various efforts areunderway to gain a better understanding of thedevelopment of the hedge fund industry and what itimplies for the financial system at large, and theBSC study aims at providing a contribution to this.However, because of the survey’s limitations andthe work underway by various international andEuropean fora in this area, the conclusionspresented in this report can only be considered aspreliminary and indicative of prevailing exposuresand risk management practices. Nevertheless, anumber of findings are worth highlighting.

The survey demonstrated that large EU banks’exposures to hedge funds varied across countries.Direct exposures seemed to be growing rapidly,although generally they were not large in relation tobanks’ balance sheets and total revenue or similarexposures undertaken by US peers. In most EUcountries, exposures took the form of directinvestments in hedge funds although more or lesssizeable financing and trading exposures wereobserved in DE, ES, FR, NL, SE and UK. It is notpossible to provide a firmer conclusion about thesize of EU banks’ direct exposures owing to the factthat only a limited number of large EU banksprovided comparable quantitative data. However, itis very likely that the absolute and relative size ofexposures to hedge funds will increase further inline with the continuing expansion of the hedgefund industry, and in particular its Europeansegment.

Most banks extensively dealing with hedge fundshad specific guidelines for this interaction andadvanced risk management systems or were in theprocess of improving them further. The surveyprovided some evidence that generally large EUbanks surveyed had stringent requirements forexposures to hedge funds with a strong emphasison collateralisation, although there was alsoevidence that banks quite often traded with hedgefunds on variation margin only.

The survey also highlighted some key areas forfurther improvement of banks’ risk management.These shortcomings carry the risk of turning into acause of concern, particularly if the current ratherbenign market conditions would change abruptly.They are:

– Counterparty discipline. The survey providedsome evidence that market discipline, as appliedby banks, shows signs of weakening owing tohighly competitive market conditions. Inparticular, several banks reported that hedgefunds were, to some extent, successful inachieving more beneficial business terms.Indeed, the largest hedge funds seemed to haveenough financial clout to negotiate someconcessions, including less rigorous collateralterms, lower lending spreads or higher NAVdecline triggers.

– Stress testing. Most of the banks’ stress tests,particularly the regular ones, included onlyhistorical scenarios. Moreover, banks werenormally only stress testing individualexposures to hedge funds rather than trying toestimate aggregate effects on all exposures oreffects on separate hedge fund strategies. Alsothe stress-testing of collateral was less commonand offers scope for further improvement.

– Aggregation of banks’ exposures to hedgefunds. Some larger banks reported havingdifficulties in aggregating hedge fund exposuresacross the entire financial group and/or differentbusiness areas/geographical regions. Thatapplies in particular to banks’ internal riskmanagement practices that were often differentin prime brokerage divisions from otherbusiness areas of the bank.

– Hedge fund disclosures and information onleverage. Despite some progress, the surveyrevealed certain shortcomings regarding thequantity, quality and timeliness of data providedby hedge funds to banks. For example, bankstypically had only lagged monthly informationon several key variables such as NAV, NAV pershare and hedge fund risk profile. It seems that inmany cases hedge funds still provided banks


with relatively crude measures of leverage,although an increasing number of hedge fundswere supplying more advanced risk-basedmeasures of leverage such as VaR/NAV.Moreover, banks generally did not haveinformation on off-balance sheet leveragearising from trading in derivatives.

– Analysis of hedge funds’ financial positions.Banks’ descriptions of their risk managementpractices also raised questions about whetherbanks were sufficiently taking into account and/or had enough timely information on the wholeportfolio structure of hedge funds, particularlyof the larger ones with financing and tradingrelationships with several counterparties.Despite the diversity of hedge fund disclosures,banks could also consider implementing theaggregate analysis of hedge funds’ financialpositions by adapting their systems to harnessdetailed position data received from hedge fundsor their administrators. If implemented, suchanalysis could also allow spotting crowdedtrades among hedge fund clients.

All in all, hedge funds are a moving target forbanks’ risk managers, as risks posed by hedgefunds are changing with the evolution of financialmarkets and the expansion of the hedge fundindustry. Hence, it is rather difficult to judge,if at all possible ex ante, whether current riskmanagement practices are adequate.

