Making Better Informed Securities Lending Decisions

Transcription

1 Making Better Informed Securities Lending Decisions Defining Best Practice in developing a Risk Adjusted Returns Framework Mark C Faulkner Head of Innovation Mark.Faulkner@dataexplorers.com David Carruthers Head of Quantitative Services David.Carruthers@dataexplorers.com Ed Oliver Head of Data Explorers Consulting Ed.Oliver@dataexplorers.com March 2010 New York 75 Rockefeller Plaza 19 th Floor New York NY London 2 Seething Lane London EC3N 4AT +44 (0)

2 About the Authors Mark Faulkner is the Founder and Head of Innovation at Data Explorers. After graduating from the London School of Economics, Mark Faulkner spent the majority of his career specializing in International Securities Finance. Since 1987, he has held management responsibility at L.M. (Moneybrokers) Ltd., Goldman Sachs, Lehman Brothers and more recently at Securities Finance International Limited. Whilst occupying these different posts he has gained experience as a lender, borrower, conduit borrower and prime broker. During his career he has worked closely with the UK Inland Revenue and has represented firms at the Securities Lending and Repo Committee and the London Stock Exchange's securities lending committees. Being an independent advisor since 1995 has provided Mark with a unique insight into the operation of the securities financing market. David Carruthers is Head of Quantitative Services at Data Explorers. Completed a thesis in econometric modeling in 1988 and has since worked as a fund manager, risk manager, and risk consultant with experience in equities, fixed income, currencies, cash management and derivatives. Since 1997 he has carried out consultancy work for the securities lending and cash reinvestment industry, and has been involved in designing and calibrating various frameworks for the analysis and management of securities lending risk. Ed Oliver is Head of Data Explorers Consulting. Ed has spent his whole career working in Financial Services initially at Lloyds Bank Global Custody Services and Barings Asset Management before spending 13 years at Northern Trust. Since 2002, Ed has focused on securities lending, initially as a member of Northern Trust s securities lending management team, leading the Product Management function for the EMEA region. In this role, Ed took a lead position in the International Securities Lending Association (ISLA), chairing their regulatory sub-committee for 2 years. Ed joined Data Explorers consulting arm in September 2007 and is responsible for advising beneficial owners and practitioners on commencing and reviewing their securities lending programs. Data Explorers Data Explorers provides its global client base with quantitative measurement of securities lending, performance and risk. The company is based in New York and London and collects data from market participants to create the world s most comprehensive and up-to-date database of information and research on stock lending and short interest. To download a free copy of this book or to contact Mark, David or Ed please visit: Page 2 of 85

3 Index Preface 4 Acknowledgements 8 Executive Summary 9 Section 1 Securities Lending - An Industry Overview 10 Section 2 Securities Lending Returns 25 Section 3 Securities Lending Risk 36 Section 4 Securities Lending and Risk Adjusted Returns 43 Section 5 Conclusions 60 Glossary of Terms 61 Frequently Asked Questions 70 Reference Sources 75 Appendices 76 Page 3 of 85

4 Preface The purpose of this publication is to provide the reader with a context within which to explore and better understand the risks and rewards associated with the activity of securities lending. The significant challenges faced by the global capital markets since 2008 and the response to those challenges by governments, regulators, investors and traders has meant that the securities lending industry has found itself at the epicenter in a global debate which has grown to involve subjects such as the efficient operation of markets, the appropriateness of short selling and the possible need for more regulation and transparency. In perhaps one of the more positive developments in the global debate, in March 2009 The International Organization of Securities Commissions (IOSCO) Technical Committee published a consultation report entitled Regulation of Short Selling 1 prepared by its Task Force on Short Selling, which contains proposed principles designed to help develop a more consistent international approach to the regulation of short selling. The report recommended that effective regulation of short selling should be based on the following four principles: 1. Short selling activities should be subject to appropriate controls to reduce or minimize the potential risks that could affect the orderly and efficient functioning and stability of financial markets; 2. Short selling should be subject to a reporting regime that provides timely information to the market or to market authorities; 3. Short selling should be subject to an effective compliance and enforcement 1 Available here Page 4 of 85

5 system; and 4. Short selling regulation should allow appropriate exceptions for certain types of transactions for efficient market functioning and development. Martin Wheatley, Chairman of the Task Force on Short Selling, said: We believe that short selling should operate in a well structured regulatory framework in the interests of maintaining a fair, orderly and efficient market. The objective of such regulation being to reduce the potential destabilizing effect that short selling can cause without exerting undue impact on its legitimate benefits in capital formation and volatility reduction. He went on to say that While IOSCO encourages a concerted move towards a consistent approach to short selling, it recognizes that the case for the regulation of this activity varies from jurisdiction to jurisdiction and depends on a range of domestic factors. These principles will provide guidance to market authorities and assist them in assessing and developing their short selling regulatory framework. The mandate to the short selling taskforce emphasized the need to minimize any adverse impact of short selling regulations on legitimate activities such as securities lending. At the same time there has also been a significant increase in the volatility and, for some participants, a sea change in attribution of revenue coming from their securities lending programs. As a result some participants have found themselves exposed to risks that they might not have fully understood or appreciated. Against this backdrop some beneficial owners have quite understandably questioned whether it is appropriate for them to continue to lend securities and, should they do so, what changes they might implement in their lending programs to ensure that they are configured in a way that is appropriate for them. Their challenge in making an informed decision has been heightened by a marked lack of appropriate information and context within which to make these important decisions. Page 5 of 85

6 This need for information gave rise to a series of client driven discussions concerning the level of risk and returns associated with securities lending. From the underlying client side there is a determination to better understand the industry and to tackle questions such as where the revenue comes from; how to quantify the risks; how to structure and communicate the configuration of a securities lending program that meets an individual organizations requirements. From the perspective of the practitioners the questions focused on how to measure, communicate and put the risks and returns into context. This context is important at a time of great financial uncertainty with some clients considering reconfiguring their programs or even withdrawing from the securities lending market altogether. Recognizing the coincidence of want from underlying clients and practitioners for a public discussion of these issues we brought together a group of practitioners and asset managers with the following objective. to agree a framework that facilitates the communication of Securities Lending Risk- Adjusted-Returns to Beneficial Owners, demonstrating best practice and aiding the education process. 2 For many years there have been organizations addressing the measurement and communication of the risks related to securities lending. However, they have found it challenging to get organizations to focus on the topic of risk when they have preferred to focus upon revenue. Recent market conditions have brought about a global re-evaluation of the priorities facing financial organizations and it is not at all surprising that the subject of risk has risen up the agenda. There is a growing realization that there needs to be a context 2 Mark C Faulkner on Data Explorers Risk-Adjusted-Return webinar 11 th February 2009 Page 6 of 85

