Market Consistency E-Book

Malcolm Kemp

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Serie: Wiley Finance Series
• Sprache: Englisch

Seitenzahl: 891

Veröffentlichungsjahr: 2009

Beschreibung

Achieving market consistency can be challenging, even for the most established finance practitioners. In Market Consistency: Model Calibration in Imperfect Markets, leading expert Malcolm Kemp shows readers how they can best incorporate market consistency across all disciplines. Building on the author's experience as a practitioner, writer and speaker on the topic, the book explores how risk management and related disciplines might develop as fair valuation principles become more entrenched in finance and regulatory practice.

This is the only text that clearly illustrates how to calibrate risk, pricing and portfolio construction models to a market consistent level, carefully explaining in a logical sequence when and how market consistency should be used, what it means for different financial disciplines and how it can be achieved for both liquid and illiquid positions.  It explains why market consistency is intrinsically difficult to achieve with certainty in some types of activities, including computation of hedging parameters, and provides solutions to even the most complex problems.

The book also shows how to best mark-to-market illiquid assets and liabilities and to incorporate these valuations into solvency and other types of financial analysis; it indicates how to define and identify risk-free interest rates, even when the creditworthiness of governments is no longer undoubted; and it explores when practitioners should focus most on market consistency and when their clients or employers might have less desire for such an emphasis.

Finally, the book analyses the intrinsic role of regulation and risk management within different parts of the financial services industry, identifying how and why market consistency is key to these topics, and highlights why ideal regulatory solvency approaches for long term investors like insurers and pension funds may not be the same as for other financial market participants such as banks and asset managers.

Leseprobe

Table of Contents

Dedication

Title Page

Copyright Page

Preface

Acknowledgements

Abbreviations

Notation

Chapter 1 - Introduction

1.1 MARKET CONSISTENCY

1.2 THE PRIMACY OF THE ‘MARKET’

1.3 CALIBRATING TO THE ‘MARKET’

1.4 STRUCTURE OF THE BOOK

1.5 TERMINOLOGY

Chapter 2 - When is and when isn’t Market Consistency Appropriate?

2.1 INTRODUCTION

2.2 DRAWING LESSONS FROM THE CHARACTERISTICS OF MONEY ITSELF

2.3 REGULATORY DRIVERS FAVOURING MARKET CONSISTENT VALUATIONS

2.4 UNDERLYING THEORETICAL ATTRACTIONS OF MARKET CONSISTENT VALUATIONS

2.5 REASONS WHY SOME PEOPLE REJECT MARKET CONSISTENCY

2.6 MARKET MAKING VERSUS POSITION TAKING

2.7 CONTRACTS THAT INCLUDE DISCRETIONARY ELEMENTS

2.8 VALUATION AND REGULATION

2.9 MARKING-TO-MARKET VERSUS MARKING-TO-MODEL

2.10 RATIONAL BEHAVIOUR?

Chapter 3 - Different Meanings given to ‘Market Consistent Valuations’