The main policy conclusions that can be drawnfrom the survey can be summarised as follows.First, on the question of whether recentdevelopments in the hedge fund industry pose adirect threat to financial stability in the EU throughthe banking channel, the evidence collected throughthe BSC survey indicates that this may not benecessarily the case. This is mainly due to the factthat the prime brokerage market is dominated by USbanks so that the direct exposures of EU banks aregenerally limited in absolute and relative terms.However, as indicated above, there are some areasof concern affecting EU banks’ exposures to hedgefunds that warrant further monitoring.

Furthermore, in some respects direct exposures mayunderestimate the true risks stemming from hedgefunds. Given their very large trading volumeactivities, there may be significant risks related toadverse developments affecting institutions withpotentially high leverage. These risks could beexacerbated when the positioning of individualhedge funds becomes more similar or “crowded”.Consequently, adverse market dynamics could havesignificant spillover effects on banks, which are notevident from direct exposures. For banks sellinghedge funds or hedge fund-related products to theirclients, reputation risk may be another potentialsource of indirect risk.

Second, the main recommendation put forward bypublic authorities in the aftermath of the LTCMcase – according to which adequate managementby banks of risks associated with hedge fundsshould be put in place – still remains relevantfor large EU banks surveyed, as some specificareas of risk management seem to require furtherimprovements, even though this is a matter forsupervisors to judge.

In particular, where individual banks’ links withhedge funds are material, supervisors shouldbecome aware of the need to review whether banks’risk management practices for the related risks aresound. More generally, the survey showed thedifficulties for banks of estimating hedge fundrisks in a comprehensive and exhaustive manner.The still limited transparency of hedge funds –taken together with the complex interactions ofthe different sources of risk – always makesaddressing hedge fund risks by banks somewhatincomplete. As a minimum, however, banks shouldbe able to aggregate their overall exposure toindividual hedge funds across books and riskbuckets and limit it to a prudent percentage of theirown funds. Where doubts remain over thediversification of risks underlying the exposures toall hedge funds taken together and the interactionamong the various risks concerned, the overallexposure to hedge funds in general should also belimited prudently.

Third, risk management guidance developed bysupervisors and the capital adequacy regime

49ECB


4 CONCLUSIONS

constitute the appropriate framework forauthorities to deal with risks resulting from banks’interactions with hedge funds. In particular thesupervisory review process provided for underBasel II allows supervisors to take any additionalmeasures, specifically with regard to capitaladequacy, needed to address such risks. Regardingthe supervisory issues arising from hedge fundactivity, it is also worth recalling that at theEuropean level the convergence of supervisoryreview practices in general is an important strand ofwork of the Committee of European BankingSupervisors (CEBS).


Country summaries

Common elements of banks’ definitions of hedge funds are as follows:– unregulated form of investment partnership;– situated in offshore centres;– use a large spectrum of different instruments – including leverage and short-selling;– intended for institutional investors and high-net-worth investors;– no or at least lower regulatory requirements;– absolute return strategies;– performance-based fee structure.

Banks define hedge funds as an unregulated form of investment partnership that uses leverage, derivative instruments, short-selling, arbitrageand other high-risk strategies to achieve targeted levels of risk and return. Hedge funds have to fulfil only weak disclosure and regulatoryrequirements, especially regarding capital requirements. Most hedge funds are located in offshore centres. In comparison with standardinvestment funds, they have longer contractual liquidation, redemption or withdrawal periods of associated certificates/shares. In order to gainhigh returns, they trade large volumes with a high leverage factor and their investments are sometimes concentrated in speculative, illiquid ornarrow markets/products, e.g. emerging markets, distressed securities or convertible arbitrage. Especially funds with arbitrage or one-directional strategies have a large refinance/repo ratio.

Most banks use some internal definition of a (fund of) hedge fund, formulated in their internal policy documents. These definitions are, ingeneral, closely related to the working definition used in the survey, containing elements such as “loosely regulated, few restrictions withregard to the products in which they can invest, the trading strategies they can pursue or special investment techniques (e.g. short-selling,leverage) they can apply, absolute return driven, managers receive performance-related fees”, etc.