7 or language developed to facilitate an informed and considered discussion of the riskadjusted returns associated with the securities lending industry. It is this need that this publication seeks to address. There are already a number of publications which seek to address the broader subject of securities lending and its position within the global capital markets such as An Introduction to Securities Lending 3 and Securities Lending - your questions answered 4. We will leave the general discussion of the industry to those publications and focus upon three core subjects within this booklet: the returns associated with the activity their source scale and attribution; the risks associated with the activity - measuring, managing and mitigating them; and finally the challenge of communicating and expressing the risk-adjusted-returns associated with the activity in a clear and explicit manner to facilitate the making of properly informed and considered decisions concerning participation and program configuration. The target audience for this publication is primarily those organizations that may or may not participate in securities lending but that would never consider themselves practitioners. The supply side of the industry is fuelled by organizations for which securities lending is not considered a core activity but without which there would not be an industry. These organizations include: Pension funds Insurance and assurance companies Mutual funds/ucits/unit Trusts We hope that any organization that is interested or involved in the activity will also find the publication useful e.g. regulators, trade associations and consultants. 3 An Introduction to Securities Lending Data Explorers Consulting 4 Securities Lending - your questions answered Page 7 of 85

8 It is the intention of the authors and those working with us to prepare the publication, that it is made freely available as a public good to help promote better understanding of this industry. Although Data Explorers has made every effort to ensure the information in this publication is correct, nevertheless no guarantee is given as to its accuracy or completeness. All opinions, views and estimates expressed in this booklet are those of Data Explorers on the date it was prepared and are subject to change without notice; however no such opinions, views or estimates constitute legal, investment or other advice. You must therefore seek independent legal, investment or other appropriate advice from a suitably qualified and/or authorized and regulated adviser prior to making any legal, investment or other decision. This publication is intended for information purposes only and is not intended as an offer or recommendation to buy, sell or otherwise deal in securities. Acknowledgements The views expressed in this publication are those of the authors. Every care has been taken to ensure that the contents are accurate. However, neither the author, nor the commissioning bodies nor the publishers can accept any responsibility for any errors or omissions. The publishers or authors can accept no responsibility for loss occasioned to any person acting or refraining from action as a result of any material in this publication. The authors would like to thank the individuals and organizations who gave of their time during the series of webinars and conference calls that helped formulate their thinking on the subject matter. Page 8 of 85

9 Page 9 of 85

10 Executive Summary Most beneficial owners understand that securities lending is not a riskless activity. A series of market crises in 1987, 1994, 1998, and 2001 have periodically focused their attention on the potential downside. However, these episodes have been relatively brief and contained. The more protracted events associated with the Credit Crunch have really brought the potential risks to their attention. The problems range widely and include; temporary concern for fund assets (many of which were safely returned); major counterpart defaults; frozen assets due to legal bottlenecks or a lack of liquidity; and in a number of cases mainly cash reinvestment outright losses The challenge for the industry now is to provide transparency around the risks being faced by beneficial owners, and to quantify and communicate those risks, putting securities lending in a general context alongside other investment activities, and in particular showing how the risks compare with the rewards. For many beneficial owners this is a quid pro quo for them to return to normal lending activity. In this publication we describe the industry, the returns it generates, and the associated risks in securities lending and cash reinvestment. We propose a framework for quantifying, comparing and communicating risk and return to non specialist beneficial owners. This framework, and the guidelines and analytics in this report have been developed at the request of some of the main custodian banks, agent lenders and asset managers active in the securities lending industry. The result of the industry wide debate, facilitated by Data Explorers, has been the production of a framework which: Gives a clear understanding of risk and return Better communicates risk / return profiles to clients Helps builds optimal portfolios reflecting client objectives Establishes industry-wide metrics Positions securities lending as a recognized asset class Page 10 of 85

11 Section 1 Securities Lending - An Industry Overview Introduction Notwithstanding our stated focus upon return, risk and risk-adjusted returns for the purposes of this booklet, it is important to provide some background on the industry. Below we provide an overview of the securities lending industry that draws extensively from An Introduction to Securities Lending. Those readers requiring more background information can download complimentary copies of this publication from a number of websites in various languages. 5 Securities Lending Securities lending is a long-established practice which plays an important role in today's capital markets by providing liquidity that reduces the cost of trading and promotes price discovery in rising as well as falling markets. The resultant increase in efficiency benefits the market as a whole - from the securities dealers and end investors through to the corporate issuers which depend on efficient, liquid markets to raise additional capital. Securities lending markets allow market participants to sell securities that they do not own in the confidence that they can be borrowed prior to settlement. They are also used for financing, through the lending of securities against cash, forming an important part of the money markets. The ability to lend and borrow securities freely underpins the services that securities dealers offer their customers and the trading strategies of dealers, hedge funds and other asset managers. On the lending side, securities lending forms a growing part of 5 Page 11 of 85

12 the revenue of institutional investors, custodian banks and the prime brokerage arms of investment banks. Lenders and intermediaries The supply of securities into the lending market comes mainly from the portfolios of beneficial owners, such as pension and other funds, and insurance companies. Underlying demand to borrow securities begins largely with the trading activities of dealers and hedge funds. In the middle are a number of intermediaries. The importance of intermediaries in the market partly reflects the fact that securities lending is a secondary activity for many of the beneficial owners and underlying borrowers. Intermediaries provide valuable services, such as credit enhancement and the provision of liquidity, by being willing to borrow securities at call while lending them for term. They also benefit from economies of scale, including the significant investment in technology required to run a modern operation. Intermediaries include custodian banks and asset managers lending securities as agents on behalf of beneficial owners, alongside the other services provided to these clients. Some specialist securities lending agents and other 3 rd parties have also emerged. Such organizations agree to split securities lending revenues with lenders and may offer indemnities against certain risks, such as borrower default. Another category of intermediary is dealers trading as principals. Dealers intermediate between lenders and borrowers, but they also use the market to finance their own wider securities trading activities. They may seek returns by taking collateral, counterpart credit or liquidity risk, for example, by lending securities to a client for a period while borrowing them on an open basis with a risk of early recall by the lender. Through their prime brokerage operations, they also meet the needs of their hedge fund clients. Page 12 of 85

13 For beneficial owners, there are a number of different possible routes into the market including: a) Discretionary/Opportunistic Daily trading/negotiating of individual securities and/or baskets of securities to the borrowers on a best efforts basis. This route to market is typically facilitated by custodial agents, third party agents and/or self-lenders. b) Direct/Principal Exclusives A portfolio of securities or subset of a portfolio made available to lend to a borrower exclusively for a defined period of time, in exchange for a guaranteed fee payment. In the direct model, the beneficial owner lends directly to the broker/dealer without an agent in the middle managing the administration or providing indemnification. This route is typically facilitated by broker/dealers. c) Agent Managed Exclusives Exclusive arrangement managed by an agent who provides full program management, indemnification and administrative oversight. This route is typically facilitated by custodians as agent and/or third party agents. Typically, the allocation of exclusive arrangements is coordinated through the use of an auction. Beneficial owners can either chose to auction their portfolios to borrowers themselves or they can utilize a lending agent to manage this process. The borrowing motivation Reading the generalist financial press one might conclude that the most common reason to borrow securities is to cover a fundamental short equity position. But not all securities lending is motivated by short selling. Financing drives many transactions the lender is seeking to borrow cash against the lent securities, whether using repo, buy/sell backs or cash-collateralized securities lending. Fixed income assets account for 60% of borrowed Page 13 of 85