3.1 INTRODUCTION

3.2 THE UNDERLYING PURPOSE OF A VALUATION

3.3 THE IMPORTANCE OF THE ‘MARGINAL’ TRADE

3.4 DIFFERENT DEFINITIONS USED BY DIFFERENT STANDARDS SETTERS

3.5 INTERPRETATIONS USED BY OTHER COMMENTATORS

Chapter 4 - Derivative Pricing Theory

4.1 INTRODUCTION

4.2 THE PRINCIPLE OF NO ARBITRAGE

4.3 LATTICES, MARTINGALES AND ÎTO CALCULUS

4.4 CALIBRATION OF PRICING ALGORITHMS

4.5 JUMPS, STOCHASTIC VOLATILITY AND MARKET FRICTIONS

4.6 EQUITY, COMMODITY AND CURRENCY DERIVATIVES

4.7 INTEREST RATE DERIVATIVES

4.8 CREDIT DERIVATIVES

4.9 VOLATILITY DERIVATIVES

4.10 HYBRID INSTRUMENTS

4.12 WEIGHTED MONTE CARLO AND ANALYTICAL ANALOGUES

4.13 FURTHER COMMENTS ON CALIBRATION

Chapter 5 - The Risk-free Rate

5.1 INTRODUCTION

5.2 WHAT DO WE MEAN BY ‘RISK-FREE’?

5.3 CHOOSING BETWEEN POSSIBLE MEANINGS OF ‘RISK-FREE’

Chapter 6 - Liquidity Theory

6.1 INTRODUCTION

6.2 MARKET EXPERIENCE

6.3 LESSONS TO DRAW FROM MARKET EXPERIENCE

6.4 GENERAL PRINCIPLES

6.5 EXACTLY WHAT IS LIQUIDITY?

6.6 LIQUIDITY OF POOLED FUNDS

6.7 LOSING CONTROL

Chapter 7 - Risk Measurement Theory

7.1 INTRODUCTION

7.2 INSTRUMENT-SPECIFIC RISK MEASURES

7.3 PORTFOLIO RISK MEASURES

7.4 TIME SERIES-BASED RISK MODELS

7.5 INHERENT DATA LIMITATIONS APPLICABLE TO TIME SERIES-BASED RISK MODELS

7.6 CREDIT RISK MODELLING

7.7 RISK ATTRIBUTION

7.8 STRESS TESTING

Chapter 8 - Capital Adequacy

8.1 INTRODUCTION

8.2 FINANCIAL STABILITY

8.3 BANKING

8.4 INSURANCE

8.5 PENSION FUNDS

8.6 DIFFERENT TYPES OF CAPITAL

Chapter 9 - Calibrating Risk Statistics to Perceived ‘Real World’ Distributions

9.1 INTRODUCTION

9.2 REFERRING TO MARKET VALUES

9.3 BACKTESTING

9.4 FITTING OBSERVED DISTRIBUTIONAL FORMS

9.5 FAT-TAILED BEHAVIOUR IN INDIVIDUAL RETURN SERIES

9.6 FAT-TAILED BEHAVIOUR IN MULTIPLE RETURN SERIES

Chapter 10 - Calibrating Risk Statistics to ‘Market Implied’ Distributions

10.1 INTRODUCTION

10.2 MARKET IMPLIED RISK MODELLING

10.3 FULLY MARKET CONSISTENT RISK MEASUREMENT IN PRACTICE

Chapter 11 - Avoiding Undue Pro-cyclicality —in Regulatory Frameworks

11.1 INTRODUCTION

11.2 THE 2007-09 CREDIT CRISIS

11.3 UNDERWRITING OF FAILURES

11.4 POSSIBLE PRO-CYCLICALITY IN REGULATORY FRAMEWORKS

11.5 RE-EXPRESSING CAPITAL ADEQUACY IN A MARKET CONSISTENT FRAMEWORK

11.6 DISCOUNT RATES

11.7 PRO-CYCLICALITY IN SOLVENCY II

11.8 INCENTIVE ARRANGEMENTS

11.9 SYSTEMIC IMPACTS OF PENSION FUND VALUATIONS

11.10 SOVEREIGN DEFAULT RISK

Chapter 12 - Portfolio Construction

12.1 INTRODUCTION

12.2 RISK-RETURN OPTIMISATION

12.3 OTHER PORTFOLIO CONSTRUCTION STYLES

12.4 RISK BUDGETING

12.5 REVERSE OPTIMISATION AND IMPLIED VIEW ANALYSIS

12.6 CALIBRATING PORTFOLIO CONSTRUCTION TECHNIQUES TO THE MARKET

12.7 CATERING BETTER FOR NON-NORMALITY IN RETURN DISTRIBUTIONS

12.8 ROBUST OPTIMISATION

12.9 TAKING DUE ACCOUNT OF OTHER INVESTORS’ RISK PREFERENCES

Chapter 13 - Calibrating Valuations to the Market

13.1 INTRODUCTION

13.2 PRICE FORMATION AND PRICE DISCOVERY

13.3 MARKET CONSISTENT ASSET VALUATIONS

13.4 MARKET CONSISTENT LIABILITY VALUATIONS

13.5 MARKET CONSISTENT EMBEDDED VALUES

13.6 SOLVENCY ADD-ONS

13.7 DEFINED BENEFIT PENSION LIABILITIES

13.8 UNIT PRICING

Chapter 14 - The Final Word

14.1 CONCLUSIONS

14.2 MARKET CONSISTENT PRINCIPLES

Bibliography

Index

For other titles in the Wiley Finance Series please see www.wiley.com/finance

© 2009 John Wiley & Sons Ltd

Registered office John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com.

The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Library of Congress Cataloging-in-Publication Data

Kemp, Malcolm H. D.

Market consistency : model calibration in imperfect markets / Malcolm H. D. Kemp. p. cm. - (Wiley finance series)

eISBN : 978-0-470-68489-4

1. Capital market. 2. Banks and banking. 3. Risk management. I. Title.

HG4523.K46 2009

332’.0415-dc22

2009021636

A catalogue record for this book is available from the British Library.

Typeset in 10/12pt Times by Aptara Inc., New Delhi, India

Preface

This book grew out of a desire on my part to offer constructive help to the many different individuals who need to take account of ‘what the market has to say’ in their day-to-day activities. It builds on several papers and talks that I have given that explore how risk management and related disciplines might develop as fair valuation principles become more entrenched in finance and regulatory practice.

The most obvious starting point for such a journey, and probably the one most focused on by previous authors, is the problem of how to place a market value on a financial instrument or other asset or liability. This process is typically known as marking-to-market in the banking, market making and asset management worlds. It is known as adopting a market consistent or fair valuation (if applied to less liquid assets and liabilities) within the insurance and accounting worlds.

Marking-to-market has become a core business discipline for many different types of financial organisation in recent years. Significant further growth in its use can be expected as a result of the implementation of the European Union’s Solvency II proposals for EU insurance companies. The underlying conceptual framework of these proposals is similar to the framework underlying the Basel Capital Accords applicable to banks.

However, marking-to-market has also been blamed by some for exacerbating the worldwide recession that was developing as this book was being written. We call this economic event the ‘2007-09 credit crisis’ throughout this book even though it was no longer just a ‘credit’ crisis by the time this book was completed in early 2009 and it did not then look particularly likely to have drawn to a close by the end of 2009.

In this book we therefore first explore the theoretical justifications for (and against) market consistency. This allows us to identify from first principles when such an approach might be of most relevance and what it might then look like. Special attention is given to how market consistency can be adopted in ways that do not result in undue pro-cyclicality in regulatory frameworks, given the importance of this to society at large.

Merely deciding that market consistency might be a useful concept doesn’t actually make it easy to apply. For example, when implementing Solvency II, EU insurance companies will face the practical challenge of identifying suitable market consistent valuations to place on their insurance liabilities even though markets in such liabilities are typically poorly developed, if they exist at all. How can you in practice develop models for valuing such instruments in these circumstances?

Moreover, in today’s interconnected world, the challenges that one financial sector might face are often mirrored in other sectors. At the time of writing, we were in the middle of one of the most severe credit crunches in recent times. Markets in many types of commercial paper, asset backed securities and collateralised debt obligations had largely dried up. Bid-offer spreads on even relatively straightforward bonds had reached historical wides. More complex instruments were being ascribed very different marks-to-market by different market makers.