Hedge funds were defined as funds which generally manage third-party funds and which have the ability to leverage their resources. Inaddition, they target absolute returns (non-correlated to markets) and use a diversified array of instruments to achieve these results.

Most definitions included criteria such as: lightly regulated entities; the use of leverage; the ability to take short positions; investment strategiesaimed at maximising returns through the use of a range of financial instruments; investors deemed to be sophisticated.

Individual bank replies

Hedge funds are defined as those investment vehicles that use gearing.

One of the main characteristics of hedge funds is that they are scarcely-regulated and absolute return oriented leveraged funds with limitedliquidity.

An asset class with the following characteristics: it is usually domiciled offshore, returns are uncorrelated with major indices (absolute returns),it makes intensive use of financial derivatives in order to maximise their performance and charges high fees (up to 20% performance fee).

Typical characteristics: limited or no regulation, limited disclosure requirements and the resulting lack of transparency, significantmanagement and performance fees, frequently changing investment portfolios, infrequent investor redemptions.

Several criteria are used:– absence or quasi-absence of specific rules defined by the regulatory authorities under whom the fund operates (e.g. lack of limitations

related to leverage, short-selling and diversification);– specific investment objectives (emphasis on absolute returns rather than on relative returns);– use of very wide range of investment techniques – including leverage, short-selling and other hedging strategies in an attempt to

achieve absolute returns).

Hedge funds are defined as unregulated or lightly regulated investment pools, set up for qualified investors seeking high return and/or littlecorrelation to financial markets evolution. Notable characteristics of hedge funds include:

– privately organised investment vehicle domiciled outside G10 countries;– access limited to qualified investors only;– subject to very little or no direct regulatory oversight;– lack of transparency and of daily prices;– use of balance sheet leverage (cash borrowing to NAV > 50% or repo/stock lending to NAV > 100%) and/or short securities (net

short position above 20% of NAV);– performance-based fees;– restrictive redemption policies, which are also subject to change.

Hedge funds are a subcategory of unregulated funds, which are investment vehicles that are not formally regulated by a government-approvedregulator accepted by our institution.

Investment vehicles that try to achieve above market returns using leverage. Their strategies might involve both long and short positions andthe universe of their investments can be quite broad. They can be either asset class specific, i.e. fixed income, FX, equity, etc. or region/countryspecific.

ANNEX 1HEDGE FUND DEFINITIONS, AS PROVIDED BY BANKS OR SUMMARISED BY COUNTRY AUTHORITIES

51ECB


ANNEX 1

Funds whose managers have very limited restrictions on the use of leverage, short positions, derivatives and other strategies in order to achievepositive absolute returns while preserving capital.

Funds that trade and invest mainly in “fixed income”, including the use of OTC derivatives, but with the possibility to deal in other markets likecommodity, currency and equity. The use of derivatives allows fund manager to create significant leverage, so that traded assets can be higherthan assets under management. For this reason, hedge funds have highly speculative activities in OTC markets.

An investment fund that uses alternative investment strategies i.e. is not a conventional mutual fund.

Other funds than mutual funds, insurance funds or pension funds are to be regarded as hedge funds according to internal instructions.

Hedge funds are investment partnerships that seek above-average returns through superior portfolio management and whose primarycompensation is percentage of profits.

All funds not equivalent to mutual funds are defined as hedge funds.

Hedge funds apply market neutral strategies to achieve performance; this means that the performance is independent from the performanceof equity or bond markets. Fund of funds invests with multiple managers through funds or managed accounts. The strategy designs a diversifiedportfolio of managers with the objective of significantly lowering the risk (volatility) of investing with an individual manager. A fund of fundsmanager has discretion in choosing which strategies to invest in for the portfolio. A manager may allocate funds to numerous managers withina single strategy, or with numerous managers in multiple strategies. The minimum investment in a fund of funds may be lower than an investmentin an individual hedge fund or managed account. The investor has the advantage of diversification among managers and styles with significantlyless capital than investing with separate managers.

Note: Country summaries and individual bank replies do not overlap and always refer to different banks.

Documents

LARGE EU BANKS’ EXPOSURES TO HEDGE FUNDS