14 securities, of which 85% are Government bonds. The majority of bond lending is driven by financing requirements, and is not directional in nature. Equities represent most of the remaining 40%, and it is headline grabbing short sellers of equities who receive media coverage. But directional fundamental shorting is actually only around 10% of the volume (Source: Data Explorers Hedge Fund Clients). It is rare for lending demand to be driven by a simple speculative bet that the value of a security will fall, with the borrower hoping to buy it more cheaply at the maturity of the loan. More commonly, the short position is part of a larger trading strategy, typically designed to profit from perceived pricing discrepancies between related securities. For example: Convertible bond arbitrage: buying a convertible bond and simultaneously selling the underlying equity short. Pairs trading: seeking to identify two companies, with similar characteristics, whose equity securities are currently trading at a price relationship that is out of line with the historical trading range. The apparently undervalued security is bought, while the apparently overvalued security is sold short. Merger arbitrage: for example, selling short the equities of a company making a takeover bid against a long position in those of the potential acquisition company. Index arbitrage: selling short the constituent securities of an equity price index against a long position in the index future. Tax Arbitrage: a slightly misleading but commonly used term for the process of ensuring that tax is paid when it is due rather than incurring premature crystallization. There are a number of variations but this source of revenue is focused on stocks which pay high dividends, so it is especially prevalent in European equities and during the dividend season it can double the volume of stock on loan in Europe. Page 14 of 85

15 Short positions also arise as a result of failed settlement (with some securities settlement systems arranging for automatic lending of securities to prevent chains of failed trades), or where dealers need to borrow securities in order to fill customer buy orders in securities where they quote two-way prices. The History Securities lending began as an informal practice among brokers who had insufficient share certificates to settle their sold bargains, commonly because their selling clients had mislaid their certificates or just not provided them to the broker by the settlement date of the transaction. Once the broker had received the certificates, they would be passed on to the lending broker. This business arrangement was not subject to any formal agreement and there was no exchange of collateral. Securities lending is now an important and significant business that describes the market practice whereby securities are temporarily transferred by one party (the lender) to another (the borrower). The borrower is obliged to return the securities to the lender, either on demand or at the end of any agreed term. For the period of the loan the lender is secured by acceptable assets delivered by the borrower to the lender as collateral. The Law Under most legal jurisdictions, absolute title to the securities lent passes to the borrower, who is obliged to return equivalent securities. Similarly the lender receives absolute title to the assets received as collateral from the borrower, and is obliged to return equivalent collateral. Securities lending today plays a major part in the efficient functioning of the securities markets worldwide. Yet it remains poorly understood by many of those outside the market. Page 15 of 85

16 The Scale of the Activity The scale of the securities lending market is difficult to assess due to the over the counter ( OTC ) nature of the activity but the charts below show how much inventory is made available and how the balance on loan has changed recently. It is to be remembered that the impact of market price movements and foreign exchange (the numbers quoted are in US dollars) has been extreme. The impact upon inventory made available by clients of Data Explorers 6 since September 2007 can be seen in the graph below. Inventory has declined from just over USD 15tn to USD 8tn in March 2009 whilst averaging USD 12.6tn during the period. Chart 1.1 Average Lendable Assets 18,000,000 16,000,000 Group Lendable (M) 14,000,000 Lendable (M) 12,000,000 10,000,000 8,000,000 6,000,000 4,000,000 2,000,000 0 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers The value of securities on loan has seen a more dramatic decline as a result of a number of factors including a strengthening dollar, declining market prices, restrictions on lending 6 Data Explorers ( collate up to 80% of global securities lending data. Page 16 of 85

17 certain securities by some regulators and as a result of the global de-leveraging in the global capital markets. The value of securities on loan peaked at almost USD 4tn during the European dividend season in 2008 and is now increasing but remains below USD 2tn, whilst averaging USD 2.7tn during the period. Chart 1.2 Average Total Balance 4,500,000 4,000,000 Group Total Balance (M) 3,500,000 Total Balance (M) 3,000,000 2,500,000 2,000,000 1,500,000 1,000, ,000 0 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers Despite the reduction in scale of the activity the rewards for participation remain self evident. The average daily return over the period of USD 40m equates to around USD 14bn per annum based on information provided to Data Explorers globally the annual revenues are likely to exceed USD 20bn. Definitions In some ways, the term securities lending is misleading and factually incorrect. From a legal perspective in many jurisdictions, the transaction commonly referred to as securities lending is, in fact... Page 17 of 85

18 a transfer of title of securities linked to the subsequent reacquisition of equivalent securities by means of an agreement. Such transactions are collateralized and the rental fee charged, along with all other aspects of the transaction, are dealt with under the terms agreed between the parties. It is entirely possible and very commonplace that securities are borrowed and then sold or onlent. There are some consequences arising from this clarification: 1. Absolute title over both the securities on loan and the collateral received passes between the parties. 2. Some of the economic benefits associated with ownership e.g. dividends, coupons etc. are manufactured back to the lender, meaning that the borrower is entitled to these benefits as owner of the securities but is under a contractual obligation to make equivalent payments to the lender. 3. A lender of equities surrenders its rights of ownership, e.g. voting. Should the lender wish to vote on securities on loan, it has the contractual right to recall equivalent securities from the borrower. 4. Appropriate documentation of lending ensures that securities lending transactions do not incur taxes that would negate any financial benefit. Different types of securities loan transaction As a by-product of being appropriately documented, and as a result of prudent risk management, securities loans in today s markets are made against collateral in order to protect the lender against the possible default of the borrower. Page 18 of 85

19 This collateral can be cash, or other securities or other assets. (a) Transactions collateralised with other securities or assets Loan Process Non cash Lending Lending Agent Agent Specific Securities Collateral Loan Loan Market Market Fee/Spread Income Specific Securities Client Client Transaction date 13th June Settlement date 16th June Term Open Security XYZ Limited Security price per share Quantity 100,000 shares Loan value 1,000, Lending fee 50 basis points (100ths of 1 per cent) Collateral UK FTSE 100 Concentrated DBVs Margin required 5% At the expiry of the transaction, the flows in this chart are reversed; although daily marking to market will result in collateral flows in either direction. Non-cash collateral would typically be drawn from the following collateral types: Government Bonds: OECD debt, or restricted to G7, G20, etc Corporate Bonds: Typically a minimum credit rating Convertible Bonds: Usually matched to the securities being lent Equities: Of specified Indices Letters of Credit: From banks of a specified credit quality Certificates of Deposit: Drawn on institutions of a specified credit quality Warrants: Atypical, but usually matched Other money market instruments Page 19 of 85