Sometimes marking-to-market is explicitly contrasted with marking-to-model. Such an approach involves defining a model the output of which is then deemed to be the relevant (market) price to use for the asset or liability in question. Some of the challenges seen in some of the markets mentioned above seem in part to reflect inconsistent or inappropriate design of models used for these purposes. However, there is no clear dividing line between marking-to-market and marking-to-model. Indeed, it is unrealistic to expect to be able to avoid marking-to-model especially for less liquid and less fungible assets. For example, market prices used in the property, i.e. real estate, markets are typically the output from surveyors’ ‘models’, based on rental income, etc., which have been ‘calibrated’ to market information by reference to recent market transactions on similar but not identical properties. When tested by actual market transactions they may prove unreliable.

I strongly believe in the merits of adopting a holistic perspective. The recent trend towards enterprise risk management suggests that others also think viewing things holistically has merit. So, it seems to me that limiting market consistency merely to the task of valuing assets and liabilities is to diminish its true worth. There are many other areas of financial practice that can benefit from better incorporation of information extracted from market prices or other market observables.

In particular, it seems to me that market consistency is of great importance in risk management and even provides some key insights that can enhance portfolio construction. Indeed, I would argue that this must be the case. When thinking about the process of managing assets and liabilities end to end, we must start where we currently are. However, it would be a poorly managed organisation that wasn’t then interested in understanding what might happen going forwards. A natural follow-on question to ask is what might happen if we changed the asset and liability mix. By implication we are trying to identify what might be ‘good’ changes to make to the asset/liability profile. Of course, applying market consistency to these disciplines involves somewhat different techniques to those applicable to the valuation problem. However, the underlying philosophy is the same, i.e. to make appropriate use of whatever information can be inferred from market observables.

However sophisticated is the analysis, market consistency cannot answer all questions. The world in which we live includes ‘unknown unknowns’. There are certain problems that cannot even in principle be fully addressed by reference to market observables. Irrespective of how well developed a market might be, it will not provide all the information that might be needed to tackle all of these problems. For example, over the last few decades there has been a huge growth in derivatives markets. Core to this has been the development of associated hedging algorithms. Indeed, one can argue that a key reason for the wide adoption of, say, the Black-Scholes option pricing formula is that it provides a framework for hedging risk, as well as for pricing risk. Less well appreciated is that it is intrinsically harder to ensure hedging metrics are market consistent than it is to ensure that pricing metrics are market consistent. The market observables that we would need for the former are more complex and therefore less prevalent than those needed for the latter.

Given the holistic perspective that I think market consistency deserves, I have tried to write this book bearing in mind the varying needs of the different types of practitioner who should be interested in market consistency. Actuaries, auditors and other professionals are being required as part of modern financial regulation to focus increasingly on market consistent valuations. They may find the earlier chapters and the suggestions in the final chapter on how to tackle less liquid instruments particularly helpful. Traders and investment managers (and quants working alongside them) may be most interested in the later chapters relating to portfolio construction and hedging. Risk managers are likely to be interested in both of the above but may be particularly interested in the chapters in the middle of the book dealing with the practical ramifications of market consistency on risk management activities and on regulatory capital frameworks. Structurers, salespeople and others developing or marketing products to financial services organisations may find helpful the entire spectrum of material covered in the book. Every part of it may help them help their clients better in the changing world in which we live. Back office professionals preparing mark-to-market valuations may be most interested in the sections on marking-to-market less liquid instruments. Regulators may value the insights and the principles that the book sets out when exploring whether and how to place greater focus on market pricing in regulatory frameworks. And I hope that students, academics and others will find the systematic treatment of all of this material interesting and helpful in their ongoing studies and research.

The actual structure of the book naturally reflects some preferences on my part. I appreciate that not all potential readers will have a strong mathematical background. More technical mathematical aspects of market consistency are therefore concentrated into a few chapters. They are developed in ways that I hope will prove relatively intuitive to follow even for those who are not ‘rocket scientists’ at heart. The book aims to blend together a systematic development of underlying principles with their practical application. Core principles are highlighted as they arise in the text. My hope is that clear articulation of these principles will lead to long and fruitful interactions that will enhance the work that we all do within the financial community.

Acknowledgements

I would like to thank Pete Baker, Aimee Dibbens and their colleagues at Wiley for encouraging me to embark on writing this book. Thanks are also due to Colin Wilson, Seamus Creedon, Stephen Bashforth and others who have read parts of this manuscript and provided helpful comments on how it might be improved. A special appreciation goes to my wife and family for supporting me as this book took shape. I would also like to record my thanks to colleagues at Threadneedle Asset Management for providing an environment in which its ideas could develop and take root.

However, whilst I am very grateful for the support I have received from various directions during the writing of this book, I still take sole responsibility for any errors and omissions that it contains.

Abbreviations

Notation

1

Introduction

1.1 MARKET CONSISTENCY

This book is about market consistency, a term that we use throughout this book as a catch-all for the activity of taking account of ‘what the market has to say’ in financial practice. We will explore, from first principles, when it is (and when it isn’t) most appropriate to listen to what the market is saying. We will also explore how in practice we might extract information from ‘the market’ (and also what ‘the market’ is) in those circumstances where market consistency has merit.

We can characterise the incorporation of market consistency in a piece of financial analysis as involving the creation of a suitable model or idealised abstraction of how something works, which is then calibrated using appropriate market derived information. We focus in this book on models applicable to three core areas of financial practice, namely:

(a) Valuation methodologies, i.e. the placing of values or prices on positions in financial instruments or other sorts of assets or liabilities;

(b) Risk management processes, i.e. the assessment and management of the (financial) risks that are created by holding such positions; and

(c) Portfolio construction techniques, i.e. the selection of which sorts of such positions or risks it is most desirable to hold (or avoid), and in what quantities.