20 The eligible collateral will be agreed between the parties, as will other key factors including: Notional limits: The absolute value of any asset to be accepted as collateral Initial margin: The margin required at the outset of a transaction Maintenance margin: The minimum margin level to be maintained throughout the transaction Concentration limits: o The maximum percentage of any issue to be acceptable, e.g. less than 5% of daily traded volume o The maximum percentage of collateral pool that can be taken against the same issuer, i.e. the cumulative effect where collateral in the form of letters of credit, CD, equity, bond and convertible may be issued by the same firm In a large number of securities lending transactions, collateral is held by a Tri Party Agent. This specialist agent (typically a large custodian bank or International Central Securities Depository) will receive only eligible collateral from the borrower and hold it in a segregated account to the order of the lender. The Tri Party Agent will mark this collateral to market, with information distributed to both lender and borrower. Typically the borrower pays a fee to the Tri Party agent. There is debate within the industry as to whether lenders that are flexible in the range of non-cash collateral they are willing to receive are rewarded with correspondingly higher fees. Some argue that they are, others claim that the fees remain largely static but that borrowers are more prepared to deal with a flexible lender and therefore balances and overall revenue rise. The agreement on a fee is reached between the parties and would typically take into account the following factors: - Page 20 of 85

21 Demand and supply: The less of a security available, other things being equal, the higher the fee a lender can obtain Collateral flexibility: See above the cost to a borrower of giving different types of collateral varies significantly, so that they might be more willing to pay a higher fee if the lender is more flexible Dividend issues: The size of the manufactured dividend required to compensate the lender for the post-tax dividend payment that it would have received had it not lent the security: Different lenders have varying tax liabilities on income from securities; the lower the manufactured dividend required by the lender, the higher the fee it can negotiate The term of a transaction: Securities lending transactions can be open to recalls or fixed for a specified term; there is much debate about whether there should be a premium paid or a discount for certainty. If a lender can guarantee a recall-free loan then a premium will be forthcoming. One of the attractions of repo and swaps is the transactional certainty on offer from a counterpart Certainty: There are trading and arbitrage opportunities, the profitability of which revolves around the making of specific decisions. If a lender can guarantee a certain course of action, this may mean it can negotiate a higher fee Fees vary from security to security and over time. (b) Transactions collateralised with cash Cash collateral is, and has been for many years, an integral part of the securities lending business, particularly in the United States. The lines between two distinct activities - Securities lending and Cash reinvestment - have become blurred; and to many US investment institutions securities lending is virtually synonymous with cash reinvestment. This is much less the case outside the United States but consolidation of the custody Page 21 of 85

22 business and the important role of US custodian banks in the market means that this practice is becoming more prevalent. Loan Process Cash loans Money Money Markets Markets Cash 130 bp Transaction date 13th June Settlement date 16th June Term Open Security XYZ Limited Security price per share Quantity 100,000 shares Loan value 1,000, Rebate rate 80 basis points Collateral USD cash Margin required 5% Collateral required $1,718, ( 1,050, x 1.637) Reinvestment rate 130 basis points 2. Cash Reinvestment 1. Repo / Loan Fee/Spread Income Lending Lending Agent Agent Specific Securities Client Client Specific Securities Collateral Rebate@ 80bp P&L: Reinvestment Rate: Rebate Rate: Repo/Loan Repo/Loan Market Market +130bp (- 80bp) 50 bps Not present in all cash loans At the expiry of the transaction, the flows in this chart are reversed. The importance of this point lies in the very different risk profiles of these increasingly intertwined activities. Crucially, cash reinvestment is not covered by an indemnity, which can be argued to create a conflict of interest for the cash manager they earn fees but do not share the direct financial risk of this activity. They do, however, run considerable reputational and commercial risk if they do not manage this potential conflict. Cash reinvestment was traditionally dominated by unitized funds which pooled the collateral for ease of management and to achieve the various economies of scale available in money market investment. However, segregated accounts with client specific risk profiles are now becoming much more common. Page 22 of 85

23 Note that the securities lending loan term (i.e. time to maturity) will determine the benchmark rate that is to be paid on the cash. Most loans can be recalled at any time, so the cash will generally have an overnight rate benchmark. It is common for the interest to be physically settled monthly. Below we provide an example of relative importance of cash and non-cash to different fiscal locations and as you can see the US domiciled lenders are overwhelmingly taking cash as collateral whilst other jurisdictions have a predilection for non-cash collateral. The relative scale and importance of the US lending community brings the percentage of collateral taken as cash up to 59.5%. These statistics owe a great deal to historic tax legislation and inertia but have served the non US lenders well in recent turbulent times. Table 1.1 Use of Cash as Collateral Domicile of % of collateral taken as lender cash US 97.9 Canada 20.7 Luxembourg 19.9 UK 8.9 All Lenders 59.5 Source: Data Explorers The revenue generated from cash-collateralised securities lending transactions is derived in a different manner from that in a non-cash transaction. It is made from the difference or spread between interest rates that are paid and received by the lender. Reinvestment guidelines are typically communicated in words by the beneficial owner to their lending agent, and some typical guidelines might be as follows: Conservative Page 23 of 85

24 Overnight G7 Government Bond repo fund Maximum effective duration of 1 day, or any term agreements are a small proportion Floating-rate notes and derivatives are not permissible Restricted to overnight repo agreements Quite Conservative AAA rated Government Bond repo fund Maximum average maturity of 90 days Maximum remaining maturity of any instrument is 13 months Quite Flexible Maximum effective duration of 90 days Maximum remaining effective maturity of 2-3 years Floating-rate notes and eligible derivatives are permissible Credit quality: Short-term ratings: A1/P1, long-term ratings: A-/A3 or better Flexible Maximum effective duration of 90 days Maximum remaining effective maturity of 5 years Floating-rate notes and eligible derivatives are permissible Credit quality: Short-term ratings: A1/P1, long-term ratings: A-/A3 or better Some securities lending agents offer bespoke reinvestment guidelines whilst others offer reinvestment pools. Other transaction types Securities lending is part of a larger set of interlinked securities financing markets. These transactions are often used as alternative ways of achieving similar economic outcomes, Page 24 of 85

25 although the legal form and accounting and tax treatments can differ. The other transactions include: Sale and repurchase agreements ( Repos ) Swaps Buy/sell backs Conclusion The scale and significance of the securities lending industry should now be apparent to the reader. It is important to note that this is not purely a business that drives short selling there are many reasons to borrow securities and securities lending is an important tool in a variety of capital market functions. In the next chapter we will focus on the issue of the returns that the activity generates. Page 25 of 85

26 Section 2 Securities Lending Returns Introduction The returns generated from securities lending make a significant contribution to the overall returns generated from long portfolios. On a global basis well over 25,000 individual funds lend securities to generate revenue. The total annual revenue from securities lending is greater than USD 20bn per annum. In this section we examine revenue scale, attribution and trends using the Data Explorer Group data for all securities. This gives us an excellent insight into global trends. However, it must be recognized that individual portfolios, securities and transactions generate their own revenue profile. Returns The Total Return from securities lending is defined as Revenue from the lending of securities + Revenue from re-investing cash collateral = Total Securities Lending Return Some lenders do not accept cash as collateral and as a result their revenue source is derived 100% from securities lending. Page 26 of 85

27 The chart below shows how total returns have become both more volatile and have grown over the period. Prior to the rise in returns during the spring 2007 dividend season, the total return was stable at around 25 Basis Points. Since then it has risen and become much more volatile although settled back in recent months. 2.1 Average Total Return Total Return (Bp) 120 Total Return (Bp) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers The volatility of the securities lending fee increased during the credit crisis that commenced in The traditional equity dividend season peak has become less discernible since 2007 when the peak in Q2 was obvious. The average of 32 Bps is obviously heavily influenced by the peaks shown in the chart below. Page 27 of 85

28 2.2 Average SL Fee 250 SL Fee (Bp) 200 SL Fee (Bp) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers The agreement on a fee is reached between the parties based on factors mentioned in Section One. However, things are changing. In a post credit crunch environment we may see loan pricing reflecting variables such as credit worthiness, margin taken and some perception of risk. Hitherto there has been little or no price differentiation in loan pricing driven by such concerns. Loan fees may also be rising in response to a reduction in supply as some lenders have chosen to withdraw their inventory from the market. This is much less of an issue in the more liquid, or General Collateral ( GC ) names. Withdrawal by many of the smaller lenders has had little or no impact upon supply which can be easily replaced. In the next chart, the reinvestment return looks very low before the credit crunch; but it was still greater than the returns from securities lending. In 2008 and early 2009 it was extremely volatile, but the rising trend reflected an adjustment of credit spreads from a very low base; and an associated repricing of credit and liquidity risk. Recently we have seen reinvestment returns reduce to long term levels. Page 28 of 85

29 2.3 Average Reinvestment Return Reinvestment Return (Bp) Reinvestment Return (Bp) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers Reinvestment returns have three main sources term, credit, and liquidity. A fund with short term liabilities (for example, overnight securities lending transactions collateralized by cash) may choose to buy longer dated securities to lock in a higher yield to maturity. Effectively borrowing short and lending long, the associated risk is compensated by the term premium. Credit risk is compensated in credit spreads, which reflect the risk of downgrades or defaults. In either case, the fund is compensated for the possibility that at the time of sale, the credit risky bond is worth less either crystallizing a temporary mark to market loss, or eventually receiving some recovery of defaulted assets under the terms of the bond covenant. Both of these risks are a form of liquidity risk in the sense that the fund may have to borrow to cover the resulting shortfalls; but in addition many fixed income assets have only patchy Page 29 of 85

30 liquidity. Fixed income managers often use the rule of thumb that around half of the credit spread actually reflects the liquidity risk. Program structures and sources of revenue can be viewed as Styles, in the same way that asset management style is now widely understood and possesses a standard language. The securities lending industry has an opportunity to define such standards to facilitate accurate discussions with underlying clients. By way of example, one style dimension in securities lending revolves around the trade off between fee and utilization. The key inputs to consider are the level at which securities are utilized and the price at which they are lent. In the language of the supermarket one can choose to lend high volumes of securities at discounted prices or to lend lower volumes at a higher value. The former represents volume lending and the latter value lending. Below, we can see how utilizations have changed over the period and one can observe a gradual reduction to below 15%. 2.4 Average Utilization 25 Utilization (%) 20 Utilization (%) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers Page 30 of 85

31 The gross daily total return 7 for the group of lenders monitored by Data Explorers is shown below and the average of USD 39.3mn per day equates to over USD 14bn. 2.5 Average Total Daily Return 140, ,000 Total daily return (K) Total Daily Return (K)) 100,000 80,000 60,000 40,000 20,000 0 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers The total daily return is derived from the securities lending return and the re-investment return which can be seen below. 7 The gross daily total return is usually split between the lender and its lending agent with the lender typically getting the majority of revenue. The size of the fee split depends on a variety of factors including potentially the bundled fee arrangements with other services such as custody. Page 31 of 85

32 2.6 Average Daily SL Return 160, ,000 Daily SL return (K) Daily SL return (K) 120, ,000 80,000 60,000 40,000 20,000 0 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers 2.7 Average Daily RI Return 60,000 40,000 Daily RI return (K) Daily RI return (K)) 20, ,000-40,000-60,000-80, ,000 Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers Depending upon certain factors which one might consider genetic or pre-determined e.g. portfolio composition, tax status and others that might be termed environmental or determinable e.g. risk appetite, collateral flexibility or lending style, the portfolio will generate a return to lendable which can be measure in basis points. Below we can see the 8.6bp total return to lendable generated on the inventory made available to lend. Page 32 of 85

33 2.8 Average Total Return to Lendable Total return to Lendable (Bp) Total return to Lendable (Bp)) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers 2.9 Average Revenue Share from SL 250 Revenue share from SL (%) Revenue share from SL (%) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers This shows that the share of revenues from Securities Lending has passed through some major cycles. The recent recovery reflects the push by agents to maximize the utilization of their client assets, with the aim of receiving large volumes of cash as collateral. In some cases, large cash collateral balances help to finance outright assets until they repay at maturity. Page 33 of 85

34 Much has been made of the short term perspective of those on the demand side of the securities lending market. But in fact, the average tenure of securities loan transactions remains above 3 months, even in these extraordinarily volatile times Average Tenure 140 Tenure (days) 120 Tenure (days) Mar-07 Sep-07 Mar-08 Sep-08 Mar-09 Aug-09 Feb-10 Source: Data Explorers Having looked at the broader picture we can describe two of the main styles of lending in the market Value and Volume. The diagram below illustrates the general concept the green line is the idealized line of equal fee income. Page 34 of 85

35 Chart 2.11 The Fee-Utilization trade off for Value and Volume Lenders Value Lender Fee Volume Lender Utilisation Source: Data Explorers Value Lending This approach focuses on pricing individual loans at levels that reflect their true value in the securities lending market, and generating the most income possible on securities that are in high demand in the borrower community. Example: 120bp, Utilization 11%; Return = 13bp Volume Lending This approach focuses on generating the most income possible on a portfolio of securities by increasing utilization and maximizing the volume of assets on loan and cash reinvestment. Example: 40bp, Utilization 32%; Return = 13bp To decide which form of lending makes most sense, each lender needs to understand how the agent manages and allocates the distribution of lending balances to counterparts with consequences for risk and return. Current fashion favors value lending but the pendulum swings regularly between the two approaches. Page 35 of 85

36 Conclusion The revenues from securities lending are substantial. It may not be a core business for most lenders but it provides sufficient revenue, particularly in falling equity markets, to justify focused attention. Having considered how securities lending returns are generated, the different possible styles of lending, and how to quantify the key factors and outputs with which one can communicate them we will now consider risk. Page 36 of 85