We do so because these three disciplines are closely allied, both in theory and in practice. We do not value something in complete isolation. Instead, there must be reasons for doing so. One key reason is to understand better the characteristics of the positions we currently have. But what is the point of gaining this understanding? Surely, it is to be better placed to understand the potential behaviour that the positions might exhibit in the future. We want to understand the risks and possible rewards attaching to them, i.e. how they might behave in adverse and favourable circumstances.1 The natural next step is then to consider and take decisions about how best to manage the risks and potential rewards we face. This in turn naturally leads to the question of which exposures we should adopt in the first place.

1.2 THE PRIMACY OF THE ‘MARKET’

Why might we care about what the market has to say? The world of finance is rarely far away from anyone these days. For some, life is a day-to-day struggle to make ends meet. Others, more fortunate, may have surplus funds deposited with banks or invested in the multitude of financial products now emanating from the world’s financial centres. But even they will often have been in debt at some stage in their lives. Perhaps this will have been to finance the purchase of a car or house, or, for the more entrepreneurially minded, to support a new business venture. The same is true on a larger scale for companies, charities, even entire countries. For better or worse, money, as a means of storing and transferring value, has proved to be one of humankind’s more important inventions. Indeed, it has been like this for many centuries, at least for the wealthier end of society. Julius Caesar built up huge debts (like several other Roman politicians of his day) and then amassed huge wealth on his way to seizing supreme power in the Roman Republic. Banking provided the wealth that enabled the Medici patronage of the Renaissance.

The last few decades have arguably seen a spurt in financial innovation. There has been huge growth in derivatives markets and in the range and sophistication of financial products and instruments now available to individuals, corporations and financial entities. In part, this reflects the technological innovation, economic growth and capital accumulation that large parts of the world have seen since the Second World War. It also, in my opinion, reflects the particular focus given during this period by financial theory and practice to the concept of the market. By this we mean some possibly hypothetical construct in which whatever we are interested in can be bought or sold without (too much) difficulty. Economic theory has always argued that properly functioning marketplaces are important for effective competition and hence efficient allocation of resources across an economy. The core innovation over the last few decades has been to apply this more general economic insight to finance itself.

We have seen, for example, changes in the underlying business models that many banks have adopted, away from a ‘borrow and hold’ business model (in which a bank would raise money from its depositors and then itself lend the proceeds, or some multiple of them, to some of its other customers) and towards an ‘originate and sell’ business model (in which the bank’s assets are repackaged and sold to other capital markets participants). Some reversal of this trend is also now apparent as more straightened economic times loom ahead.

The ‘sell’ element of such a business model ultimately involves transfer of risk exposures to third parties. It is of course facilitated if there are ready markets in such exposures. It places particular importance on the (market) prices at which transfers of these exposures can take place. Many banks and bank-like entities have historically been involved to some extent in market making activities. Increasingly, they have become ‘market makers’ of their own core cash flow streams.

Other players in the financial markets, such as pension funds and insurance companies, have perhaps been less affected by this fundamental change in financial mindsets. However, they too do not operate in a vacuum. The mere possibility of transferring blocks of pension benefits or policyholder entitlements reminds (or ought to remind) them that they too live not only surrounded by markets but also, in some sense, within them. For better or for worse, a focus on ‘what the market has to say’ is likely to be here to stay.

This does not mean that enthusiasm for ‘the market’ will not wax and wane over time. As I wrote this book, a number of pillars of the financial services industry were revealing eye-watering sized losses. These were arguably a consequence of their overexuberant hope that markets would continue to operate in ways to which they had previously become accustomed. Some of these firms subsequently defaulted. Others had to be rescued in mammoth government sponsored bailouts as governments endeavoured to bring stability to their financial systems. Enthusiasm for listening to what the market has to say often diminishes when what it is saying is unpalatable!

1.3 CALIBRATING TO THE ‘MARKET’

Merely observing that there may be useful information extractable from the ‘market’ doesn’t actually help us extract this information. Nor does it help us work out when such information is of most use or how to use it in ways that do not exacerbate systemic, i.e. economy-wide, risks. This is particularly true when the ‘market’ does not have all of the characteristics of the perfectly behaved construct of economic theory. In this book we explore how market consistency can be applied in the world in which we actually live, where markets are imperfect.

As mentioned above we focus in this book on three interrelated disciplines, namely valuation, risk management and portfolio construction. In each case, we focus on practical ways of incorporating greater market consistency, whilst simultaneously providing a systematic treatment of the underlying theory. For example, when dealing with valuation, we explore the theory and regulatory drivers currently favouring greater use of marking-to-market (as well as describing some of the current countervailing drivers) and we explore the differences, or often the lack thereof, between mark-to-market and mark-to-model valuations. However, there are also sections offering more practical guidance on how to determine market consistent valuations for assets or liabilities where markets are limited, illiquid or even almost non-existent.

In the field of risk management, market consistency can mean different things to different people. In my view, we are likely in due course to see a paradigm shift towards greater use of market implied risk measures. The principles involved are explored in some detail, because of the important ramifications this would have for practitioners in this field. However, I accept that this view has yet to achieve wide acceptance (not least because of the practical challenges involved and the unpalatable answers it might generate). The book therefore also provides a full treatment of the more limited types of market consistency that are incorporated in current risk measurement and management paradigms. These might best be described as attempting to quantify (probabilistically or otherwise) the ‘real world’ likelihood of some risk materialising over a given timeframe.