37 Section 3 Securities Lending Risk Introduction In risk terms, intrinsic value securities lending is, inherently, a conservative product. However, there are risks which must be understood and managed. Both market risks and operational / legal risks can result in financial losses for the lenders, but the financial risks are more readily quantifiable than those in the non-financial risk category. All of the risks identified can be measured to some degree and managed and mitigated a subject that we explore in detail in this section. The aftermath of the Lehman Brothers default showed that financial risks can be more difficult to measure than previously thought; and that non-financial risks can play a significant part in the total risk taken. Non Financial Risks These include: Legal Risk covers both contractual risk and enforceability risk. Contractual risk refers to ensuring the terms and conditions for securities lending between two institutions are comprehensive and appropriate. One recent important example: problems following the Lehman default in the legal set off calculation failing to match the reality of the sales and purchases required to close positions. Enforceability risk refers to the enforceability of the contract under prescribed national laws. In the event of a default situation, the Page 37 of 85

38 right of the non-defaulting party to net the collateral value and the underlying loan value is arguably the most important example of this. Operational Risk the risk that the securities lending agent does not have adequate controls and infrastructure in place to manage a securities lending program. It should be noted that if the lender does not or cannot engage in the process of collateral liquidation and loan repurchase, they end up with a de facto change in asset allocation; this may be dramatic and unexpected to some beneficial owners. Recall Risk the risk that the borrower does not return recalled securities in time to enable a sale or corporate action to be met. Reputational the risk that by virtue of being engaged in a securities lending program the reputation and standing of the lender is somehow damaged. This may result from an operation failure or more likely in recent times, the headline risk of being in the paper on the wrong side of a story. Tax risks for example, crystallized profits on disposals of assets following a default may attract Capital Gains Tax. Non financial risks can be mitigated by appropriate legal agreements industry standard documentation gives more security that they are enforceable, particularly when reinforced by external legal opinion. Netting is an important element of risk management; market participants will often have many outstanding trades with a counterpart. If there is a default, standard industry master agreements ensure that, post-default, various payments are accelerated, i.e. payments become due at current market values. Page 38 of 85

39 Financial Risks This covers a number of areas and is best summarised in a table showing possible mitigants: Table 3.1 Financial Risks and Mitigants Financial Risk Default or Credit risk - the risk that a counterpart or issuer cannot meet its obligations. In a collateralised loan, a default triggers the process of collateral liquidation and loan repurchase, exposing the program to market risk; in outright reinvestment the loss is immediate and prospects of asset recovery are usually poor. Market Risk (Mismatch risk, occurring only in the event of a default) - the risk that the market price of the underlying security or collateral moves adversely in a short period of time. This can arise because of changes in yield curves (i.e. interest rates), currencies, credit spreads, or equity markets. - the risk that the collateral cannot be realized at the price valued due to illiquidity or volatility in the security, Liquidity Risk - Non Cash - the risk that collateral could not be realized at all (due to lack of volume in the asset markets) or because the lender is holding a high proportion of the total issue - the risk that the inventory of lendable assets is not, in general, in demand by the lending market; so it becomes difficult to raise additional cash collateral to offset existing illiquidity in cash reinvestment assets Mitigant Counterparty or Issuer Selection / Approval by Lender - an area of increasing focus and one which needs to be kept under constant review. Agents who can demonstrate a dynamic approach and remove most of the burden of counterparty vetting are finding that their client relationships are strengthened this is particularly relevant where the agent may be unable to offer a strong indemnity. Over-collateralisation - the provision of collateral with a market value in excess of the market value of the securities lent. The amount the loan is overcollateralised can be varied to take into account market, credit, FX, liquidity and mismatch risk as well as the costs of liquidating the collateral securities in the event of default. Collateral Schedules - Defining a schedule that is within the lender s risk parameters. This includes choice of asset classes acceptable, the credit quality and markets of those securities, as well as concentration limits. This should be respectful of the underlying securities lending regulations that apply to the lender. Some regulators will specify the allowable collateral for their regulated entities. Some are more conservative than others, for example the Irish, Luxembourg and UK regulators have clear guidance on what collateral is allowed. Risk Modelling discussed in later sections Liquidity Risk Cash - the risk that reinvestment assets cannot be liquidated to meet client demands for the return of collateral; or that they can be liquidated but at significant discounts to par value. Risk modelling discussed in later sections Cash Reinvestment Risk in cash collateral securities lending programs, there is a risk that the instrument bought with the borrower s cash falls significantly in value or defaults. So this is a combination of a number of risks - yield curve, mismatch risk, spread risk, and default or credit risk. Page 39 of 85

40 Indemnification A major source of risk mitigation can also be the construction of a securities lending program with an appropriate indemnity, provided that the indemnity terms are clear and provide for appropriate interaction with any Service Level Agreement. Below we explore in more detail questions specifically around indemnification. As discussed earlier, Indemnification is one of the key risk mitigation factors in a securities lending program. It is effectively an insurance policy written to the lenders of securities to provide them with additional security and risk reduction. It is an important factor to consider when exploring the configuration of a lending program. The key things to remember when addressing indemnities are: 1. Who is providing the indemnity? A lender needs to understand exactly which legal entity is providing the indemnification, their financial strength, and the relationship between the indemnifier and the lending provider (if any) is also an important consideration. 2. What is being indemnified? There are many differing types of indemnities being offered in the market and the lender needs to understand exactly what risks and under which circumstances they are protected. Options on indemnities range as follows: Indemnification in respect of all securities loans Indemnification in respect of distributions on loaned securities Indemnification in respect of US securities loans only Page 40 of 85

41 Exclusion from Indemnification in respect of non-us securities loans Indemnification in respect of loans to selected borrowers Indemnification in respect of cash reinvestment Indemnification in respect of reverse repo transactions on cash collateral 3. How clear and strong is the legal framework? Indemnification language is often tailored to individual provider and client requirements and it is imperative that the indemnification language is clear and all parties understand that what will happen in a default situation. Given the risk conscious environment prevalent today and the scarcity and expense of providing the capital necessary to support indemnification it is to be expected that indemnification may well become less homogeneous and potentially more of a differentiator in the future. In addition to the extent of the coverage of the indemnification there are also different forms as we shall see below: - 1) No indemnification Here the Agent offers no financial indemnity to the client but either: a) acts for the client in the market to sell collateral and re-purchase loaned securities or b) the client is required to unwind the collateral/loan position themselves. Page 41 of 85

42 Some sophisticated lenders may opt for this option if they prefer taking control themselves and/or if the cost of indemnification is deemed too high for their liking. However, effectively dealing with a default crisis requires considerable preparation with dealing lines and accounts in place. 2) Credit of Proceeds When a default event occurs some Agents choose to warrant to clients that they will provide the cash value of the loaned securities as at the close of business the night before the default was filed. The cash will come from the sale of collateral provided by the defaulting borrower. In practice most Agents offering this indemnity do so as part of an option in the securities lending agreement. The clients must read the language carefully and in particular understand the tax consequences of receiving cash rather than securities post default. 3) Full Replacement In a Full Replacement indemnification, the Agent is promising to make the client good following borrower default as if the loan had never taken place. The Agent is warranting that the equivalent securities are returned to the client s portfolio, regardless of cost. Some Agents offer this indemnity but with the codicil that, should a security cease to trade, rather than be difficult to purchase, the indemnity lapses into a Credit of Proceeds format. Examining documentation in the marketplace over a number of years has shown there to be a number of variations to what might be called more commonplace indemnification language. Page 42 of 85