The application of market consistency to portfolio construction is simultaneously both core and peripheral. Decisions about what to buy or sell should take account of how much they might cost to buy or sell. However, a world in which the ‘market’ is taken to be the sole arbiter of knowledge is not one that can be fully reconciled with the concept of active management. By this we mean taking views about when the market is right and when it is wrong and acting accordingly. The need here is to understand and take due notice of the market but not to let it be the sole input into your decision-making process.

1.4 STRUCTURE OF THE BOOK

For practical reasons, this book is, in the main, segmented between the interrelated disciplines described above. Most larger financial organisations segment their business by activity. The day-to-day working life of readers within the financial services industry will therefore tend to have a bias towards one of these three disciplines. For example, asset management and investment banking businesses are often subdivided into three parts, a back office, a middle office and a front office. Typically, individuals working in the back office have day-to-day responsibility for the processes used to administer and value relatively simple instruments. Those working in the front office are responsible for deciding what positions to buy or sell. Those working in the middle office may provide a bridge between back and front offices. It is also becoming the norm for risk management to be explicitly differentiated from the front (and back) office and thus to fall within the remit of the middle office function. Similar types of role distinctions do also apply to insurers and pension funds, but in many cases the activities in question are outsourced or given different names.

The structure adopted by this book is as follows.

Chapters 2 to 6 cover in the main core material applicable to all of the disciplines being considered. Chapter 2 focuses on when market consistency is and isn’t appropriate, extrapolating from the properties of money to establish what sorts of properties we might expect monetary values to exhibit. From these properties it deduces when it is most and least relevant to be market consistent. It also explores at a high level some of the main drivers currently for and against market consistent approaches, and how this can influence what in practice is actually meant by ‘market consistency’. Chapter 3 focuses more on valuation activities, and on how in practice different standards setters and commentators interpret market consistency and other similar terms. Chapter 4 provides a primer on derivative pricing theory. It cannot be claimed that it does so without reference to any mathematics. However, hopefully even those less welcoming of complex mathematical arguments will feel that they have gained some useful insights after reading it. In this context ‘derivative pricing theory’ is really a catch-all for virtually all of the financial theory that underpins the rest of the book (apart from the theory relating to the interaction of risk and return, which we introduce principally in Chapter 12). Chapter 5 explores a particularly important issue in practice when applying market consistent principles to less liquid instruments. This is how to understand and identify a suitable ‘risk-free’ interest rate or yield curve. For a book that is focused in part on how to handle less liquid markets it is also natural to include in this part of the book a chapter specifically on liquidity theory, i.e. Chapter 6.

Chapters 7 to 10 consider market consistency in the context of risk management. Chapter 7 covers the fundamental theory, such as the description of different sorts of risk measures and how they are typically calculated. Chapter 8 focuses on capital adequacy. It provides examples of how current regulatory frameworks try to identify the appropriate minimum capital that an organisation should hold to protect itself against risks that might lead to insolvency. Chapter 9 explores how to apply market consistency in what might be called the current risk management paradigm. It focuses principally on how market valuations might vary in the future, and hence how the risks being expressed in these positions might be best managed. Chapter 10 considers how these approaches would need further refinement if we want them to adopt a fully market consistent paradigm. This involves applying market consistency not merely to the valuations used within the risk assessment but also to the probability distributions ascribed to the future movement of these valuations. Reasons for adopting this paradigm are also covered in this chapter.

One claim some commentators made during the 2007-09 credit crisis was that inappropriate use of marking-to-market techniques can create a lack of confidence in the financial soundness of banks in stressed market conditions, which can then become a self-fulfilling prophecy. It behoves a book on market consistency to consider carefully this logic. We explore in Chapter 11 ways in which as a society we may best protect against system-wide concerns whilst not diluting other benefits that may come from a greater focus on ‘what the market has to say’.

Chapter 12 focuses on portfolio construction techniques. As noted earlier, the application of market consistency to portfolio construction is less direct than for valuation and risk management, but arguably no less important. There are also several analogies that can be drawn out between market consistency as applied to portfolio construction and market consistency as applied to other financial disciplines.

Chapter 13 draws together many of the strands developed in earlier chapters. It provides case studies exploring how to incorporate market consistency in various types of computations relating to different types of assets and liabilities. It also explores questions like what to do when all available market observable prices relate to relatively illiquid instruments with relatively large or uncertain bid-offer spreads. There would be less need for this book if markets were always ‘perfect’.2

Finally, Chapter 14 summarises and repeats in one place all of the market consistency principles highlighted elsewhere in the book.

Throughout the book we draw out principles, i.e. guidance to practitioners, that have relatively universal application, independent of any particular regulatory or current market practice drivers. Within the text these principles are indented and shown in bold, and are referenced by P1, P2, etc.

Each chapter contains at least one such principle, along with many other insights. Thus readers only interested in certain aspects of market consistency should still find this book worthwhile reading, even if they limit their attention just to those chapters particularly relevant to their own specialisms. Conversely, there are valuable insights throughout the book, including in the more mathematical sections (even for readers who don’t wish to follow the mathematics in detail). I would therefore encourage all readers to consider exploring parts of the book that they might have assumed were only tangentially relevant to their own specialisms, because of the greater depth of understanding that this might bring.

1.5 TERMINOLOGY

In parts of the book focused on valuation, we use as essentially interchangeable terms such as market consistency and ‘marking-to-market’. Another term with much the same meaning is the accounting concept of fair value. When financial services regulators use the term ‘realistic valuation’ they also normally have a similar concept in mind. We define the market consistent value of an asset or liability to mean:

(a) Its market value, if it is readily traded on a market at the point in time that the valuation is struck; or

(b) A reasoned3 best estimate of what its market value would have been had such a market then existed, in all other situations.

Such a definition is similar to the more standard accounting definition of fair value as ‘the value at which an arm’s-length transaction involving willing, knowledgeable counterparties would take place’. However, explicit inclusion of the word ‘market’ within the terminology has the advantage of highlighting that for a non-traded asset or liability we are not wanting to focus on the valuer’s own intrinsic assessment of its value. Rather we are interested in modelling how some actual or hypothetical market would be expected to value the asset or liability. By implication, we also demand that any such model should, if possible, be calibrated back to market prices of instruments that are more readily traded.4

More generally, choosing to use the term ‘market consistency’ has the advantage of not unwittingly guiding the reader towards an overly valuation centric focus, thus downplaying other disciplines to which market consistency may be applicable. Focusing too much on terms like marking-to-market (or indeed any other phrase involving the word ‘mark’) runs this risk because a ‘mark’ is typically associated with the price we place on an instrument in our books. As we have already noted (and we will stress subsequently), we do not value things in isolation. Valuations ultimately gain their wider meaning and context from the purposes to which they are put.

Given that this book covers risk management, it is also appropriate to include in this section terminology to help categorise the main sorts of risks faced by financial services entities. We shall explore in more detail later what we (and others) might mean by ‘risk’. However, at this juncture, a helpful subdivision often used in practice is the following. It should be noted that it is not always easy (or even useful) to identify clear boundaries between some of these different types of risk.

(a) Market risk, i.e. the risk of loss due to adverse market movements. More generally, we might focus on adverse market movements affecting the entity’s asset/liability position, although this might be called asset-liability risk. Market movements in this context would typically include movements in equity values and in interest rates.

(b) Credit risk, i.e. the risk that the creditworthiness of a name or counterparty to which an entity is exposed declines, causing the entity loss. At one extreme would be actual default of the counterparty. A subtlety here is whether credit risk should be deemed to include only default risk (i.e. some intrinsic assessment now of the risk that the counterparty or issuer might default in the future) or whether it should also include ratings migration risk or spread risk. The spread on a bond-like or cash-like instrument is the difference in the redemption yield available on the instrument versus the corresponding yield available on some standard reference instrument. For example, people refer to ‘spread’ versus government bond yields, as the difference between the yield on the instrument in question and the yield on government bonds of equivalent duration, type and currency.5 However, spread could be measured versus Libor, see Section 5.2, or some other interest rate or yield measure, etc. The market price of a bond subject to default risk is influenced by likelihood of future default. We might attempt to proxy this by some statistic based on the credit rating that a credit rating agency or an internal credit ratings team ascribes to the instrument. However, the market price, and hence spread, will also be influenced by the market’s expectation of how likelihood of default might change over time. Even if one ignores market prices and focuses on some perceived ‘intrinsic’ likelihood of default derived from credit ratings, these can also change through time (even if the instrument has not defaulted). A rating ascribed to a particular instrument can migrate up or down. At issue is whether spread risk is a form of credit risk (i.e. defining credit risk as risks associated with ‘credit’ instruments) or whether it is a form of market risk (i.e. defining market risk as anything relating to movements in market prices whatever the instrument type).

(c) Liquidity risk, which the UK’s Financial Services Authority (FSA) defines as the risk that a firm, although balance sheet-solvent, cannot maintain or generate sufficient cash resources to meet its payment obligations in full as they fall due, or can only do so at materially disadvantageous terms, see FSA (2007). Some view a part of the spread payable on non-default-free instruments as relating to their liquidity characteristics, again highlighting the difficulties in rigidly demarcating between different types of risk.

(d) Insurance risk, i.e. risks specific to insurance companies, typically relating to the uncertain outcome of insurance contingencies. These would typically include life contingencies, i.e. risks linked to mortality, morbidity or longevity.6 They would also include property/ casualty and other sorts of non-life insurance risk. Non-life insurance is called ‘general’ insurance in some jurisdictions. It is not always easy to differentiate what risks are ‘insurance-related’ and what are not, other than by falling back onto the practical but partly circular definition that insurance risks are ones that are carried by insurance companies.

(e) Operational risk, i.e. the risk of loss resulting from inadequate or failed internal processes, people and systems or from external events. In the six-way classification of risk described in this section a wide range of risks would be deemed to fall into this category, including legal risk and, possibly, strategic risk and reputational risk.7

(f) Group risk, i.e. the additional risk to a particular legal entity caused by it being within a larger group structure. For example, resources may be diverted from the entity in question to other group companies if the latter companies suffer a large loss, which can have adverse knock-on effects which would not have arisen had the entity been stand-alone.

For convenience, we also follow the convention, adopted by many other writers in this field, of using the term firm to encompass not just bodies with an explicit corporate form and purpose but also other entities that operate within the financial services arena, such as pension funds. Where the context demands, we clarify the specific type of ‘firm’ or ‘entity’ on which our attention is focused.

2

When is and when isn’t Market Consistency Appropriate?

2.1 INTRODUCTION

In this chapter we seek to identify when focusing on market consistency is most relevant. There is little point in expending substantial effort on being market consistent when it is less relevant or even misguided to try to do so.

To do this, we start by formulating from first principles what we ought to understand by ‘market consistency’. Any such logic needs to link back to the underlying economic theory underpinning the functioning of markets. So, we focus on intrinsic characteristics exhibited by money itself, since the existence of money is a necessary precursor to the placing of a monetary value on anything. This powerful analogy helps us identify how to place a (monetary) value on assets or liabilities, and how this value may vary depending on the purpose of the valuation.