43 Indemnifiers can employ a variety of legal means to diminish their exposure at the expense of the lenders. As with all contracts therefore, reading the small print is important. Conclusion Securities lending is a low-risk activity if managed correctly. As we have demonstrated in this section, there are many different risks, but similarly, there are also many aspects of risk mitigation. We can now turn our attention to consider how we might combine return and risk outputs to create a risk-adjusted-return framework. Page 43 of 85

44 Section 4 Securities Lending and Risk Adjusted Returns Introduction In this section we discuss the rationale for bringing financial risks and returns together, and the ways in which these can be measured and compared with one another, across programs, and through time. In this section, we will focus on ways of comparing programs their structures, risks and returns. The details of individual risk calibration are discussed in Appendix A and in a supplementary methodology document. Having looked at risk measurement, we will also look at the process of monitoring and managing risk. Most importantly, we will concentrate on how the comparison of programs and the level of risk can be communicated to beneficial owners who may not be risk experts. Risk and Return As an asset class in its own right, Securities Lending seems to provide superior risk adjusted returns which are uncorrelated with traditional bond and equity asset classes. Since 2006, the average annualized return to lendable has been 8.6bp, with a standard deviation of 4bp (Source: Data Explorers); giving a Sharpe Ratio in excess of 2. The return is consistently positive, so it has the characteristics of a cash asset, although the standard deviation is proportionately higher. However, in absolute terms the return is very low, so the lender has to place a large proportion of their assets at risk. This means that, as an asset class, securities lending has some of the characteristics of a corporate bond portfolio the entire principal is at risk in Page 44 of 85

45 order to achieve an additional excess income without the prospect of capital gain that is offered by equities. For this reason, any analysis of securities lending risk adjusted return has to primarily focus on the scope for capital loss, which will completely outweigh the volatility of returns as a source of risk. To calculate risk adjusted returns, the actual return needs to be compared with the ever present but uncrystallized risk that follows a counterpart or issuer default. The following equation illustrates this by decomposing the program return: Program Return = Sum across all counterparts of: + SL Fee%* Utilization% * $Lendable - Crystallized Losses following Counterpart Default [=Min(Final Collateral Final Loan),0)] + [Reinvestment Return% - Benchmark%] * $Cash Collateral - Crystallized Losses following Issuer Default The essential dilemma for owners and lenders is that the likely scale and frequency of these possible losses, highlighted in red, is somewhere between uncertain and unknowable. So while owners will lend stock in order to achieve additional and well understood returns, they run variable risks, which cannot be measured in any agreed way 8. These risks are always present, but may not be crystallized for many years, if at all. How can the asset owner decide whether the returns justify the risks? The owner is in the same position as an insurance company, effectively writing insurance in return for a fee, a fee which must be reserved against the day when a large loss is realized. If we pursue this insurance analogy, then the objectively justifiable approach is a direct arithmetic 8 The traditional measure of return volatility is nevertheless a useful analytic for comparing efficiency across programs. Page 45 of 85

46 comparison of the return and the losses that are expected to occur over a suitably long time period. Consider the typical non cash program over the past few years. It has been earning 8 basis points per year on average; it has experienced no losses. Suppose that a risk analysis of the program suggests that the theoretical expected loss is around 1bp pa i.e. the program is expected to lose, on average, 1bp per year from counterpart and issuer defaults. So the owner is on average making a net return of (8 1) each year, and has been reserving 1bp pa. So after 10 years, the owner has a buffer of 10bp. If a counterpart defaults, the loss will depend on the overall exposure to that counterpart, the haircut, and the behavior of loan and collateral market values between the date of the default event and date of the closure of the position. Will 10bp be enough? This is the essential risk measurement problem. Calculation of this expected loss requires calibration using historic data on asset movements and counterpart defaults; a task which we now know is as much art as science. This is discussed in more detail in the appendix, but for this discussion we can say that risk estimation is subject to error and that the error range needs to be quantified and communicated. The owner might decide that the potential risks are too high; this is why the impact of indemnification is so important. After indemnification, the risks to the owner will be much lower; but indemnifications can be expensive (the fee split may be altered in favor of the agent). So the return will be lower. Comparison of risk and return, pre- and postindemnity, will allow the owner and agent to decide which route they prefer. Page 46 of 85

47 Risk Adjusted Return Outputs This section details some of the risk and return outputs which can be used to understand and compare program structures and styles, along with their associated risks and returns. Structure Outputs Measures of program structure are aimed at identifying the potential fault lines in a program. The program structure determines the risk and return characteristics. While it is usually difficult to achieve consensus on the best way to measure risk, program structure is largely objective; so structure metrics can provide a quick (but possibly blunt) form of comparison, especially when there is a need to present to a non-specialist audience. For example, asset backed securities might be rated as AAA or even Super Senior, but a concentration of these assets in the program should, in itself, require those assets and the risk assumptions underlying them to be scrutinized in more detail. Some examples (this list is not exhaustive) of structure measures and their importance are as follows: Securities Lending (Non cash) specific: % Equities vs. Bonds In a crisis, equities usually drop in value and high quality bonds increase in value. So a book which lends equities and takes high quality bonds as collateral is naturally hedged. The fees on such trades will usually be lower to reflect this. However, liquidity can dry up quickly even in high quality corporate bonds, so apparently hedged trades may be a poor form of collateral in a crisis. Page 47 of 85

48 % Corporate Bonds vs. Government Bonds Credit spread movements and rising default risks can create a gap in the performance of what might otherwise be viewed as a monolithic bond exposure % Investment Grade vs. Non Investment Grade The fault line here may lie somewhere other than the official dividing line, but it is a useful starting point to assess overall quality. See comments on asset backed exposures in the Reinvestment section, below. % cash vs. non cash collateral Cash collateral can be deceptive intuitively it seems to be the safest form of collateral, but in practice it needs to be managed which exposes the holder to a range of additional default and yield curve risks. Weighted Average Expected Default Frequency of Counterparts; Concentration Index Default frequencies are difficult to predict despite the growing number of agencies addressing the task. Rating is a starting point and the RMA regularly publish mappings between ratings and defaults using data from Moody s KMV. An exposure weighted average of these probabilities is a good starting point, but knowing the distribution of counterparty exposures is also important. The Herfindahl 9 index of industrial concentration is a useful summary statistic in this context. 9 This is widely used in anti-trust cases. If market shares are equally distributed across many firms, the index has a value close to zero; if one firm completely dominates then the index is close to 1. The actual value indicates the degree of concentration. A similar approach can be used for portfolio diversification. Page 48 of 85