However, we do not live in a world governed merely by economic theory. So, later on in this chapter we also look at some of the more practical drivers favouring greater (and in some instances lesser) adoption of market consistency. In particular we explore current regulatory and accounting pressures. Since these drivers will change from time to time, we also explore what are the underlying mindsets that have influenced and are likely to continue to influence these drivers.

Towards the end of the chapter we develop from these underlying principles a conceptual framework for valuation, risk management and capital adequacy to which we can refer whenever we need greater clarity on what we mean by ‘market consistency’. This framework necessarily includes elements covering both valuation and risk management. The amount of assets that a financial entity needs to hold to demonstrate ‘solvency’ depends not just on the value of its liabilities but also on whatever additional capital is considered necessary by regulators, markets and others.

2.2 DRAWING LESSONS FROM THE CHARACTERISTICS OF MONEY ITSELF

2.2.1 The concept of ‘value’

As humans, we value things that are worth something to us and, if we are altruistic, worth something to society (or life more generally). What we value can be many and varied, e.g. our relationships with others (particularly our family), the pleasure we get from a nice meal or seeing a beautiful sunset, the benefits we may get from the ownership or use of a good or provision of a service. Taking a utilitarian philosophical stance we might say that the value of such things depends on the utility that they provide to the recipient (or conversely the disutility they generate to the provider). Those with a more religious perspective, like me, might argue that some things like human life, a just and free society, etc. have an intrinsic value (or value to God) irrespective of the utility that they might bring to any specific individual.

Principle P1: ‘Value’ is a term that has many different meanings. Unless the context makes clear the intended meaning, it should be used with a qualifier.

Early in the history of civilisation, humankind invented money. Individuals (or groups of individuals) can barter one good (or service, etc.) for another. Implicit in such bartering is some valuation process that each party is putting on the goods (or services) being swapped. It was found to be convenient to set aside special ‘commodities’ in restricted supply, i.e. ‘money’, which could provide a common means of exchange. Individuals no longer needed to ‘sell’ and ‘buy’ from the same person. They could disaggregate the barter into two separate transactions that could then be carried out with different parties.

Work relating to financial matters will typically involve placing monetary values on things. Strictly speaking, all such values must include a numeraire in which the value is expressed. We do not say that the monetary value of a loaf of bread is, say, 60 without some qualifier. Instead we might say that it costs £0.60 or $0.60, etc. Numeraires are often currencies (e.g. GBP, USD, EUR) but can also be, say, units in a fund, shares, barrels of oil, etc. In this context a currency would typically be the normal medium of exchange within a given economy, although occasionally bespoke currencies are created that are not in general circulation (e.g. the International Monetary Fund’s Special Drawing Rights).

Principle P2: When value is expressed in monetary terms and is immediate in nature then people will expect it to correspond to the amount of money that would need to be exchanged to buy or sell the item being valued, because the most important characteristic of money is that it is a medium of exchange.

2.2.2 The time value of money

Most of finance involves not just consideration of the present but also consideration of the future. As soon as money was invented, it actually became not just a medium of exchange but also a store of value. Rarely do we obtain money via one transaction and instantaneously use it up in another. The greater the time difference, the more money is being used as a store of value or wealth. This means that in practice the numeraire in which we express value needs to include a time dimension. This concept goes under various names, including the time value of money.

Most currencies exhibit a positive time value of money most of the time. This reflects the greater utility that people typically gain by consumption now versus consumption later, which typically works itself out economically in a positive interest rate being paid on whatever is the normal medium of exchange within a given economy. However, if an economy is undergoing sufficiently large negative inflation, the increase in purchasing power of a unit of its currency may provide a sufficient consumption deferral incentive without any positive interest rate. In some circumstances, it is even possible for interest rates on the normal medium of exchange to go negative. For example, if money is costly to store physically, then after taking these storage costs into account the net interest rate may be negative. This is one reason why societies typically now favour paper or electronically operated currencies that involve little if any physical storage costs.

It might be thought helpful to circumvent this issue by using a numeraire on which the relevant future ‘interest rate’ is set exactly equal to zero and then to re-express all future payments using this numeraire. This is the basic concept behind the present value of a future payment. It is the amount of money that you would need to set aside now to provide for a future payment, when compounded with interest in the meantime.

However, this basic concept is not by itself sufficient because it does not necessarily take fully into account the risk characteristics of the payments in question. Suppose that the 1 year interest rate on 1 year deposits is 10 % (for a given currency). Then the (current) market value of a payment of 110 in that currency in 1 year’s time is 100 currency units (if we ignore the risk of default), since this is the amount that you would need to put into a bank now to receive the relevant payment in the future.

This does not mean that an asset that involves a payment in 1 year’s time that we estimate to have a 50 % probability of being 100 and a 50 % probability of being 120 (and so currently to have an average or expected size in 1 year’s time of 110) will necessarily also have a market value of 100. Baxter and Rennie (1996) make a particularly eloquent case for why this so-called strong law of expectations which we might expect to apply in these situations actually leads to erroneous conclusions. For example, the circumstances when the payment is lower might correspond with situations when the typical recipient of the payment is also worse off in other ways, and may therefore be given a higher weighting in the formation of the market price than 50 %. Moreover, we will rarely know with certainty the exact probabilities involved since in practice the payment will depend on the outcome of contingent events the characteristics of which are not fully predictable in advance. Any one individual’s estimates of the probabilities involved may not correspond with others, again leading to a different price.

Properly taking into account the time value and risk characteristics of future payments requires the introduction of the concept of a pay-off. This is shorthand for the description of the complete tabulation of what the future payments might be (and when) in every possible outcome for the contingencies on which the payments depend. Any asset or liability can then be characterised by a suitable pay-off. Another equivalent term is a (contingent) claim.