49 ReInvestment (Cash) specific: % Floating Rate Notes (FRNs) This is a more direct measure. It can be useful to separate out Asset backed from traditional FRNs. Duration and Gap risk (Weighted Average Maturity) to next reset Because the reinvestment program is funded by cash which can be recalled at any time and is usually paying some variation on Libor, it is common to neutralize the main interest rate risk using floating rate notes. This measure shows the extent of any pure rate mismatch. Duration and Gap risk (Weighted Average Maturity) to Final Maturity This indicates the overall mismatch in duration, which is relevant because FRNs expose the fund to credit spread movements, which may have a noticeable effect on mark to market valuations and FRN resale values. Comparison of the above two measures indicates the overall effect of FRN exposure. % Asset Backed Securities This category covers a broad range of asset types, from vanilla securitizations to CDOsquared. It now tends to consist of the more traditional assets but it is important to know the proportion and subdividisions will again be very useful. If some assets are difficult to identify or to price, that fact n itself tends to indicate the presence of risk, either in the form of an exotic asset structure or lack of liquidity. Histograms showing the distribution of Maturity and Credit exposures Page 49 of 85

50 These go beyond the single numbers such as weighted average duration and weighted average rating. They particularly highlight the presence of skew which may be due to a small but significant exposure to very long dated or very low quality assets. Weighted Average Expected Default Frequency of Issuers The same comments apply here as for the non Cash counterparts. The Herfindahl concentration index can again provide a single number summary of the diversification of the Reinvestment pool in its exposure to individual issuers. In the next two sections, we look at some of the most common Returns and Risk outputs. These outputs can be split into the four most important dimensions: - Securities Lending (Non Cash) vs. Cash Reinvestment and - Pre- and Post- Indemnified Return Outputs Returns can be presented in a number of ways: Gross of revenue split the split between owner and agent is subject to negotiation and will be dictated by a range of factors, especially the indemnification arrangement. In the interests of like for like comparisons, it is advisable to look at revenues on a gross basis. Total, Securities Lending (Intrinsic), Reinvestment it is important to split the returns in this way the intrinsic return is the proceeds from offering stock (equities and bonds) to Page 50 of 85

51 those who need it; the reinvestment return is the proceeds from offering cash (held as collateral) to those who need it. Daily running fee the snapshot of the program s current yield. This may vary considerably from day to day, so it is important to do some smoothing or assess returns over longer periods. Cumulative Fee the returns from securities lending may seem small, but the cumulative effect can be substantial. Losses sustained in the Lehman collapse were, in many cases, only a small proportion of the cumulative returns from securities lending and reinvestment. Risks can be expressed in terms of years of cumulative revenue at risk. Yield to Maturity (Reinvestment only) as the name suggests, this assumes that assets are held to maturity. If they are not due to early recalls, then there may be substantial variance between the current market price and the maturity value. Similarly, if liquidity evaporates then the position needs to be funded. Risk Outputs Risk is in essence about the expected position and shape of the future mark to market changes, including those which are conditional on counterpart default. This can be calibrated and communicated in an increasingly complex variety of ways, some of which are well beyond the scope of this publication. Texts such as Dowd, Grinold and Kahn, Jorion, and Taleb 10 cover the spectrum of views and approaches to risk measurement. Risk measures can be viewed as following a natural progression: 10 Dowd Measuring Market Risk (2005); Grinold and Kahn Active Portfolio Management (1999), Jorion Value at Risk (1996), Taleb Dynamic Hedging (1997), The Black Swan (2007) Page 51 of 85

52 Market measures: Volatility how wide is the distribution? This can mean the volatility of day to day returns, reflecting the management of the securities lending and reinvestment portfolios; but also the mark to market risk following a counterpart default. Probability of Loss how likely is it that the return is even slightly negative? [Note that this can be a large number but refer to a small loss; often a normal distribution is assumed to assign probabilities] Value at Risk how long is the negative tail of the distribution? [VaR is the minimum loss with a specified probability and time horizon, so we need to know the shape of the distribution and its properties over time e.g. path dependency or independence and it only selects one point on the distribution] Percentiles what are all the possible VaR values? [Generalization of the VaR concept] Expected Loss what is the probability weighted sum of all possible losses (i.e. VaR values)? [This is effectively an option pricing calculation (a numerical integration)] Default adjusted measures: Where any of the above losses are conditional on a default, then they can be multiplied by the probability of a counterpart default. For example: Risk premium one name for the product of Expected Loss and Default Probability Break Even Fee another name for the same estimate Page 52 of 85

53 All of these measures can be recorded directly or they can be adjusted by an average gain or buffer either the haircut (margin) for non cash programs or the reinvestment yield for cash programs. This will have the effect of reducing the apparent risk in each case. Stress Test - these measures can be augmented by detailed Stress Tests. These ask the question What is the worst loss that can happen? To calculate them, it is necessary to estimate sensitivities for each asset class to a range of stress inputs, such as a) equity markets b) yield curves c) credit spreads The question is How far will this asset move if the underlying factor moves by x%? This can be estimated by regression analysis, or through more detailed simulation from the structure of the asset. The Derivatives Policy Group in the US has developed standard stress tests which closely match the worst case experiences for the past 100 years; these include volatility; and should probably now include house prices. The main challenge with stress testing is to assign probabilities to the outcomes; by their nature, the worst case scenarios do not happen often enough to give historic frequencies and no two crises are exactly alike. Stress testing is discussed in more technical detail in Appendix A and in more detail in the supplementary methodology document. Combined measures of Risk and Return There are a number of technical measures for risk adjusted returns: Page 53 of 85

54 Sharpe Ratio (and related measures) these seek to calculate the ratio of returns to various measures of risk (i.e. returns per risk unit). They can be interpreted as measures of statistical significance but have the drawback of all ratio measures - they have a broad distribution and the scale can be misleading. Risk Adjusted Return (1) the Return minus the Risk Premium. This converts risk directly into returns space; risk premium can be calculated in a number of ways. It can be interpreted as the minimum fee that must be achieved by a program to compensate for the risks. Risk Adjusted Return (2) Years of accumulated revenue at risk this again translates risk into return space. The challenge is to decide which risk measure to use e.g. the worst case, or a weighted average of worst cases. This allows the owner to decide whether it is time to remove some exposure, and puts the magnitude of the current risk number in some context. However, we need to articulate these types of measures in a way that is understandable to non-experts. In the next section, we look at ways of doing that. Page 54 of 85

55 Communicating, Monitoring and Managing Risk Adjusted Returns In this section we describe, in general terms, how to communicate and use these outputs in the monitoring and management process. The following chart shows one method for combining these approaches. In the interests of space, only a subset of the outputs is displayed. Monitoring Chart 1: Program positioning within peer group Source: Data Explorers Note that the y-axis scale is a peer group ranking The chart shows program exposure to risk and return in a number of dimensions, relative to a chosen peer group. The columns form a number of sections from left to right return measures (including risk adjusted return); structure measures; risk measures, stress test Page 55 of 85