2.2.3 Axioms of additivity, scalability and uniqueness

In this context, money has yet another characteristic that will strongly influence a lay person’s understanding of (monetary) value, which is its fungibility. By this we mean that any two separate units of a currency are typically interchangeable with each other, and thus will be expected to have exactly equal value. More generally, the market value of money adheres to the mathematical axioms of additivity and scalability and hence lay people will generally expect (monetary) values to do likewise.

By these axioms we mean that if the ‘value’ we place on a pay-off a is V(a) then the value of a pay-off k(a + b) (i.e. a pay-off which is k times as large as a + b in all circumstances) satisfies the relationship:

(2.1)

The market value of money exhibits these characteristics. If 1 unit of currency A is worth a in currency C, and 1 unit of currency B is worth b in currency C, then the combination of k units of currency A and k units of currency B will be worth k(a + b) in currency C.

We note that implicitly we have also adopted an assumption of uniqueness, i.e. that V(a) is single-valued.

Principle P3: Most recipients of advice from financial practitioners will take for granted that monetary value, however measured, adheres to the basic axioms of additivity and scalability unless the context clearly justifies otherwise, because money itself adheres to these axioms.

The existence of a suitable risk-neutral distribution as per (c) above follows if we decompose the overall pay-off into scalar multiples of many component parts. Each component involves a payment of unity only when the contingencies driving the original pay-off result in a particular outcome. There is one such component for each possible outcome. The risk-neutral distribution is then found by rescaling the values of these components in a manner that results in a unit payout in all circumstances being valued in a manner that corresponds with the time value of money for the particular currency in which the valuation framework is expressed.

Conversely, a valuation framework will satisfy additivity and scalability if:

(a) There exists a probability distribution covering all possible outcomes; and

(b) The values ascribed to pay-offs under a given valuation framework equal the expected value of these pay-offs calculated by reference to this probability distribution.

Of course, we do not actually need to express a valuation (even one within a framework that satisfies additivity and scalability) in the form of a discounted value of an expected outcome, with the expectation being based on risk-neutral probability distributions. A widely used alternative is to carry out the expectation using our own ‘real world’ estimates of the probabilities of occurrence of the different possible outcomes and to adjust the time value element, i.e. to use a suitable risk discount rate. More accurately, the risk discount rate would in this formulation be a risk discount yield curve, to reflect how the time value of money varies by term to payment. In some disciplines such a discount rate is also called a deflator or a state price density.

A potential weakness in such a presentation is that lay people (and even some experts) might be misled into assuming that use of the term ‘risk discount’ somehow implies that the resulting discount rate should be higher than the corresponding risk-free discount rate, since pay-offs involving uncertain payments might be expected in some sense to be riskier (i.e. less certain) than ones involving certain payments. The difference between the risk discount rate and the risk-free discount rate is typically referred to as the risk premium, a term which again might be thought to imply that risk premia should normally be positive.

However, if a valuation framework is to adhere fully to axioms of additivity and scalability then the appropriately averaged risk premium across all possible pay-offs must be zero. Thus, there will be some pay-offs the values of which when presented in this fashion need to incorporate negative risk premia, as well as some that will need to exhibit positive risk premia.8 Perhaps this is why practitioners in this field may prefer to use more technical terms such as ‘deflators’ which are less easy to misinterpret in this manner (or perhaps use of such terms is merely an example of jargon creep).

Principle P4: Financial practitioners wishing to derive valuations of future cash flows by calculating the expected values of these cash flows using their own (or other people’s) ‘real world’ probabilities of outcomes should carefully consider the risk adjustments that then need to be incorporated in the discount rates used in these computations. In particular they should bear in mind that the correct ‘risk premium’ to incorporate in the discounting computation (if they want their valuation framework to adhere to additivity and scalability axioms) may be negative and may bear no obvious link to the degree of uncertainty or ‘risk’ applicable to the cash flows being valued.

An important ramification of these axioms that we explore further in Section 3.3 is that they in effect force the focus to be on the ‘marginal’ trade.

2.2.4 Market consistent valuations

Principles P3 and P4 are generic to valuation frameworks that satisfy additivity and scalability. They do not by themselves place any requirement on the valuation framework to exhibit market consistency. However, they do guide us towards defining a market consistent valuation framework as one where:

(a) the (present) values of pay-offs that are actively traded in a deep and liquid market match (at all times) their observed market prices;

(b) the (present) values of pay-offs that are not actively and transparently traded in such a market are reasoned best estimates of the prices at which such pay-offs would trade in such a market, were such a market to exist; and

(c) valuations adhere to the basic axioms of additivity and scalability.

It is implicit in such a definition that the market must be suitably transparent, as otherwise the market prices would not be observable to those involved in preparing the relevant valuations. Also implicit is that any market to which (a) is applicable is sufficiently deep and liquid (and responds sufficiently rapidly to new information) to adhere to the principle of no arbitrage. Otherwise (a) and (c) are inconsistent with each other.

We note that in practice there will be subjectivity in (b). Indeed there will even be some potential subjectivity in (a), given the existence of bid-ask spreads and time delays in price formation within markets. The less the pay-offs can be decomposed into things that are actively traded in deep and liquid markets the greater is this subjectivity, which might be expressed in terms of widened implied bid-ask spreads.

Principle P5: Whilst a core aim of market consistency is objectivity by being faithful to market prices, market consistent valuations will often still involve subjective input. The more material is this subjectivity the more important it becomes for practitioners preparing these valuations to explain the judgements involved. They should be careful to minimise the impact that their own biases might have in selection of these subjective inputs.